<%BANNER%>

Essays on Capital Structure

Permanent Link: http://ufdc.ufl.edu/UFE0024199/00001

Material Information

Title: Essays on Capital Structure International Evidence
Physical Description: 1 online resource (150 p.)
Language: english
Creator: Oztekin, Ozde
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: adjustment, agency, approach, blundell, bond, bruno, capital, cash, constraints, corporate, corrected, cost, costs, credit, deviation, differencing, distress, dummy, dynamic, economic, external, financing, future, gmm, governance, industry, institutional, institutions, international, least, legal, leverage, long, market, markets, order, panel, partial, pecking, political, rebalance, research, securities, size, social, square, structure, tangibility, taxes, theory, timing, tradeoff, variable
Finance, Insurance and Real Estate -- Dissertations, Academic -- UF
Genre: Business Administration thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: My dissertation examines international differences in the determinants of capital structure and firms? adjustment to optimal capital structure across 37 countries over the period from 1991 to 2006. I first explore which leverage factors are consistently important for the capital structure decisions of firms around the world. The results highlight past leverage as the most important factor in all sample countries. Next, I examine the adjustment to optimal leverage. A firm?s ability to adjust its leverage is greatly influenced by economic, legal, and political institutions, corporate governance, tax systems, and the structure of credit and securities markets of the countries. In institutional environments with higher adjustment costs, due to the severity of external financing costs and regulatory cash constraints, the adjustment is significantly slower; in settings with greater adjustment benefits as implied by higher tax shields and the ability to prevent distress and deviation costs, the adjustment is considerably faster. Collectively, my research indicates that a country?s legal and institutional heritage is a significant factor in determining a firm?s choice of capital structure and the adjustment process to optimal leverage around the world.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Ozde Oztekin.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Flannery, Mark J.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024199:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024199/00001

Material Information

Title: Essays on Capital Structure International Evidence
Physical Description: 1 online resource (150 p.)
Language: english
Creator: Oztekin, Ozde
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: adjustment, agency, approach, blundell, bond, bruno, capital, cash, constraints, corporate, corrected, cost, costs, credit, deviation, differencing, distress, dummy, dynamic, economic, external, financing, future, gmm, governance, industry, institutional, institutions, international, least, legal, leverage, long, market, markets, order, panel, partial, pecking, political, rebalance, research, securities, size, social, square, structure, tangibility, taxes, theory, timing, tradeoff, variable
Finance, Insurance and Real Estate -- Dissertations, Academic -- UF
Genre: Business Administration thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: My dissertation examines international differences in the determinants of capital structure and firms? adjustment to optimal capital structure across 37 countries over the period from 1991 to 2006. I first explore which leverage factors are consistently important for the capital structure decisions of firms around the world. The results highlight past leverage as the most important factor in all sample countries. Next, I examine the adjustment to optimal leverage. A firm?s ability to adjust its leverage is greatly influenced by economic, legal, and political institutions, corporate governance, tax systems, and the structure of credit and securities markets of the countries. In institutional environments with higher adjustment costs, due to the severity of external financing costs and regulatory cash constraints, the adjustment is significantly slower; in settings with greater adjustment benefits as implied by higher tax shields and the ability to prevent distress and deviation costs, the adjustment is considerably faster. Collectively, my research indicates that a country?s legal and institutional heritage is a significant factor in determining a firm?s choice of capital structure and the adjustment process to optimal leverage around the world.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Ozde Oztekin.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Flannery, Mark J.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024199:00001


This item has the following downloads:


Full Text

PAGE 1

1 ESSAYS ON CAPITAL STRUCTURE : INTERNATIONAL EVIDENCE By ZDE ZTEK N A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 zde ztekin

PAGE 3

3 To my parents, Ay e Yarman ztekin and Erbil ztekin; my sister, lke ztekin; my brother, Bilgehan Uygar ztekin; and my husband, Ali G ng rayd no lu for their endless support and love they have given me through the years

PAGE 4

4 ACKNOWLEDGMENTS First and foremost, I acknowledge Mark J. Flannery, my dissertation committee ch air, for his endless mentoring in turning a collection of ideas into a practicable thesis He has always been very interested in my work and has always been available to advise me. I am grateful for his patience, enthusiasm and immense knowledge in finance I would like to thank him for his inspiring guidance throughout my work, without his help, this work would not be possible. I owe many thanks to my graduating committee Jay R. Ritter Mahendrarajah Nimalendran, and Chunrong Ai for their support througho ut my graduate work, for their assistance during committee meetings, for attending my seminars, for reading my dissertation and providing valuable feedback. I would also like to thank Andy Naranjo, David Brown, Evan Dudley, Jason Karceski, Joel Hou ston, Michael Ryngaert, Miles Livingston, and Olivier De Jonghe for their helpful comments. I am grateful to the seminar participants of the University of Kansas, University of Mississippi, IIT Stuart School of Business, Auburn University, University of San Dieg o, California State University at Fullerton, University of Memphis, Oregon State University, Villanova University, Miami University, and Iowa State University for their comments and suggestions. I would like to thank all my colleagu es at the University of Florida with whom w e shared great moments, supported each other and had fruitful scientific discussions together. I would like to thank my mother, Ay e Yarman ztekin; my father, Erbil ztekin; my sister, lke ztekin; my brother, Bilgehan Uygar ztekin for their faith and support. They were always with me and offered a hand during this long journey. I am indebted to my husbands family for their valued support and help. Last but not least, I would like to thank the most important person in my life, my husband Ali G ng rayd no lu He has always been with me with his love and unlimited patience, making me laugh and strongly supporting me.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 4 TABLE OF CONTENTS ..................................................................................................................... 5 LIST OF TABLES ................................................................................................................................ 7 LIST OF FIGURES .............................................................................................................................. 9 ABSTRACT ........................................................................................................................................ 10 CHAPTER 1 INTRODUCTION ....................................................................................................................... 11 2 THE DETERMINANTS OF CAPITAL STRUCTURE IN AN INTERNATIONAL CONTEXT ................................................................................................................................... 15 Motivation and Main Findings ................................................................................................... 15 Data .............................................................................................................................................. 19 Empirical Methodology .............................................................................................................. 20 Test Design .................................................................................................................................. 24 Separate Method .................................................................................................................. 24 Pooled Method ..................................................................................................................... 25 World model ................................................................................................................. 25 Institutional model ....................................................................................................... 25 Analysis and Results ................................................................................................................... 27 Leverage Factors and the Capital Structure Theories ........................................................ 27 Which capital structure determinants are reliably important? ................................... 27 The empirical relevance of the capital structure theories .......................................... 28 Leverage Factors, Capital Structure Theories, and the Effect of Conditioning on the Institutional Setting .......................................................................................................... 33 Which capital structure determinants are reliably important? ................................... 36 The empirical relevance of the capital structure theories .......................................... 37 Summary ...................................................................................................................................... 41 3 PARTIAL ADJUSTMENT TOWARD OPTIMAL CAPITAL STRUCTURE AROUND THE WORLD ........................................................................................................... 68 Motivation and Main Findings ................................................................................................... 68 An I nternational Theory of Partial Adjustment ......................................................................... 70 Adjustment Costs ................................................................................................................. 72 External financing costs ............................................................................................... 72 Cash constraints ............................................................................................................ 76

PAGE 6

6 Adjustment Benef its ............................................................................................................ 77 Costs of financial distress ............................................................................................ 77 The value of tax shields ............................................................................................... 78 Costs of deviation ......................................................................................................... 78 Data .............................................................................................................................................. 79 Empirical Methodology .............................................................................................................. 79 Test Designs ................................................................................................................................ 82 Separate Tests ...................................................................................................................... 83 Pooled Tests ......................................................................................................................... 83 Comparison of Separate and Pooled tests .......................................................................... 84 Analysis and Results ................................................................................................................... 84 Partial Adjustment Around the World ................................................................................ 84 International adjustment speeds .................................................................................. 84 Capital structure adjustments ...................................................................................... 87 The Impact of the Legal and Financial Traditions on the Adjustment Speeds ................ 90 Legal traditions ............................................................................................................. 90 Financial system s ......................................................................................................... 91 Institutional Determinants of the Adjustment Speed ......................................................... 93 Individual indices ......................................................................................................... 93 Alternative formulations of the adjustment factors .................................................... 99 Robustness ................................................................................................................................. 102 Alternative Definition of Leverage and Estimation Method .......................................... 103 Variable Adjustment Speed Model ................................................................................... 103 Asymmetric Response to Adjustment Factors ................................................................. 107 Relative leverage ........................................................................................................ 107 Firm size ..................................................................................................................... 108 Other considerations .................................................................................................. 110 Summary .................................................................................................................................... 111 4 DIRECTIONS FOR FUTURE RESEARCH .......................................................................... 136 Motivation .................................................................................................................................. 136 Empirical Methodology ............................................................................................................ 136 Test Design ................................................................................................................................ 138 Direct Effect of the Firm, Industry, and Macroeconomic Determinants of Leverage .. 138 Effects of the Institutional Determinants on Leverage .................................................... 139 Direct institutional effect ........................................................................................... 139 Indirect institutional effect ......................................................................................... 139 5 CONCLUSION ......................................................................................................................... 141 LIST OF REFERENCES ................................................................................................................. 143 BIOGRAPHICAL SKETCH ........................................................................................................... 150

PAGE 7

7 LIST OF TABLE S Table page 2 1 Description and Sources of Firm, Industry, and Macroeconomic Determinants of Capital Structure ..................................................................................................................... 43 2 2 Description and Sources of Institutional Characteristics Influencing Firm, Industry, and Macroeconomic Determinants of Capital Structure ...................................................... 44 2 3 Compustat Global Sample and Macroeconomic Variables ................................................. 46 2 4 Institutional Characteristics ................................................................................................... 47 2 5 Correlations for the Individual Institutional Indices ............................................................ 49 2 6 Separate Country Regressions Using Book Leverage as the Dependent Variable ............ 50 2 7 Separate Country Regressions Using Market Leverage as the Dependent Variable ......... 53 2 8 Pooled Country Regressions .................................................................................................. 56 2 9 Correlations between Leverage and Firm, Industry, Macroeconomic Determinants ........ 57 2 10 Empirical Relevance of the Capital Structure Theories ....................................................... 58 2 11 The Effect of Conditioning on the Institutional Setting ...................................................... 59 2 12 Correlations between Leverage and its Determinants for Weak, Strong, and All Institutional Settings .............................................................................................................. 63 2 13 Empirical Relevance of the Capital Structure Theories for Weak, Strong, and All Institutional Settings .............................................................................................................. 64 2 14 Bankruptc y Costs, Agency Costs, and Tax Benefits of Debt and the Determinants of Capital Structure ..................................................................................................................... 65 2 15 Agency Costs of Equity and the Determinants of Capital Structure ................................... 66 2 16 Information Asymmetry Costs and the Determinants of Capital Structure ........................ 67 3 1 Description and Sources of Variables Determining the Adjustment Speed ..................... 113 3 2 Description and Sources of Variables Determining the Optimal Capital Structure ........ 115 3 3 Adjustment Speed Estimates ............................................................................................... 116 3 4 The Frequency and Size of the Capital Structure Adjustments ......................................... 117

PAGE 8

8 3 5 The Impact of the Legal and Financial Traditions on the Adjustment Speed .................. 118 3 6 Institutional Determinants of the Adjustment Speed Individual Indices .......................... 119 3 7 Institutional Determinants of the Adjustment Speed Principal Components ................... 120 3 8 Alternative Definition of Leverage and Estimation Method ............................................. 121 3 9 Variable Adjustment Speed Estimates ................................................................................ 124 3 10 Institutional Determinants of the Variable Adjustment Speed Individual Indices .......... 125 3 11 Institutional Determinants of the Variable Adjustment Speed Principal Components ... 126 3 12 Asymmetric Response to Adjustment Costs and BenefitsIndividual Indices ................. 127 3 13 Asymmetric Response to Adjustment Factors -Principal Components ............................. 129

PAGE 9

9 LIST OF FIGURES Figure page 3 1 Adjustment Speed Estimates. .............................................................................................. 131 3 2 Capital Structure Adjustments ............................................................................................. 132 3 3 Adjustment Speeds and Adjustment Cost Factors ............................................................. 133 3 4 Adjustment Speeds and Adjustment Benefit Factors. ........................................................ 134 3 5 Variable Adjustment Speed Estimates. ............................................................................... 135

PAGE 10

10 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ESSAYS ON CAPITAL STRUCTURE : INTERNATIONAL EVIDENCE By zde ztekin May 2009 Chair: Mark J. Flannery Major: Business Administration M y dissertation examines international differences in the determinants of capital structure and firms adjustment to optimal capital structure across 37 countries over the period from 1991 to 2006. I first explore which leverage factors are consistently important for the capital structure decisions of firms around the world. The results highlight past leverage as the most important factor in all sample countries. Next, I examine the adjustment to optimal leverage. A firms ability to adjust its leverage is greatly influenced by economic, legal, and political institutions, corporate governance, tax systems, and the structure of credit and securities markets of the countries. In institutional environments with higher adjustment costs, due to the severity of external financing costs and regulatory cash constraints, the adjustment is significantly slower; in settings with greater adjustment benefits as implied by higher tax shields and the ability to prevent distress and deviation costs, the adjustment is considerably faster Collectively, my research ind icate s that a countrys legal and institutional heritage is a significant factor in determining a firms cho ice of capital structure and the adjustment process to optimal leverage around the world.

PAGE 11

11 CHAPTER 1 INTRODUCTION A central challenge facing financial economics is integrating finance theory with legal, social and political frameworks so that cross -country comparisons of financial data are facilitated and our understanding of the impact of different institutional features on the firms financing decis ions can be improved. My study is concerned with one such question, the impact of the combination of legal political, and social institutions on firms capital structures. The law and finance and re lated literature (La Porta, Lopez -de -Silanes, Shleifer, a nd Vishny (1997, 1998, 1999, 2000a, 2000b, 2002), Djankov, McLiesh, and Shleifer (2007), and others) assumes that institutional structure varies across countries and suggests that it would systematically affect the financing choices of firms. However, to d ate, to my knowledge, there has not been an attempt to systematically classify and quantify the relevance of country level institutional factors on the determinants of capital structure and the adjustment to optimal leverage This study takes a first step to fill this major gap in the capital structure literature by testing whether a firms adjustment speed reflects international cross -sectional variation in adjustment costs and benefits and whether country level institutional determinan ts of capital structure are important for explaining global capital structure activities. The static trade off theory posits that market frictions originate a link between leverage and firm value, and that firms pro actively and fully adjust to deviations from their optimal debt ratio (Graham and Harvey (2001)). The implicit assumption behind this theory is that a firms adjustment speed is equal to unity, and consequently that lagged leverage is redundant. The partial adjustment model (also referred to as the dynamic model or the dynamic tradeoff theory) recognizes that capital market imperfections create rebalancing costs that hamper the speed of adjustment (Hovakimian, Opler, and Titman (2001), Leary and Roberts (2005), Flannery and

PAGE 12

12 Rangan (2006), Strebul aev (2007) Huang and Ritter (2009) ). Accordingly, a firms ability to reverse the deviations from its optimal leverage varies by the cost and benefits of adjusting the leverage. On the other hand, the pecking-order, market timing, and inertia theories of capital structure do not propose any proactive effort s to reverse changes in leverage. The dynamic tradeoff, pecking order, market timing and inertia theories predict that a firms adjustment speed should lie between zero and one. While the dynamic trade off theory suggests that the adjustment speed systematically varies across firms as a function of the adjustment costs and benefits and is strictly different from both zero and one ; the pecking order, market timing, and inertia theories assume an adjustmen t speed close to zero, or equivalently a coefficient equal to unity on lagged leverage. In general, t he comparative analyses of capital structure in an international scope have ignored the dynamic nature of a firms financial decisions and/ or ignored unobserved firm heterogeneity inhe rent in capital structure data. Rajan and Zingales (1995), Demirguc -Kunt and Maksimovic (1996, 1999), Booth, Aivazian, Demirguc -Kunt, and Maksimovic (2001), Giannetti (2003), Fan, Titman and Twite (2006) and Jong, Kabir and Nguyen (2007) are some of the papers that use the observed debt ratio as a proxy for the optimal debt ratio in an international context Yet several studies show that imposing such constraints may bias coefficient estimates (Hovakimian, Opler, and Titm an (2001), Leary and Roberts (2005), Flannery and Rangan (2006), Strebulaev (2007), and Lemmon, Roberts, and Zender (2008 )). These papers also emphasize that i gnoring unobserved firm heterogeneity result s in a nontrivial omitted variable bias in the capit al structure estimations. Haas and Peeters (2006), Song and Philippatos (2004), and Farhat, Cotei and Abugri (2006) explore the dynamic nature of the firms capital structure in an international context, but fail to appropriately take into account firm he terogeneity Antoniou,

PAGE 13

13 Guney, and Paudyal (2008) explore the dynamic adjustment to optimal leverage in an international context by employing a system GMM and incorporating firm fixed effects in to the target equation However their sample is limited to a sm all -cross section of G5 countries and they dont investigate the determinants of the adjustment process. I use a large panel data set from 37 countries and address the econometric challenges involved in modeling the capital structure decisions of firms by employ ing Blundell and B onds (1998) dynamic system GMM, and simultaneously taking into account the heterogeneous nature of the data Undertaking a comprehensive investigation of the adjustment speeds and the determinants of capital structure around the wo rld, this study contributes to the capital structure literature in two ways. First I examine how firm, industry, and macroeconomic determinants relate to the capital structure of firms around the world in Chapter 2 The most reliable determinants of lever age are past leverage, tangibility, firm size, a research and development dummy depreciation expenses industry median leverage, and liquidity. The most significant factor is lagged leverage in all countries and its effect on capital structure is consiste nt with the dynamic tradeoff theory alone. The coefficient estimate on the lagged dependent variable lies in zero -one interval in each country and significantly differs from zero and one indicating that the capital structure converges to its optimal level over time. This finding confirms the presence of dynamism in the capital structure decisions of firms from all sample countries; in the sense that firms pro actively alter their leverage to reach their optimal target and that the actual leverage would be an inappropriate proxy for the optimal leverage. The signs of the remaining determinants give also consistent support to the dynamic tradeoff theory. T he relative efficiency and the degree of impact of the leverage factors on capital structure are driven by cross -country

PAGE 14

14 differences in the quality of institutions that affect bankruptcy costs, agency costs, tax benefits of debt; agency costs of equity ; and information asymmetry costs Second I explore the ad justment to optimal leverage in Chapter 3 My analyses indicate that the corporate financing decisions around the world are consistent with dynamic rebalancing of capital structure. The average firm in every country in my sample has a target capital struct ure and adjust s to it in the long run. More im portantly, the large variation in adjustment speeds is driven by the country -specific difference s in the adjustment costs and benefits that are able to explain a large portion of the worldwide frequency and siz e of the capital structure rebalancing, establishing the relevance of the dynamic tradeoff theory. Overall, the dynamic tradeoff theory provides a natural account for the observed empirical regularities in the international data. My results imply that fur ther research is needed to establish an institutional framework that enables systematic comparisons for the differences in the determinants of capital structure across countries. In Chapter 4, I suggest directions for future research. I propose a comprehen sive econometric model to explore how firm, industry, macroeconomic, and institutional arrangements simultaneously affect corporate financial leverage around the world.

PAGE 15

15 CHAPTER 2 THE DETERMINANTS OF CAPITAL STRUCTURE IN AN INTERNATIONAL CONTEXT M otivation and Main Findings The empirical literature on corporate finance has investigated whether there is support for the tradeoff, pecking order, market timing and inertia theories of capital structure in the context of a single country. Recently, a growing body of literature has started to employ cross -country comparisons to tes t theories of capital structure. For instance, Rajan and Zingales (1995) investigate whether the factors identified as correlated with firm leverage in the United States are s imilarly correlated across G 7 counties. Furthermore, Booth, Demirg -Kunt and Maksimovic (2001) assess if capital structure theory is portable to firms in 10 developing countries. Similarly, Anoniou, Guney, and Paudyal (2008) identify the firm -specific an d macroeconomic factors that managers consider in their capital structure choices in two capital market oriented (the United States and the United Kingdom) and three bank oriented (Germany, Japan, and France) economies and measure their speed of adjustment to optimal leverage The main theories of capital structure make precise statements on how leverage relate s to the observable firm attributes If a particular theory is empirically relevant, the direction and the significance of the observed relationships between leverage and its determinants should be consistent with its predictions. By investigating how firm features relate to capital structure across many countries, these studies aim to identify the economic forces underlying leverage factors and revea l more information about the strengths and shortcomings of a particular theory. The comparative empirical relevance of alternative theories is still not established (Frank and Goyal (2007) and Huang and Ritter (2009) ). The main motivation of this chapter is to evaluate which capital structure theories can account for the reliable patterns identified in the international data. The contribution of my current investigation is threefold: First, I go beyond a

PAGE 16

16 static model and take into account the potentially d ynamic nature of the firms capital structure and its unobserved heterogeneity to examin e the relations of several firm, industry, and macroeconomic determinants and leverage in the sample countries for two reasons. The pattern of firm financing decisions might not be stable over time. Furthermore, firms may allow their leverage to drift when exposed to shocks if it is optimal to do so due to binding refinancing costs. In fact, my results indicate that a firms past leverage is the most important leverage factor for all sample countries, confirming that the dynamic model is the right model. Studies that have used a static model have unfortunately ignored the most crucial factor in the capital structure decisions. Second, I assess why each determinant has th e observed relationship with leverage based on the main capital structure theories. Finally, I provide large sample evidence from a panel of firms from 37 countries with different institutional characteristics to increase the power of my analyses and to examine the effect of conditioning on firm circumstances on the relative reliability of leverage determinants. My findings reveal that not all leverage factors are equally reliable because firms are exposed to different circumstances as a consequence of the intrinsic characteristics of their countrys legal and economic structure. Institutional factors change how selected determinants affect leverage. I start by exploring which leverage factors are consistently important for the capital structure decisions of firms around the world. Frank and Goyal (2007) document for U.S. firms that the core factors that have reliable signs and statistical significanc e across many treatments of data are industry median leverage, market to -book assets ratio, tangibility, profits, firm size, and inflation. Similarly, for major industrialized G 7 countries, Rajan and Zingales (1995) report that the dominant factors are th e market to -book assets ratio, tangibility, profits, and firm size. An intriguing question is whether these factors are equally reliable in a large panel of countries for

PAGE 17

17 firms subject to similar institutional features as well as special circumstances tha t arise from differences in the quality of institutions governing their country. The results indicate that among the factors that Frank and Goyal (2007) and Rajan and Zingales (1995) propose, industry median leverage, tangibility and firm size have consi stent signs and statistical significance across a large number of countries, using both book and market definitions of leverage. In addition to these factors emphasized by Frank and Goyal (2007) and Rajan and Zingales (1995), the results highlight past leverage as the most significant factor in all sample countries. R&D Dummy liquidity, and depreciation expenses are some other firm features that have robust and statistically significant influence on leverage around the world. Furthermore, the signs of thes e seven reliable firm -specific determinants, lagged leverage (+ and strictly bounded between 0 and 1), size (+), tangibility (+), R&D Dummy (+), industry median leverage (+), liquidity ( ), and depreciation expenses ( -) give consistent support to the dynam ic tradeoff theory. The effects of the market to -book assets ratio, profitability, inflation, and GDP growth are statistically significant, but their correlations with leverage are different across book and market definitions of leverage. There is additional support to the dynamic tradeoff theory that predicts a negative relationship between inflation and leverage using the book leverage definition. There is some support for the pecking order theory, which predicts a negative relationship between profitabi lity and leverage using the book leverage definition. This finding is also consistent with the dynamic tradeoff theory (Strebulaev (2007)). There is also some support for the mark et timing theory which predicts respectively negative and positive correlati ons with market leverage for the market to -book assets ratio and inflation, similar to the trade -off theory. Overall,

PAGE 18

18 the international empirical evidence I uncover is convincingly consistent with the dynamic tradeoff theory of capital structure. I then tu rn to investigating the implications of institutional differences for the importance and relevance of leverage factors. Several papers (La Porta, Lopez de Silanes, Shleifer, and Vishny (1997, 1998), Demirg -Kunt and Maksimovic (1999, 2002), Booth, Demirg -Kunt and Maksimovic ( 2001), and Bancel and Mittoo (2004) ) find that a firms capital structure is not only influenced by firm -specific features but also by country-specific institutional characteristics Country characteristics (bankruptcy procedures, tax codes, legal protections for minority investors, accounting standards, functionality of the capital and bond markets etc.) influence firms costs and benefits of leverage Countries differ in the quality of institutions that may potentially affect the tradeoff between bankruptcy costs and tax benefits, agency costs, and information asymmetry costs imposed on firms. If the capital structure theories are indeed conditional and not general ( e.g. Myers (2003)), t he correlation of firm -specific, industry specific, and macroeconomic features with leverage may vary across different legal and economic environments. To establish the robustness of the selected dominant factors to special firm circumstances imposed by the country features, I explore how the firm industry, and macroeconomic l evel covariates relate to capital structure in different institutional environments. Specifically, I examine the implications for capital structure of the firm, industry, and macroeconomic attributes sepa rately for countries with strong and weak institutions. Results indicate that the relative efficiency and the degree of impact of leverage factors on capital structure are driven by the differences in the quality of institutions that shape the degree of th e bankruptcy costs agency costs, tax benefits and information asymmetry costs imposed on firms

PAGE 19

19 Overall, the dynamic tradeoff theory provides a natural account for the observed empirical regularities in the international data. I also analyze how institu tions affect the within country correlations between leverage and its determinants. The fact that the core factors are robust and that the direction of their impact is similar across firms from diverse institutional environment s is encouraging. However, so me differences emerge as well. For instance, t he lower efficiency (inconsistent signs across countries ) of the remainder factors is mostly driven by the institutional characteristics of the country the firm is subject to. Furthermore, the magnitude of the impact for the robust factors varies according to the institutional environment the firm operates in. In this study, I uncover the impact of institutions on within country correlations of leverage and its factors. My results emphasize the importance of country features on capital structure decisions. Apart from understanding the impact of institutions on within country correlations of leverage and its factors, it would be beneficial to further understand their effect on the between country differences. F urther re search is needed to establish a comprehensive institutional framework enabling systematic comparisons of the differences in the determinants of capital structure across countries simultaneously controlling for firm industry, macroeconomic, and c ountry features This could allow the researcher to better discriminate among alternative theories and uncover the true forces underlying the leverage determinants. I hope to explore this issue in future work. This issue is further discussed in Chapter 4. Data For the tests of my hypotheses, I acquire firm -level data as well as data on countries institutional characteristics.1 I construct my firm -level sample from all non -financial firms 1 I am grateful to Andrei Shleifer for making several of my proxies freely available on his webpage ( http://www.economics.harvard.edu/faculty/shleifer/dataset ). Since several country attributes are obtained from LLSV (1998), the sample is limited to the subset of the Compustat Global Va ntage that overlaps with LLSV (1998) coverage.

PAGE 20

20 included in the Compustat Global Vantage database between the years 1 991 2006. As my regression specification includes lagged variables, I exclude any firm with fewer than two consecutive years of data from my analyses. In addition, I exclude firm year observations with missing financial and accounting data required for the firm level analysis. To minimize the potential impact of outliers, I winsorize t he firm -level variables at the 1st and 99th percentiles. The final sample consists of 15,177 firms from 37 countries, totaling 105,568 firm -years an average of 7 years per fi rm The country level structural and institutional data comes from different sources. The description of the firm, industry, and macroeconomic variables is reported in Table 2 1. The description of the countrylevel institutional characteristics is provide d in Table 2 2. The information on firm and country level variables that are employed in my analysis is documented in Table 2 3 and Table 2 4. Table 2 5 reports the correlations for the individual institutional indices. Empirical Methodology The major theo ries of capital structure make different predictions on how the theories relate to observable firm characteristics The tradeoff, pecking order, agency theories, and to some extent the market timing theory make specific predictions about the influence of f actors like bankruptcy costs, agency costs, and information asymmetry costs on firms capital structures.2 Several firm, industry, and macroeconomic p roxies have been proposed to account for the relationship between these factors and leverage. In this chapter, I evaluate the impact of the leverage factors on the choice of firms capital structure in many countries around the world. 2 Welch (2004) proposes a managerial inertia theory under which observed changes in market leverage are the result of general movements in equity values rather than specific managerial actions affecting firm d ebt levels. Since Welchs theory makes no specific predictions about the relationship between the firm specific variables and leverage, I dont explicitly incorporate it into the framework of my study. Furthermore, to the degree that Welchs inertia theory is an important determinant of the observed market debt ratios, the speed of adjustment to target book ratios should be much faster than the speed of adjustment to target market values. However, my analyses do not provide any support to this proposition.

PAGE 21

21 The merits of each capital structure theory can be judged by comparing the consistency of the signs on the determinants with it s predictions. I also conduct the same analysis for firms from weak and strong institutional environments separately. It is conceivable that a firm feature be dominant for some firms under certain conditions, yet unimportant for other firms, elsewhere (Mye rs (2003)). I condition the firms circumstances on the institutional setting the firm operates in since a certain theory may be dominant for certain type of institutional settings. I employ a general partial adjustment model that permits each firms opti mal leverage to vary over time and according to its characteristics and that allows the deviations from target leverage not to be offset immediately. It would be unreasonable to impose the assumption that firms always operate at their optimal leverage rati os. Following Flannery and Rangan (2006) and others, I use a partial adjustment model that incorporates rebalancing costs that may slow down the firms adjustment to its optimal leverage. 1= 1 + (2 1 ) where is firm is debt ratio in year t and in country j; is firm is desired debt ratio in country j and in year t; In Equation 2 1, is the adjustment parameter representing the magnitude of adjustment during one period. If equals one, then all the gap between the observed and optimal leverage is closed for that firm that period, leading the actual leverage to equal the o ptimal leverage. In the presence of market frictions, will be less than one and the firm will only be able to close a proportion of the gap between its actual and optimal leverage. Many papers examine the factors that determine the target leverage including Hovakimian, Opler, and Titman (2001), Frank and Goyal (2005, 2007), Flannery and Rangan (2006) and Huang and Ritter (2009) I follow the existing literature for the selection of firm specific factors

PAGE 22

22 affecting target leverage but also incorporat e country -specific macroeconomic factors that are theoretically important in a firms determination of target leverage (Frank and Goyal (2007), Korajczyk and Levy (2003)). Furthermore, Flannery and Rangan (2006), Huang and Ritter (2009) and Lemmon, Roberts and Zender (2008) stress the importance of including firm dummies for an unbiased estimation of firm targets. Accordingly, each firms optimal leverage in a particular country is modeled as a function of both the observed firm characteristics, and the unobserved firm heterogeneity, .3 I model the target leverage ratio allowing for the possibility that it might differ across firms and over time: = 1+ (2 2 ) where and are coefficient vectors to be estimated. 1 is a vector of firm industry, and macroeconomic characteristics related to the costs and benefits of operating with various leverage ratios. Substituting Equation 2 2 into Equation 2 1 and re arranging yields: = ( ) 1+ ( 1 ) 1+ + (2 3 ) Equation 2 3 constitutes a typical partial adjustment model of capital structure. The base adjustment speed can be obtained from the coefficient on the lagged dependent variable 1, by simply subtracting it from one ( 1 ( 1 ) This specification assumes that all s ample firms adjust uniformly at the constant rate, The pecking order, market timing, and inertia theories predict that firms do not make any pro active efforts to adj ust their leverage, 3 Frank and Goyal (2007), in the context of a single country, the U.S., argue that macroeconomic factors like expected inflation are an important factor for leverage determination especially for market leverage, but not as important as firm specific factor s like tangibility or firm size. However, in a cross sectional comparative study of a large panel of countries, it is reasonable to presume that excluding these macroeconomic factors from the estimation equation can cause nontrivial bias in calculating le verage targets. As Frank and Goyal (2007) illustrate, the use of the term spread instead of inflation or the use of the stock market index instead of GDP growth yield very similar results for the target estimation. I only report the results with inflation and GDP growth.

PAGE 23

23 hence is assumed to be trivial, close to zero. In contrast to the predictions of the static tradeoff theory, in the presence of market frictions the lagged leverage is a reliable factor, hence 1 should hold. If managers have targ et debt ratios and make pro active efforts to reach it 0 should hold. Therefore, i f the dynamic tradeoff theory is relevant, should be strictly bounded between zero and one. To test which additional leverage determinants have a robust impact on capital structure according to Equation 2 3, one can conduct the following test: = 0 If the leverage factor in question is reliable, 0 should hold. The dynamic panel model in Equation 2 3 requires instruments for the endogenous transf ormed lagged dependent variable (Baltagi (2001)) and a correction for the short panel bias (Bruno (2005))4. Flannery and Hankins (2007) investigate the efficiency of several dynamic panel estimators in the presence of firm fixed effects and short unbalance d panels5 and conclude that Blundell Bond (1998) generalized method of moments estimation (BB) appear s least sensitive to the panel length and yield s superior estimates of the adjustment speeds. Accordingly, I use two -step system GMM to estimate Equation 2 3. 4 The lagged dependent variable is necessarily correlated with the error term in the presence of firm fixed effects and this correlation does not vanish as the number of the firms in the sample gets larger. The transformation of the estima tion specification by expressing the observations on each firm as deviations from their individual means or first differencing eliminates this source of inconsistency by removing individual firm effects since their mean is time invariant. However, for shor t panels, the transformation induces a negative correlation between the transformed lagged dependent variable and the transformed error term that leads to biased and inconsistent results. This bias decreases with panel length (Nickell (1981)), but can be quite significant even for panels with 30 observations (Judson and Owen (1999)). 5 Flannery and Hankins (2007) do not evaluate the long difference estimator of Hahn, Hausman, and Kuersteiner (2007). Hahn, Hausman, and Kuersteiner (2007) show that the long difference estimator does better than the system GMM estimator of Blundell and Bond (1998) when the true adjustment speed parameter is sufficiently close to zero, in which case the system GMM estimator is upward biased. Unfortunately, the panel length for most countries does not allow me to use the long differencing estimator.

PAGE 24

24 My analysis employs measures of both book leverage and market leverage.6 I define book leverage using the book value of firm assets: = + (2 4 ) Market leverage is defined using market valued instead of book valued assets: = + + (2 5 ) Test Design To examine the relative importance of many firm -specific, industry -specific, and macroeconomic features for the leverage decisions of firms around the world, I analyze which determinants of capital structure are reliably signed and reliably important (statistically significant), for explaining firms leverage choices. To evaluate which capital structure theories are empirically relevant, I assess whether the sign of the particular determinan t is consistent with the predictions of a certain capital structure theory. I perform two types of analyses, SEPARATE and POOLED, to evaluate the impact of the leverage determinants on capital structure around the world. Separate Method For the SEPARATE me thodology, I es timate Equation 2 3 separately for every single country in the sample to obtain an estimate of each countrys capital structure determinants, s and s using both book and market definitions of leverage The estimation results are provided in Table 2 6 and Table 2 7 for book and market leverage respectively. I then compute the number of countries (out of 37 in total) where a particular leverage determinant is of a particular 6 The results for book and market leverage in U.S. context are generally comparable, even though in general researchers have focused more on market valued debt ratios ( e.g. Hovakimian et al. (2001), Fama and French (2002), Welch (2004), and Leary and Roberts (2005) ) Since no other study has ever examined the adjustment speed and the determinants of capital structure for such a large panel of countries, I report the estimates using bot h book and market leverage measures for comparative purposes

PAGE 25

25 sign and statistically significant at the 90% or higher confidence level If the correlations consistently hold for the sample countries, I infer that the firm feature in question is a reliable or a dominant factor for the leverage decisions around the world. As a rule of thumb, I require a leverage factor to be significant with a consistent sign in at least 18 countries Pooled Method World model For the POOLED methodology, I es timate Equation 2 3 as a world model where by I combine the data on all sample countries to obtain an estimate of the overall sam ples capital structure determinants7, a single and a single using both book and market definitions of leverage. The estimation results are provided in Table 2 8, in columns 1 and 2 for book and market leverage, respectively. If a leverage determina nt is of a particular sign and statistically significant at the 90% or higher confidence level in the world model I infer that it is a reliable factor across all firms around the world, and assign a score of 37 to it On the other hand, I assign a score of 0 to the determinants that have ins ignificant coefficient estimate in the world model. As before, I use the simple rule requiring a minimum score of 18 to consider a factor as reliable. Institutional model In assessi ng the effect of institutional characteristics for the empirical relevance of the capital structure theories, I conduct two separate analyses. To conduct my tests, I first classify 7 The portfolio regressions assume slope and error variance homogeneity suggesting that the gains from pooling would outweigh any costs imposed for not taking into account the inherent heterogen eity in the slope estimates. I conducted few statistical tests using the bootstrapping procedure. The null hypothesis that the coefficient on a capital structure determinant is equal across all 37 countries is rejected with 99% confidence. However, the null hypothesis that the coefficient on a capital structure determinant is equal across countries sharing the same institutional characteristics fails to be rejected using conventional significance levels (1%, 5%, 10%). This is consistent with Antoniou et al. (2008)s results illustrating that the capital structure slope estimates vary according to countries legal and financial traditions. I only make inferences about the coefficient estimates using the POOLED methodology where I control for the institutional environment by estimating Equation 23 for weak and strong portfolios separately. The common world model is only investigated to infer the direction of the relationship between leverage and its determinants, and not the magnitude of the coefficient est imates.

PAGE 26

26 my sample countries into two distinct portfolios according to the median va lue of various indices representing the quality of legal, financial economic, and political institutions. If the institutional attribute consists of an indicator variable, I group the sample countries according to the presence or absence of that attribute hence according to whether the indicator variable is equal to zero or one. My goal is to control for the economic and legal environment in exploring the determinants of capital structure around the world. I then estimate Equation 23 separately for each individual institutional characteristic for both strong and weak institution portfolios The estimation results are provided in Table 2 11 for book leverage.8 First, t o examine the robustness of the relative importance of many firm -specific, industry speci fic, and macroeconomic features for the leverage decisions of firms around the world to conditioning on firms circumstances, I analyze which determinants of capital structure are reliably signed and reliably important (statistically significant), for expl aining a firms leverage choices controlling for various institutional characteristics that affect bankruptcy costs, agency costs, tax benefits of debt; agency costs of equity; and information asymmetry costs. If the correlations consistently hold for the sample countries for partitioning of the data regardless of the institutional attribute controlled for, I infer that the firm feature in question is a reliable or a dominant factor for the leverage decisions around the world. On the other hand, institutio nal characteristics may affect the effectiveness of a determinant making it a dominant factor in a particular institutional environment, either weak or strong, and not in the other. As a rule of thumb, I require a leverage factor to be significant with a consistent sign at least 50% of the time. There is a total of 18 different partitioning of the data according to the country characteristics, indicating that the rule of thumb requires at least 9, 9, and 18 consistent and significant signs on 8 I unreported results, I conduct the tests for market leverage using the same methodology and obtain similar conclusions.

PAGE 27

27 a leverage de terminant to consider it reliable for we ak, strong, and all institutional settings respectively. Next, to e valuate the relative empirical importance of the theories under different institutional settings, I also compare the magnitude of the firm, i ndustry, and macroeconomic level determinants of capital structure ( s and s ) in environments with either strong or weak institutions. If an institutional characteristic matter s for capital structure choices, I expect to obtain a o r estimate that is different for the strong institution portfolio and the weak institution portfolio.9 Analysis and Results Leverage Factors and the Capital Structure Theories The leading capital structure theories have different relative emphasis on the determinants of leverage In this section, I aim to a ssess the relative importance of the leverage factors for capital structure decisions by evaluating the ir explanatory power. The relative applicability of the capi tal structure theories around the world can then be established by comparing the signs on the leverage determinants with their predictions. Which capital structure determinants are reliably important? I start by reporting the correlations between the firm, industry, and macroeconomic determinants of capital structure and book and market measures of leverage, using SEPARATE and POOLED methods in Table 2 9. Panel A documents the consistency of the direction of the relationship between leverage and each determ inant by report ing the number of instances out of a 9 I conduct a bootstrapping procedure for testing the equality of the coefficients across two groups. Bootstrapping provide s consistent standard errors from the regression model by resampling the original data, applying the regression model to obtain a sample of coefficient estimates using this new dataset by repeating this procedure a number of times. The empirical p value obta ined by the bootstrapping procedure estimates the probability of obtaining observed differences in coefficient estimates, if the true coefficients are in fact equal using 100 iterations I have experimented with 50, 100, and 250 iterations, and obtained q ualitatively similar results.

PAGE 28

28 total of 37 sample countries that the given determinant of leverage has a particular sign at the 90% confidence level or higher. Panel B evaluates whether or not the leverage determinant is a dominant fac tor by using the simple rule of thumb requiring a minimum score of 18 for each factor Columns 4 and 5 of Table 2 10 summarize the signs observed on each determinant using book and market leverage, respectively A particular sign is assigned to column 4 or 5 as long as the sign implications of the SEPARATE and POOLED methods in Table 2 -9 do not conflict. For instance, for GDP growth, no sign is assigned to book leverage, since the SEPARATE method implies a positive coefficient, whereas the POOLED method yields a negative coefficient. The results on factor selection documented in columns 4 and 5 of Table 2 10 along with the observed signs indicate that lagged leverage, profits, depreciation expenses, firm size, tangibility, liquidity, R&D Dummy, industry m edian leverage, and inflation are dominant factors across all firms around the world, using both definitions of leverage. Market to -book assets ratio and GDP growth, on the other hand, are selected as core factors for firms around the world only using the market value definition of leverage.10 Finally, it is also worthwhile to note that profits and inflation are relatively less reliable factors, with opposite implications across alternative definitions of leverage. The empirical relevance of the capital stru cture theories Table 2 10 evaluates the empirical relevance of capital structure theories for firms around the world. The first three columns summarize the hypothesized predictions of the main theories of capital structure : the dynamic tradeoff theory, the pecking order theory, and the market timing 10 Since the dependent variable using market leverage has the market value of equity in the denominator and market to book assets ratio has the market value of equity in the numerator, a mechanical negative relation also exists between the two variables.

PAGE 29

29 theory, about how leverage relates to the observable firm, industry, and macroeconomic attributes. The tradeoff theory maintains that the capital structure of a firm is the outcome of the tradeoff between the be nefits of debt and the costs of debt. Classic arguments for the tradeoff between the costs and benefits of debt are based on bankruptcy costs tax benefits and agency costs related to asset substitution (Jensen and Meckling (1976)), underinvestment (Myers (1977)) and overinvestment (Jensen (1986), Stulz (1990)). E ach firm has a value -maximizing target leverage ratio, which it strives to reach at all times, indicating that th e speed of adjustment, equals one A growing literature advocates a dynamic version of the tradeoff theory (Frank and Goyal (2005)). The partial adjustment tradeoff model recognizes that capital market imperfections create rebalancing costs that hamper the speed of adjustment to optimal capital structure Accordingly, in contrast to the predictions of the static tradeoff theory, in the presence of market frictions, the lagged leverage is a reliable factor, and the speed of adjustment, is predicted t o be strictly less than one. Furthermore, since managers have target debt ratios and make pro active efforts to reach it, the speed of adjustment, is predicted to be strictly greater than zero. The dynamic tradeoff theory predicts that higher bankruptc y costs will decrease a firms leverage. Accordingly, larger and more profitable firms; firms with fewer growth opportunities and R&D expenses (or trivial /missing R&D expense as captured by R&D Dummy) ; more tangible assets and liquidity; firms operating in regulated industries or industries with higher leverage ; in economies with higher growth and lower inflation that face lower bankruptcy costs should carry more debt in their capital structure. Furthermore, since the tradeoff theory maintains that a higher value of tax shields would lead a firms leverage to increase; higher effective tax

PAGE 30

30 rates and inflation should have a positive impact on leverage, while depreciation expenses, a substitute mechanism for taxes, should have a negative impact. Finally, more profitable firms, firms with more liquid assets and less growth opportunities, and firms operating in a low economic growth environment that face higher agency costs of equity; and larger firms, firms with more tangible assets, fewer growth opportunities, and less liquidity that face lower agency costs of debt11 should carry on more debt. A negative sign on profitability is consistent with a dynamic version of the tradeoff theory (Strebulaev (2007)), since an increase in profitability would imply an increase in future profitability and value of the firm, which would lower the firms leverage12 unless it refinances. A refinancing would only occur if adjustment costs outweigh adjustment benefits. Furthermore, Hennessy and Whited (2005) argue that each dollar of debt issued by high liquidity and/or high profitability firm s would serve to incr ease the distribution to share holders, rather than replacing external equity. Their model shows that the marginal increase in debt (reduction in savings) is more attractive wh en it serves as a replacement for exte rnal equity and is less attrac tive when it finances an increase in distributions to shareholders. Since high liquidity and/or high profitability firms are more likely to be at the latter financing margin, they issue le ss debt. According to the pecking order theory, the adverse selection cost s of issuing risky securities either due to asymmetric information (Myers (1984), Myers and Majluf (1984) ) or managerial optimism (Heaton (2002)), lead to a preference ranking over financing sources by 11 When the firms liquidity is sufficient to allow investment using internal funds only, so that no additional external financing is needed, asset substitution can take place, in the sense that the riskfree liquid funds are r eplaced by risky investment projects. This results in overinvest ment problem where the manager undertakes projects that would not be otherwise pursued from a firm value maximization perspective (Jensen and Meckling (1976)). 12 Fama and French (2002) posit that the negative relationship between profitability and leverage holds theoretically only for book leverage. In empirical regressions, Strebulaev (2007) shows that the value of the coefficients estimates on profitability are very similar for both book and market definitions of leverage. Strebulaev (2007) argues that while for book leverage the result is likely to hold under a broader set of conditions than for market leverage, it is unlikely that the definition of leverage drives his conclusions.

PAGE 31

31 creating a wedge between internal and external financing costs and by increasing the difficulty of issuing securities. In order to minimize adverse selection costs, firms first issue internal funds, followed by debt and then equity. Conversely, firms with excess internally generated funds will tend to retire debt in order to preserve their options to borrow again (Lemmon and Zen der, 2004). In contrast to the tradeoff theory, t he pecking order theory implies no leverage target because l everage simply reflects the past imbalances between internal cash flows and investment opportunities. As a result the speed of adjustment, is predicted to be trivial, close to zero. The p ecking order theory maintains that more internal funds and fewer investment opportunities lead to less debt. This argument imp lies that more profitable firms and firms with more liquidity, fewer growth opportunities and R&D expenses (or trivial/missing R&D expense as captured by R&D Dummy) and firms operating in economies with higher growth should have a lower amount of debt in their capital structure. Moreover, the theory implies that higher adverse selection costs would result in more debt. If smaller firms and firms with fewer tangible assets are more prone to adverse selection costs, they should carry more debt in their capital structure. Alternatively, if adverse selection is about assets in place, tangibility may increase adverse selection costs and result in higher debt (Frank and Goyal (2007)). Similarly, t he impact of adverse selection on larger firms with more assets in place can be more severe, requiring more debt in their capital structure. Therefore, the effect of firm size and tangibility on adverse selection costs is ambiguous. Market timing theory posits that managers issue securities considering the time varying relative costs of issuances for debt and equity (Myers (1984), Graham and Harvey (2001), Hovakimian, Opler, and Titman (2001), Baker and Wurgler (2002) Huang and Ritter (2009) ). Accordingly, similar to the pecking order theory, the firms capital structure merely reflects the

PAGE 32

32 cumulative longlasting impact of the history of attempts to time the market and the optimal leverage ratio is irrelevant implying that the speed of adjustment, is close to zero. Unlike the pecking order theory which assumes semi -strong form market efficiency, the market timing theory does not necessarily impose this assumption. Even though there is a room for the role of information asymmetry costs in determining th e relative cost of issuances for debt and equity according to the market timing theory, other factors like psychological patterns may also play a role. According to the market timing theory, firms reduce their (market) leverage when they prefer to issue e quity securities to exploit equity overpricing opportunities. As long as the market to book assets ratio is a reasonable proxy for stock overpricing opportunities, it should be negatively associated with leverage. Furthermore, since higher economic growth is associated with higher stock valuations, it should also result in less debt. On the other hand, higher expected inflation makes debt issuances cheaper, implying more debt in firms capital structure. Additionally, equities may be undervalued in the pres ence of inflation if investors suffer from inflation illusion (Ritter and Warr (2002)), resulting in higher leverage. The comparison of the observed signs on leverage determinants documented in columns 4 and 5 with the predictions of the capital structure theories summarized in columns 1 to 3 in Table 2 10 yield s consistent support to the dynamic tradeoff theory. Five determinants have signs that are only consistent with the dynamic tradeoff theory: lagged leverage (+ and strictly greater than zero, but str ictly less than one ) 13, depreciation expenses ( -), R&D Dummy (+), industry median 13 The co efficient estimate on the lagged dependent variable lies in zero one interval in each country and significantly differs from zero and one at the 99% confidence level in most countries using both book and market definitions of leverage consistent with the dynamic tradeoff theory. This result indicates that the capital structure converges to its optimal level over time. The static tradeoff theory is refuted in the sample since the speed of adjustment strictly differs from one in all countries. The pecking or der, market timing, and inertia theories are refuted for 34 countries (out of 37 countries in total) as the speed of adjustment strictly differs from zero in all countries, except Argentina, Columbia, Peru, and Portugal. The dynamic tradeoff theory may sti ll dominate in those

PAGE 33

33 leverage (+), and GDP growth (+). T he signs on firm size (+), tangibility (+), profits ( for book and + for market leverage) and liquidity ( ) provide support to both the pecking order and dynamic tradeoff theories. Finally, the signs on market -to -book assets ratio ( ) and inflation (+) give support to both the dynamic tradeoff and market timing theories. Leverage Factors, Capital Structure Theories, and the Effect of Conditioning on the Institutional Setting Implicit in the capital structure theories is the prediction that certain type of firms has a capital structure different from other types of firms in certain institutional settings. By testing the importance of leverage determinants across different institutional environments, I may be able to infer additional information about which capital structure theory applies to what type of firms. I concentrate on three main categories of institutional determinants of capital structure: bankruptcy costs, agency costs of debt and tax benefits; agency costs of equity ; and information asymmetry costs. Insolvency codes and court mechanisms governing default on debt contracts should affect the effectiveness of resolution of systemic and non-systemic financial distress. I use the design of the bankruptcy codes and debt contracts, including the attached creditor rights and the associated enforcement mechanisms governing default on debt contracts as determinants of the bankru ptcy costs. In countries where lenders can easily force repayment, repossess collateral, gain control of the firm, or enforce debt contracts, the ex ante distress costs should be less prominent Furthermore, firms from countries that administer the bankruptcy process in court in a manner that is less time consuming, less costly and more efficient should have lower four countries, as long as their slow adjustment is driven by the tradeoff between their adjustment costs and benefits. This issue is further explored in Chapter 3.

PAGE 34

34 deadweight costs associated with the insolvency process implying lower ex -post financial distress costs. T he agency costs of debt14 could be mitigated by adjusting the properties of the debt contracts, such as including covenants that restrict the distribution of dividends, disposition of assets (Smith and Warner (1979)); issuing secured debt via collateralization of tangible assets (Stulz and Johnson (1985)); issuing convertible debt or debt with warrants (Jensen and Meckling (1976)); shorten ing the maturity of debt (Myers (1977)). The creditors would be able to modify and enforce the debt contracts if and only if they are granted the legal rights. I use the creditor rights attached to debt contracts and the quality of their enforcement to proxy for the degree of agency costs of debt, since well protected creditors will have more power against shareholder expropriation. Debt tax shiel ds play an important role in determining the capital structure (Graham (1996)). I use the effective corporate tax rates to evaluate the effect of the value of tax shields on the adjustment decision .15 The degree of the agency costs of equity16 should depend to a great extent on the rights attached to equity securities and their enforcement. Shareholders would not be granted dividends 14 The use of debt might lead to a moral hazard problem arising from t he divergence of interests between the shareholders and the debtholders. First, shareholders have an incentive to engage in excessive risk taking, asset substitution, given their limited liability (Jensen and Meckling (1976)). Second, shareholders will be reluctant to exercise growth options and even undertake the positive value projects in cases where the increase in risky debt reduces the total market value of the firm, leading to suboptimal investment (Myers, (1977). Third, claims dilution through new borrowings may conflict with the interests of existing creditors. Finally, shareholders may engage in direct wealth transfers from debtholders. 15 The effective corporate tax rate is the rate the company has to pay on marginal income, taking into account fe deral, state, local taxes and all available deductions. DeAngelo and Masulis (1980) find that the nondebt tax shields such as net operating loss carryforwards, depreciation expense, and investment tax credits are substitute mechanisms for the tax benefits of leverage. There are limitations to using the effective tax rate. Even though this rate is calculated factoring some of the non debt tax shields like depreciation, it fails to consider all of them and only firms that exceed their non debt tax shield can benefit from a high tax shield potential. 16 A particular type of agency problem in the capital structure literature is the conflict of interests between the manager and the outside shareholders. This moral hazard problem arises due to the separation of ow nership and

PAGE 35

35 unless they could vote out the directors who do not pay them Equity i nvestors would be reluctant to provide fun ds unless they get their return in exchange. Security laws may exist in countries yet not be effectively enforced. In addition to appropriately designed legal code, there is a need for an efficient enforcement system to implement these rights, or at least to act as a credible threat. I use measures of shareholder rights and the quality of private enforcement to evaluate the impact of agency costs on the capital structure adjustment. The perseverance of agency costs of equity also depends on the existence of internal and external disciplinary and monitoring mechanisms such as the legal rules, the effectiveness of the jurisdictions, and the quality of government that limit managerial discretion. I use the executive quality the strength of law and order the q uality of government and the quality of contract enforcement to proxy for internal and external control mechanisms to correct any conflict between the manager and shareholders. I use several measures to proxy for information asymmetry costs. I use an estimate of a countrys quality of accounting standards to proxy for asymmetric information arising from lack of corporate transparency. T he quality of accounting standards and the quality of disclosure in general decrease the external financing co sts by moderating the degree of information asymmetry among the financial agents (Verrecchia (2001), Amihud and Mendelson (1986), Merton (1987), Lombardo and Pag ano (2002), Lambert, Leuz, and Verrecchia (2007)). In control, leading the manager to seek private benefits rather than maximize firm value and can arise in several different forms. The manager may attempt to maximize the job security rather than the firm value or overdiversify so they can lessen the uncertainty of their human capital investment (Amihud and Lev (1981)). Alternatively, the manager may simply not exert the optimal effort to maximize firm value for the sake of pursuing its own objectives (Jensen and Meckling (1976), Ramakrishnan and T hakor (1984)). The free cashflow problem occurs when the manager wastes firms resources through consumption of perquisites or engages in excessive compensation and excessive growth of the firm beyond its optimal size known as empire building ((Murphy ( 1985), Jensen (1986)). Similarly, the overinvestment problem refers to the inclination of the manager to accept projects with negative present value (Jensen (1986), Stulz (1990)).

PAGE 36

36 addition to accounting standards, I focus on information asymmetry proxies for equity and debt markets separately. For the equity markets, I use the regulation of security laws governing initial public offerings, with a focus on mandatory disclosure liability standards, and public enforcement I also use a measure of the quality of capital market governance with a focus on the insider trading laws in individual exchanges around the world For the debt markets, I use the presence of public credit registries that collect information on credit histories and current indebtedness of various borrowers and share it with lenders. Stiglitz and Weiss (1981) propose that when lenders have knowledge about the borrowers or other lenders to the firm, the moral hazard problem of financing non -viable projects wil l be less prominent. The stylized relationships between the country level institutional characteristics and the bankruptcy costs, agency costs of debt and tax benefits, agency costs of equity and information asymmetry costs are summarized in Table 2 1. Which capital structure determinants are reliably important? I report the correlations between the firm, industry, and macroeconomic determinants of capital structure and book leverage for we ak, strong, and all institutional settings in Table 2 12. As befor e, Panel A documents the consistency of the direction of the relationship between leverage and each determinant this time by report ing the number of instances out of a total of 18 different partitioning that the given determinant of leverage has a particu lar sign at the 90% confidence level or higher. Panel B evaluates whether or not the leverage determinant is a dominant factor by using the simple rule of thumb requiring a minimum score of 9, 9, and 18 for each factor in each category of weak, strong, and all institutions respectively T he columns 4, 5 and 6 of Table 2 13 summarize the signs observed on each determinant using book leverage for we ak, strong, and all institutional environments respectively

PAGE 37

37 The results on factor selection are documented with their respective signs in columns 4 to 6 in Table 2 13 and indicate that lagged leverage depreciation expenses, firm size, tangibility, liquidity, R&D Dummy, and industry median leverage continue to be the core factors across all firms around the world (weak, strong, and all institutional settings) even controlling for the institutional environment The fact that the core factors are mostly robust and that the direction of their impact is similar across firms from diverse institutional environment is encouraging. However, some differences also emerge across weak and strong institutional settings. Profits ( ) is a core factor only for firms in weak institutional settings, whereas market -to book assets ratio ( -), R&D expenses (+) and inflation ( ) ar e core factors only for firms in strong institutional settings. Finally, GDP growth is also selected as a core factor However, the effect of GDP growth on leverage is not robust (+ in weak but in strong) across weak and strong institutional settings. The empirical relevance of the capital structure theories The comparison of the observed signs on leverage determinants documented in column s 4 to 6 in Table 2 13 with the predictions of the capital structure theories continue to yield consistent and unive rsal support to the dynamic tradeoff theory, regardless of the institutional setting for the following determinants: lagged leverage (+ and strictly greater than zero, but strictly lesss than one )17, depreciation expenses ( ), size (+), tangibility (+), liq uidity ( ), R&D Dummy (+), and industry median leverage (+). The effect of GDP (+) on leverage is consistent with dynamic tradeoff theory but is supported only in weak institutional settings. T he predictions 17 The coefficient estimate on the lagged dependent variable lies in zero one interval and significantly differs from zero and one at the 99% confidence level for all institutional settings around the world. This finding is consistent with the dynamic tradeoff theory alone, and indicates that the capital structure converges to its optimal level over time in all sample countries. The static tradeoff theory is refuted in the sample since the speed of adjustment strictly differs from one in all 18 different partitioning of the data based on country characteristics. Similarly, the pec king order, market timing, and inertia theories are refuted, as the speed of adjustment strictly differs from zero in all institutional environments around the world. This point is further explored in Chapter 3.

PAGE 38

38 of dynamic tradeoff and pecking order theories a re supported only in weak institution al environments through profits ( ) but its negative sign also dominate in general around the world. The effect of inflation ( ) in the strong institution portfolio is consistent with dynamic tradeoff theory of capital structure. The predictions of the dynamic tradeoff theory and the market timing theory are only supported in strong institutional settings through market to -book assets ratio ( ). Furthermore, the predictions of the pecking order theory are supported only in strong institutional settings through research and development expenses (+). Finally, the negative sign on GDP growth is observed in strong institutional settings and dominates also in general around the world, consistent with pecking order and market timing theories. Table 2 14, Table 2 15, and Table 2 1 6 further investigate the impact of the institutional features that affect the bankruptcy costs, agency costs, and tax benefits of debt on the firm, industry, and macroeconomic determinants of capital structure and assess the relative empirical importance of the capital structure theories across different institutional settings by comparing the magnitude of the impact of each determinant on leverage acr oss weak and strong institutional environments (wea k minus strong) The results for profits confirm the earlier results in that this factor is more reliable in weak institutional setting s In countries where the bankruptcy procedures are less efficient and more costly, and creditor rights enforcement is weaker, the negative relationship between leverage and profitability is stronger. According to the dynamic trade -off theory, an increase in profitability would imply an increase in future profitability and value of the firm, which would lower its leverage unless it refinances. A refinancing would only occur if adjustment costs outweigh adjustment benefits. The results of Table 2 14 and Table 2 15 are consistent with the predictions of the dynamic tradeoff theory as to the effects of profits. I n Table 2 14, firms in weak

PAGE 39

39 institutional settings that are subject to lower adjustment benefits as implied by the increased (lower) bankruptcy and agenc y costs (tax benefits) of debt do not refinance as frequently compared to firms in the strong institution portfolio. Similarly, in Table 2 15, firms in the strong institution portfolio that are subject to lower agency costs of equity have lower (higher) in centives to refinance by issuing debt (equity). In Table 2 16, firms in weak institutional settings that are subject to higher information asymmetry costs refrain more from external financing with an increase in their retained earnings consistent with the pecking order theory. The results for market -to -book assets also verify the earlier results, in that this factor is more reliable in strong institutional environments. The differences across weak and strong institutional settings are highly significant us ing most institutional proxies when agency costs of equity and information asymmetry costs are considered in Table 2 15 and Table 2 16 respectively. Market to -book assets ratio alleviates the agency costs of equity and leads to a reduction in debt but mor e so in institutional settings that are not severely prone to these costs cons istent with the agency view of the dynamic tradeoff theory. Furthermore, firms from strong institutional settings are presumably able to benefit more from mispricing opportuniti es compared to firms from weak institutional environments leading to bigger reduction in their leverage with a given increase in market to book assets ratio according to the market timing theory Similarly, consistent with previous results documented in Table 2 13, the positive effect of GDP growth on leverage is more pronounced for firms from weaker institutional settings. In Table 2 14, a decrease in bankruptcy costs via higher economic growth is more beneficial in countries with more binding bankruptcy costs, leading to a bigger increase in debt, consistent with the dynamic tradeoff theory. Similarly, the results in Table 2 15 indicate that GDP growth

PAGE 40

40 alleviates the agency costs of equity and lea ds to a reduction in debt, but more so for firms in institutional settings that are not severely prone to these costs, consistent with the agency view of the dynamic tradeoff theory Finally, T able 2 1 6 indicate s that the predictions of the pecking order a nd market timing theories as to the effect of GDP growth are strengthened in strong institutional settings. For instance, firms from strong institutional environments might be able to benefit more from mispricing opportunities compared to firms from weak i nstitutional environments, leading to bigger reduction in their leverage with a given increase in their stock valuations due to higher economic growth, according to the market timing theory. A newer finding from Table 2 14, Table 2 15, and Table 2 16 is th at the effects of the robust factors such as lagged leverage, depreciation expens es, and tangibility differ in their magnitude acr oss weak and strong institutional environments as well In Table 2 1 4, depreciation expenses and tangibility have stronger neg ative and positive coefficient s respectively for firms in weak er institutional settings. In other words, tax benefits of debt are more valued in weak institutional settings subject to high bankruptcy and agency costs of debt so that a substitution of the tax benefits by depreciation expenses lead s to a reduction in debt relatively more compared to strong institutional environments In addition, t he need for collateral is more pronounced in countries with higher distress costs Therefore, th e predictions of the dynamic tradeoff theory are strengthened when bankruptcy costs and agency costs of debt are more severe. Fin ally, in Table 2 14, Table 2 15, and Table 2 1 6, some systematic differences can be observed for the lagged leverage across wea k and strong institutional settings I explore this issue in detail in Chapter 3. Overall, the empirical evidence illustrates that institutions play an important role on the impact of the firm, industry, an d macroeconomic determinants on capital structure and have important implications for the relative empirical relevance of the capital structure theories.

PAGE 41

41 Summary This chapter examines international differences in the firm level determinants of capital structure across 37 countries over the period from 1991 to 2006. First, I examine which characteristics, at the firm, industry, and macroeconomic level, are reliably important for explaining the choice of leverage around the world using a large panel of countries The signs on the five dominant leverage fac tors, lagged leverage (+), depreciation expenses ( ), R&D Dummy (+), industry median leverage (+), and GDP growth (+) give consistent support to the dynamic tradeoff theory. The signs on the remaining six dominant factors give partial support to each theor y. The signs on profits ( for book + for market leverage) firm size (+), tangibility (+), and liquidity ( ) provide support both to the pecking order and dynamic tradeoff theories. Finally, the signs of market to -book assets ratio ( ) and inflation (+) g ive support to both the dynamic tradeoff and market timing theories. Not all factors are equally reliable. The effects of profits, market to book assets ratio, GDP growth, and inflation d epend on how leverage is defined Overall, the international empirical evidence I uncover is convincingly consistent with the dynamic tradeoff theory of capital structure. Second, I examine how firms operating in different institutional settings determine their capita l structure. Specifically, I explore whether the interactions of firm, industry, and macroeconomic level featu res and country characteristics affect capital structure decisions by conducting a systematic comparison of the implications on leverage of the fi rm level attributes across countries with strong and weak institutions. There is a significant impact of countryspecific factors on leverage, through their inf luence on the roles of the firm, industry, and macroeconomic level factors determining capital s tructure. The correlations of some leverage factors (e.g. profits, market to -book assets, inflation, and GDP growth ) have different implications on firms from dissimilar institutional settings. Furthermore, t he effects of the

PAGE 42

42 dominant reliable factors (e.g lagged leverage, tangibility, and firm size) are either strengthened or moderated by country specific factors These result s indicate that further research is needed to understand the true economic forces underlying the determinants of leverage across co untries and assess the relative empirical relevance of the main capital structure theories .

PAGE 43

43 Table 2 1. Description and Sources of Firm, Industry, and Macroeconomic Determinants of Capital Structure Variable Description and Source BLEV Book Leverage. Ratio of book value of total debt to book value of total assets. (Long term debt[106]+Short term debt[94])/Total assets[89] Source: Compustat Global Vantage. MLEV Market Leverage. Ratio of book value of total debt to book value of assets minus book value of equity plus market value of equity (Long term debt[106]+Short term debt[94])/(Total assets[89] Book equity[135]+Market Value of Equity[PRCCI *SHOI]. Source: Compustat Global Vantage. PROFIT Earnings before interest and taxes as a proportion of total assets. (Operating Income[14]+ Interest and related expense[15]+Current income taxes[24])/Total assets[89] Source: Compustat Global Vantage MB The ratio of assets market to book values. (Long term debt[106]+Short term debt[94]+Preferred capital[119]+Market Value of Equity[PRCCI*SHOI])/Total assets[89]. Source: Compustat Global Vantage. DEP_TA Depreciation expense as a proportion of total assets Total Depreciation and Amortization[11]/To tal assets[89]. Source: Compustat Global Vantage. LnTA Log of total book assets (a measure of firm size). Log[89]. Source: Compustat Global Vantage. FA_TA Fixed assets as a proportion of total assets. Fixed assets[76]/Total assets[89]. Source: Compustat Global Vantage. R&D_TA Research and development expenses as a proportion of total assets. Firms with more intangible assets in the form of R&D expenses will prefer to have more equity. Research and Development Expense [52]/Total assets[89]. Sour ce: Compustat Global Vantage. RD_DUM A dummy variable equal to one if R&D expenditures are not reported; otherwise zero. About 65% of the sample firm years do not report R&D expenses. For these firms, I set R&D expense to zero and set R& D_DUM equal to one. Source: Compustat Global Vantage IND_MED The prior years median leverage ratio for the firms industry. Industry classifications are based on the 48 industry categories in Fama and French (1997 Source: Compustat Global Vantage TAXES Ratio of total income taxes to pre tax income Current income taxes[24]/Income before income taxes[21]. Source: Compustat Global Vantage. LIQUID Total current assets as a proportion of total current liabilities. Total Current Assets[75]/Total current liabilities[104]. Source: Compustat Global Vantage. REG A dummy variable equal to one for firms operating in regulated industries A dummy variable equal to one for firms operating in regulated industries. SIC [49004999]. Source: Compustat Global Vantage. INF Annual Inflation rate. Growth in CPI. Source:WDI (World Development Indicators). GDPG Annual GDP growth. Growth in nominal GDP. Source:WDI (World Development Indicators).

PAGE 44

44 Table 2 2. Description and Sources of Institutional Characteristics Influencing Firm, Industry, and Macroeconomic Determinants of Capital Structure Variables Des cription Predicted Sign BANKRUPTCY COSTS, AGENCY COSTS OF DEBT AND TAX BENEFITS BANKRUPTCY COSTS TIME Time to resolve Time to resolve the insolvency process. Source: Djankov, Hart, McLiesh and Shleifer (2008). + COST Cost of Bankruptcy The costs to complete the insolvency proceeding, expressed as a percentage of the bankruptcy estate at the time of entry to the bankruptcy. Source: Djankov, Hart, McLiesh and Shleifer (2008). + EFFICIENCY Efficiency of Bankruptcy A dummy variable for whether the bankruptcy outcome is efficient. Source: Djankov, Hart, McLiesh and Shleifer (2008). TAX BENEFITS TAX 1st Year Effective Tax Rate (%) The tax rate obtained by dividing the total corporate tax TaxpayerCo pays by its pretax earnings. Source: Djankov et al. (2008) + AGENCY COSTS OF DEBT CREDITOR Creditor rights An index aggregating creditor rights. A score of one is assigned when each of the following rights of secured lenders are defined in laws and regulations: First, there are restrictions, such as creditor consent or minimum dividends, for a debtor to file for reorganization. Second, secured creditors are able to seize their collateral after the reorganization petition is approved, i.e. there is no "automatic stay" or "asset freeze." Third, secured creditors are paid first out of the proceeds of liquidating a bankrupt firm, as opposed to other creditors such as government or workers. Finally, if management does not retain administration of its property pending the resolution of the reorganization. The index ranges from 0 (weak creditor rights) to 4 (stron g creditor rights). Source: La Porta et al. ( 1998) FORMALISM Debt Enforcement The index measures substantive and procedural statutory intervention in judicial cases at lower level civil trial courts, and is formed by adding up the following indices: (i) professionals vs. laymen, (ii) written vs. oral elements, (iii) legal justification, (iv) statutory regulation of evidence, (v) control of superior review, (vi) engagement formalities, and (vii) independent procedural actions. The index ranges from 0 t o 7, where 7 means a higher level of control or intervention in the judicial process. Source: Djankov et al. (2003) + AGENCY COSTS OF EQUITY ANTIDIR Shareholder rights This index of anti director rights is formed by adding one when: (1) the country allows shareholders to mail their proxy vote; (2) shareholders are not required to deposit their shares prior to the General Shareholders Meeting; (3) cumulative voting or pr oportional representation of minorities on the board of directors is allowed; (4) an oppressed minorities mechanism is in place; (5) the minimum percentage of share capital that entitles a shareholder to call for an Extraordinary Shareholders Meeting is l ess than or equal to 10% (the sample median); or (6) when shareholders have preemptive rights that can only be waived by a shareholders meeting. The range for the index is from 0 to 5. Source: La Porta et al. (1998). PRENF Equity enforcement Average of ex ante and ex post private control of self dealing. Source: Djankov et al. (2008) EXECUTIVE Executive Quality Index of constraints on the executive power based on the number of effective veto points in a country. Veto points include (1) an effective legislature (represents two veto points in the case of bicameral systems); (2) an independent judiciary; and (3) a s trong federal system. Average of the years 1945 through 1998. Source: Henisz (2001), and Djankov et al. (2002) LAW&ORDER Law and Order Integrity of legal system in 2000. This component is based on the Political Risk Component 1 (Law and Order) from the PRS Groups International Country Risk Guide (various issues). Rankings are modified to a 10 point scale. Source: Economic Freedom of the World (2002). Djankov et al. (2003) ENFORCE Enforceability of contracts The relative degree to which contractual agreements are honored and complications presented by language and mentality differences. Scale for 0 to 10, with higher scores indicating higher enforceability. Source: Business Environmental Risk Intelligence. Exact definition in Knack, Stephen and Phi lip Keefer, 1995. Djankov et al. (2003)

PAGE 45

45 Table 2 2. Continued Variables Des cription Predicted Sign GOVERNANCE Quality of government Principal Components of corruption in government, risk of expropriation and repudiation of contracts by government. GOVERNANCE = 0.92 CORRUPT + 0.95 EXPROPR + 0.95 REPUDR Corruption of the government has low ratings to indicate that high government officials are likely to demand special payments and illegal payments are generally expected throughout lower levels of government in the form of bribes connected with import and export licenses, exchange controls, tax assessment, policy protection, or loans. Scale from 0 to 10. Average of the months of April and October in the monthly index between 1982 and 1995. Source : International Country Risk Guide (ICRG). La Porta et al. (1999) Risk of expropriation is the ICRs assessment of the risk of outright confiscation or forced nationalization. Average of the months of April and October of the monthly index between 1982 and 1995. Scale from 0 to 10, with lower scores for higher risks. Source: La Porta et al. ( 1998 ) Repudiation of contracts by government is ICRs assessment of the risk of a modification in a contract taking the form of repudiation, postponement, or scaling down due to budget cutbacks, indigenization pressure, a chang e in government, or a change in government economic and social priorities. Average of the months of April and October of the monthly index between 1982 and 1995. Scale from 0 to 10, with lower scores for higher risks. Source: La Porta et al. ( 1998 ) INFORMATION ASYMMETRY COSTS ACCSTDS Transparency Index created by examining and rating companies 1990 annual reports on their inclusion or omission of 90 items. These items fall into seven categories (general information, income statements, balance sheets, funds flow statement, accounting standards, stock data and special items). Source: La Porta et al. ( 1998 ) International accounting and auditing trends, Center for International Financial Analysis and Research. EDISCL Disclosure requirements The index of disclosure equals the arithmetic mean of (1) prospectus; (2) compensation; (3) shareholders; (4) inside ownership; (5) contracts irregular; and (6) transactions. Source: La Porta et al. (2006) ELIABS Liability Standards The index of liability standards equals the arithmetic mean of (1) liability standard for the issuer and its directors; (2) liability standard for distributors; and (3) liability standard for accountants. Source: La Porta et al. (2006) EPUBENF Securities Market Enforcement The index of public enforcement equals the arithmetic mean of (1) supervisor characteristics index; (2) rule making power index; (3) investigative powers index; (4) orders index; and (5) criminal index. Source: La Porta et al. (2006) CMG Insider Trading Prevalence of Insider Trading (1=pervasive ; 7=extremely rare). Source: Schwab, Klaus et al., eds., 1999, The Global Competitiveness Report 1999, (Oxford University Press, New York, NY). Source: La Porta et al. (2006) PUBBUREAU Information Sharing The variable equals 1 if a public credit registry operates in the country, 0 otherwise. A public registry is defined as a database owned by public authorities (usually the Central Bank or Banking Supervisory Authority), that collects information on the standing of borrowers in the financial syst em and makes it available to financial institutions. The variable is constructed as at January for every year from 1978 to 2003. Source: Djankov et al. (2007) The table summarizes the relationships between the country level institutional characteristics and the bankruptcy costs, agency costs of debt and tax benefits; agency costs of equity; and information asymmetry costs.

PAGE 46

46 Table 2 3. Compustat Global Sample and Macroeconomic Variables Country BLEV MLEV PROFIT MB DEP_TA LnTA FA_TA RD_TA TAXES PROFIT LIQUID INF GDPG Argentina 0.35 0.39 0.06 0.52 0.05 7.28 0.60 0.00 0.03 0.08 0.97 0.05 0.04 Australia 0.14 0.17 0.04 0.49 0.04 4.11 0.28 0.00 0.04 0.05 1.55 0.03 0.03 Austria 0.25 0.15 0.05 0.85 0.05 6.92 0.35 0.00 0.13 0.07 1.48 0.02 0.02 Belgium 0.25 0.16 0.06 0.80 0.05 7.32 0.28 0.00 0.19 0.07 1.35 0.02 0.02 Brazil 0.27 0.16 0.08 12.11 0.04 7.96 0.49 0.00 0.11 0.05 1.22 0.09 0.03 Canada 0.18 0.23 0.00 0.43 0.06 4.91 0.54 0.00 0.00 0.03 1.55 0.02 0.03 Switzerland 0.24 0.29 0.07 0.20 0.04 6.48 0.33 0.00 0.18 0.08 1.76 0.01 0.01 Chile 0.22 0.39 0.07 0.24 0.03 11.51 0.55 0.00 0.06 0.08 1.54 0.04 0.03 Columbia 0.12 0.27 0.07 0.08 0.03 13.18 0.47 0.00 0.09 0.05 1.49 0.10 0.03 Germany 0.16 0.15 0.06 0.34 0.06 5.64 0.24 0.00 0.19 0.07 1.80 0.02 0.01 Denmark 0.27 0.31 0.08 0.26 0.05 7.11 0.34 0.00 0.22 0.08 1.61 0.02 0.02 Spain 0.24 0.14 0.07 1.99 0.04 8.98 0.38 0.00 0.19 0.08 1.22 0.03 0.03 Finland 0.26 0.23 0.08 0.43 0.05 6.34 0.30 0.00 0.21 0.08 1.51 0.02 0.03 France 0.23 0.19 0.07 0.37 0.04 6.35 0.18 0.00 0.24 0.08 1.35 0.02 0.02 U nited K ingdom 0.17 0.16 0.10 0.45 0.04 4.48 0.28 0.00 0.20 0.08 1.37 0.03 0.03 Greece 0.30 0.25 0.08 0.44 0.04 7.78 0.34 0.00 0.20 0.08 1.45 0.04 0.04 Hong Kong 0.17 0.25 0.04 0.28 0.02 7.64 0.30 0.00 0.08 0.03 1.52 0.02 0.05 Indonesia 0.39 0.41 0.07 0.22 0.04 13.23 0.41 0.00 0.06 0.05 1.34 0.10 0.04 India 0.28 0.32 0.11 0.32 0.03 9.19 0.39 0.00 0.13 0.09 1.50 0.05 0.06 Ireland 0.26 0.30 0.09 0.30 0.04 5.32 0.34 0.00 0.15 0.08 1.50 0.03 0.06 Israel 0.25 0.26 0.05 0.30 0.03 6.72 0.22 0.01 0.11 0.05 1.56 0.04 0.03 Italy 0.25 0.16 0.06 0.47 0.04 9.25 0.24 0.00 0.20 0.06 1.37 0.03 0.01 Japan 0.23 0.16 0.05 0.76 0.03 10.49 0.30 0.00 0.43 0.05 1.34 0.00 0.01 S. Korea 0.27 0.36 0.07 0.19 0.04 13.86 0.40 0.00 0.06 0.07 1.18 0.03 0.05 Mexico 0.26 0.16 0.09 3.30 0.03 9.37 0.56 0.00 0.12 0.08 1.42 0.08 0.03 Malaysia 0.22 0.23 0.05 0.38 0.03 5.63 0.38 0.00 0.15 0.05 1.45 0.02 0.06 Norway 0.28 0.28 0.06 0.30 0.05 7.25 0.30 0.00 0.05 0.07 1.59 0.02 0.03 New Zealand 0.27 0.32 0.10 0.41 0.04 5.51 0.41 0.00 0.12 0.09 1.42 0.02 0.03 Pakistan 0.24 0.30 0.12 0.27 0.04 8.85 0.48 0.00 0.19 0.10 1.17 0.05 0.04 Peru 0.23 0.37 0.11 0.64 0.04 6.67 0.47 0.00 0.11 0.11 1.37 0.03 0.04 Philippines 0.25 0.31 0.04 0.18 0.03 8.60 0.44 0.00 0.05 0.03 1.14 0.06 0.04 Portugal 0.34 0.17 0.05 10.09 0.05 8.61 0.40 0.00 0.15 0.07 1.12 0.03 0.02 Singapore 0.19 0.20 0.04 0.48 0.03 5.06 0.32 0.00 0.16 0.05 1.47 0.01 0.06 Thailand 0.33 0.30 0.06 0.51 0.04 7.73 0.44 0.00 0.04 0.07 1.22 0.03 0.05 Turkey 0.17 0.21 0.12 6.28 0.06 12.89 0.36 0.00 0.07 0.11 1.64 0.13 0.05 U nited S tates 0.23 0.23 0.09 0.39 0.04 6.00 0.26 0.00 0.25 0.08 1.86 0.03 0.03 South Africa 0.13 0.16 0.13 0.34 0.04 7.63 0.27 0.00 0.10 0.11 1.42 0.06 0.04 Each year, the median value of each variable is calculated for each country. The reported statistic is the mean of these timeseries medians. Table 2 1 provides the description of variables and their sources.

PAGE 47

47 Table 2 4. Ins titutional Characteristics TIME COST EFFICIENCY TAX CREDITOR FORMALISM ANTIDIR PRENF EXECUTIVE ENFORCE Argentina 2.75 0.12 35.8 0 23.54 1 5.4 0 4 0.34 3.76 5.04 Australia 0.58 0.08 87.8 0 21.96 1 1.8 0 4 0.76 7 .00 7.71 Austria 0.92 0.18 78.00 20.86 3 3.52 2 0.21 7 .00 8.25 Belgium 0.92 0.04 90.80 16.71 2 2.73 0 0.54 7.00 8.16 Brazil 3.67 0.12 13.4 0 15.49 1 3.06 3 0.27 3.37 5.3 0 Canada 0.75 0.04 93.2 0 21.78 1 2.09 5 0.64 7 .00 8.38 Switzerland 3 .00 0.04 60.4 0 13.74 1 3.13 2 0.27 7 .00 8.94 Chile 5.08 0.15 40.9 0 15.09 2 4.57 5 0.63 4.02 5.2 0 Columbia 3 .00 0.01 64.8 0 24.28 0 4.11 3 0.57 6.1 0 4.76 Germany 0.92 0.08 57.00 23.5 3 3.51 1 0.28 7 .00 8.4 0 Denmark 2.5 0 0.09 76.7 0 21.94 3 2.55 2 0.46 7 .00 8.22 Spain 1 .00 0.15 82 .00 18.52 2 5.25 4 0.37 4.69 6.23 Finland 0.92 0.04 92.4 0 16.3 0 1 3.14 3 0.46 7 .00 7.5 0 France 1.69 0.09 54.1 0 14.06 0 3.23 3 0.38 4.95 6.36 United Kingdom 0.5 0 0.06 92.3 0 18.61 4 2.58 5 0.95 7 .00 8.5 0 Greece 1.92 0.09 53.8 0 19.78 1 3.99 2 0.22 5.44 5.81 Hong Kong 0.63 0.09 88.3 0 0 .00 4 0.73 5 0.96 Indonesia 5.5 0 0.18 25.1 0 20.84 4 3.9 0 2 0.65 2.48 4.29 India 20.28 4 3.34 5 0.58 6.95 4.53 Ireland 0.42 0.09 89.9 0 9.62 1 2.63 4 0.79 7 .00 7.78 Israel 1.5 0 0.23 66.2 0 25.72 4 3.3 0 3 0.73 7 .00 7.3 0 Italy 1.17 0.22 45.3 0 23.82 2 4.04 1 0.42 7 .00 5.18 Japan 0.58 0.04 95.5 0 28.66 2 2.98 4 0.5 0 7 .00 7.57 South Korea 1.5 0 0.04 88.1 0 14.94 3 3.37 2 0.47 5.52 Mexico 1.83 0.18 72.6 0 22.21 0 4.71 1 0.17 3.52 4.92 Malaysia 2.25 0.15 48.4 0 10.5 4 2.34 4 0.95 5.17 5.71 Norway 0.92 0.01 91.8 0 18.5 2 2.95 4 0.42 7 .00 8.48 New Zealand 0.67 0.04 90.7 0 26.44 3 1.58 4 0.95 7 .00 Pakistan 31.28 4 3.76 5 0.41 4.24 3.85 Peru 3.08 0.07 41.8 0 22.03 0 5.6 0 3 0.45 3.87 4.29 Philippines 5.67 0.38 17.5 0 22.08 0 5 .00 3 0.22 4.1 0 4.84 Portugal 2 .00 0.09 82.3 0 16.03 1 3.93 3 0.44 4.75 4.54 Singapore 0.58 0.01 96.1 0 10.25 4 2.5 0 4 1 .00 3.3 0 7.64 Thailand 2.67 0.36 54.9 0 22.04 3 3.14 2 0.81 3.46 5.61 Turkey 5.88 0.07 6.6 0 16.92 2 2.53 2 0.43 6.07 4.79 United States 2 .00 0.07 85.8 0 18.19 1 2.62 5 0.65 7 .00 8.73 South Africa 1.92 0.18 39.8 0 18.1 0 3 1.68 5 0.81 7 .00 6.87

PAGE 48

48 Table 2 4 Continued LAW& ORDER CORRUPT EXPROPR REPUDR ACCSTDS EDISCL ELIABS EPUBLICENF CMG PUBBUREAU Argentina 8.33 6.01 5.91 4.91 45 0.5 0 0.22 0.58 3.5 1 Australia 10 .00 8.51 9.27 8.71 75 0.75 0.66 0.9 0 5.7 1 Austria 10 .00 8.57 9.69 9.6 0 54 0.25 0.11 0.17 5.5 1 Belgium 8.33 8.81 9.63 9.48 61 0.42 0.44 0.15 5.1 1 Brazil 3.33 6.31 7.62 6.3 0 54 0.25 0.33 0.58 4 .0 1 Canada 10 .00 10 .00 9.67 8.96 74 0.92 1 .00 0.8 0 5.2 1 Switzerland 10 .00 10 .00 9.98 9.98 68 0.67 0.44 0.33 5.3 1 Chile 8.33 5.3 0 7.5 0 6.8 0 52 0.58 0.33 0.6 0 4.3 1 Columbia 1.67 5 .00 6.95 7.02 50 0.42 0.11 0.58 4 .0 1 Germany 8.33 8.93 9.9 0 9.77 62 0.42 0 .00 0.22 4.9 1 Denmark 10 .00 10 .00 9.67 9.31 62 0.58 0.55 0.37 5.5 1 Spain 6.67 7.38 9.52 8.4 0 64 0.5 0 0.66 0.33 4.1 1 Finland 10 .00 10 .00 9.67 9.15 77 0.5 0 0.66 0.32 5.5 1 France 8.33 9.05 9.65 9.19 69 0.75 0.22 0.77 5.1 0 United Kingdom 10 .00 9.11 9.71 9.63 78 0.83 0.66 0.68 6.2 1 Greece 5 .00 7.26 7.12 6.62 55 0.33 0.5 0 0.32 3.2 1 Hong Kong 8.33 8.52 8.29 8.82 69 0.92 0.66 0.87 4.4 1 Indonesia 3.33 2.14 7.16 6.09 0.5 0 0.66 0.62 2.8 0 India 6.67 4.58 7.75 6.11 57 0.92 0.66 0.67 3.5 0 Ireland 10 .00 8.51 9.67 8.96 0.67 0.44 0.37 5.4 1 Israel 8.33 8.33 8.25 7.54 64 0.67 0.66 0.63 4.9 1 Italy 10 .00 6.13 9.35 9.17 62 0.67 0.22 0.48 4.2 1 Japan 8.33 8.51 9.67 9.69 65 0.75 0.66 0 .00 5.1 1 South Korea 5.3 0 8.31 8.59 62 0.75 0.66 0.25 4.4 1 Mexico 3.33 4.76 7.29 6.55 60 0.58 0.11 0.35 3.8 1 Malaysia 5 .00 7.38 7.95 7.43 76 0.92 0.66 0.77 4.4 1 Norway 10 .00 10 .00 9.88 9.71 74 0.58 0.39 0.32 4.1 1 New Zealand 10 .00 10 .00 9.69 9.29 70 0.67 0.44 0.33 5.6 1 Pakistan 5 .00 2.98 5.62 4.87 0.58 0.39 0.58 0 Peru 5 .00 4.7 0 5.54 4.68 38 0.33 0.66 0.78 3.5 1 Philippines 3.33 2.92 5.22 4.8 0 65 0.83 1 .00 0.83 2.9 1 Portugal 8.33 7.38 8.9 0 8.57 36 0.42 0.66 0.58 4.9 1 Singapore 10 .00 8.21 9.3 0 8.86 78 1 .00 0.66 0.87 5.5 0 Thailand 8.33 5.18 7.42 7.57 64 0.92 0.22 0.72 3.3 0 Turkey 6.67 5.18 7 .00 5.95 51 0.5 0 0.22 0.63 3.8 0 United States 10 .00 8.63 9.98 9 .00 71 1 .00 1 .00 0.9 5.5 1 South Africa 3.33 8.91 6.88 7.27 70 0.83 0.66 0.25 4.3 1 Each year, the median value of each variable is calculated for each country. The reported statistic is the mean of these time series medians. Table 2 2 provides the description of variable s and their sources.

PAGE 49

49 Table 2 5 Correlations for the Individual Institutional Indices COMMON MARKET ANTIDIR PRENF CREDITOR FORMALISM ACCSTDS EDISCL ELIABS EPUBLICENF CMG PUBBUREAU MDIV RESERVE TIME COST EFFICIENCY TAX EXECUTIVE LAW&ORDER ENFORCE CORRUPT EXPROPR REPUDR COMMON 1 MARKET 0.5541 1 ANTIDIR 0.5103 0.4022 1 PRENF 0.8444 0.4538 0.5268 1 CREDITOR 0.5019 0.0419 0.1053 0.5933 1 FORMALISM 0.5958 0.2163 0.1924 0.555 0.4569 1 ACCSTDS 0.5726 0.1885 0.3057 0.4799 0.3745 0.6325 1 EDISCL 0.7485 0.4738 0.4878 0.6691 0.289 0.4592 0.7018 1 ELIABS 0.478 0.3272 0.5694 0.3767 0.009 0.2312 0.3716 0.5162 1 EPUBLICENF 0.5213 0.6178 0.4122 0.447 0.1237 0.0452 0.1018 0.4886 0.3709 1 CMG 0.3273 0.132 0.1923 0.3664 0.3568 0.6003 0.4875 0.2437 0.2259 0.0477 1 PUBBUREAU 0.1712 0.2228 0.0927 0.2422 0.1384 0.2067 0.1067 0.3062 0.2291 0.3469 0.0702 1 MDIV 0.2501 0.0139 0.0242 0.1578 0.2759 0.1653 0.3367 0.4285 0.2684 0.004 0.3121 0.1499 1 RESERVE 0.639 0.4507 0.3314 0.4974 0.3856 0.3798 0.512 0.5074 0.4654 0.3722 0.2973 0.1274 0.4874 1 TIME 0.2964 0.3436 0.0425 0.2441 0.3087 0.3273 0.4654 0.136 0.0915 0.2851 0.5918 0.1714 0.3377 0.2729 1 COST 0.1188 0.2321 0.1591 0.0731 0.0946 0.2936 0.0774 0.1787 0.0171 0.1536 0.5005 0.0772 0.1219 0.265 0.3529 1 EFFICIENCY 0.2544 0.2726 0.1571 0.291 0.1944 0.309 0.4418 0.2195 0.285 0.193 0.6862 0.1695 0.3207 0.1672 0.8131 0.4636 1 TAX 0.1225 0.2492 0.1612 0.248 0.1421 0.2731 0.2369 0.1492 0.1029 0.2418 0.1933 0.3258 0.034 0.1395 0.1215 0.239 0.0474 1 EXECUTIVE 0.1244 0.3807 0.0258 0.0879 0.2422 0.5451 0.3844 0.0582 0.1343 0.3784 0.588 0.3475 0.2572 0.0625 0.4381 0.3448 0.4225 0.2467 1 LAW&ORDER 0.1839 0.257 0.0786 0.2013 0.3457 0.3384 0.3587 0.2494 0.1159 0.0083 0.6788 0.1102 0.5231 0.2426 0.4757 0.2698 0.5644 0.066 0.4454 1 ENFORCE 0.3215 0.1974 0.1213 0.2104 0.4008 0.6121 0.655 0.2808 0.2361 0.2557 0.7677 0.1405 0.3571 0.2533 0.6025 0.3959 0.6555 0.0494 0.6974 0.662 1 CORRUPT 0.2388 0.367 0.1619 0.2017 0.3525 0.6413 0.5564 0.1592 0.2078 0.3098 0.7846 0.1311 0.3208 0.1873 0.6855 0.5432 0.6521 0.1878 0.6877 0.6433 0.871 1 EXPROPR 0.111 0.3807 0.0139 0.1213 0.3146 0.5047 0.521 0.1604 0.0623 0.273 0.8158 0.0346 0.3125 0.1173 0.6902 0.4546 0.7274 0.1605 0.6043 0.7443 0.7897 0.8248 1 REPUDR 0.1278 0.4125 0.0544 0.1656 0.3611 0.5403 0.53 0.1958 0.0154 0.3502 0.7952 0.0317 0.3241 0.0811 0.7071 0.4011 0.7338 0.1084 0.6585 0.7054 0.7998 0.8215 0.9677 1 The table illustrates the correlation for the individual indices. Table 2 2 provides the description of variables and their sources.

PAGE 50

50 Table 2 6 Separate Country Regressions Using Book Leverage as the Dependent Variable A rgentina A ustralia A ustria B elgium Brazil Canada Switzerland C hile Columbia Germany Denmark Spain Finland BLEV 0.8719*** 0.6171*** 0.8090*** 0.8492*** 0.8671*** 0.7069*** 0.8453*** 0.7866*** 0.9597*** 0.7413*** 0.7399*** 0.7989*** 0.7657*** (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) EBIT_TA 0.2992*** 0.0037 0.1217*** 0.1788*** 0.0943*** 0.0697*** 0.0310*** 0.0232 0.2202 0.0155 0.0597*** 0.0394*** 0.0908*** (0.0000) (0.4123) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.1683) (0.2506) (0.2462) (0.0000) (0.0000) (0.0000) MB 0.0027 0.0028*** 0.0005* 0.0005*** 0 .0000 0.0035*** 0.0060*** 0.0037*** 0.0355 0.0013 0.0052*** 0.0002*** 0.0078*** (0.9179) (0.0031) (0.0645) (0.0001) (0.6693) (0.0000) (0.0000) (0.0000) (0.5795) (0.4192) (0.0000) (0.0000) (0.0000) DEP_TA 0.2461 0.1126*** 0.3481*** 0.3110*** 0.0141 0.5403*** 0.3655*** 1.2554*** 0.0167 0.0471 0.5952*** 0.2172*** 0.1059** (0.5852) (0.0022) (0.0000) (0.0000) (0.6193) (0.0000) (0.0000) (0.0000) (0.9861) (0.3284) (0.0000) (0.0000) (0.0255) LnTA 0.0004 0.0122*** 0.0137*** 0.0035*** 0.0092*** 0.0248*** 0.0056*** 0.0096*** 0.0296 0.0002 0.0042*** 0.0018*** 0.0041*** (0.9908) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.3266) (0.8776) (0.0000) (0.0000) (0.0000) FA_TA 0.1828 0.0202* 0.0697*** 0.0969*** 0.1296*** 0.0531*** 0.0767*** 0.0354*** 0.0846 0.0323 0.2896*** 0.0454*** 0.0769*** (0.4108) (0.0616) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.3632) (0.1196) (0.0000) (0.0000) (0.0000) TAXES 0.0809*** 0.0152*** 0.0008 0.0042* 0.0095*** 0.0301*** 0.0156*** 0.0087* 0.0572* 0.0006 0.0077*** 0.0011** 0.0057*** (0.0034) (0.0000) (0.4957) (0.0654) (0.0000) (0.0000) (0.0000) (0.0784) (0.0746) (0.8224) (0.0000) (0.0286) (0.0005) LIQUID 0.008 0.0002 0.0039*** 0.0026** 0.0261*** 0.0012*** 0.0033*** 0.0076*** 0.0163 0.0015** 0.0007*** 0.0103*** 0.0038*** (0.4660) (0.7582) (0.0000) (0.0119) (0.0000) (0.0000) (0.0000) (0.0000) (0.4111) (0.0397) (0.0004) (0.0000) (0.0009) RD_DUM 0.0305 0.0233*** 0.0093*** 0.0440*** 0.0528*** 0.0157*** 0.0240*** 0.0141*** 0.4488 0.0146*** 0.0736*** 0.0170*** 0.0093*** (0.8542) (0.0000) (0.0065) (0.0000) (0.0000) (0.0011) (0.0000) (0.0000) (0.2694) (0.0002) (0.0000) (0.0000) (0.0000) R&D_TA 7.7175 0.2662*** 0.1926* 0.3809*** 2.4349*** 0.4286*** 0.5474*** 3.7656*** 0.0772 0.0945*** 0.5956*** 0.7682*** (0.6739) (0.0000) (0.0885) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.1302) (0.0000) (0.0000) (0.0000) IND_MEDIAN 0.0157 0.0604*** 0.1581*** 0.0306*** 0.0247*** 0.3801*** 0.0907*** 0.0037 0.0676 0.0003 0.0675*** 0.0112*** 0.0422*** (0.9672) (0.0000) (0.0000) (0.0005) (0.0000) (0.0000) (0.0000) (0.7864) (0.8493) (0.9889) (0.0000) (0.0047) (0.0000) REG 0.0508 0.0215 0.1470*** 0.0093 0.0908*** 0.0681*** 0.0420*** 0.0140** 0.1399 0.0171 0.0723*** 0.0907*** 0.0187 (0.7986) (0.1262) (0.0000) (0.6744) (0.0000) (0.0000) (0.0000) (0.0150) (0.7109) (0.4504) (0.0000) (0.0000) (0.4968) INF 0.097 0.1896*** 0.5894*** 0.2378*** 0.2783*** 1.4192*** 0.5499*** 0.1449*** 0.0893 0.0376 2.9622*** 0.0812*** 0.6377*** (0.7160) (0.0019) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.8707) (0.5581) (0.0000) (0.0001) (0.0000) GDPG 0.0187 0.1300*** 0.046 0.1863*** 0.2006*** 1.5291*** 0.4429*** 0.2011*** 0.7052 0.3734*** 0.0914*** 0.5806*** 0.1361*** (0.9473) (0.0070) (0.3277) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.2567) (0.0000) (0.0000) (0.0000) (0.0000) Observations 154 4,934 688 853 832 740 1,772 440 107 4,938 1,169 1,162 975

PAGE 51

51 Table 2 6. Continued F rance United Kingdom G reece H ong Kong I ndia I ndonesia I reland I srael I taly J apan Korea M exico BLEV 0.7234*** 0.6855*** 0.8860*** 0.6418*** 0.7637*** 0.7810*** 0.8435*** 0.6714*** 0.9054*** 0.8493*** 0.6721*** 0.7957*** (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) EBIT_TA 0.0639*** 0.0166 0.1422*** 0.0051 0.1662*** 0.1121*** 0.0017 0.3367*** 0.0441*** 0.0257 0.1707*** 0.0591*** (0.0039) (0.2032) ( 0.0000 ) (0.1903) ( 0.0000 ) ( 0.0000 ) (0.8972) ( 0.0000 ) ( 0.0000 ) (0.3459) ( 0.0000 ) (0.0002) MB 0.0012** 0.0034** 0.0001 0.0103*** 0.0015** 0.0123*** 0.0160*** 0.0081* 0.0006*** 0.0037** 0.0037 0.0009*** (0.0218) (0.0239) (0.6811) ( 0.0000 ) (0.0109) ( 0.0000 ) ( 0.0000 ) (0.0632) ( 0.0000 ) (0.0466) (0.3860) (0.0049) DEP_TA 0.2818*** 0.1872*** 0.3772*** 0.4017*** 0.4343*** 0.9269*** 0.0033 0.2429 0.0692 1.0095*** 0.6173*** 0.3058*** ( 0.0000) (0.0044) ( 0.0000) ( 0.0000) ( 0.0000) ( 0.0000) (0.9699) (0.5713) (0.1314) ( 0.0000) ( 0.0000) ( 0.0000) LnTA 0.0064*** 0.0059*** 0.0029*** 0.0081*** 0.0004 0.0117*** 0.0062*** 0.0178*** 0.0004 0.0077*** 0.0093*** 0.0039*** ( 0.0000) (0.0002) ( 0.0000) ( 0.0000) (0.3510) ( 0.0000) (0.0037) ( 0.0000) (0.1462) ( 0.0000) ( 0.0000) (0.0093) FA_TA 0.0075 0.0688*** 0.1711*** 0.0051 0.2054*** 0.0041 0.0183 0.2408* 0.0199** 0.0719*** 0.0155 0.0337** (0.6635) ( 0.0000 ) ( 0.0000 ) (0.1390) ( 0.0000 ) (0.5996) (0.3803) (0.0535) (0.0232) (0.0004) (0.2764) (0.0228) TAXES 0.0017 0.0023 0.0180*** 0.0052*** 0.0093*** 0.0043* 0.0082* 0.0032 0.0049*** 0.0015 0.0111* 0.0015 (0.5201) (0.4200) ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) (0.0741) (0.0589) (0.5878) (0.0030) (0.2432) (0.0995) (0.1639) LIQUID 0.002 0.0009 0.0040*** 0.0009*** 0.0115*** 0.0010* 0.0008 0.0065*** 0.0010** 0.0015 0.0038* 0.0060*** (0.1122) (0.2965) ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) (0.0684) (0.4339) (0.0010) (0.0428) (0.3702) (0.0631) ( 0.0000 ) RD_DUM 0.0016 0.0013 0.0084** 0.0127*** 0.0181*** 0.0002 0.0237*** 0.0455** 0.0554*** 0.0056* 0.0186*** 0.0317*** (0.6945) (0.7642) (0.0254) ( 0.0000 ) ( 0.0000 ) (0.7813) (0.0006) (0.0492) ( 0.0000 ) (0.0577) ( 0.0000 ) ( 0.0000 ) R&D_TA 0.061 0.0773 1.6617*** 1.7098*** 4.8900*** 0.2581*** 0.2776* 1.0689*** 0.4359*** 0.7705*** 0.5858*** 3.0686 (0.4100) (0.2717) (0.0021) ( 0.0000 ) ( 0.0000 ) (0.0019) (0.0981) ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) (0.4643) IND_MEDIAN 0.0979*** 0.0724*** 0.0174 0.0643*** 0.1148*** 0.0022 0.1479*** 0.0663 0.0412*** 0.0741*** 0.0781*** 0.2929*** ( 0.0000 ) (0.0051) (0.1444) ( 0.0000 ) ( 0.0000 ) (0.7754) ( 0.0000 ) (0.3264) ( 0.0000 ) ( 0.0000 ) (0.0012) ( 0.0000 ) REG 0.8399*** 0.3117*** 0.0702* 0.008 3.8696*** 0.0087** 0.6038 0.0527 0.0034 0.0498*** 0.7397*** ( 0.0000 ) ( 0.0000 ) (0.0714) (0.2056) ( 0.0000 ) (0.0139) (0.2351) (0.7765) (0.2517) (0.0013) (0.0004) INF 0.1916 0.2725*** 0.1036 0.0912*** 0.0399*** 0.1587*** 0.0079 0.0191 0.3716*** 0.2451*** 0.4580*** 0.0958*** (0.1182) (0.0007) (0.1878) ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) (0.8953) (0.9085) ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) GDPG 0.4633*** 0.2225*** 1.1044*** 0.0818*** 0.7775*** 0.6254*** 0.3820*** 0.2429 0.0613*** 0.4700*** 0.0805** 0.0374* ( 0.0000 ) (0.0005) ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) (0.2470) (0.0033) ( 0.0000 ) (0.0229) (0.0548) Observations 4766 11 978 448 1 080 1 513 1 721 515 198 1 543 17 623 1 178 587

PAGE 52

52 Table 2 6. Continued M alaysia N orway N ew Zealand P akistan P eru Philippines P ortugal S ingapore T hailand T urkey U nited States South Africa BLEV 0.8704*** 0.7396*** 0.5939*** 0.8663*** 0.9058*** 0.8332*** 0.9360*** 0.7436*** 0.7580*** 0.8773*** 0.7590*** 0.7293*** (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) EBIT_TA 0.0959*** 0.0668*** 0.0011 0.0498 0.4463 0.1845*** 0.5170*** 0.0363*** 0.2212*** 0.1791*** 0.0107 0.0300*** ( 0.0000 ) ( 0.0000 ) (0.8246) (0.3139) (0.1282) ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) (0.4071) (0.0018) MB 0.0061*** 0.0010*** 0.0060*** 0.0115*** 0.0477 0.0017*** 0.0006 0.0034*** 0.0038*** 0.0001 0.0015 0.0036** ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) (0.0087) (0.2705) ( 0.0000 ) (0.4044) (0.0040) (0.0016) (0.7858) (0.3606) (0.0113) DEP_TA 0.6830*** 0.0210*** 0.021 0.6491 4.1174** 0.0513** 1.5126*** 0.2301*** 0.7169*** 0.3413 0.2267*** 0.1312** ( 0.0000 ) (0.0002) (0.2385) (0.1974) (0.0127) (0.0218) (0.0008) ( 0.0000 ) ( 0.0000 ) (0.1578) (0.0013) (0.0110) LnTA 0.0102*** 0.0035*** 0.0130*** 0.0009 0.064 0.0095*** 0.0019 0.0055*** 0.0047** 0.0157*** 0.0028** 0.0064*** ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) (0.8691) (0.2438) ( 0.0000 ) (0.5120) ( 0.0000 ) (0.0231) ( 0.0000 ) (0.0136) ( 0.0000 ) FA_TA 0.0437*** 0.1166*** 0.1073*** 0.1740** 0.7430** 0.1590*** 0.0104 0.0657*** 0.2387*** 0.0898*** 0.0722*** 0.0365** (0.0005) ( 0.0000 ) ( 0.0000 ) (0.0365) (0.0328) ( 0.0000 ) (0.8517) ( 0.0000 ) ( 0.0000 ) (0.0025) (0.0001) (0.0167) TAXES 0.0106*** 0.0064*** 0.0345*** 0.0499** 0.3380*** 0.0069*** 0.0036 0.0109*** 0.0187** 0.0584*** 0 0.0075** (0.0002) ( 0.0000 ) ( 0.0000 ) (0.0112) (0.0026) ( 0.0000 ) (0.4306) ( 0.0000 ) (0.0110) ( 0.0000 ) (0.9894) (0.0390) LIQUID 0.0031*** 0.0015*** 0.0010** 0.0088*** 0.016 0.0047*** 0.0026 0.0020*** 0.0019*** 0.0039*** 0.0002 0.0054** ( 0.0000 ) ( 0.0000 ) (0.0213) ( 0.0000 ) (0.2850) ( 0.0000 ) (0.5497) ( 0.0000 ) (0.0037) (0.0012) (0.8316) (0.0311) RD_DUM 0.0071 0.0099*** 0.0442*** 0.0340* 0.3764** 0.0136*** 0.2043*** 0.0022 0.0035 0.0204** 0.0092 0.0175*** (0.1764) ( 0.0000 ) ( 0.0000 ) (0.0686) (0.0219) (0.0049) (0.0057) (0.1750) (0.8424) (0.0192) (0.4355) ( 0.0000 ) R&D_TA 0.3522*** 0.1854*** 0.2053*** 2.1623*** 7.3823 0.3399*** 0.2164*** 1.1529 0.4265 0.0089 0.1690*** (0.0002) ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) (0.1100) (0.0016) (0.0084) (0.5693) (0.2486) (0.8601) ( 0.0000 ) IND_MEDIAN 0.0176 0.0910*** 0.0609*** 0.0179 0.1636 0.0633*** 0.0016 0.1037*** 0.0544*** 0.3088*** 0.1382*** 0.0460** (0.2144) ( 0.0000 ) ( 0.0000 ) (0.8265) (0.8051) ( 0.0000 ) (0.9730) ( 0.0000 ) ( 0.0000 ) (0.0002) ( 0.0000 ) (0.0105) REG 0.0278*** 0.0301*** 0.0788*** 0.0819 0.3429 0.0396** 0.2686* 0.0527*** 0.0442 0.1368 0.0531* 0.0513 (0.0001) ( 0.0000 ) ( 0.0000 ) (0.2639) (0.1230) (0.0460) (0.0855) ( 0.0000 ) (0.1799) (0.5445) (0.0729) (0.1093) INF 0.0414 0.7521*** 0.7792*** 0.4706*** 0.0589 0.3572*** 0.7241*** 0.3631*** 0.3825*** 1.6349*** 0.0245 0.0145 (0.4389) ( 0.0000 ) ( 0.0000 ) (0.0092) (0.9340) ( 0.0000 ) (0.0044) ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) (0.7971) (0.4611) GDPG 0.0343*** 0.3622*** 0.2653*** 0.2368 0.711 0.3808*** 0.3190** 0.0114* 0.4881*** 0.0556 0.3568*** 0.2454*** (0.0032) ( 0.0000 ) ( 0.0000 ) (0.5146) (0.3798) ( 0.0000 ) (0.0140) (0.0616) ( 0.0000 ) (0.5851) ( 0.0000 ) ( 0.0000 ) Observations 4, 909 973 542 230 136 507 323 2, 701 2, 371 247 29, 693 1, 014 The definitions of the variables in the regressions are provided in Table 2 1. is firm is debt ratio in year t and in country j, is the adjustment parameter, 1 is a vector of firm, industry, and macroeconomic characteristics related to the costs and benefits of operating with various leverage ratios, is the unobserved firm heterogeneity captured by the firm dummies, and is the error term. T he t op row for each determinant reports the coefficient estimates. *, **, *** indicate significance at the 90%, 95%, and 99% confidence levels respectively The p values of the two tailed tests for the significance of the coefficients are reported beneath the coefficient estimates in parentheses I report the estimation results from Blundell and Bonds (1998) twostep system GMM of the following model, run separately for each country using book leverage : = ( ) 1+ ( 1 ) 1+ + (2 3)

PAGE 53

53 Table 2 7. Separate Country Regressions Using Market Leverage as the Dependent Variable A rgentina A ustralia A ustria B elgium Brazil Canada Switzerland C hile Columbia Germany Denmark Spain Finland MLEV 0.8462*** 0.5992*** 0.8009*** 0.8152*** 0.7730*** 0.4714*** 0.7547*** 0.7310*** 0.8774* 0.7245*** 0.7189*** 0.7643*** 0.7643*** (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0687) (0.0000) (0.0000) (0.0000) (0.0000) EBIT_TA 0.5939*** 0.0014 0.1265*** 0.1270*** 0.1103*** 0.0147*** 0.0236*** 0.2757*** 0.5102 0.0255** 0.0393*** 0.2322*** 0.0934*** (0.0014) (0.7832) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.3448) (0.0160) (0.0000) (0.0000) (0.0000) MB 0.0192 0.0013* 0.0005** 0.0035*** 0.0003*** 0.0022*** 0.0020*** 0.0001 0.0188 0.0019 0.0150*** 0.0026*** 0.0011*** (0.3432) (0.0594) (0.0216) (0.0000) (0.0001) (0.0000) (0.0000) (0.9008) (0.8313) (0.1598) (0.0000) (0.0000) (0.0002) DEP_TA 1.7743** 0.3855*** 0.3369*** 0.4824*** 0.3403*** 0.5136*** 0.2855*** 1.6995*** 3.4587* 0.1275*** 0.5505*** 0.0800*** 0.8309*** (0.0277) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0536) (0.0020) (0.0000) (0.0000) (0.0000) LnTA 0.0107 0.0096*** 0.0119*** 0.0063*** 0.0159*** 0.0248*** 0.0065*** 0.0027*** 0.0057 0.0011 0.0031*** 0.0011*** 0.0105*** (0.6698) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0016) (0.8944) (0.4184) (0.0000) (0.0000) (0.0000) FA_TA 0.7387* 0.0453*** 0.0245*** 0.0613*** 0.2075*** 0.1969*** 0.1349*** 0.0782*** 0.0641 0.0216 0.3028*** 0.0835*** 0.1187*** (0.0650) (0.0007) (0.0010) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.7675) (0.2860) (0.0000) (0.0000) (0.0000) TAXES 0.087 0.0078*** 0.0063*** 0.0026 0.0062*** 0.0296*** 0.0218*** 0.1921*** 0.1463 0.001 0.0168*** 0.0368*** 0.0025 (0.3437) (0.0015) (0.0002) (0.3808) (0.0000) (0.0000) (0.0000) (0.0000) (0.4737) (0.7303) (0.0000) (0.0000) (0.3927) LIQUID 0.0111 0.0012*** 0.0001 0.0133*** 0.0327*** 0.0070*** 0.0030*** 0.0093*** 0.0344 0.0002 0.0017*** 0.0071*** 0.0036*** (0.2625) (0.0003) (0.5333) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.1862) (0.6117) (0.0000) (0.0000) (0.0001) RD_DUM 0.0057 0.0116*** 0.0235*** 0.0477*** 0.0974*** 0.0353*** 0.0201*** 0.0165*** 0.0181 0.0044 0.0594*** 0.0163*** 0.0194*** (0.9566) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0033) (0.9752) (0.3680) (0.0000) (0.0000) (0.0000) R&D_TA 12.9854 0.0493 0.5843*** 0.0531 3.8015*** 0.0615*** 0.6683*** 2.3164*** 0.1130** 0.3753*** 0.4964*** 0.2795*** (0.5678) (0.1260) (0.0000) (0.1547) (0.0000) (0.0006) (0.0000) (0.0031) (0.0372) (0.0000) (0.0000) (0.0000) IND_MEDIAN 0.0271 0.0384*** 0.1213*** 0.0536*** 0.1457*** 0.1382*** 0.0119*** 0.2154*** 0.1889 0.0518*** 0.0130*** 0.0293*** 0.0317*** (0.8384) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.6307) (0.0029) (0.0000) (0.0000) (0.0000) REG 0.0236 0.0055 0.0321* 0.0026 0.0048 0.0173*** 0.0920*** 0.2341*** 0.1788 0.1210** 0.0103 0.1007*** 0.1244*** (0.9430) (0.7240) (0.0863) (0.9177) (0.4931) (0.0053) (0.0000) (0.0000) (0.6881) (0.0100) (0.6905) (0.0000) (0.0000) INF 0.6964*** 0.4703*** 0.1773*** 0.2303*** 0.6905*** 0.6176*** 0.6171*** 1.5628*** 0.3319 0.2195*** 2.5371*** 0.3569*** 0.5085*** (0.0017) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.6949) (0.0011) (0.0000) (0.0000) (0.0000) GDPG 0.7705* 0.4716*** 0.0051 0.5080*** 0.6640*** 0.1202*** 1.1466*** 0.9156*** 1.0931 0.7474*** 0.3149*** 0.0009 0.2615*** (0.0935) (0.0000) (0.8179) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.4230) (0.0000) (0.0000) (0.9416) (0.0000) Observations 154 4,934 688 853 832 740 1,772 440 107 4,934 1,169 1,162 975

PAGE 54

54 Table 2 7. Continued F rance United Kingdom G reece H ong Kong I ndia I ndonesia I reland I srael I taly J apan Korea M exico MLEV 0.7062*** 0.6129*** 0.8264*** 0.6752*** 0.7010*** 0.7788*** 0.8295*** 0.6161*** 0.7264*** 0.7694*** 0.6027*** 0.7664*** (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) EBIT_TA 0.0773*** 0.0142 0.1940*** 0.0198*** 0.0370*** 0.0419*** 0.0577*** 0.2708*** 0.0677*** 0.0351* 0.1284*** 0.0740*** (0.0000) (0.1680) (0.0000) (0.0001) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0948) (0.0001) (0.0000) MB 0.0033*** 0.0018* 0.0021*** 0.0006 0.0034*** 0.0031*** 0.0124*** 0.0107*** 0.0029*** 0.0033** 0.0098** 0.0002 (0.0000) (0.0811) (0.0000) (0.2448) (0.0000) (0.0000) (0.0000) (0.0062) (0.0000) (0.0147) (0.0262) (0.5631) DEP_TA 0.0943* 0.3598*** 1.7484*** 0.4729*** 0.5172*** 1.0866*** 0.3252*** 0.3796 0.1805*** 0.4616*** 0.6927*** 0.0816 (0.0874) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.4926) (0.0000) (0.0001) (0.0000) (0.2069) LnTA 0.0106*** 0.0081*** 0.0107*** 0.0107*** 0.0039*** 0.0037*** 0.0140*** 0.0128*** 0.0003* 0.0073*** 0.0121*** 0.0049*** (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0834) (0.0000) (0.0000) (0.0000) FA_TA 0.0007 0.0263* 0.1797*** 0.0326*** 0.1473*** 0.0223*** 0.1099*** 0.185 0.0899*** 0.0734*** 0.1050*** 0.0197*** (0.9502) (0.0632) (0.0000) (0.0000) (0.0000) (0.0003) (0.0000) (0.1498) (0.0000) (0.0000) (0.0000) (0.0007) TAXES 0.0009 0.0007 0.0624*** 0.0182*** 0.0028*** 0.0025 0.0276*** 0.0104 0.0295*** 0.0026** 0.0335*** 0.0034*** (0.7165) (0.7900) (0.0000) (0.0000) (0.0000) (0.1290) (0.0000) (0.2075) (0.0000) (0.0144) (0.0001) (0.0009) LIQUID 0.0060*** 0.0031*** 0.0079*** 0.0010*** 0.0048*** 0.0005 0.0028** 0.0035* 0.0001 0.0065*** 0.0084*** 0.0047*** (0.0000) (0.0015) (0.0000) (0.0002) (0.0000) (0.3109) (0.0166) (0.0761) (0.6679) (0.0000) (0.0018) (0.0000) RD_DUM 0.0095*** 0.0101*** 0.0601*** 0.0056** 0.0240*** 0.0208*** 0.0027 0.1038*** 0.0345*** 0.0064** 0.0382*** 0.0602*** (0.0024) (0.0008) (0.0000) (0.0182) (0.0000) (0.0000) (0.7543) (0.0000) (0.0000) (0.0214) (0.0000) (0.0000) R&D_TA 0.2866*** 0.0416 1.9941*** 1.5472*** 5.1876*** 0.2058*** 0.0821 0.5097* 0.5795*** 0.4520*** 0.3317** 16.1119*** (0.0000) (0.4329) (0.0092) (0.0000) (0.0000) (0.0012) (0.4450) (0.0649) (0.0000) (0.0000) (0.0485) (0.0000) IND_MEDIAN 0.0019 0.0588*** 0.0504*** 0.0221*** 0.0555*** 0.1098*** 0.0125 0.2067*** 0.0662*** 0.1139*** 0.1249*** 0.0262*** (0.8805) (0.0044) (0.0000) (0.0000) (0.0000) (0.0000) (0.6106) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) REG 0.3143*** 0.1999*** 0.5674*** 0.1471*** 0.7662 0.0573*** 0.0059 0.094 0.004 0.0492*** 0.4055** (0.0066) (0.0000) (0.0012) (0.0000) (0.1792) (0.0000) (0.9912) (0.5778) (0.2008) (0.0005) (0.0499) INF 0.0673 0.4662*** 0.1254 0.0217 0.2895*** 0.0055 0.0326 0.4833*** 0.8696*** 0.3387*** 1.7416*** 0.2184*** (0.4939) (0.0000) (0.1307) (0.2302) (0.0000) (0.7000) (0.7259) (0.0003) (0.0000) (0.0000) (0.0000) (0.0000) GDPG 0.1445** 0.5259*** 0.5803*** 0.1559*** 0.3091*** 0.5052*** 0.8373*** 0.8501*** 0.3773*** 0.0173 1.1350*** 0.1747*** (0.0132) (0.0000) (0.0001) (0.0000) (0.0000) (0.0000) (0.0000) (0.0009) (0.0000) (0.4533) (0.0000) (0.0000) Observations 4 765 11 974 448 1 080 1 513 1 721 514 198 1 543 17 623 1 178 586

PAGE 55

55 Table 2 7. Continued M alaysia N orway N ew Zealand P akistan P eru Philippines P ortugal S ingapore T hailand T urkey U nited States South Africa MLEV 0.7790*** 0.7187*** 0.5254*** 0.8620*** 0.8913*** 0.7688*** 0.8167*** 0.6325*** 0.7141*** 0.8726*** 0.7468*** 0.6935*** (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) EBIT_TA 0.0023 0.0213*** 0.0329*** 0.0381 1.1190** 0.0856*** 0.2002 0.0178*** 0.0423*** 0.2273*** 0.0236** 0.0124 (0.7953) (0.0000) (0.0009) (0.4827) (0.0306) (0.0000) (0.2945) (0.0001) (0.0068) (0.0018) (0.0103) (0.2908) MB 0.0032*** 0.0043*** 0.0077*** 0.0151*** 0.2045* 0.0001 0.0050*** 0.0018*** 0.0039*** 0.0017** 0.0026*** 0.0038** (0.0010) (0.0000) (0.0000) (0.0015) (0.0942) (0.7408) (0.0000) (0.0023) (0.0000) (0.0101) (0.0027) (0.0356) DEP_TA 0.5976*** 0.0367*** 0.0752 1.0752*** 3.5348 0.5753*** 0.5662** 0.3195*** 0.6839*** 0.5917** 0.3657*** 0.1002 (0.0000) (0.0000) (0.1372) (0.0031) (0.1340) (0.0000) (0.0208) (0.0000) (0.0000) (0.0254) (0.0000) (0.1688) LnTA 0.0183*** 0.0011*** 0.0308*** 0.0049 0.0132 0.0184*** 0.0045 0.0106*** 0.0091*** 0.0198*** 0.0060*** 0.0149*** (0.0000) (0.0002) (0.0000) (0.3648) (0.8409) (0.0000) (0.2833) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) FA_TA 0.0570*** 0.0993*** 0.0233*** 0.1615* 1.7477* 0.1575*** 0.2692*** 0.0736*** 0.1105*** 0.0396 0.0211 0.0376** (0.0000) (0.0000) (0.0062) (0.0628) (0.0570) (0.0000) (0.0000) (0.0000) (0.0000) (0.1930) (0.1420) (0.0250) TAXES 0.0082*** 0.0086*** 0.0558*** 0.0339*** 0.2512* 0.0104*** 0.0073 0.0069*** 0.0255*** 0.0898*** 0.0044** 0.0014 (0.0031) (0.0000) (0.0000) (0.0021) (0.0749) (0.0000) (0.3414) (0.0000) (0.0024) (0.0000) (0.0193) (0.6852) LIQUID 0.0045*** 0.0010*** 0.0043*** 0.0193*** 0.0066 0.0100*** 0.0190* 0.0019*** 0.0004 0.0164*** 0.0003 0.0075*** (0.0000) (0.0000) (0.0000) (0.0000) (0.6766) (0.0000) (0.0550) (0.0000) (0.5851) (0.0004) (0.7066) (0.0014) RD_DUM 0.0191*** 0.0302*** 0.0383*** 0.0458* 0.2102 0.0324*** 0.0169 0.0017 0.0589*** 0.0359* 0.0161 0.0417*** (0.0000) (0.0000) (0.0000) (0.0668) (0.1768) (0.0000) (0.7846) (0.1570) (0.0000) (0.0642) (0.1092) (0.0000) R&D_TA 0.4780*** 0.3498*** 1.0809*** 1.7683** 12.7399 0.2076 0.0976* 1.1350*** 0.9483 0.0216 0.1615*** (0.0000) (0.0000) (0.0000) (0.0120) (0.3625) (0.4888) (0.0610) (0.0026) (0.1958) (0.6214) (0.0001) IND_MEDIAN 0.0472*** 0.0255*** 0.0797*** 0.0275 0.293 0.0722*** 0.1957*** 0.0088* 0.0190** 0.2437*** 0.1373*** 0.0224** (0.0000) (0.0000) (0.0000) (0.3575) (0.2995) (0.0000) (0.0014) (0.0683) (0.0308) (0.0001) (0.0000) (0.0259) REG 0.0131 0.0302*** 0.0849*** 0.0718 1.9269 0.0979*** 0.2431 0.1041*** 0.0664*** 0.6479 0.1406*** 0.0705*** (0.3119) (0.0000) (0.0000) (0.2055) (0.3995) (0.0036) (0.1069) (0.0000) (0.0000) (0.3034) (0.0002) (0.0006) INF 0.6891*** 1.8262*** 1.5670*** 0.9274*** 0.1422 0.3387*** 0.4307** 1.5886*** 0.1865*** 2.3295*** 0.1235 0.9637*** (0.0000) (0.0000) (0.0000) (0.0003) (0.8901) (0.0000) (0.0321) (0.0000) (0.0000) (0.0000) (0.1696) (0.0000) GDPG 0.0889*** 2.0432*** 1.3192*** 0.6881 0.5167 0.4851*** 0.5587*** 0.1326*** 0.2753*** 0.2748*** 0.5425*** 1.6230*** (0.0000) (0.0000) (0.0000) (0.1624) (0.5430) (0.0000) (0.0000) (0.0000) (0.0000) (0.0034) (0.0000) (0.0000) Observations 4, 907 973 542 229 136 507 323 2, 701 2, 371 247 29, 687 1, 013 The definitions of the variables in the regressions are provided in Table 2 1. is firm is debt ratio in year t and in country j, is the adjustment parameter, 1 is a vector of firm, industry, and macroeconomic characteristics related to the costs and benefits of operating with various leverage ratios, is the unobserved firm heterogeneity captured by the firm dummies, and is the error term. T he t op row for each determinant reports the coefficient estimates. *, **, *** indicate significance at the 90%, 95%, and 99% confidence levels respectively The p values of the two tailed tests for the significance of the coefficients are reported beneath the coefficient estimates in parentheses I report the estimation results from Blundell and Bonds (1998) two step system GMM of the following model, run se parately for each country using market leverage : = ( ) 1+ ( 1 ) 1+ + (2 3)

PAGE 56

56 Table 2 8 Pooled Country Regressions BOOK LEVERAGE MARKET LEVERAGE (1) (2) LEV 0.8484*** 0.7028*** (0.0000) (0.0000) EBIT_TA 0.0235** 0.0145** (0.0116) (0.0206) MB 0.0001 0.0021*** (0.5710) (0.0000) DEP_TA 0.2689*** 0.3435*** (0.0000) (0.0000) LnTA 0.0025*** 0.0056*** (0.0000) (0.0000) FA_TA 0.0844*** 0.0475*** (0.0000) (0.0000) TAXES 0.0011 0.0007 (0.3180) (0.4900) LIQUID 0.0014** 0.0019*** (0.0109) (0.0004) RD_DUM 0.0084*** 0.0105*** (0.0018) (0.0000) R&D_TA 0.0215 0.0123 (0.6070) (0.6930) IND_MEDIAN 0.0759*** 0.0360*** (0.0000) (0.0000) REG 0.0004 0.0083 (0.9720) (0.4940) INF 0.1495*** 0.0794** (0.0001) (0.0272) GDPG 0.1174*** 0.0471** (0.0000) (0.0410) Observations 105,560 105,539 *The definitions of the variables in the regressions are provided in Table 2 1. is firm is debt ratio in year t and in country j, is the adjustment parameter, 1 is a vector of firm, industry, and macroeconomic characteristics related to the costs and benefits of operating with various leverage ratios, is the unobserved firm heterogeneity captured by the firm dummies, and is the error term. The top row for each determinant reports the coefficient estimates. *, **, *** indicate significance at the 90%, 95%, and 99% confidence levels, respectively. The p values of the two tailed tests for the significance of the coefficients are reported beneath the coeff icient estimates in parentheses. I report the estimation results from Blundell and Bonds (1998) two step system GMM of the following model, run separately for book and market definitions of leverage, pooling the data on all sample countries: = ( ) 1+ ( 1 ) 1+ + (2 3)

PAGE 57

57 Table 2 9. Correlations between Leverage and Firm, Industry, Macroeconomic Determinants SEPARATE POOLED BOOK MARKET BOOK MARKET (1) (2) (3) (4) (5) (6) (7) (8) + + + + PANEL A. Number of Significant Correlations LEV 37 0 37 0 37 0 37 0 EBIT_TA 2 23 15 15 0 37 37 0 MB 16 11 9 21 0 0 0 37 DEP_TA 4 23 2 30 0 37 0 37 LnTA 25 4 26 5 37 0 37 0 FA_TA 21 7 27 4 37 0 37 0 TAXES 11 17 12 14 0 0 0 0 LIQUID 9 18 10 18 0 37 0 37 RD_DUM 19 9 19 10 37 0 37 0 R&D_TA 13 13 14 12 0 0 0 0 IND_MEDIAN 19 7 18 13 37 0 37 0 REG 10 11 10 11 0 0 0 0 INF 16 10 19 10 0 37 37 0 GDPG 19 11 18 13 0 37 37 0 SEPARATE POOLED BOOK MARKET BOOK MARKET (1) (2) (3) (4) (5) (6) (7) (8) + + + + PANEL B. Core Factors LEV YES YES YES YES EBIT_TA YES YES YES MB YES YES DEP_TA YES YES YES YES LnTA YES YES YES YES FA_TA YES YES YES YES TAXES LIQUID YES YES YES YES RD_DUM YES YES YES YES R&D_TA IND_MEDIAN YES YES YES YES REG INF YES YES YES GDPG YES YES YES YES The definitions of the variables documented in this table are provided in Table 2 1. Panel A provides a summary for the consistency of the direction of the relationship between leverage and each determinant Panel B evaluates whether or not the leverage de terminant is a dominant factor. The columns (1) (4) refer to the core estimation model (Equation 2 3) run separately for each country. The columns (1) and (2) summarize the results of Table 2 6 which uses book leverage. The columns (3) and (4) summarize the results of Table 2 7 which uses market leverage. The columns (5) (8) refer to the core model estimated pooling all 37 sample countrie s. The columns (5) and (6) summarize the results of Table 2 8, column (1) which uses book leverage. The columns (7) and (8) summarize the results of Table 2 8, column (2) which uses market leverage. F irst or second (third or fourth) column in Panel A repor t s the number of insta nces out of 37 sample countries in Table 2 6 (Table 2 7) that the given determinant of leverage has a positive or negative significant coefficient at the 90% or higher confidence level Fifth or sixth (seventh or eighth ) column in Panel A report s 37 if the given determinant of leverage has a positive or negative significant coefficient at the 90% or higher confidence level in Table 2 8 in column (1) (column (2)) ; and reports 0 otherwise Panel B assigns a leverage determinant as a core factor using a YES indicator if the score reported in Panel A is at least 18. = ( ) 1+ ( 1 ) 1+ + (2 3)

PAGE 58

5 8 Table 2 10. Empirical Relevance of the Capital Structure Theories Predicted Signs by the Capital Structure Theories Empirically Observed Signs (1) (2) (3) (4) (5) DYNAMIC TRADEOFF PECKING ORDER MARKET TIMING BOOK MARKET LEV + and (0,1) + and equal to 1 + and equal to 1 + and (0,1) + and (0,1) EBIT_TA + or + MB + DEP_TA LnTA + + or + + FA_TA + + or + + TAXES + LIQUID + or RD_DUM + + + R&D_TA + IND_MEDIAN + + + REG + INF + or + + GDPG + or + *The definitions of the variables in the regressions are provided in Table 2 1. The columns (1) (3) summarize the hypothesized predictions of the main theories of capital structure about how leverage relates to the observable firm industry, and macroeconomic attributes The columns (4) and (5) summarize the direction of the observed relationships between leverage and its determinants that is reported in Panel B of Table 2 9. A particular sign is assigned to column (4) or (5) as long as the implications of the SEPARATE and POOLED methods do not conflict. In other words, a + sign is assigned to column (4) if at least one of the columns (1) or (5) is assi gned a YES indicator, and neither of columns (2) and (6) is assigned a YES indicator. Similarly, a sign is assigned to column (4) if at least one of the columns ( 2 ) or (6 ) is assigned a YES indicator, and neither of columns ( 1) and (5 ) is assigned a YES indicator. A + sig n is assigned to column (5 ) if at least one of the columns ( 3 ) or (7 ) is assigned a YES indicator, and neither of columns ( 4) and (8 ) is assigned a YES indicator. Similarly, a sign is assigned to column ( 5 ) if at least one of the columns ( 4 ) or ( 8 ) is assigned a YES indicator, and neither of columns ( 3 ) and ( 7 ) is assigned a YES indicator.

PAGE 59

59 Table 2 11. The Effect of Conditioning on the Institutional Setting TIME COST EFFICIENCY TAX CREDITOR Strong Weak Strong Weak Weak Strong Low High Weak Strong BLEV 0.8697*** 0.8488*** 0.8307*** 0.9546*** 0.9371*** 0.8215*** 0.8375*** 0.8997*** 0.8642*** 0.8536*** (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) EBIT_TA 0.0153 0.0339*** 0.0013 0.1096*** 0.0626*** 0.0056 0.0056 0.0358** 0.0240** 0.0105 (0.1580) (0.0063) (0.8930) (0.0000) (0.0001) (0.5920) (0.5930) (0.0202) (0.0367) (0.3100) MB 0.0001 0.0001 0.0006 0.0003 0.0004 0.0019*** 0.0003 0.0001 0.0006*** 0.0015 (0.6520) (0.6990) (0.1980) (0.3090) (0.2320) (0.0001) (0.2040) (0.8720) (0.0058) (0.2050) DEP_TA 0.1701*** 0.3444*** 0.2596*** 0.3404*** 0.2295*** 0.2393*** 0.2186*** 0.2246*** 0.4082*** 0.1764*** (0.0006) (0.0000) (0.0000) (0.0001) (0.0001) (0.0000) (0.0000) (0.0003) (0.0000) (0.0005) LnTA 0.0020*** 0.0032*** 0.0018*** 0.0028*** 0.0019** 0.0021*** 0.0032*** 0.0021*** 0.0021*** 0.0027*** (0.0009) (0.0011) (0.0010) (0.0067) (0.0410) (0.0001) (0.0002) (0.0068) (0.0011) (0.0078) FA_TA 0.0505*** 0.1060*** 0.0761*** 0.0872*** 0.1078*** 0.0550*** 0.0667*** 0.0647*** 0.0973*** 0.0795*** (0.0001) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0006) (0.0000) (0.0000) TAXES 0.0004 0.0023 0.0001 0.0048 0.0001 0.0002 0.0023 0.0003 0.0006 0.0032 (0.7750) (0.2110) (0.8960) (0.1160) (0.9790) (0.8550) (0.1380) (0.8450) (0.6280) (0.1360) LIQUID 0.0015** 0.0009 0.0013** 0.0020** 0.0021** 0.0009 0.0007 0.0028*** 0.0021*** 0.0002 (0.0189) (0.2590) (0.0287) (0.0469) (0.0148) (0.1510) (0.3010) (0.0009) (0.0068) (0.7340) RD_DUM 0.0082*** 0.0003 0.0116*** 0.0152** 0.0128*** 0.0132*** 0.0013 0.0068* 0.0128*** 0.0075* (0.0052) (0.9600) (0.0001) (0.0147) (0.0033) (0.0001) (0.7600) (0.0546) (0.0000) (0.0806) R&D_TA 0.1169** 0.0062 0.0711* 0.1283 0.0685 0.1053** 0.0847* 0.0438 0.0334 0.1570** (0.0496) (0.9040) (0.0824) (0.1790) (0.2560) (0.0175) (0.0595) (0.5920) (0.4780) (0.0140) IND_MEDIAN 0.0787*** 0.0802*** 0.0986*** 0.0259 0.0406** 0.1029*** 0.0399** 0.0526*** 0.0741*** 0.0301* (0.0000) (0.0000) (0.0000) (0.2020) (0.0242) (0.0000) (0.0164) (0.0008) (0.0000) (0.0762) REG 0.0170 0.0011 0.0048 0.0598* 0.0118 0.0037 0.0208 0.0137 0.0078 0.0300* (0.1170) (0.9680) (0.6970) (0.0675) (0.6580) (0.7820) (0.2920) (0.2640) (0.6170) (0.0508) INF 0.1245** 0.0534 0.1720*** 0.0482 0.0420 0.1604*** 0.0639 0.1924*** 0.2115*** 0.0312 (0.0120) (0.2750) (0.0005) (0.2650) (0.3730) (0.0020) (0.1810) (0.0001) (0.0000) (0.5600) GDPG 0.2514*** 0.0931*** 0.1245*** 0.0185 0.0508 0.1229*** 0.0657** 0.3698*** 0.1915*** 0.0150 (0.0000) (0.0100) (0.0000) (0.6300) (0.1950) (0.0000) (0.0298) (0.0000) (0.0000) (0.6330) Observations 57,387 46,222 87,691 15,918 26,482 77,127 65,859 39,701 70,231 35,329

PAGE 60

60 Table 2 11. Continued FORMALISM ANTIDIR PRENF EXECUTIVE ENFORCE Strong Weak Weak Strong Weak Strong Weak Strong Weak Strong BLEV 0.8329*** 0.9593*** 0.9432*** 0.8290*** 0.9275*** 0.8301*** 0.9073*** 0.8355*** 0.9213*** 0.8324*** (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) EBIT_TA 0.0000 0.0823*** 0.0822*** 0.0007 0.0529*** 0.0052 0.0940*** 0.0001 0.1105*** 0.0039 (0.9970) (0.0000) (0.0000) (0.9480) (0.0010) (0.6150) (0.0000) (0.9920) (0.0000) (0.6840) MB 0.0008* 0.0001 0.0008*** 0.0028*** 0.0002 0.0020** 0.0006* 0.0013*** 0.0006** 0.0031*** (0.0640) (0.6210) (0.0072) (0.0000) (0.4210) (0.0222) (0.0910) (0.0021) (0.0301) (0.0001) DEP_TA 0.2935*** 0.1624** 0.2808*** 0.2489*** 0.1209** 0.3170*** 0.4302*** 0.1892*** 0.4435*** 0.2351*** (0.0000) (0.0101) (0.0000) (0.0000) (0.0304) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) LnTA 0.0021*** 0.0015* 0.0003 0.0031*** 0.0001 0.0022*** 0.0031*** 0.0014*** 0.0022** 0.0015*** (0.0002) (0.0994) (0.7770) (0.0000) (0.8790) (0.0001) (0.0019) (0.0078) (0.0196) (0.0046) FA_TA 0.0726*** 0.0907*** 0.1468*** 0.0588*** 0.0693*** 0.0877*** 0.0984*** 0.0748*** 0.0911*** 0.0788*** (0.0000) (0.0000) (0.0000) (0.0000) (0.0008) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) TAXES 0.0007 0.0002 0.0029 0.0013 0.0013 0.0010 0.0022 0.0011 0.0056* 0.0002 (0.5780) (0.9360) (0.2270) (0.2870) (0.5490) (0.4080) (0.4240) (0.3240) (0.0901) (0.8380) LIQUID 0.0008 0.0024* 0.0010 0.0009 0.0002 0.0009 0.0023*** 0.0012** 0.0024** 0.0010* (0.1720) (0.0556) (0.3940) (0.1390) (0.8170) (0.1470) (0.0073) (0.0353) (0.0100) (0.0831) RD_DUM 0.0149*** 0.0144*** 0.0163*** 0.0121*** 0.0162*** 0.0141*** 0.0053 0.0129*** 0.0186*** 0.0114*** (0.0000) (0.0006) (0.0003) (0.0001) (0.0001) (0.0000) (0.2770) (0.0000) (0.0009) (0.0001) R&D_TA 0.0736 0.0813 0.0471 0.1004** 0.0681 0.0783* 0.2118** 0.0594 0.0437 0.0829** (0.1030) (0.1930) (0.4140) (0.0292) (0.2320) (0.0955) (0.0299) (0.1450) (0.7480) (0.0378) IND_MEDIAN 0.0923*** 0.0642*** 0.0544*** 0.0831*** 0.0543*** 0.0760*** 0.0468** 0.0813*** 0.0326* 0.0866*** (0.0000) (0.0007) (0.0014) (0.0000) (0.0016) (0.0000) (0.0131) (0.0000) (0.0882) (0.0000) REG 0.0087 0.0099 0.0034 0.0064 0.0219** 0.0145 0.0071 0.0041 0.0341 0.0066 (0.5350) (0.4980) (0.8800) (0.6330) (0.0441) (0.2910) (0.8150) (0.7300) (0.1880) (0.6100) INF 0.1420*** 0.0935* 0.0579 0.0763* 0.0486 0.1154*** 0.0931** 0.1014** 0.1140** 0.0416 (0.0008) (0.0653) (0.2530) (0.0899) (0.3720) (0.0091) (0.0413) (0.0319) (0.0108) (0.3770) GDPG 0.1242*** 0.0748 0.2855*** 0.1632*** 0.2115*** 0.1221*** 0.0579* 0.1821*** 0.0128 0.1219*** (0.0000) (0.1520) (0.0000) (0.0000) (0.0000) (0.0000) (0.0775) (0.0000) (0.7270) (0.0000) Observations 84,921 20,639 25,151 80,409 22,628 82,932 23,154 80,148 18,408 85,530

PAGE 61

61 Table 2 11. Continued LAW&ORDER GOVERNANCE ACCSTDS EDISCLOSE Weak Strong Weak Strong Weak Strong Weak Strong BLEV 0.9356*** 0.8119*** 0.9311*** 0.8340*** 0.9404*** 0.8299*** 0.9434*** 0.8286*** (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) EBIT_TA 0.0452*** 0.0116 0.1039*** 0.0020 0.0745*** 0.0042 0.0509*** 0.0052 (0.0045) (0.2650) (0.0000) (0.8350) (0.0000) (0.6730) (0.0023) (0.6080) MB 0.0009*** 0.0006 0.0007** 0.0024*** 0.0001 0.0030*** 0.0001 0.0027*** (0.0022) (0.1820) (0.0130) (0.0000) (0.7260) (0.0030) (0.8130) (0.0086) DEP_TA 0.2732*** 0.1814*** 0.4010*** 0.2217*** 0.2888*** 0.2836*** 0.1342** 0.3392*** (0.0000) (0.0004) (0.0000) (0.0000) (0.0000) (0.0000) (0.0188) (0.0000) LnTA 0.0021*** 0.0037*** 0.0019* 0.0023*** 0.0012 0.0024*** 0.0005 0.0022*** (0.0027) (0.0002) (0.0539) (0.0000) (0.1590) (0.0000) (0.5900) (0.0001) FA_TA 0.0735*** 0.0674*** 0.0984*** 0.0610*** 0.1087*** 0.0692*** 0.1107*** 0.0831*** (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) 0.0000 TAXES 0.0007 0.0013 0.0054* 0.0003 0.0022 0.0009 0.0012 0.0004 (0.6170) (0.4130) (0.0983) (0.8220) (0.4250) (0.4560) (0.6400) (0.7580) LIQUID 0.0033*** 0.0001 0.0025*** 0.0012** 0.0007 0.0010* 0.0017 0.0008 (0.0004) (0.9020) (0.0072) (0.0417) (0.5030) (0.0728) (0.1410) (0.1860) RD_DUM 0.0066** 0.0004 0.0189*** 0.0100*** 0.0207*** 0.0137*** 0.0122*** 0.0126*** (0.0433) (0.9400) (0.0005) (0.0009) (0.0000) (0.0000) (0.0087) (0.0001) R&D_TA 0.1555* 0.0457 0.0887 0.0927** 0.1126* 0.0998** 0.0017 0.0861* (0.0536) (0.3050) (0.4560) (0.0223) (0.0874) (0.0238) (0.9760) (0.0604) IND_MEDIAN 0.0451*** 0.0640*** 0.0311* 0.0901*** 0.0230 0.0946*** 0.0140 0.0916*** (0.0020) (0.0002) (0.0935) (0.0000) (0.2180) (0.0000) (0.4000) (0.0000) REG 0.02 0.0036 0.0444* 0.0014 0.0234 0.0075 0.0080 0.0128 (0.2980) (0.7960) (0.0602) (0.9100) (0.1900) (0.5910) (0.6260) (0.3760) INF 0.1125*** 0.0177 0.0700* 0.0732 0.0040 0.0812* 0.0660 0.1196*** (0.0069) (0.7750) (0.0742) (0.1300) (0.9290) (0.0862) (0.1840) (0.0055) GDPG 0.2324*** 0.1410*** 0.0011 0.1394*** 0.1187*** 0.1375*** 0.1597*** 0.1185*** (0.0000) (0.0002) (0.9770) (0.0000) (0.0097) (0.0000) (0.0009) (0.0000) Observations 46 159 58,223 18,458 87,102 19,095 84,207 20,345 85,215

PAGE 62

62 Table 2 11. Continued ELIABS EPUBLICENF CMG PUBBUR Weak Strong Weak Strong Weak Strong Weak Strong BLEV 0.9102*** 0.8343*** 0.9479*** 0.8205*** 0.8997*** 0.8348*** 0.8954*** 0.8445*** (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) EBIT_TA 0.0524*** 0.0094 0.0348** 0.0193* 0.1032*** 0.0072 0.1269*** 0.0135 (0.0008) (0.3650) (0.0337) (0.0662) (0.0000) (0.4630) (0.0000) (0.1630) MB 0.0006* 0.0018*** 0.0004 0.0001 0.0006** 0.0027*** 0.0015*** 0.0003 (0.0667) (0.0010) (0.1230) (0.9180) (0.0306) (0.0001) (0.0062) (0.2070) DEP_TA 0.2196*** 0.2799*** 0.1754*** 0.3006*** 0.3838*** 0.2300*** 0.5454*** 0.2151*** (0.0002) (0.0000) (0.0018) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) LnTA 0.0009 0.0024*** 0.0010 0.0039*** 0.0024*** 0.0012** 0.0020 0.0030*** (0.3370) (0.0000) (0.1520) (0.0000) (0.0069) (0.0263) (0.1080) (0.0000) FA_TA 0.1380*** 0.0776*** 0.0502*** 0.0991*** 0.0780*** 0.0782*** 0.1416*** 0.0614*** (0.0000) (0.0000) (0.0067) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) TAXES 0.002 0.0009 0.0015 0.0030* 0.0074** 0.0003 0.0016 0.0016 (0.3990) (0.4270) (0.2490) (0.0585) (0.0189) (0.8090) (0.6090) (0.1400) LIQUID 0.0005 0.0009 0.0019* 0.0002 0.0023** 0.0012** 0.0002 0.0015*** (0.5730) (0.1210) (0.0947) (0.7640) (0.0108) (0.0483) (0.8770) (0.0075) RD_DUM 0.0103** 0.0142*** 0.0114*** 0.0007 0.0162*** 0.0131*** 0.0021 0.0106*** (0.0166) (0.0000) (0.0007) (0.8780) (0.0016) (0.0000) (0.6320) (0.0004) R&D_TA 0.0673 0.0749 0.1684** 0.0725 0.0269 0.0876** 0.0949 0.0296 (0.2350) (0.1060) (0.0259) (0.1190) (0.8390) (0.0295) (0.3100) (0.4860) IND_MEDIAN 0.0295* 0.0822*** 0.0343** 0.0518*** 0.0324* 0.0968*** 0.0696*** 0.0790*** (0.0844) (0.0000) (0.0219) (0.0011) (0.0832) (0.0000) (0.0002) (0.0000) REG 0.0072 0.0116 0.0016 0.0139 0.0367 0.0016 0.0172 0.0015 (0.7430) (0.4020) (0.9070) (0.3800) (0.1020) (0.8990) (0.4410) (0.9070) INF 0.0422 0.1943*** 0.0475 0.0309 0.0359 0.0798 0.0668 0.1577*** (0.4260) (0.0000) (0.3370) (0.5100) (0.3930) (0.1110) (0.2280) (0.0002) GDPG 0.2240*** 0.1411*** 0.2540*** 0.0344 0.0012 0.1114*** 0.0919** 0.1863*** (0.0000) (0.0000) (0.0000) (0.2470) (0.9740) (0.0001) (0.0263) (0.0000) Observations 23, 175 82,385 37,626 67,934 20,922 84,408 13,549 92,011 The definitions of the variables in the regressions are provided in Table 2 1. is firm is debt ratio in year t and in country j, is the adjustment parameter, 1 is a vector of firm, industry, and macroeconomic characteristics related to the costs and benefits of operating with various leverage ratios, is the unobserved firm heterogeneity captured by the firm dummies, and is the error term. The top row for each determinant reports the coefficient estimates. *, **, *** indicate significance at the 90%, 95%, and 99% confidence levels, respectively. The p values of the two tailed tests for the significance of the coefficients are reported beneat h the coefficient estimates in parentheses. I report the estimation results from Blundell and Bonds (1998) two step system GMM of the following model, run separately for each portfolio of countries form ed based on the sample median of the individual insti tutional indices using book leverage : = ( ) 1+ ( 1 ) 1+ + (2 3)

PAGE 63

63 Table 2 12. Correlations between Leverage and its Determinants for Weak, Strong, and All Institutional Settings WEAK INSTITUTIONS STRONG INSTITUTIONS ALL INSTITUTIONS (1) (2) (3) (4) (5) (6) + + + PANEL A. Number of Significant Correlations LEV 18 0 18 0 36 0 EBIT_TA 0 17 0 2 0 19 MB 7 2 0 11 7 13 DEP_TA 0 18 0 18 0 36 LnTA 11 0 18 0 29 0 FA_TA 18 0 18 0 36 0 TAXES 0 3 0 1 0 4 LIQUID 0 10 0 9 0 19 RD_DUM 14 0 16 0 30 0 R&D_TA 3 2 11 0 14 2 IND_MEDIAN 15 0 18 0 33 0 REG 2 1 1 0 3 1 INF 0 6 0 12 0 18 GDPG 9 3 1 15 10 18 WEAK INSTITUTIONS STRONG INSTITUTIONS ALL INSTITUTIONS (1) (2) (3) (4) (5) (6) + + + PANEL B. Core Factors LEV YES YES YES EBIT_TA YES YES MB YES DEP_TA YES YES YES LnTA YES YES YES FA_TA YES YES YES TAXES LIQUID YES YES YES RD_DUM YES YES YES R&D_TA YES IND_MEDIAN YES YES YES REG INF YES YES GDPG YES YES YES *Th e definitions of the variables documented in this table are provided in Table 2 1. Panel A provides a summary for the consistency of the direction of the relationship between leverage and each determinant. Panel B evaluates whether or not the leverage determinant is a dominant factor. The rows in Panel A report the number of instances out of a total of 18 different partitioning of the data made according to the institutional indices that the given determinant of leverage has a positive or negative significant coefficient at the 90% confidence level or higher in Table 2 11. The columns (1) (6 ) refer to the core estimation model (Equation 2 3) run separately for each portfolio of countries formed based on the sample median of the individual institutional indices using book leverage The columns (1) and (2) summarize the results of Table 2 11, in the weak institution column The columns (3) and (4) summarize the results of Table 2 11, in the strong institution column The columns (5) and (6 ), in Panel A, give the gross total of (1), (3) and (2), (4) respectively. Panel B assigns a leverage determinant as a core factor using a YES indicator if the score reported in Panel A is at least 9 for columns (1) to (4), and at least 18 for columns (5) to (6) = ( ) 1+ ( 1 ) 1+ + (2 3)

PAGE 64

64 Table 2 13. Empirical Relevance of the Capital Structure Theories for Weak, Strong, and All Institution al Settings Predicted Signs by the Capital Structure Theories Empirically Observed Signs (1) (2) (3) (4) (5) (6) DYNAMIC TRADEOFF PECKING ORDER MARKET TIMING WEAK STRONG ALL LEV + and (0,1) + and equal to 1 + and equal to 1 + and (0,1) + and (0,1) + and (0,1) EBIT_TA + or MB + DEP_TA LnTA + + or + + + FA_TA + + or + + + TAXES + LIQUID + or RD_DUM + + + + R&D_TA + + IND_MEDIAN + + + + REG + INF + or + GDPG + or + *The definitions of the variables in the regressions are provided in Table 2 1. The columns (1) (3) summarize the hypothesized predictions of the main theories of capital structure about how leverage relates to the observable firm, industry, and macroeconom ic attributes. The columns (4) (6 ) summarize the direction of the observed relationships between leverage and its determinants as is reported in Panel B of Table 2 12, for weak, strong, and all institutional settings respectively,

PAGE 65

65 Table 2 14. Bankruptcy Costs, Agency Costs, and Tax Benefits of Debt and the Determinants of Capital Structure (1) (2) (3) (4) (5) (6) Time Cost Efficiency Tax Creditor Formalism Leverage 0.0209 0.1239** 0.1156** 0.0622*** 0.0106 0.1265*** ( 0.6329 ) ( 0.0480 ) ( 0.0221 ) ( 0.0039 ) ( 0.2714 ) ( 0.0073 ) Profit 0.0186 0.1109 *** 0.0682 *** 0.0302 0.0136 0.0823 *** ( 0.3591 ) ( 0.0000 ) ( 0.0003 ) ( 0.2298 ) ( 0.4234 ) ( 0.0005 ) Market to Book 0.0003 0.0003 0.0016 ** 0.0003 0.0022 0.0007 ( 0.5738 ) ( 0.6925 ) ( 0.0133 ) ( 0.6049 ) ( 0.1307 ) ( 0.3596 ) Depreciation 0.1742 ** 0.0808 0.0098 0.0060 0.2318 *** 0.1311 ( 0.0436 ) ( 0.5012 ) ( 0.9121 ) ( 0.9481 ) ( 0.0075 ) ( 0.1779 ) Size 0.0012 0.0010 0.0002 0.0010 0.0006 0.0007 ( 0.3467 ) ( 0.4539 ) ( 0.8908 ) ( 0.4715 ) ( 0.6738 ) ( 0.6048 ) Tangibility 0.0554 *** 0.0111 0.0529 ** 0.0020 0.0179 0.0181 ( 0.0179 ) ( 0.6195 ) ( 0.0421 ) ( 0.9412 ) ( 0.4076 ) ( 0.6199 ) Tax 0.0020 0.0046 0.0002 0.0025 0.0026 0.0009 ( 0.4337 ) ( 0.2116 ) ( 0.9469 ) ( 0.2383 ) ( 0.3257 ) ( 0.7805 ) Liquidity 0.0006 0.0007 0.0013 0.0034 *** 0.0019 0.0016 ( 0.5316 ) ( 0.6337 ) ( 0.3282 ) ( 0.0051 ) ( 0.0677 ) ( 0.3336 ) R&D Dummy 0.0086 0.0035 0.0004 0.0081 0.0053 0.0005 ( 0.2851 ) ( 0.6559 ) ( 0.9433 ) ( 0.2069 ) ( 0.4179 ) ( 0.9333 ) R&D Expense 0.1230 0.0572 0.0368 0.1285 0.1904 ** 0.0077 ( 0.1914 ) ( 0.6778 ) ( 0.6826 ) ( 0.1635 ) ( 0.0290 ) ( 0.9367 ) Industry Median 0.0015 0.0727 0.0623 0.0127 0.0440 0.0281 ( 0.9563 ) ( 0.0565 ) ( 0.0567 ) ( 0.6113 ) ( 0.0504 ) ( 0.3991 ) Regulated Industry 0.0180 0.0646 0.0081 0.0071 0.0378 0.0186 ( 0.5725 ) ( 0.1602 ) ( 0.8103 ) ( 0.7916 ) ( 0.1334 ) ( 0.3920 ) Inflation 0.0712 0.1238 0.1184 0.1285 0.1803 ** 0.0485 ( 0.3331 ) ( 0.0898 ) ( 0.1016 ) ( 0.1451 ) ( 0.0157 ) ( 0.5005 ) GDP Growth 0.3445 *** 0.1431 ** 0.1737 *** 0.4355 *** 0.2064 *** 0.1990 *** ( 0.0000 ) ( 0.0255 ) ( 0.0032 ) ( 0.0000 ) ( 0.0001 ) ( 0.0001 ) *The definitions of the variables documented in this table are provided in Table 21. The table compares the magnitude of the impact of each determinant on leverage across weak and strong institutions. The columns (1) (6) refer to the institutional indices representing bankruptcy costs, agency costs, and tax benefits of debt. The t op row for each determinant reports the difference in the coefficient estimates across low and high quality of institutions (weak minus strong). The coefficient estimates are repor ted in Table 2 11 as a pair of columns for weak and strong institution portfolios for each particular institutional characteristic. *, **, *** indicate significance at the 90%, 95%, and 99% confidence levels, respectively. The p values of the two tailed tests for the significance of the difference in the coefficient estimates between weak and strong institutional settings are reported beneath the estimates. The empirical p values are obtained by the bootstrapping procedure and refer to the probability of obtaining observed differences in the coefficient estimates, if true coefficients are in fact equal, using 100 iterations.

PAGE 66

66 Table 2 15. Agency Costs of Equity and the Determinants of Capital Structure (1) (2) (3) (4 ) (5 ) (6 ) Antidir Prenf Executive Enforce Law&Order Governance Leverage 0.1141 *** 0.0974 *** 0.0718* 0.0890** 0.1237*** 0.0971** (0.0005) (0.0040) (0.0701) (0.0101) (0.0002) (0.0320) Profit 0.0816 *** 0.0478 0.0939 *** 0.1144 *** 0.0336 0.1059 *** ( 0.0002 ) ( 0.0424 ) ( 0.0003 ) ( 0.0000 ) ( 0.1614 ) ( 0.0002 ) Market to Book 0.0020 0.0018 0.0019 *** 0.0037 *** 0.0015 ** 0.0031 *** (0.0380) (0.1437) (0.0039) (0.0009) (0.0185) (0.0001) Depreciation 0.0319 0.1960 ** 0.2410 ** 0.2084 0.0918 0.1793 (0.7400) (0.0452) (0.0298) (0.1045) (0.3233) (0.1834) Size 0.0033 ** 0.0020 0.0017 0.0007 0.0016 0.0004 (0.0130) (0.0989) (0.1298) (0.5888) (0.2767) (0.7604) Tangibility 0.0880 ** 0.0184 0.0236 0.0124 0.0061 0.0374 (0.0109) (0.5811) (0.3071) (0.6214) (0.8029) (0.1426) Tax 0.0041 0.0024 0.0011 0.0053 0.0020 0.0051 (0.1279) (0.3544) (0.7150) (0.1450) (0.3998) (0.1951) Liquidity 0.0001 0.0007 0.0011 0.0014 0.0032 ** 0.0013 (0.9526) (0.6308) (0.3343) (0.3209) (0.0231) (0.3533) R&D Dummy 0.0042 0.0021 0.0076 0.0072 0.0062 0.0089 (0.4550) (0.7123) (0.2409) (0.3256) (0.3292) (0.1820) R&D Expense 0.0533 0.0102 0.1524 0.0392 0.2011 0.0039 (0.5374) (0.9103) (0.2973) (0.8709) (0.1308) (0.9806) Industry Median 0.0287 0.0217 0.0345 0.0540 0.0189 0.0591 (0.2583) (0.4337) (0.2948) (0.0981) (0.4630) (0.0546) Regulated Industry 0.0030 0.0364 0.0030 0.0276 0.0237 0.0430 (0.9253) (0.0864) (0.9387) (0.4884) (0.4163) (0.2456) Inflation 0.0185 0.0668 0.0083 0.0724 0.1302 0.0032 (0.8050) (0.3899) (0.9089) (0.3264) (0.1029) (0.9633) GDP Growth 0.4487 *** 0.3336 *** 0.2400 *** 0.1347 ** 0.3733 *** 0.1405 ** (0.0000) (0.0000) (0.0000) (0.0372) (0.0000) (0.0157) *The definitions of the variables documented in this table are provided in Table 21. The table compares the magnitude of the impact of each determinant on leverage across weak and strong institutions. The columns (1) (6) refer to the institutional indices representing agency costs of equity. The top row for each determinant reports the difference in the coefficient estimates across low and high quality of institutions (weak minus strong). The coefficient estimates are reported in Table 2 11 as a pair of col umns for weak and strong institution portfolios for each particular institutional characteristic. *, **, *** indicate significance at the 90%, 95%, and 99% confidence levels, respectively. The p values of the two tailed tests for the significance of the di fference in the coefficient estimates between weak and strong institutional settings are reported beneath the estimates. The empirical p values are obtained by the bootstrapping procedure and refer to the probability of obtaining observed differences in th e coefficient estimates, if true coefficients are in fact equal, using 100 iterations.

PAGE 67

67 Table 2 16. Information Asymmetry Costs and the Determinants of Capital Structure (1) (2) (3) (4) (5) (6) Accstds EDisclose ELiabs EPublicenf Insider PubBureau Leverage 0.1104 0.1148 *** 0.0759 *** 0.1274 *** 0.0649 ** 0.0509 (0.0804) (0.0002) (0.0024) (0.0000) (0.0134) (0.1489) Profit 0.0787 *** 0.0457 0.0430 0.0156 0.1104 *** 0.1134 *** ( 0.0031 ) ( 0.1020 ) ( 0.0655 ) ( 0.5736 ) ( 0.0000 ) ( 0.0006 ) Market to Book 0.0029 ** 0.0027 ** 0.0013 0.0003 0.0033 *** 0.0018 ( 0.0252 ) ( 0.0250 ) ( 0.0758 ) ( 0.7293 ) ( 0.0004 ) ( 0.0581 ) Depreciation 0.0051 0.2050 0.0602 0.1253 0.1538 0.3303 *** ( 0.9558 ) ( 0.0547 ) ( 0.5485 ) ( 0.1781 ) ( 0.1434 ) ( 0.0084 ) Size 0.0036 *** 0.0017 0.0033 ** 0.0028 ** 0.0013 0.0011 ( 0.0035 ) ( 0.2148 ) ( 0.0178 ) ( 0.0427 ) ( 0.2683 ) ( 0.5480 ) Tangibility 0.0395 0.0275 0.0604 0.0488 0.0002 0.0802 ** ( 0.2040 ) ( 0.4407 ) ( 0.1026 ) ( 0.1181 ) ( 0.9931 ) ( 0.0489 ) Tax 0.0031 0.0008 0.0029 0.0045 0.0072 0.0032 ( 0.3187 ) ( 0.7842 ) ( 0.2715 ) ( 0.0603 ) ( 0.0736 ) ( 0.4484 ) Liquidity 0.0004 0.0010 0.0004 0.0017 0.0011 0.0012 ( 0.8104 ) ( 0.5174 ) ( 0.7605 ) ( 0.1348 ) ( 0.3588 ) ( 0.4891 ) R&D Dummy 0.0069 0.0004 0.0039 0.0120 ** 0.0031 0.0127 ( 0.2203 ) ( 0.9486 ) ( 0.5422 ) ( 0.0394 ) ( 0.6158 ) ( 0.0993 ) R&D Expense 0.0127 0.0878 0.0076 0.2409 0.1144 0.0653 ( 0.8912 ) ( 0.3269 ) ( 0.9185 ) ( 0.0394 ) ( 0.5366 ) ( 0.6714 ) Industry Median 0.0716 ** 0.0776 *** 0.0527 ** 0.0175 0.0644 ** 0.0094 ( 0.0127 ) ( 0.0090 ) ( 0.0329 ) ( 0.4356 ) ( 0.0437 ) ( 0.7571 ) Regulated Industry 0.0159 0.0208 0.0188 0.0124 0.0351 0.0157 ( 0.5573 ) ( 0.4566 ) ( 0.6047 ) ( 0.5565 ) ( 0.3395 ) ( 0.6381 ) Inflation 0.0852 0.0536 0.2365 *** 0.0165 0.0439 0.0909 ( 0.2774 ) ( 0.4504 ) ( 0.0055 ) ( 0.8275 ) ( 0.5518 ) ( 0.2816 ) GDP Growth 0.2563 *** 0.2782 *** 0.3651 *** 0.2883 *** 0.1103 0.2782 *** ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) ( 0.0000 ) ( 0.0663 ) ( 0.0000 ) *The definitions of the variables documented in this table are provided in Table 21. The table compares the magnitude of the impact of each determinant on leverage across weak and strong institutions. The columns (1) (6) refer to the institutional indices representing information asymmetry costs The top row for each determinant reports the difference in the coefficient estimates across low and high quality of institutions (weak minus strong). The coefficient estimates are reported in Table 2 11 as a pair o f columns for weak and strong institution portfolios for each particular institutional characteristic. *, **, *** indicate significance at the 90%, 95%, and 99% confidence levels, respectively. The p values of the two tailed tests for the significance of t he difference in the coefficient estimates between weak and strong institutional settings are reported beneath the estimates. The empirical p values are obtained by the bootstrapping procedure and refer to the probability of obtaining observed differences in the coefficient estimates, if true coefficients are in fact equal, using 100 iterations.

PAGE 68

68 CHAPTER 3 PARTIAL ADJUSTMENT T OWARD OPTIMAL CAPITAL STRUCTURE AROUND T HE WORLD Motivation and Main Findings It is widely accepted that at least some firms have a target debt ratio and that at least some firms issue debt or equity with a target in mind (Graham and Harvey (2001)). The partial adjustment tradeoff model recognizes that capital market imperfections c reate rebalancing costs that hamper the speed of adjustment. Several studies indicate that dynamic tradeoff models dominate capital structure decisions, with firms actively pursuing target debt ratios, although market imperfections might lead to incomplete adjustment in any given period (Hovakimian, Opler, and Titman (2001), Leary and Roberts (2005), Flannery and Rangan (2006), and Strebulaev (2007)). The tradeoff theory maintains that market frictions originate a link between leverage and firm value, and that firms will pro actively adjust to deviations from their optimal debt ratio. The firms ability to reverse the deviations from their optimal leverage varies by the cost s and benefits of adjusting the leverage. On the other hand, the pecking order mark et timing, and inertia theories of capital structure do not propose any proactive effort to reverse changes in leverage. If firms around the world seek to attain an optimal capital structure, then we would observe differential adjustment speeds toward the ir optimal level conditional on the institutional setting they operate in. Estimating the adjustment speeds for diverse institutional environments that have differing adjustment costs and benefits for firms is an important step to improve our understanding of the competing theories of capital structure. This chapter aims to provide further insight to the theory of adjustment to optimal capital structure by examining the adjustment speeds across a large panel of countries with distinct institutional structur e s

PAGE 69

69 Using a dynamic panel dataset spanning 37 countries and 16 years, I illustrate that the adjustment is significantly faster in countries with better protection and enforcement of investor rights, stronger legal and financial systems, better capital market s and better corporate and political governance. I report that the m ost essential components of the adjustment costs are disclosure in the capital markets, corporate transparency, and shareholder protection; while the least critical factors are creditor rights, information sharing in debt markets, and mandatory dividends. I also document that the least significant component of the adjustment benefits is the tax shields, while other components are equally important. Among the adjustment cost factors, both the severity of external financing costs and the presence of cash con straints are important in determining the adjustment speeds, with differing impact on debt and equity contracts While investor protection in equity contracts is more crucial for the adjustment process the information asymmetry in debt contracts is more d etrimental for the cross country adjustment speed estimates. The evidence indicates that cash constraints on debt contracts such as reserve requirements have a more universal impact on adjustment speeds among different type of firms than the constraints on equity contracts such as mandatory dividends. Overall, in institutional environments where the adjustment costs are higher (above median), due to the severity of external financing costs and regulatory cash constr aints, the adju stment speed is 11%12% slower translating into a gain of 1.9 to 6.5 years in closing half of the distance to the target. Among the adjustment benefit factors, distress costs and deviation costs systematically and significantly affect the adjustment process, whereas tax shields have a relatively less important impact. In general, firms in institutional settings where the adjustment benefits are greater as implied by a greater ability to prevent distress and deviation costs and higher tax shields adjust

PAGE 70

70 8%9 % fas ter translating i nto a gain of 1.4 to 5.5 years in closing half of the distance to the target. These differences are economically and statistically highly significant. My findings suggest that the adjustment costs and benefits play an important role in the speed of convergence to optimal capital structure around the globe, consistent with the dynamic tradeoff theory, and that an important determinant of the capital structure adjustment process is the countrys legal and institutional heritage. An International Theory of Pa rtial Adjustment Several recent papers advocate that the financing policies are influenced by the institutional environment. Demirg -Kunt and Maksimovic (1999) compare the capital structure of firms from developed countries and developing countries and re port that institutional differences explain some of the variation in long term debt ratios In addition, Bancel and Mittoo (2004) illustrate in their survey of European firms that the capital structure choice in each country may be the result of many insti tutional features including its legal environment. Finally, Rajan and Zingales (1995) and Booth, Aivazian, Demirgc Kunt, and Maksimovic (2001 ) compare capital structures across different countries and conclude that further research is needed regarding th e influence of different institutional features on capital structure choices of firms. A notable feature of most existing studies on international capital structure is their implicit but unrealistic assumption that firms are always at equilibrium, operati ng at their optimal leverage, thereby assuming a constant and uniform speed of adjustment equal to unity. I drop this assumption and examine the cross -sectional determinants of the adjustment to optimal capital structure around the world.1 The purpose of t his chapter is to explore the importance of country 1 To understand my motivation, it is useful to illustrate the importance of taking into account the dynamic nature of the capital structure decisions and firm level heterogeneity in estimating adjustment speeds. A regression of firm leverage, measured as the book value of debt over total assets, on firm specific variables including the lagged dependent variable, year dummy variables and country dummy variables, has an adjusted R square of 0.78. When

PAGE 71

71 characteristics and the effect of the different types of institutional settings in explaining the adjustment to optimal leverage. Both intuition and theory suggest that institutional factors matter for t he adjustment to optimal capital structure. S ome institutional environments might be more conducive to writing and enforcing financial contracts than others. I directly test how differences in the contracting environment affect the adjustment to optimal leverage. The conjecture of this chapter is tha t firms might deviate from their optimal capital structures in the aftermath of a leverage shock and that the speed of adjustment back to the optimal level varies across different institutional environments based on the size of adjustment costs and benefit s in reaching and maintaining the target imposed by legal, financial, political, and regulatory differences. The adjustment to target leverage is expected to be faster in institutional settings with better protection of investors and better functioning of the markets, as significantly lower market frictions and inefficiencies would imply lower adjustment costs and/or higher adjustment benefits. The static tradeoff theory argues there is a link between leverage and firm value, and that managers will pro act ively rebalance leverage to fully offset any deviation from their optimal debt ratio. According to the dynamic tradeoff theory, t he speed at which managers will reverse the deviations from their optimal leverage varies by the cost and benefits of adjusting the leverage. To the degree that institutional features affect adjustment costs and benefits, variations the regression is run with all the variables except for the lagged dependent variable, the adjusted R square is substantially reduced to 0.37. This little exercise shows that the consideration of dynamic nature of financial leverage is crucial to be able to effectively explain the capital structure decisions. The inclusion of the lagged dependent variable doubles the overall explanatory power of the capital structure regressions, regardless of the exclusion of year and country dummies. Furthermore, a n examination of the effect of a one period lagged leverage on the current leverage would shed light on whether firms have a target capital structure and, if so, what is the speed of adjustment. A positive and below unity coefficient would imply that firms have a target leverage ratio and adjust their capital structure over time A coefficient equal to one w ould imply that firms always operate at their optimal leverage. A coefficient that is zero or greater than one would imply that firms do not have any target debt equity ratio.

PAGE 72

72 in these factors should influence the speed of adjustment. I now turn to the specific institutional factors of interest that might potentially affect the adjustment speed by recognizing that the costs and benefits of adjustment would vary with the institutional environment. Adjustment Costs When deviated from the optimum, a firm has two rebalancing options to revert back. If over -leveraged, it can retir e debt or issue equity. If under leveraged, it can repurchase shares or issue debt. In the presence of capital market frictions, these actions necessitate either external financing or financial flexibility. To the extent that the institutional environment is more conducive to easy access to the capital markets or greater financial flexibility, altering the leverage ratio to revert back to target becomes less burdensome, implying faster adjustment. External f inancing c osts The cost of external financing mig ht negatively affect the adjustment speed in cases where rebalancing requires debt and/or equity issuance. If managers are concerned about the direct costs of external financing, they might not rebalance frequently to reach the target capital structure, re sulting in slower adjustment. The cost of external financing would possibly vary with the ease of access to capital markets and the information asymmetry. The ease of access to capital markets is an important determinant of the costs of securities issuan ce (Faulkender and Petersen (2006). When firms require securities issuance to actively manage their capital ratio to achieve the optimal level, the cost of external financing may be a hurdle slowing the adjustment speed. If access to capital markets is bet ter, then firms can repeatedly adjust their debt or equity to reach their optimal leverage, rather than having to wait until access becomes available or relatively cheaper.

PAGE 73

73 The ease of access to capital markets should depend to a great extent on the rights attached to securities and their enforcement.2 Shareholders would not be granted dividends unless they could vote out the directors who do not pay them; creditors would not be paid interest and principal unless they could have power to seize the collater al or unless the firm waits to borrow again in the future. Investors would be reluctant to provide funds unless they get a return on their investment in exchange. Security laws may exist in countries yet not be effectively enforced. In addition to an appro priately designed legal code, there is a need for an efficient enforcement system to implement these rights, or at least to act as a credible threat. I use measures of shareholder and creditor rights and the quality of their enforcement to evaluate the imp act of easier access to capital markets on the capital structure adjustment. There is a substantial amount of research confirming that investor rights and their enforcement are important determinants for external financing costs. The growing law and finan ce literature argues that capital markets would function properly only when good secur ity laws exist (La Porta, Lopezde -Silanes, Shleifer, and Vishny (1998)). A common premise in this literature is that property rights directly influence the cost of exter nal financing ( La Porta, Lopez de -Silanes, Shleifer, and Vishny (1997), (1998), (2000 a), and (2000b) ). Well -functioning legal systems increase firms ability to raise external finance and to exploit growth opportunities by providing protection for outside investors and consequently reducing the parameter uncertainty and the estimation risk that can be non diversifiable to a certain extent (Barry and Brown (1985), Coles, Loewenstein, and Suay (1995), Lam bert, Leuz, and Verrecchia (2007 )). Rajan and Zingales (1995) suggest that strong creditor rights enhance exante contractibility. Claessens, Djankov, and Larry (2000) and La Porta, Lopez -de Silanes, Shleifer, 2 Reputation effects could substitute for legal rights. However, La Porta, Lopez de Silanes, Shleifer, and Vishny (2002) show that legal rights are empirically important.

PAGE 74

74 and Vishny (2002) argue that investors pay more for equity when legal institutions effectively prote ct their rights. Qian and Strahan (2007) examine the influence of creditor protection and legal origins on the terms and pricing of bank loans and show that loans have more concentrated ownership, longer maturities, and lower interest rates in countries wi th stronger creditor protection. Bae and Goyal (2004) examine how property rights affect loan spreads and find that banks charge higher loan spreads when property rights are weaker and conclude that the improvement in the cost of external financing will be greater with policies that improve property rights protection at the country level than with policies that aim at improving governance mechanism at the firm level. Furthermore, several studies emphasize the enforcement of investor rights as an effective m echanism that reduces external financing costs ( La Porta, Lopez -de Silanes, Shleifer, and Vishny (1997, 1998), Levine (1999), and Hail and Leuz (2006)) and argue that the effective enforcement of laws matters more than the quality of laws itself (Berkowitz, Pistor, and Richard (2003)). Information asymmetry is another factor of interest as it negatively affects capital structure rebalancing by increasing the difficulty of issuing securities and creating a wedge between internal and external financ ing costs. Myers (1984) and Myers and Majluf (1984) advocate that information asymmetry between managers and outside investors leads to a preference ranking over financing sources.3 In order to minimize adverse selection costs, firms first issue internal f unds, followed by debt and then equity. Myers (2003) posits that the pecking order behavior can also be generated by agency costs stemming from asymmetric information. When firms depend on external financing to correct their deviation from optimal leverage subsequent to a leverage 3 Heatons (2002) analysis also generates a pecking order from the simple assumption of managerial optimism.

PAGE 75

75 shock, the costlier financing under information asymmetry makes continuous adjustment suboptimal, hampering the adjustment to optimal leverage I use an estimate of a countrys quality of accounting standards to proxy for asymmet ric information arising from lack of corporate transparency. High quality accounting standards are needed to render company disclosures transparent and interpretable. Any financial contract between managers and investors usually relies on some verifiable a ccounting measure such as firms income or assets. Also, better accounting information may help investors distinguish between good and bad investments and managers, lowering the adverse selection costs and consequently decreasing the cost of external finan cing. There are several papers suggesting that the quality of accounting standards and the quality of disclosure in general decrease the external financing costs by moderating the degree of information asymmetry among the financial agents (Verrecchia (2001 ), Amihud and Mendelson (1986), Merton (1987), Lomb ardo and Pagano (2002), Lambert, Leuz, and Verrecchia (2007 )). In addition to accounting standards, I focus on information asymmetry proxies for equity and debt markets separately. For the equity markets, I use the regulation of security laws governing initial public offerings, with a focus on mandatory disclosure liability standards, and public enforcement The law in some countries can require the disclosure of particular information in the prospectus t o facilitate the evaluation of companies by investors. The law can also specify liability standards for issuers and intermediaries who fail to reveal material information mandated to be disclosed. Finally, an independent public enforcer such as the Securit ies and Exchange Commission can secure information from issuers and investors and impose sanctions. I also use a measure of the quality of capital market governance with a focus on the insider trading laws in individual exchanges around the world. Daouk Lee, and Ng

PAGE 76

76 (2006) argues that with weak insider trading laws large controlling shareholders would have incentives to profit from stock tips from the management rather than monitoring. Enforcement of the insider trading laws as well as its existence is req uired to deter the insiders. F urthermore, Bhattacharya and Daouk (2002) show that the transaction costs are higher in stock markets where insiders trade with impunity. For the debt markets, I use the presence of public credit registries that collect inform ation on credit histories and current indebtedness of various borrowers and share it with lenders. Stiglitz and Weiss (1981) propose that when lenders have knowledge about the borrowers or other lenders to the firm, the moral hazard problem of financing no n -viable projects will be less prominent. Cash c onstraints Internal cash availability may affect the adjustment speed. If cash is ample, the only relevant decision for adjustment is whether to repurchase equity or repay debt. However, if internal financing constraints are binding, it may impede the ability to reba lance, depending on the severity of the external financing costs. Moreover, as with free cash flow financial slack removes the need or reduces the cost of external financing, which may affect the speed of adjustment. A cash constraint or financing constra int denotes the amount of cash flow that is available for financing activities meaning that when it is binding, the firm would need to raise external funds to cover the shortfall. Dividends are a principle financing constraint reducing cash availability a nd restricting the ability to modify leverage (Myers (1984), Jensen (1986), Fama and French (2002)). In some countries, firms are required by law to pay out a certain fraction of their stated earnings as dividends. Cash constraints would be more binding in institutional settings where the company or commercial code requires firms to distribute some percentage of net income as mandatory dividends among ordinary shareholders. I use the existence of mandatory dividend payments to

PAGE 77

77 evaluate if and how financia l constraints affect capital structur e adjustment. La Porta, Lopez -de Silanes, Shleifer, and Vishny (2000a ) have shown that mandatory dividend policy can be recourse for constraining agency problems by limiting the free cash flow to managers. As an additional measure, I use the existence of a legal reserve requirement a requirement that forces firms to maintain a certain level of capital to avoid automatic liquidation in order to protect creditors before all the capital is stolen or wasted by the insiders. Adjustment Benefits The speed of rebalancing also depends on the benefits of adjustment to optimal leverage. Convergence to optimal leverage is most valuable in institutional settings where the suboptimal leverage leads to either higher distress costs, e specially for the over leveraged firms, or prevents gains from tax shelters, in particular for the under leveraged firms. Alternatively, loan and debt covenants may increase the benefits of adjustment to the optimal leverage by penalizing deviations from t he target. Costs of financial d istress Financial distress and mechanisms for its resolution are an important element of the adjustment process. Adjustment benefits should be higher in countries imposing higher financial distress costs on firms ex ante sin ce the divergence from target is likely to be more costly. Similarly, the value of maintaining and reaching the target should be higher in countries that manage more efficient outcomes for firms in default ex post since the deadweight costs will be less si gnificant, sacrificing only a smaller proportion of a firms assets that can be devoted to capital structure adjustments if the firm is to continue to operate. Insolvency codes and court mechanisms governing default on debt contracts should affect the effe ctiveness of resolution of systemic and non-systemic financial distress. I use the design of the bankruptcy codes and debt contracts, including the attached creditor rights and the associated enforcement mechanisms

PAGE 78

78 governing default on debt contracts as determinants of the financial distress costs. In countries where lenders can easily force repayment, repossess collateral, gain control of the firm, or enforce debt contracts, the value of approaching the target and m aintaining it must be higher due to more binding ex ante distress costs. Furthermore, firms from countries that administer the bankruptcy process in court in a manner that is less time consuming, less costly and more efficiently will adjust more rapidly to their target due to lower deadweight costs associated with the insolvency process. The value of tax s hields Debt tax shields play an important role in the adjustment to optimal capital structure (Graham (1996)). The tax benefits of leverage should increas e the value of reaching and maintaining the leverage target especially for those firms that are under -leveraged I use the effective corporate tax rates to evaluate the effect of the value of tax shields on the adjustment decision. The effective corporate tax rate is the rate the company has to pay on marginal income, taking into account federal, state, and local taxes and all available deductions. Costs of d eviation Adjustment benefits should be higher when the value of maintaining and reaching the targe t is higher due to more binding costs imposed when diverged from it. The loan and debt covenants can be considered as mechanisms that would increase the benefits of adjustment to the target for some firms, for example by imposing penalties for over leverag ed situations. The penalties for deviation would be effective only to the degree they are efficiently enforced. I use the executive quality, the strength of law and order the quality of government ( referred to as governance) and the quality of contract en forcement to proxy for external pressure to correct any suboptimal leverage situation.

PAGE 79

79 Data For the tests of my hypotheses, I use the same data as in Chapter 2. I construct my firm level sample from all non-financial firms included in the Compustat Globa l Vantage database between the years 19912006. As my regression specification includes lagged variables, I exclude any firm with fewer than two consecutive years of data from my analyses. In addition, I exclude firm -year observations with missing financia l and accounting data required for the firm level analysis. To minimize the potential impact of outliers, I winsorize the firm -level variables at the 1st and 99th percentiles. The final sample consists of 15,177 firms from 37 countries, totaling 105,568 fi rm -years. The country -level structural and institutional data comes from different sources. Empirical Methodology I employ a general partial adjustment model that permits each firms optimal leverage to vary over time and according to its characteristics and that allows the deviations from target leverage not to be offset immediately. It would be unreasonable to impose the assumption that firms always operate at their optimal leverage ratios. Following Flannery and Rangan (2006) and others, I use a partial adjustment model that incorporates rebalancing costs that may slow down the firms adjustment to its optimal leverage. 1= 1 + (3 1 ) where is firm is debt ratio in year t and in country j; is firm is desired debt ratio in country j and in year t; In Equation 3 1, is the adjustment parameter representing the magnitude of the adjustment If equals one, then all the gap between the observed and optimal leverage is closed for that firm that period, leading the actual leverage to equal the optimal leverage. In the

PAGE 80

80 presence of market frictions, will be less than one and th e firm will only be able to eliminate a proportion of the gap between its actual and optimal leverage. Many papers examine the factors that determine the target leverage including Hovakimian, Opler, and Titman (2001), Frank and Goyal (2005, 2007), Flannery and Rangan (2006). I follow the existing literature for the selection of firm specific factors affecting target leverage but also incorporate country -specific macroeconomic factors that are theoretically important in a firms determination of target leverage (Frank and Goyal (2007), Korajczyk and Levy (2003)). Furthermore, Flan nery and Rangan (2006), Huang and Ritter (2009) and Lemmon, Roberts and Zender (2008) stress the importance of including firm dummies for an unbiased estimation of firm targets. Accordingly, each firms optimal leverage in a particular country or instituti onal setting is modeled as a function of both the observed firm characteristics, X and the unobserved firm heterogeneity, .4 I model the target leverage allowing the possibility that target leverage might differ across firms and over time by specifying a target leverage ratio of the form: = 1+ (3 2 ) where and are coefficient vectors to be estimated. 1 is a vector of firm characteristics related to the costs and benefits of operating with various leverage ratios. Under the tradeoff theory, 0 and the variation in is non trivial. Substituting Equation 2 2 into Equation 3 1 and re arranging yields: 4 Frank and Goyal (2007), in the context of a single country, the U.S., argue that macroeconomic factors like expected inflation are an important factor for leverage determination, but not as important as firm specific factors like tangibility or firm size. However, in a cross sectional comparative study of a large panel of countries, it is reasonable to presume that excluding these macroeconomic factors from the estimation equation can cause non trivial bias in calculating leverage targets. As Frank and Goyal (2007) illustrate, the use of the term spread instead of inflation or the use of the stock market index instead of GDP growth yield very similar results for the target estimation. I only report the results with inflation and GDP growth.

PAGE 81

81 = ( ) 1+ ( 1 ) 1+ + (3 3 ) Equation 3 3 constitutes a typical partial adjustment model of capital structu re. The base adjustment speed can be obtained from the coefficient on the lagged dependent variable 1, by simply subtracting it from one ( 1 ( 1 ) This specification assumes that all s ample firms adjust uniformly at the constant rate, The d ynamic panel model in Equation 3 3 requires instruments for the endogenous transformed lagged dependent variable (Baltagi (2001)) and a correction for the short panel bias (Bruno (2005)). Flannery and Hankins (2007) investig ate the efficiency of several dynamic panel estimators in the presence of firm fixed effects and short unbalanced panels and conclude that Blundell Bond (1998) generalized method of moments estimation (BB) and Brunos (2005) bias corrected least squares du mmy variable approach (LSDVC) appear least sensitive to the panel length and yield superior estimates of the adjustment speeds.5 Accordingly, I use two -step system GMM and the bias -corrected least squares dummy variabl e approach to estimate Equation 3 3 My analysis employs measures of both book leverage and market leverage.6 I define book leverage using the book value of firm assets: = + (3 4 ) 5 Flannery and Hankins (2007) do not evaluate the long difference estimator of Hahn, Hausman, and Kuersteiner (2007). Hahn, Hausman, and Kuersteiner (2007) show that the long difference estimator does better than the system GMM estimator of Blundel l and Bond (1998) when the true adjustment speed parameter is sufficiently close to zero, in which case the system GMM estimator is upward biased. Unfortunately, the panel length for most countries does not allow me to use the long differencing estimator. It is worthwhile to note that my results would be even stronger using the long differencing estimator since for the slow adjusters the long differencing estimator would yield even smaller adjustment speed estimates, resulting in sharper differences across slow and fast adjusters. 6 The results for book and market leverage in U.S. context are generally comparable, even though in general researchers ha ve focused more on market value debt ratios ( e.g. Hovakimian et al. (2001), Fama and French (2002), Welch (20 04), and Leary and Roberts (2005). Since no other study has ever examined the adjustment speed and the determinants of capital structure for such a large panel of countries, I report the estimates using both book and market leverage measures for comparativ e purposes

PAGE 82

82 Market leverage is defined using market valued instead of book valued assets: = + + (3 5 ) Test D esigns To conduct my tests, I first classify the sample countries into two distinct portfolios according to the median value of various indices representing the quality of legal, financial or po litical institutions. My aim is to be able to compare the speed of adjustment in environments with either strong or weak institutions. If the institutional attribute consists of an indicator variable, I group the sample countries according to the presence or absence of that attribute, hence according to whether the indicator variable equals to zero or one. One of the many institutional features I test in this study is the level of investor protection. To test whether the level of investor protection matters for the adjustment process, I allocate a country into either a portfolio of strong (above median) investor protection or weak (below median) investor protection. I then perform two types of analyses, SEPARATE and POOLED, to assess the direct impact of the investor protection on the adjustment speed ( ) around the world. The SEPARATE method allows different adjustment speeds ( s ) and determinants of capital structure ( s ) for all sample countries but ignores the similarities of countries sharing same institutional features, whereas the POOLED method allows different s and s for countries dissimilar in their institutional attribute, but constrains them to be equal if they share the same institutional characteristics.7 7 In unreported results, I have experimented with a third method to test s across different institutional settings as a variant of the SEPARATE analysis, by using the estimated adjustment speeds obtained from E quation 33 as the dependent v ariable in cross sectional regressions on the institutional adjustment cost and benefit factors and found similar results (available upon request). Note that this methodology may be subject to error in variables problem and therefore is likely to be inferior to the SEPARATE or POOLED method.

PAGE 83

83 Separate Tests For the SEPARATE methodology, I use one countryat a time estimates of the adjustment speed to evaluate the average speed of adjustment in countries with strong investor protection and weak investor protecti on. This method therefore involves estimating separately regression s using Equation 3 3 to obtain an estimate of each countrys and then averaging these estimates for each group formed based on that particular institutional characteristic. I expect to f ind that the average is significantly different across strong and weak investor protection portfolio s .8 Pooled Tests For POOLED methodology, I re run my estimation specification Equation 3 3 separately for the strong and the weak investor protection portfolios and compare the estimated s from the two regressions. This method involves combining the data across countries, imposing common slopes within each portfolio, allowing for varying intercept by firms and estimating pooled regressions for the c ountries involved in the same portfolio to arrive at a single for all countries with a given economic/legal conditi on.9 Again, I expect to obtain a estimate that is different for the strong investor protection portfolio and the weak investor protecti on portfolio.10 8 For the separate tests, I report the results from t tests with the assumption of unequal variances after confirming that the assumption of equal variances would be inappropriate for the data and confirm the robustness of the results using nonparametric rank (Wilco xonMann Whitney) and K sample median tests. 9 I experimented augmenting the regression specification by including country dummies in the estimation equation. The country dummi es turn out to be insignificant and the results are very similar with the case of their exclusion. This is consistent with the two factor analysis of leverage ratio which shows that the between country variation (3%) in the sample is trivial, and that most of the variation stems from between firm variation (70%). 10 For the pooled tests, I conduct Chow tests and the bootstrapping procedure for testing the equality of the coefficients across two groups. Bootstrapping provide consistent standard errors from the regression model by resampling the original data, applying the regression model to obtain a sample of coefficient estimates using this new dataset by repeating this procedure a number of times. T he empirical p values obtained by the bootstrapping procedure estimates the probability of obtaining observed differences in coefficient estimates, if true coefficients are in fact equal using 100 iterations I have experimented with 50, 100, and 250 iterations, and obtained qualitatively similar results. For eas e of exposition, after confirming the robustness of the results using either method, I report the results using the Chow test.

PAGE 84

84 Comparison of Separate and Pooled tests Each test design has its own merits. Averaging individual country regressions as in the SEPARATE method allows for full heterogeneity in parameter estimates and error variances, but forfeits any effici ency gains from pooling countries with similar institutional characteristics since certain parameters might be equal across countries within the same group. The SEPARATE method could also be weaker for countries with small cross -sectional dimension, since dynamic panel estimators are designed for situations with large cross -sectional and small time dimensions. In countries where the number of firms is limited, the parameter estimates might be biased and unstable.11 The portfolio regressions as in the POOLED method assumes slope and error variance homogeneity suggesting that the gains from pooling would outweigh any costs imposed for not taking into account the inherent heter ogeneity in the slope estimates Analysis and Results Partial Adjustment Around the World International adjustment s peeds To explore the existence of dynamism in capital structure decisions, I start by documenting the individual country estimates of adjustment speed to optimal leverage around the world. A major contribution of this study is to provide an extensive cross -country comparison of the adjustment speeds for a large cross -section of countries under a uniform and appropriate econometric methodology by taking into account the potentially dynamic nature of the firms capital structu re and its unobserved heterogeneity. Our knowledge of financing choices of firms 11 One source of bias is weak instruments proble m (Nelson and Startz (1990)). Another source of bias is the many instrum ents pr oblem (Tauchen (1986)). I specified the instrument set efficiently without employing all available instruments to minimize the potential bias. I also conducted my tests using the weighted average of the one country at a time estimates of the adjustment sp eeds and found similar results. Weighting of the adjustment speed estimates by the number of observations in each country should reduce the small sample bias associated with some countries that have small number of firms in their panel.

PAGE 85

85 has mostly been derived from studies based on U.S. data. The comparative analyses of capital structure in an international scope have either analyzed the static nature of firm s financial decisions implicitly using the observed debt ratio as a proxy for the optimal debt ratio, or ignored unobserved firm heterogeneity inherent in capital structure data. Using a static framework and imposing the restriction that the observed debt ratio can proxy for the optimal debt ratio has been shown to be particularly problematic if there are significant adjustment costs impeding firms from fully adjusting to their optimal long term leverage targets.12 Furthermore, unobserved firm level heterog eneity is an integral building block of the capital structure theory (Flannery and Rangan (2006), Lemmon, Roberts, and Zender (2008), Huang and Ritter (2009) ) and if ignored, it may result in a non -trivial omitted variable bias in capital structure estimat ions.13 Table 3 3 summarizes the adjustment speed estimates for each country. I document the adjustment speed estimates for book and market leverage using two alternative estimation procedures. I estimate a speed o f adjustment in the range 24%27% for book leverage and 25 % 28% for market leverage for U.S. firms. These estimates are consistent with L emmon, Roberts, and Zender (2008) who find a speed of adjustment of 25% per year for book leverage by 12 An exceptional s tudy dealing with international capital structure and documenting the dynamic adjustment to optimal leverage employing a system GMM and incorporating firm fixed effects to target equation is by Antoniou, Guney, and Paudyal (2008). However their sample is l imited to a smallcross section of G5 countries and they dont investigate the determinants of the adjustment process. Very recently I have come across a very recent working paper by Clark, Francis and Hasan (2009), investigating the relevance of country level variables on the adjustment to optimal capital structure. My study differs in its scope and focus from their work in two ways. It first explores the corporate financing patterns around the globe and provides considerable evidence for global dynamic r ebalancing and then looks at the impact on the global adjustment of core and fundamental institutional factors that are shown in the law and finance literature to determine most of the underlying proxies these authors use such as private credit to GDP or equity market size. I believe that including the institutional factors along with the output variables as in Clark, Francis and Hasan (2009) in the estimation specification may have confounding effects and bias the results since institutional variables are shown to determine financial outcomes. 13 A preliminary (but untabulated) analysis shows that the importance of firm fixed effects in the capital structure regressions as documented in the literature in the context of a U.S. similarly holds in an international context. One factor analysis of variance for book leverage shows that firm fixed effects explain about 60% 70% of the total variation in leverage in the combined sample. The contribution of firm fixed effects to the variance of book leverage in individual country samples varies from 50% (Hong Kong) to 85% (Japan). Ignoring this heterogeneity would se riously bias the estimates of optimal leverage and adjustment speed.

PAGE 86

86 employing a system GMM estimation incorporating firm fixed effects.14 The overall sample mean (median) adjustment speed estimate for book leverage is 21% (21%) and for market leverage 26% (25%) using system -GMM compared to 23% (25%) for book leverage and for market leverage 30% (31%) using LSDVC. This suggests that in an average year, the gap between actual and optimal capital structure is closed by about one -fourth in the sample. The estimation results reveal a positive and significant effect of the lagged dependent variable on the capital structure of firms in all sample countries. The coefficient estimate of the lagged dependent variable lies in zero -one interval in each country and significantly differs from zero, indicating that the capital structure converges to its optimal level over time. This finding confirm s the presence of dynamism in the capital structure decisions of firms from all of the sample countries; in the sense that firms alter their leverage to reach their optimal target and that the actual leverage would be an inappropriate proxy for the optimal leverage. Comparing the top ten and bottom ten adjusters of capital structure in the sample countries, I find that the ranking of countries for the adjustment speed is also fairly consistent across the two estimation methods, especially considering book l everage. A visual representation of the adjustment speed estimates obtained using different methodologies and definition of financial leverage is provided in Figure 3 1 Comparing the efficiency of the two estimation methods, I observe that the standard d eviation in adjustment speed estimates is also fairly similar across the two techniques. For book leverage, LSDVC y ields a standard deviation of 10% compared to 9 % with system GMM, which 14 The results are also consistent with Huang and Ritter (2009) who document a speed of adjustment of 21% (22%) for U.S. firms using the change in book (market) leverage between the end of year t and the end of year t 4 as the dependent variable, years 1972 to 2001 as the sample period, and the long differencing estimator as the empirical methodology.

PAGE 87

87 suggests that there is no significant tradeoff for efficiency using t he two alternative estimation methods.15 The individual country adjustment speeds largely reflect the magnitudes of the adjustment costs and/or benefits. Slow (fast) adjusters have high (low) adjustment costs and/or low (high) adjustment benefits on overal l or on several dimensions of my individual indices. I will frequently revisit this issue later in the study Capital structure a djustments The dynamic tradeoff theory posits that firms would wait to recapitalize and stay away from their target for prolonged periods of time if the costs of capital structure adjustments outweigh the benefits. Accordingly, the existence of transaction costs is a potential explanation for partial adjustment to optimal capital structure. The distribution of the adjustm ent spee d estimates range from 4 % to 53% in the sample. Therefore, there is a tremendous cross -sectional variation in the adjustment speeds, consistent with the view that the cost of being off target relative to the cost of adjustment is not the same across all co untries. The lack of significant differences in adjustment speeds across the countries in the sample would be questionable under dynamic tradeoff theory in the presence of varying costs and benefits of adjustments. The rebalancing behavior in the sample wo uld be consistent with dynamic tradeoff theory if slower adjustment is associated with fewer and/or less significant (successful) efforts of active rebalancing due to higher transactions costs. Furthermore, if there are substantial costs associated with ac tive rebalancing, then conditional on incurring these costs, I would observe higher adjustment speeds for firms accessing external markets since it is only through their operating 15 The overall samples cross sectional standard deviations in adjustment speed estimates are 7% for book leverage using LSDVC, and 6% using system GMM respectively. After establishing the consistency of both estimation methods in this section, I continue to report the results using only the Blundell and Bond system GMM for ease of exposition and confirm the robustness of the analysis to using the bias corrected least squares dummy variable approach.

PAGE 88

88 cash flows and payout policy that their actual leverage could have changed t o reach the target capital. To test whether fast adjusting countries make more frequent rebalancing activities and/or in bigger amounts, I perform a simple analysis, where I categorize slow (fast) adjusters as countries with adjustment speed smaller (large r) than the median of estimates that are documented in Table 3 3 column 1. I then define a capital structure adjustment as occurred when the net change in equity or debt, normalized by the value of assets at the end of the previous period is greater than a 5% of total assets.16 I categorize four basic types of financing following the a pproach used by Hovakimian, Opler, and Titman (2001), Korajczyk and Levy (2003) and Leary and Roberts (2005). The frequency of adjustments denotes the proportion of the sample for which there has been an occurrence of an external financing activity. The size of capital structure adjustments refer to the magnitudes of two basic financing activities: debt and equity either in the form of an issue or retirement. Table 3 4 and Figure 3 2 document and compare the frequency and size of capital structure adjustments across the sample countries and the two groups of slow and fast adjusters Specifically, the last two rows of Panel C of Table 3 4 illustrate that slow adjusting countries make less frequent adjustments and in smaller magnitudes consistent with presence of adjustment costs (Leary and Roberts (2005)) and these differences are statistically significant .17 A slow adjuster on overall is 8 % less likely to have recourse to external capital markets to 16 The choice of the cut off of 5% is not material for the results. Alternative cut offs (e.g. 1.5%, 2%, 2.5, 3% 3.5% ) lead to the same conclusion. 17 The standard deviation of actual and target leverage ratios as well as their absolute deviation from the optimal leverage is greater for the slow adjusters compared to the fast adjusters. In other words, the reason for slow adjustment is not the absence of leverage shocks among the slow adjusters. My analyses support the view that slow adjusters are subject to leverage shocks, they arrange their target accordingly, but they fail to adequately rebalance their deviation from the optimal for prolonged periods of time due to binding adjustment costs. Figure 22 in provides a visual representation of the frequency and size of capital structure adjustments.

PAGE 89

89 correct any sub optimal leverage situation. This effect is slightly stronger for access to equity markets (9%) compared to debt markets (7%). Furthermore, the frequency of issuances rather than retirements are more vulnerable to the presence of transaction costs and benefits in both debt and equity markets. While the debt (equity) issuances are 4 % (4%) less frequent for slow adjusters, debt retirements are only 2% (3%) less common. The overall re sults are very similar for the magnitudes of the capital structure adjustments. In line with my hypothesis, slow adjusters are able to devote a smaller proportion of their book assets to the capital structure adjustments compared to fast adjusters. The mag nitude of this effect is fairly comparable across debt and equity markets and different type of issuance and retirement activities. To assess whether the rebalancing behavior is consistent with the implications of costly adjustment, I carry out a straight forward test. I categorize the sample firms according to whether they have access to external capital markets or not and compare the estimated adjustment speeds across the two groups to determine whether transactions costs prevent firms from fully adjusting to their optimal capital structure. Access to external markets is defined as a change in absolute value of outstanding debt or equity exceeding 5% of total assets. I find that firms raising significant external capital adjust to optimal capital structur e faster than those who do not. Firms accessing the capital marke t adjust at an annual rate of 25%, versus 9 % for firms that do not access capital markets.18 This difference is statistically and economically very significant. The estimation results therefor e confirm that the fast adjusters have indeed less binding transaction costs and/or larger benefits of being near target compared to slow adjusters in the sample, providing support to the dynamic tradeoff theory. In other words, firms around the world beha ve 18 I use the POOLED method to preserve a reasonable sample size for the tests. A comparison of firms acces sing the debt and equity markets and those who do not separately yields very similar results. Firms accessing the debt (equity) markets adjust at an annual rate of 18% (20%), versus 1% (4%) for firms that do not.

PAGE 90

90 consistent with actively managing their capital structure to reach their optimal, but observed capital structures appear to deviate from desired leverage, due to the presence of significant adjustment costs and/or benefits. The Impact of the Legal and F inancial Traditions on the Adjustment Speeds Having established that managers in each country assess the tradeoff between the cost of adjustment and the cost of being away from the target while making their capital structure decisions, I next conduct a pre liminary test to see if the speed at which they alter their capital structure depends on the characteristics of the legal and financial traditions of their country. Legal t raditions The law and finance literature establishes a link between the legal and i nstitutional variables and financial outcomes. La Porta, Lopez -de Silanes, Shleifer, and Vishny ((1997), (1998)) rank countries according to their protection of investors and quality of institutions, including legal enforcement, corruption and risk of cont ract expropriation. Countries with English (common) origin provide the strongest legal protections to both shareholders and creditors and have better institutions while countries with French (civil) origin provide the weakest protection and worst instituti ons. The law and finance view implies that the adjustment costs are lower and/or adjustment benefits are higher in common law originated countries. Hence they should revert back to their optimal leverage more quickly compared to civil -law originated countr ies. To test this proposition, I first allocate countries to legal origin portfolios. In the sample, 14 countries belong to an English common law origin, 15 belong to a French law origin, 5 belong to German origin and 3 belong to a Scandinavian law origin. The civil law origin consists of counties from German, Scandinavian and French law traditions, and comprises a total of 23 countries. I then compare the adjustment speed obtained for different groupings using SEPARATE and POOLED methods.

PAGE 91

91 The results of t his analysis are illustrated in Table 3 5 Panel A. Consistent with my proposition, using the SEPARATE (POOLED) method, I find that the common law originated countries adjust to optimal capital structure with an average rate of 9 % (14%) faster than civil l aw originated countries and that this difference is strongly significant. Specifically, the English law originated countries adjust at a mean speed of 2 7 % (19 %) whereas the Civil law originated coun ties adjust at a mean speed of 18 % (5 %). The difference in adjustment speeds is largest between English law origin and French origin, consistent with La Porta, Lopez de -Silanes, Shleifer, and Vishny (1997, 1998, and 2002) that document weaker property rights, enforcement and disclosure standards in countries foll owing civil law tradition. Moreover, the ranking of the adjustment speeds is consistent with the hypothesis that better quality of institutions as implied by legal origin would result in faster adjustment. Just as La Porta, Lopez -de -Silanes, Shleifer, and Vishny (1997, 1998, 2002) find that best institutions are located in English, followed by Scandinavian, and finally by German and French Law origins, using SEPARATE method, I find fastest adjustment with book leverage in English ( 27%), followed by Scandina vian (25%), German ( 22%) and finally French ( 15%) law origin. The qualitative result is similar with POOLED method but differences are even sharper. The German (6 %) and French (6%) origins are the slowest and the English ori gin is the fastest to adjust (19 %) with Scandina vian origin lying in between (15 %). These differences are statistically and economically very significant. Financial s ystems Another debate in the finance literature is whether the bank -based or market -based financial systems lead to better functioning of the markets. There could be relative merits associated with each. The market -based view highlights the role of well -functioning markets in increasing liquidity (Holmstrom and Tirole (1993)), enhancing corporate governance (Jensen and Murphy (1990)) and facilitating risk management (Levine (1991), Obstfeld (1994)). This view

PAGE 92

92 maintains that banks may hinder innovation by protecting established firms that are their source of luxurious informational rents (Hellwsig (1991), Rajan (1992)) and that markets act to offset the inefficiencies associated with banks. The bank -based view stresses the positive role of banks in collecting information about firms and managers and thereby improving capital allocation and corporate governance (Diamond (1984), R amakrishnan and Thakor (1984)), managing cross sectional and liquidity risk (Allen and Gale (1999), Bencivenga and Smith (1991)), mobilizing capital to take advantage of economies of scale (Sirri and Tufano (1995)), ameliorating moral hazard through effect ive monitoring (Booth and Thakor (1997)). This view maintains that greater market development may hamper corporate control since in liquid markets investors will have less incentive of recourse to corporate control (Rajan and Zingales (1998)) and that bank s will reduce the inherent inefficiencies associated with financial markets. Under a market -based system, e.g., covenants may be weaker or fewer and hence there is less external pressure to correct an over levered situation. Both markets and banks can ser ve to ameliorate market imperfections and provide sounder financial services and contracting. To the extent that the overall level and quality of financial services is accomplished under market based systems, I would expect adjustment to be faster in marke t based economies compared to bank based economies. To test this proposition, I allocate countries to two distinct portfolios: market-based versus bank -based. In the sample, 15 countries belong to a market -based financial tradition, whereas 22 countries be long to a bank -based financial tradition. The test results presented in Table 3 5 Panel B suggest that market based traditions impose lower costs of adjusting and/or higher benefits of converging to optimal leverage as implied by faster adjustment to optimal leverage. Specifically, using the SEPARATE (POOLED) method, I find that firms in market based financial systems adjust at an

PAGE 93

93 average rate of 23% (19%) with book leverage while firms in bank -based financial systems adjust at an average rate of 2 0 % (3 %). The differences are highly significant using both methodologies. These results suggests that market based economies are able to reduce market frictions and inefficiencies in financial contracting better than the bank -based economies. My finding is cons istent with Lfs (2003) investigation, which illustrates that the market -based US firms adjust faster towards the optimal capital structure compared to the bank based Swedish companies. Another view that is consistent with both legal traditions and mark et based view is that financial arrangements like contracts, markets and intermediaries exist to reduce market imperfections by providing sounder financial services (Levine (2002)). Panel C of the Table 3 5 explores whether the financial sector development based on indicators of size, efficiency and overall quality positively affect the adjustment speeds.19 All indicators of size, efficiency and overall quality of the stock markets matter for the adjustment speeds. Overall, I find that financial system devel opment as measured by the combined indices of bank and market activity has even bigger impact on the adjustment speed estimates than the financial structure or the legal system. Institutional Determinants of the Adjustment Speed Individual i ndices An exam ination of the adjustment speeds across different legal origins, financial market traditions and financial development reveals the importance of legal and financial institutions for 19 The size of stock markets and intermedia ries is captured by the market capitalization ratio and the private credit ratio respectively; while the size of the overall financial sector is defined as the logarithm of the market capitalization ratio times the private credit ratio. The efficiency of s tock markets and intermediaries is proxied by total value traded ratio and overhead costs of the banking system, respectively. On the other hand, the principal component of the activity, size and efficiency of the financial sector is extracted to quantify the overall quality of the financial sector.

PAGE 94

94 dynamic capital structure decisions. I now turn to analyzing the impact of several specific institutional factors that I believe would affect the costs and benefits of adjusting to optimal leverage. The results are presented in Table 3 6 The detailed description of variables can be found in Table 3 1 Less costly access to capi tal would imply lower external financing costs which in turn would speed up the adjustment to optimal capital structure I use separate measures of shareholder and creditor rights and the quality of contract enforcement ( La Porta, Lopez -de Silanes, Shleife r, and Vishny (1998)) respectively for equity and debt contracts to evaluate the impact of ease of access to capital markets on the capital structure adjustment. I hypothesize that the rebalancing costs will be lower due to easier access to capital through better rights and enforcement of debt and/or equity contracts, leading in turn to lower external financing costs and faster adjustment to optimal capital structure. The empirical evidence from Panel A 1 of Table 3 6 indicates that the easier is the access to capital, as captured by higher quality of investor protection in debt and equity contracts, the lower are the costs of adjustment. The results for equity contracts are consistent across both SEPARATE and POOLED methods and indicate that both the rights and enforcement of these rights are crucial for adjustment to optimal leverage whereas for debt contracts, there is stronger support for the impact of creditor rights using the SEPARATE method and the enforcement of these rights using the POOLED method. Using both method s I find that the adjustment is faster in countries with stronger anti director rights, stronger creditor rights and better enforcement of these rights For equity contracts, I find that the adjustment is faster in countries with stronge r anti -director rights, and the impact is similar with shareholder rights ( 6% 11%) and enforcement of these rights (6% -

PAGE 95

95 10%) using both SEPARATE and POOLED methods For debt contracts, the effect of creditor rights per se is more pronounced relative to the enforcement of these rights using the SEPARATE method (12% versus 6%), and the opposite is true (1% versus 13%) using the POOLED method. Overall, the results suggest that the degree of investor protection for both equity and debt contracts matter for the a djustment process. While undoubtedly both equity rights and enforcement equally matter for adjustment, debt rights and their enforcement might be substitutes. Higher asymmetric information would imply higher external financing costs which in turn would slow down the adjustment to optimal leverage. I anticipate that the rebalancing costs would be lower due to lower information asymmetry in institutional settings with higher accounting, disclosure and liability standards, better capital market governance and information sharing in credit markets, leading in turn to lower external financing costs and faster adjustment to optimal capital structure The empirical evidence from Panel A 2 of T able 3 6 suggests that the asymmetric information plays an important role on the adjustment process. The results for corporate transparency and equity contracts are consistent across both SEPARATE and POOLED methods and indicate that both are crucial for a djustment to optimal capital structure whereas for debt contracts, the results are mixed. Using SEPARATE method, I document that the adjustment is faster ( 8 %) in countries with higher corporate transparency. I also find that the higher is the asymmetric information in equity markets, the higher are the costs of adjustment, hampering the speed of adjustment to capital structure. Specifically, the countries that have strong disclosure standards (8 % ), strong liability standards (5 %), stro ng capital market en forcement (5 %) adjust significantly faster. Good capital

PAGE 96

96 market governance where insider trading is trivial (6 %) is also associated with faster adjustment. This is consistent with my hypothesis that the asymmetric information costs would be lower in countr ies where insider trading laws are more effective. Surprisingly, there is not an empirical support for the proposition that the information asymmetry in debt markets has adverse effects on the adjustment speed. The results are stronger with the POOLED meth od. All of the institutions that I presume proxy for information asymmetry matter for the adjustment speed. The qua lity of accounting standards (11%), disclosure (11 %), liability standards (8%), capital market enforcement (13%) and capital market governance (6 %), all result in higher speed of adjustment. There is also evidence that the presence of a public credit bureaus result in faster adjustment (5 %) to optimal capital structure by facilitating the flow of information. I assume that the rebalancing cost s will be lower due to higher financial flexibility in institutional settings that do not impose mandatory dividends or legal reserve requirements as financial constraints upon firms, causing faster adjustment to optimal leverage I acknowledge though that these measures might not adequately capture financial constraints around the world. Mandatory dividends can also be considered as a remedial shareholder right measure. If so, the existence of mandatory dividends might lower external financing costs and re sult in faster adjustment. The effectiveness of this right would also depend on the strength of the accounting standards, since earnings could be misrepresented to avoid distributing the mandatory dividends. In countries where there is no mandatory dividen d rule, the firms could also voluntarily distribute dividends to signal for quality or to reduce agency costs associated with free cash flow. Similarly, legal reserve requirements can also be considered as a remedial creditor right. This obligation forces firms to preserve a certain level of capital to avoid automatic liquidation. It

PAGE 97

97 protects creditors who have few other powers by forcing an automatic liquidation before all the capital is exhausted by the insiders. The empirical evidence presented in Panel A 3 of Table 3 6 suggests that legal reserve requirements act as financial constraints around the world, hampering swift adjustment to capital while the evidence is mixed for mandatory dividends. I document that the adjustment is slower (10 %) in countrie s with mandatory dividends using the SEPARATE method only T he existence of mandatory dividends seems to lower external financing costs and result in faster adjustment according to the POOLED method. L egal reserve r equirements negatively affect (9% 13%) t he adjustment speed to optimal capital structure using both methods The empirical evi dence does not support that the remedial rights in debt markets are effective in reducing the external financing costs by making access to capital easier. Adjustment ben efits should be higher in countries imposing higher financial distress costs on firms and that cope more efficiently with firms in default, leading to faster adjustment. Firstly, I use the design of the bankruptcy codes and debt contracts, including the at tached creditor rights and the associated enforcement mechanisms governing default on debt contracts as determinants of the financial distress costs. I assume that firms would not only have easier access to capital hence lower rebalancing costs in the pre sence of strong creditor rights and efficient enforcement mechanisms, but that the rebalancing benefits will also be more pronounced due to binding distress costs especially in bad states, causing faster adjustment to optimal leverage Secondly, I anticipa te that firms from countries that minimize the deadweight costs by administering the insolvency process in court in a way that is less time consuming, less costly and more efficiently will adjust to their target more rapidly.

PAGE 98

98 Panel A 1 of Table 3 6 provides some support for my first hypothesis. The adjustment speed is faster in countries with either higher scores of creditor rights and their enforcement. Panel B 1 of Table 3 6 provides information on the effect of the efficiency of the bankruptcy process o n the adjustment speeds around the world and evaluates my second hypothesis. I find that the administration of bankruptcy matter for capital structure adjustment, with the ex -post bankruptcy efficiency having the most significant (7% 12%) effect. In general, the results are supportive of the relevance of the bankruptcy codes and debt contracts for the adjustment to capital structure. The rebalancing benefits will be higher due to higher incentives to shield taxes in institutional settings imposing higher effective tax rates causing higher speed of adjustment to optimal leverage There are limitations using the effective tax rate. Even though this rate is calculated factoring some of the non-debt tax shields like depreciation, it fails to c onsider all of them and only firms that exceed their non-debt tax shield can benefit from a high tax shield potential. Furthermore, higher taxes positively affect the benefits of debt only for under levered firms. Finally, this uniform country level tax ra te may fail to adequately reflect the tax status of each individual firm with distinct features. Despite these limitations of effective tax rates as proxies for tax shields, I find some support that it increases the rebalancing benefits and results in fast er adjustment but using the SEPARATE method only. In Panel B 2 of Table 3 6 using the SEPARATE method, I d ocument that the adjustment is slightly faster in countries that have higher effective tax rates. I suppose that the rebalancing benefits are higher in countries with more constraints on executive power, better quality of contract enforcement and better quality of outside control as measured by the integrity of the legal system and the strength of the governance (measured by

PAGE 99

99 the government quality) d ue to more binding costs of deviation imposed on firms. I quantify constraints on the executive power based on the number of effective veto points in a country where veto points include an effective legislature, an independent judiciary and a strong federa l system. My governance measure reflects a combination of factors, namely the level of corruption in government, risk of contract expropriation and repudiation by the government. Using the SEPARATE (POOLED) method, I document in Panel B 3 of Table 3 6 that the adjustment is 8% (7 %) faster in countries that have more constraints on executive power, 9 % (9% ) faster in countries with better enforcement of contracts, 7% (12 %) faster with stronger law and order and 8 % ( 10%) faster with higher quality of governanc e. Overall, my findings indicate that the proxies for costs of deviation from the optimal leverage strongly affect the adjustment process. Alternative formulations of the adjustment f actors Many economic and legal institutions can be considered as compleme nts or substitutes and the variables characterizing them are likely to be highly correlated.20 For this reason, the proxies I adopt should not be interpreted narrowly. If certain institutional features are complementary, separately including the indices in the model would not properly reflect their complementary nature, and if certain institutional features are substitutes, countries may be able to choose different combinations of institutions to arrive to a certain outcome. I address the concern that institutions can be either complements or substitutes, making their interpretation harder. I form more comprehensive measures by aggregating the institutional indices to allow for substitution among the various institutions and facilitate their interpretation. I accomplish this in two different ways. First, I form the principal components of the related individual indices. Second, I compute an overall score by aggregating or averaging the related indices. The results are very 20 Table 2 5 provides information on the correlation structure of the institutional variables.

PAGE 100

100 similar using the two methods. For ea se of exposition, I only report my results using the first principal components of the related indices. The results are reported in Table 3 7 Table 3 1 provides detailed information on the formulation of the composite indices to evaluate the combined effe ct of the adjustment cost and benefit factors. 21 Consistent with the results from individual indices, my data lend support for my proposition that the proxies I picked for various factors of adjustment costs and benefits significantly affect the adjustment speeds worldwide in the direction I hypothesized. The empirical evidence from Panel A 1 of Table 3 7 indicates that higher quality investor protection, the proxy for the ease of accessing the capital markets, leads to higher adjustment spe ed with an impa ct ranging from 8% to 12%. Looking at equity and debt contracts separately, I observe that both shareholder and creditor protection matter for the adjustment speed, but the former has relatively bigger impact (6% 12% versus 2%10 %). The factor loadings on my composite investor protection measure indicate that the enforcement of the shareholder (93%) and creditor (81%) rights has relatively more important impact on the adjustment process than the shareholder (74%) and creditor (42%) rights per se. The resul ts reported in Panel A 2 of Table 3 7 suggest that the information asymmetry imposes significant adverse selection costs, hampering the speed of adjustment to optimal leverage The information asymmetry in financial markets slows the adjustment speed with a magnitude ranging from 7% to 11 %. Information sharing in debt markets is now considerably important for capital structure adjustments, possibly because the corporate transparency and the information sharing in debt markets are complements, making their c ombined effect larger. Comparing the effect of information asymmetry in debt and equit y markets separately shows 21 The exact formulation of the composite indices is not crucial for the analysis; I experimented with se veral alterna tive specifications and found robust results.

PAGE 101

101 that, on average the adverse selection costs in debt markets are more detrimental to the adjustment speed estimates for the countries in the sa mple The factor loadings on my composite information asymmetry measure point out that disclosure (85%) and liability requirements (8 6 %) in the capital markets an d the corporate transparency (76%) are among the most important factors in the adjustment process, while the presence of public bureaus has the least significant effect (21 %). On overall, Panel A 3 of Table 3 7 shows that the external financing costs significantly influence (11%12 %) the adjustment to target capital. The impact of external financing costs across debt and equity markets is economically very similar. The empirical evidence presented in Panel A 4 of Table 3 7 implies that the cash co nstraints have similar effect (8 % to14%) on the adjustment speeds as the information asymmetry and ease of access proxies. My conglomerate adjustment costs proxy presented in Panel A 5 of Table 3 7 reveal that on o verall the adjustment speed is 11% to 12% higher in the presence of lower adjustmen t costs. Differentiating among debt and equity contracts does not yield significant differences across the two in the sense that both matters equally. A closer look to at my composite adjustment cost proxy reveals that the most important components of the adjustment costs are disclosure in the capital markets (83%), corporate transparency (82%), shareholder rights and their enforcement (82% and 81% ); while the least important factors are creditor rights (9 %), information sharing in debt markets (11%), and m andatory dividends (30%). Besides disclosure, additional measures of the quality of capital market governance; enforcement of creditor rights; and legal reserve requirements are among other factors that have a decent level of impact (60%75%) on the adjust ment process. The adjustment benefit proxies I choose yield consistent results with my hypotheses even when I extract the principal components. The results are presented in Panel B of Table 3 7

PAGE 102

102 Higher distress costs leads to faster adjustment on average r anging from 5% to 11%. In addition, more binding deviation costs lead also to swifter adjustment on average of magnitude 8% to 9%. Finally, higher adjustment benefits lead to an increase in adjustment speed on average from 8% to 9%. Among the individual in dices, the least important component of the adjustment benefits is the presence of tax shields, while other components are almost equally significant. Overall, the results indicate that both adjustment costs and benefits significantly affect the adjustment speeds around the world. A closer look at the individual country adjustment speeds further confirms my findings. Slow (fast) adjusters have high (low) adjustment costs and/or low (high) adjustment benefits on overall or on several dimensions of my individ ual indices. A visual representation of the relationship between the adjustment speed estimates and the adjustmen t factors is provided in Figure 3 3 and Figure 3 4 respectively for the adjustment cost and benefit factors. Robustness In this section I per form additional sensitivity analyses I first verify the robustness of the analysis to an alternative definition of capital leverage, namely market leverage and the alternative estimation method, the Corrected L east S quares D ummy V ariable technique, with b ook and market leverage, using the separate method Next, I broaden my econometric analysis to consider sample composition in each country, using the pooled method. If adjustment speed varies across firms having their own specific characteristics and circu mstances, imposing the restriction of a constant speed of adjustment in each country or institutional context may not be the most accurate econometric methodology. This could potentially bias the results to the extent that the uniform adjustment speed esti mate is a noisy approximation of the mean or the median of the adjustment speeds varying with respect to firms in each country or institutional environment. Furthermore, so far, my discussion and analyses concerned a representative

PAGE 103

103 average/median firm in each country or institutional context; however, it is reasonable to think that the structural variables I employ might differently affect certain types of firms. For example, an over levered firm would want to respond more quickly than an under levered firm to leverage shock due to more binding bankruptcy costs and stronger incentives to increase its debt capacity. To the extent that the sample composition differs across countries or institutional environments, I could see differential impact of my structura l variables on the adjustment speed estimates. To overcome the restriction of uniform adjustment, to allow for asymmetric response to adjustment factors and to be able to account for the differential impact of sample composition, I apply a variable partial adjustment speed model to the data by allowing the firms specific characteristics to determine their optimal leverage ratio and the adjustment speed toward the target. Alternative Definition of Leverage and Estimation Method The results for ma rket leverage using two-step GMM estimation method; and for book and market leverage using Corrected Least Squares Dummy Variable Approach are documented in Table 3 8 for individual and composite institutional indices and indicate that both adjustment costs and benefits significantly affect the adjustment speed s around the world consistent with my hypotheses regardless of the estimation methodology and definition of financial leverage employed. Overall, higher adjustment costs result in slowe r adjustment of magnitude 5% 11% and higher benefits lead to faster adjustment speed of magnitude 3% 7%. Variable Adjustment Speed Model I model the target leverage and the adjustment speed simultaneously allowing for the possibility that they might both differ across firms and over time in each country. As before, I model the optimal leverage ratio as a function of firm specific characteristics as in Equation (3 3 ). However, I now relax the assumption that all firms adjust at a constant rate in the countr y or institutional context they operate in. I conjecture that the speed at which the firm adjusts its

PAGE 104

104 capital structure depends on its specific conditions and hypothesize that the same variables affecting the target capital determine the adjustment speed. I model a firm -specific adjustment that varies with firm characteristics, 1 and over time: = 1 (3 6 ) where is a vector of coefficients for the adjustment speed func tion, and 1 is the set of firm characteristics that affect the adjustment speed and the optimal leverage I investigate a nonlinear model in which firm characteristics affect both the target capital ratio and the speed of adjustment22: 1= 1 1 1 + 1+ (3 7 ) As before, SEPARATE metho d involves estimating Equation 3 7 for each country to arrive at individual country s and then averaging these estimates for each group formed based on the particular institutional characteristic. According to this method, I hypothesize that the same variables affecting the target leverage determine the adjustment speed, hence 1 is the same vector as 1. For the POOLED methodology, I re run my estimation specification, Equati on 3 7 twice, separately for the two portfolios I form according to a particular institutional attribute and compare the estimated s from the t wo regressions. To implement this method, I 22 This chapter focuses exclusively on the link between legal and financial institutions and firms adjustment speed to optimal capital structure across nations and institutional environments. It is conceivabl e that in countries/environments with weak institutions firms would create their own complementary measures to ameliorate their status vis vis investors. For example, in countries with weak creditor rights, firms could rely more on collateral to attract funds. Alternatively, in countries with strong investor rights, firms could have less recourse to their own corporate governance measures. Allowing different sensitivities for the firm specific determinants of the optimal leverage and the adjustment speed helps me to account for the firms propensity to undo the impact of the institutional uniqueness they operate in.

PAGE 105

105 specify 1 as the 1 vector augmented by country fixed effects to allow the adjustment speed to vary according to both firm and country characteristics. Combining the nonlinear character of my partial adjustment model with unbalanced dynamic panel suggests that the most appropriate estimation is an iterative stepwise manner. In the first step, I estimate my partial adjustment model under the assumption that the adjustment speed is constant for a ll firms for that part icular country, using Equation 3 3 This provides me an initial average estimate for all firms residing in a particular country and an initial set of estimated s, which I use to calculate an initial estimated target capital ratio = 1+ for each firm and year in that particular country j. These initial estimates of the firms targets are likely to be biased since the adjustment speed for each firm is constrained to be constant across firms and years. In the second step, I relax the restriction that the adjustment speed is constant. First, I calculate each firms deviation from its estimated long run target debt ratio as = 1 1 (3 8 ) Substituting this back to my partial adjustment model yields: 1= 1 + (3 9 ) Estimating Equation 3 9 yields from which I am able to calculate firm and time -varying adjustment speed In the third step, I re -estimate the target capital ratios in a variable adjustment speed context by substituting the estimated adjustment speeds 1 fr om the second step and re arrange as follows: (1 ) 1 1 1= [ ( 1) 1] + 1+ (3 10)

PAGE 106

106 Finally, to come up with unbiased estimates of both adjustment speeds and targets, I repe at steps one through three iteratively until the model finally converges.23 Even though there is a large variation in adjustment speed estimates in each country, the overall mean and median is close to the uniform estimate, implying that the constant adjus tment speed is a reasonable approximation of the true sample mean or median. The estimation results of estimating Equation 3 7 are summarized in Table 3 9 Figure 3 5 compares the uniform and mean variable adjustment speed estimates. T he results as docume nted in Table 3 10 and Table 3 11 respectively for individual indices and principal components are qualitatively similar using the variable adjustment speed estimate s and indicate that there is large support for all individual indices except for mandatory dividends as cash constraints among the adjustment cost factors and tax benefits of debt among the adjustment benefit factors. The magnitudes of the impact for both individual and composite indices are generally considerably bigger. T he ease of access to capital markets results in faster adjustment and the effect ranges from 12% to 14% using SEPARATE and POOLED methods respectively. The alleviation in information asymmetry increases the adjustment speed estimate by 5% to 9%. Lower external financing costs measured as the combination of lower information asymmetry and better access to capital lead to faster adjustment of magnitude 6% to 13%. The existence of cash constraints hampers adjustment speed by an amount 12% to 17%. Adjustment costs on overall impede the adjustment speed of order 4% to 13%. Among the benefits of adjustment, tax shields continue to be insignificant for the adjustment speed. On the other hand, the distress costs and the deviation costs continue to matter for the capital structure adjust ments. The distress costs have an impact 23 I report my results using the converged value of the adjustment speed from the final step. However, the results are very similar using the two step estimates.

PAGE 107

107 of magnitude 6% to 15%. The deviation costs affect adjustment speed on overall around 1% to 12%. Combined effect of the adjustment benefits range from 2% to 11%. The overall adjustment costs and benefits captured by my composite institutional indices reveal an important impact on the adjustment speeds around the world even after allowing variable adjustment according to each firm. My robustness exercise with the variable adjustment model indicates that there is stro ng support for my international partial adjustment model, regardless of adopting a uniform adjustment rate to proxy for the mean or median adjustment from a sophisticated model that allows firms and countries to vary in regard to their speed of convergence to optimal leverage .24 Asymmetric Response to Adjustment Factors Relative l everage My theory of international partial adjustment might at first seem cavalier about asserting some structural variable will affect the adjustment speed, when it seems that some of the institutional differences might differentially affect debt versus equity.25 Even though I dont ignore such possibility, based on the results on the frequency and size of the capital structure adjustments, I think that this is unlikely to be a conse quential issue in the sample. An assessment 24 I have experimented w ith more complex variations of E quation 37, allowing asymmetric response for firms with differing characteristics (high versus low quartiles of my firm specific characteristics ) or augmenting by other variables that I believe might affect the adjustment speed (dividend payers, cash constraints etc.). In each case, my results are qualitatively and quantitatively very similar to the ones reported in my study. In addition, I have r un Equation 37 using only country fixed effects in the vector using the POOLED method The country dummies are significant but the results remain qualitatively the same, and the differences become even sharper once I relax that assumption and allow for variable adjustment across countries. 25 For example, an over levered firm might respond more or less to an adjustment factor than an under levered one depending on the circumstances. Consider a firm in a country with a high quality of investor protection that is away from its target. If the firm is over levered (under levered), it will need to issue equity (debt) and/or repurchase debt (equity) to converge to its optimal capital. As long as the easier access to capital due to the better investor protectio n reduces the cost of issuance and repurchase equally, the response of the firm is expected to be symmetric whether it is over or under levered. On the other hand, if better investor protection makes issuances (repurchases) more attractive, an over levered (under levered) firm would approach to its target more rapidly using equity (debt) in which case the response to adjustment factors is likely to be asymmetric.

PAGE 108

108 of the capital structure adjustments among the slow and fast adjusters implies that all four forms of external financing activities, namely debt issuance, debt retirement, equity issuance and equity retirement ar e uniformly important in determining the convergence to optimal leverage Nevertheless, I test the validity of my supposition by dual splits. The asymmetry results are documented in Table 3 12 and Table 3 13 respectively for the individual indices and the principal components; in columns 1 and 2 for over levered and under levered firms. As before, for each particular institutional variable I form two groups based on the median country in the sample. I then fu rther sub-divide each of these two groups by distinguishing among the over -levered and under levered firms. A firm is defined as over levered (under levered) if the actual leverage is 10% above (below) the calculated target leverage. The test results show that both over -levered and under levered firms continue to be significantly affected by the international adjustment cost and benefit factors similar to an average firm and that allowing for asymmetric response to adjustment costs and benefits does not qua litatively alter my main conclusions. Firm s ize So far, this chapter evaluated how financial systems and institutions affect the firms adjustment speed in different countries and institutional environments in general. However, it is also possible that better institutions positively affect the adjustment speed of each firm individually, although the general prediction may be different due to sample composition effects.26 26 For instance, suppose that the information sharing in debt market leads to a sharper decrease in adverse selection and moral hazard problems among the lower quality borrowers (e.g. smaller size firms or firms with fewer/smaller amount of collateral), decreasing their external financing costs on debt contracts to a greater extent. In that case, I c ould observe a trivial difference on the adjustment speed estimates across institutional contexts on overall, with differing impact only among the lower quality borrowers. This could explain my finding that some of my individual indices, namely mandatory dividends and tax shields, significantly and systematically affect the adjustment speeds using a uniform adjustment rate (ignoring sample composition), but do not seem to affect the adjustment to optimal

PAGE 109

109 To test this assertion, I form two groups for each particular institutional variable based on the median country in the sample and further sub-divide each of these two groups according to firm size, using the median firm in each institutional setting. The results are reported in Table 3 12 and Table 3 13 respectively for the individual indices and the principal components; in columns 3 and 4 for small and large firms. Consistent with my assertion, I find a significant effect of mandatory dividends on adjustment speed among the smaller firms and a non significant effect among the larger fir ms. However, tax shields continue to be inconsequential for both small and large firms. This result partly suggests that the uniform corporate tax rate that represents the average firm in each country might be an over -simplified measure of the firm level tax rates and consequently might not closely relate to firm level varying adjustment speeds. All of my institutional proxies continue to have an important impact on the adjustment speed irrelevant of firm size except for mandatory dividends and time to repa y in case of bankruptcy. Timely repayment in insolvency imposes binding costs of deviation from optimal leverage for large firms but not for small firms. The general conclusion from this robustness exercise is that the sample composition of the particular country or institutional setting is not the driving factor of my results.27 leverage using a variable adjustment rate (taking int o account the sample composition). It is possible that mandatory dividends are binding cash constraints for smaller firms, with no impact on larger firms or that tax shields are an important consideration for larger firms rather than smaller firms. 27 I hav e experimented with additional firm specific characteristics (e.g. proportion of fixed assets to total assets) to verify the robustness of my analysis to sample composition effects. More generally, I have employed a common world model of the adjustment spe ed by estimating Equation 37 for the whole sample and specifying Equation 37 in terms of both firm specific ( ) and country specific adjustment factors ( my composite indices). Note that this method constrains both s and s for all sample countries to be equal. The country level adjustment cost and benefits are important determinants of the adjustment speed even after accounting for the firm level cost and benefits. The impact of their effect is similar to the documented results.

PAGE 110

110 Other c onsiderations This chapter focuses exclusively on the link between legal, political and financial institutions and firms adjustment speed to optimal leverage across nations and institutional environments. It is conceivable that in countries/environments with weak institutions firms would create their own complementary measures to ameliorate their status vis -vis investors. For example, in countries with weak creditor rights, firms could rely more on collateral to attract funds. Alternatively, in countries with strong investor rights, firms could have less recourse to their own corporate governance measures. I implicitly control for this possibility by allowing different sensitivities for the firm specific determinants of the optimal leverage and the adjustment speed to account for the firms propensity to undo the impact of the institutional uniqueness they operate in. In addition, I explicitly control for endo geneity concerns in the empirical specifications by instrumenting for the firm specific characteristics. Many legal institutions can also be considered as complementary and the variables characterizing them are likely to be highly correlated. Therefore, s eparating out the marginal effects of various institutional attributes is challenging. For this reason, the proxies I adopt should not be interpreted narrowly. However, the results are robust to using several alternative specifications of the composite or individual indices.28 Finally, a caveat is in order. Even though the indices adopted in this study capture various aspects of institutional environments that might possibly affect the factors that influence the adjustment to optimal capital structure, I am constrained by data availability. For example there 28 For example, controlling for the overall economic development (GDP per capita or GDP growth) for my composite measures reinforces the results. Also, the choice of the specific indices is not material for the conclusions. The results employing similar varia bles (e.g. statutory or five year effective tax rates rather than first year effective tax rate; revised anti director rights rather than anti director rights) produce robust results.

PAGE 111

111 might be important cultural differences among nations reflecting into the corporate behavior around the globe but that I dont directly address in this study 29 Summary In this chapter, I empirically te st and quantify the impact of the hypothesized institutional characteristics on the adjustment speed to optimal leverage. My analyses lend strong empirical support for my hypotheses. T he institutional proxies I employ to account for the ease of access to c apital, the information asymmetry, the financial constraints, the tax benefits of debt, the costs of distress and the deviation costs all affect the speed of adjustment in a manner consistent with my theory. I find that a country from an institutional env ironment allowing easier access to capital markets via better (above median) investor protection adjusts to its capital structure up to 210% faster. This difference (17.4% or a half -life of 3.6 years30 versus 5.6 % or a half lif e of 12 years) is both economically and statistically very meaningful and translates into a gain of 8.4 years in closing half of the distance to the target. Moving a country from a higher information asymmetry to a lower information asymmetry environment would in crease its adjustment speed by up to 80%, indicating a gain of up to 3.2 years in covering half of the distance to the target. Finally, removing the cash constraints from firms imposed by mandatory dividends or legal reserve requirements would increase the ir adjustment speed of magnitude up to 330%, a boost of 12.6 years in covering half of the distance to the target. Overall, in institutional environments where the adjustment costs are lower, the adjustment speed is 160% faster. 29 John R. Graham, Campbell R. Harvey, and Manju Puri (2008) find evide nce on how US CEOs differ from nonUS CEOS, CFOs, and the general population. For example, both CEOs and CFOs from the United States tend to be less risk averse than CEOs and CFOs whose companies are not located in the United States. They are also more opt imistic, more patient and older with longer tenure. Furthermore, they show that the managerial actions are correlated with capital structure decisions. 30 The calculation is LN(0.5)/LN(1 0.1740).

PAGE 112

112 Among the adjustment benef it factors, distress costs and deviation costs systematically affect the adjustment process, whereas the evidence for the tax shields is mixed. A firm operating under an institutional environment imposing higher exante distress costs and lower ex post ine fficiency in the bankruptcy process adjusts to its target up to 180% faster or 7.2 years sooner in closing half of the distance to the target. For the institutional environments that impose higher costs of deviation from the optimal capital structure the a djustment speed is up to 130% higher 5.5 years sooner in covering half of the distance to the target. In general, in institutional settings where the adjustment benefits are greater, the adjustment speed is 128% faster. These differences are economically a nd statistically highly significant. My findings suggest that adjustment costs and benefits play an important role in the speed of convergence to optimal capital structure around the world. T he adjustment is significantly faster in countries with better p rotection and enforcement of investor rights, stronger legal, judicial, financial systems, and better capital, corporate, and political governance. The empirical evidence that I present in this chapter provides substantial support for the dynamic tradeoff theory.

PAGE 113

113 Table 3 1 Description and Sources of Variables Determining the Adjustment Speed Variable Description and Source Predicted Sign LEGOR Legal origin Identifies the legal origin of each Company Law or Commercial Code of each country. There are four possible origins: (1) English Common Law; (2) French Commercial Code; (3) German Commercial Code; (4) Scandinavian Commercial Code. Source: La Porta et al. ( 1998) ; Reynolds and Flores ( 1989) ; CIA World Factbook ( 2001 ) COMMON English Common Law origin Equals one if the legal origin is common (English), zero if it is civil (French, German or Scandinavian). Source: La Porta et al. ( 1998) ; Reynolds and Flores ( 1989 ) ; CIA World Factbook ( 2001 ) + MARKET Financial system classification Equals one if the financial system is marketbased, zero if it is bankbased. Source: Levine (2002). + FSIZE Finance Size Measure of the size of stock markets and intermediaries captured by the logarithm of the market capitalization ratio times the private credit ratio. The measures of the size of the domestic stock market and intermediaries are market capitalization ratio and the private credit ratio, respectively. Source: Levin e (2002). + FEFFICIENCY Finance Efficiency Efficiency of the financial sector as measured by the logarithm of the total value traded ratio divided by overhead costs. The measures of the efficiency of the domestic stock market and intermediaries are total value traded ratio and the overhead costs, respectively. Source: Levine (2002). + FAGGREGATE Finance Aggregate Quality Principal Component of finance activity, size and efficiency where finance activity is the logarithm of the total value traded ratio times the private credit ratio. Source: Levine (2002). + AD JUSTMENT COSTS EASE OF ACCESS TO CAPITAL MARKETS ANTIDIR Shareholder rights This index of anti director rights is formed by adding one when: (1) the country allows shareholders to mail their proxy vote; (2) shareholders are not required to deposit their shares prior to the General Shareholders Meeting; (3) cum ulative voting or proportional representation of minorities on the board of directors is allowed; (4) an oppressed minorities mechanism is in place; (5) the minimum percentage of share capital that entitles a shareholder to call for an Extraordinary Shareholders Meeting is less than or equal to 10% (the sample median); or (6) when shareholders have preemptive rights that can only be waived by a shareholders meeting. The range for the index is from 0 to 5. Source: La Porta et al. (1998). + PRENF Equity enforcement Average of ex ante and ex post private control of self dealing. Source: Djankov et al. (2008) + SHAREHOLDER PROTECTION Equity Investor Protection Principal Components of the shareholder rights and equity enforcement indices. SHAREHOLDER PROTECTION = 0.89 ANTIDIR + 0.89 PRENF + CREDITOR Creditor rights An index aggregating creditor rights. A score of one is assigned when each of the following rights of secured lenders is defined in laws and regulations: First, there are restrictions, such as creditor consent or minimum dividends, for a debtor to file for reorganization. Second, secured creditors are able to seize their collateral after the reorganization petition is appr oved, i.e. there is no "automatic stay" or "asset freeze." Third, secured creditors are paid first out of the proceeds of liquidating a bankrupt firm, as opposed to other creditors such as government or workers. Finally, if management does not retain adm inistration of its property pending the resolution of the reorganization. The index ranges from 0 (weak creditor rights) to 4 (strong creditor rights). Source: La Porta et al. ( 1998 ) + FORMALISM Debt Enforcement The index measures substantive and procedural statutory intervention in judicial cases at lower level civil trial courts, and is formed by adding up the following indices: (i) professionals vs. laymen, (ii) written vs. oral elements, (iii) legal justificat ion, (iv) statutory regulation of evidence, (v) control of superior review, (vi) engagement formalities, and (vii) independent procedural actions. The index ranges from 0 to 7, where 7 indicates a higher level of control or intervention in the judicial pro cess. Source: Djankov et al. (2003) CREDITOR PROTECTION Debt Investor Protection Principal Components of the creditor rights and debt enforcement indices. CREDITOR PROTECTION = 0.74 CREDITOR 0.74 FORMALISM + INVESTOR PROTECTION Investor Protection Principal Components of the individual indices constituting equity holder rights, creditor rights, equity enforcement and debt enforcement. INVESTOR PROTECTION = 0.74 ANTIDIR + 0.93 PRENF + 0.42 CREDITOR 0.81 FORMALISM + ASYMMETRIC INFORMATION ACCSTDS Transparency Index created by examining and rating companies 1990 annual reports on their inclusion or omission of 90 items. These items fall into seven categories (general information, income statements, balance sheets, funds flow statement, accounting standards, stock data and special items). Source: La Porta et al. ( 1998) International accounting and auditing trends, Center for International Financial Analysis and Research. + EDISCL OSE Disclosure requirements The index of disclosure equals the arithmetic mean of (1) prospectus; (2) compensation; (3) shareholders; (4) inside ownership; (5) contracts irregular; and (6) transactions. Source: La Porta et al. (2006) + ELIABS Liability Standards The index of liability standards equals the arithmetic mean of (1) liability standard for the issuer and its directors; (2) liability standard for distributors; and (3) liability standard for accountants. Source: La Porta et al. (2006) + EPUBLICENF Securities Market Enforcement The index of public enforcement equals the arithmetic mean of (1) supervisor characteristics index; (2) rule making power index; (3) investigative powers index; (4) orders index; and (5) criminal index. Source: La Porta et al. (2006) + CMG Insider Trading Prevalence of Insider Trading (1=pervasive; 7=extremely rare). Source: Schwab, Klaus et al., eds., 1999, The Global Competitiveness Report 1999, (Oxford University Press, New York, NY). Source: La Porta et al. (2006) + EQUITY INFO Asymmetric Info in Equity Markets Principal Components of the transparency, disclosure requirements, liability standards, securities market enforcement and insider trading indices. EQUITY INFO= 0.80 ACCSTDS + 0.87 EDISCL + 0.80 ELIABS + 0.75 EPUBLICENF + 0.62 CMG + PUBBUREAU Information Sharing The variable equals 1 if a public credit registry operates in the country, 0 otherwise. A public registry is defined as a database owned by public authorities (usually the Central Bank or Banking Supervisory Authority), that collects information on the standing of borrowers in the financial system and makes it available to financial institutions. The variable is constructed as at January for every year from 1978 to 2003. Source: Djankov et al. (2007) + DEBT INFO Asymmetric Info in Debt Markets Principal Components of the transparency and information sharing indices. DEBT INFO= 0.70 ACCSTDS + 0.52 PUBBUREAU + ALL INFO Asymmetric Info in All Markets Principal Components of the transparency, disclosure requirements, liability standards, securities market enforcement, insider trading and information sharing. ALLINFO = 0.76 ACCSTDS + 0.85 EDISCL +0.86 ELIABS + 0.64 EPUBLICENF + 0.65 CMG +0.21 PUBBUREAU +

PAGE 114

114 Table 3 1. Continued Variable Description and Source Predicted Sign EQUITY EXT Equity External Financing Costs Principal Components of the sub indices forming ease of access and asymmetric info in equity markets. EQUITYEXT= 0.87 ANTIDIR + 0.78 PRENF+ 0.80 ACCSTDS + 0.87 EDISCL + 0.80 ELIABS + 0.70 EPUBLICENF + 0.60 CMG + DEBT EXT Debt External Financing Costs Principal Components of the sub indices forming ease of access and asymmetric info in debt markets. DEBTEXT= 0.42 CREDITOR + 0.86 FORMALISM+ 0.87 ACCSTDS + 0.23 PUBBUREAU + ALL EXT External Financing Costs Principal Components of the sub indices forming ease of access and asymmetric info in debt and equity markets. ALLEXT= 0.85 ANTIDIR + 0.80 PRENF + 0.07 CREDITOR + 0.74 FORMALISM + 0.82 ACCSTDS + 0.84 EDISCL + 0.77 ELIABS + 0.67 EPUBLICENF + 0.62 CMG + 0.16 PUBBUREAU + CASH CONSTRAINTS MDIV Mandatory dividend Equals the percentage of net income that the Company Law or Commercial Code requires firms to distribute as dividends among ordinary stockholders. It takes a value of zero for countries without such restriction. Source: La Porta et al. ( 1998) RESERVE Legal Reserve Requirements It is the minimum percentage of total share capital mandated by Corporate Law to avoid the dissolution of an existing firm. It takes a value of zero for countries without such restriction. Source: La Porta et al. ( 1998) CASH Cash Constraints Principal Components of Mandatory dividends and legal reserve requirements. CASH= 0.77 MDIV 0.77 RESERVE + EQUITY COSTS Equity Adjustment Costs Principal Components of the sub indices forming ease of access, asymmetric info in equity markets and mandatory dividends. EQUITYCOSTS = 0.85 ANTIDIR + 0.77 PRENF+ 0.81 ACCSTDS + 0.87 EDISCL + 0.79 ELIABS + 0.69 EPUBLICENF + 0.60 CMG 0.32 MDIV + DEBT COSTS Debt Adjustment Costs Principal Components of the subindices forming ease of access, asymmetric info in debt markets and legal reserve requirements. DEBTCOSTS = 0.31 CREDITOR + 0.82 FORMALISM+ 0.88 ACCSTDS + 0.10 PUBBUREAU 0.78 RESERVE + COMBINED COSTS Equity and Debt Adjustment Costs Principal Components of the sub indices forming ease of access, asymmetric info and cash constraints in equity and markets. ALLCOSTS = 0.82 ANTIDIR + 0.81 PRENF + 0.09 CREDITOR + 0.73 FORMALISM + 0.82 ACCSTDS + 0.83 EDISCL + 0.74 ELIABS + 0.71 EPUBLICENF + 0.60 CMG + 0.11 PUBBUREAU 0.30 MDIV 0.76 RESERVE + ADJUSTMENT BENEFITS DISTRESS COSTS TIME Time to resolve Time to resolve the insolvency process. Source: Djankov, Hart, McLiesh and Shleifer (2008). COST Cost of Bankruptcy The costs to complete the insolvency proceeding, expressed as a percentage of the bankruptcy estate at the time of entry to the bankruptcy. Source: Djankov, Hart, McLiesh and Shleifer (2008). EFFICIENCY Efficiency of Bankruptcy A dummy variable for whether the bank ruptcy outcome is efficient. Source: Djankov, Hart, McLiesh and Shleifer (2008). + DISTRESS Distress Costs Principal Components of time to resolve, cost of bankruptcy and efficiency of bankruptcy. DISTRESS = 0.85 TIME 0.83 COST + 0.91 EFFICIENCY + TAX SHIELDS TAX 1st Year Effective Tax Rate (%) The tax rate obtained by dividing the total corporate tax TaxpayerCo pays by its pretax earnings. Source: Djankov et al. (2008) + DEVIATION COSTS EXECUTIVE Executive Quality Index of constraints on the executive power based on the number of effective veto points in a country. Veto points include (1) an effective legislature (represents two veto points in the case of bicameral systems); (2) an independent judiciary; and (3) a s trong federal system. Average of the years 1945 through 1998. Source: Henisz ( 2001 ) Djankov et al. (2002) + ENFORCE Enforceability of contracts The relative degree to which contractual agreements are honored and complications presented by language and mentality differences. Scale for 0 to 10, with higher scores indicating higher enforceability. Source: Business Environmental Risk Intelligence. Exact definition in Knack, Stephen and Philip Keefer, 1995. Djankov et al. (2003) + LAW&ORDER Law and Order Integrity of legal system in 2000. This component is based on the Political Risk Component 1 (Law and Order) from the PRS Groups International Country Risk Guide (various issues). Rankings are modified to a 10 point scale. Source: Economic Freedom of the World (2002). Djankov et al. (2003) + CORRUPT Corruption in government Low ratings indicate high government officials are likely to demand special payments and illegal payments are generally expected throughout lower levels of government in the form of bribes connected with import and export licenses, exchange controls, tax assessment, policy protection, or loans. Scale from 0 to 10. Average of the months of April and October in the monthly index between 1982 and 1995. Source: International Country Risk Guide (ICRG). La Porta et al. ( 1999) + EXPROPR Risk of expropriation ICRs assessment of the risk of outright confiscation or forced nationalization. Average of the months of April and October of the monthly index between 1982 and 1995. Scale from 0 to 10, with lower scores for higher risks. Source: La Por ta et al. ( 1998 ) + REPUDR Repudiation of contracts by government ICRs assessment of the risk of a modification in a contract taking the form of a repudiation, postponement, or scaling down due to budget cutbacks, indigenization pressure, a change in government, or a change in government economic and social priorities. Average of the months of April and October of the monthly index between 1982 and 1995. Scale from 0 to 10, with lower scores for higher ri sks. Source: La Porta et al. ( 1998) + GOVERNANCE Quality of government Principal Components of corruption in government, risk of expropriation and repudiation of contracts by government. GOVERNANCE = 0.92 CORRUPT + 0.95 EXPROPR + 0.95 REPUDR + DEVIATION COSTS Deviation Costs Principal Components of executive quality, quality of law and order and the sub indices of the quality of government. DEVIATIONCOSTS = 0.78 EXECUTIVE + 0.84 LAW&ORDER + 0.89 CORRUPT + 0.95 EXPROPR + 0.90 REPUDR + 0.92 ENFORCE + COMBINED BENEFITS Adjustment Benefits Principal Components of subindices forming distress costs, tax shields and deviation costs. BENEFITS = 0.71 TIME 0.78 COST + 0.84 EFFICIENCY + 0.27 TAX + 0.81 EXECUTIVE + 0.86 ENFORCE + 0.79 LAW&ORDER + 0.87 CORRUPT + 0.92 EXPR OPR + 0.92 REPUDR + The table summarizes the stylized relationships between the country level institutional characteristics and the adjustment speeds

PAGE 115

115 Table 3 2. Description and Sources of Variables Determining the Optimal Capital Structure Variable Description and Source BLEV Book Leverage. Ratio of book value of total debt to book value of total assets. (Long term debt[106]+Short term debt[94])/Total assets[89] Source: Compustat Global Vantage. MLEV Market Leverage. Ratio of book value of total debt to book value of assets minus book value of equity plus market value of equity (Long term debt[106]+Short term debt[94])/(Total assets[89] Book equity[135]+Market Value of Equity[PRCCI *SHOI]. Source: Compustat Global Vantage. PROFIT Earnings before interest and taxes as a proportion of total assets. (Operating Income[14]+ Interest and related expense[15]+Current income taxes[24])/Total assets[89] Source: Compustat Global Vantage MB The ratio of assets market to book values. (Long term debt[106]+Short term debt[94]+Preferred capital[119]+Market Value of Equity[PRCCI*SHOI])/Total assets[89]. Source: Compustat Global Vantage. DEP_TA Depreciation expense as a proportion of total assets Total Depreciation and Amortization[11]/Total assets[89]. Source: Compustat Global Vantage. LnTA Log of total book assets (a measure of firm size). Log[89]. Source: Compustat Global Vantage. FA_TA Fixed assets as a proportion of total assets. Fixed assets[76]/Total assets[89]. Source: Compustat Global Vantage. R&D_TA Research and development expenses as a proportion of total assets. Firms with more intangible assets in the form of R&D expenses will prefer to have more equity. Research and Development Expense [52]/Total assets[89]. Source: Compustat Global Vantage. RD_DUM A dummy variable equal to one if R&D expenditures are not reported; otherwise zero. About 65% of the sample firm years do not report R& D expenses. For these firms, I set R&D expense to zero and set R&D_DUM equal to one. Source: Compustat Global Vantage IND_MED The prior years median leverage ratio for the firms industry. Industry classifications are based on the 48 industry categorie s in Fama and French (1997 Source: Compustat Global Vantage TAXES Ratio of total income taxes to pre tax income. Current income taxes[24]/Income before income taxes[21]. Source: Compustat Global Vantage. LIQUID Total current assets as a proportion of total current liabilities. Total Current Assets[75]/Total current liabilities[104]. Source: Compustat Global Vantage. REG A dummy variable equal to one for firms operating in regulated industries. A dummy variable equal to one for firms operating in regulated industries. SIC [49004999]. Source: Compustat Global Vantage. INF Annual Inflation rate. Growth in CPI. Source:WDI (World Development Indicators). GDPG Annual GDP growth. Growth in nominal GDP. Source:WDI (World Development Indicators).

PAGE 116

116 Table 3 3 Adjustment Speed Estimates Book Leverage Market Leverage Panel Structure BB LSDVC BB LSDVC Firms Average Obs per Firm Argentina 12.81 14.62 15.38 28.07 27 6 Australia 38.29 43.25 40.08 39.01 939 5 Austria 19.10 20.43 19.91 22.84 92 7 Belgium 15.08 25.43 18.48 30.33 116 7 Brazil 13.29 25.90 22.70 32.41 140 6 Canada 29.31 29.56 52.86 51.64 165 4 Switzerland 15.47 10.69 24.53 22.95 91 5 Chile 21.34 23.74 26.90 42.23 21 5 Columbia 4.03 3.65 12.26 4.56 152 8 Germany 25.87 22.83 27.55 30.49 119 8 Denmark 26.01 28.87 28.11 29.96 661 7 Spain 20.11 17.55 23.57 28.69 692 7 Finland 23.43 24.22 23.57 20.60 80 6 France 27.66 28.74 29.38 33.31 131 8 United Kingdom 31.45 32.57 38.71 36.81 262 7 Greece 11.40 32.58 17.36 31.91 220 7 Hong Kong 35.82 33.82 32.48 29.36 61 8 India 23.63 23.43 29.90 31.38 41 5 Indonesia 21.90 27.39 22.12 27.77 244 6 Ireland 15.65 21.76 17.05 29.08 3,063 6 Israel 32.86 27.40 38.39 37.88 679 7 Italy 9.46 10.39 27.36 36.87 81 7 Japan 15.07 22.38 23.06 38.43 95 6 South Korea 32.79 34.21 39.73 38.86 137 7 Mexico 20.43 32.85 23.36 34.24 45 5 Malaysia 12.96 16.81 22.10 29.43 21 6 Norway 26.04 32.07 28.13 31.36 93 5 New Zealand 40.61 36.23 47.46 43.49 42 8 Pakistan 13.37 11.09 13.80 16.72 400 7 Peru 9.42 17.98 10.87 13.72 193 5 Philippines 16.68 15.36 23.12 33.15 239 5 Portugal 6.40 4.88 18.33 37.20 147 8 Singapore 25.64 27.20 36.75 32.65 207 9 Thailand 24.20 25.90 28.59 24.76 300 8 Turkey 12.27 5.95 12.74 10.76 43 6 United States 24.10 27.22 25.32 28.07 1,568 8 South Africa 27.07 28.65 30.65 33.34 3,570 8 The definitions of the variables in the regressions are provided in Table 32. I report the adjustment speed estimates (%) obtained from Blundell and Bonds (1998) twostep system GMM estimations and Brunos (2005) Corrected Least Squares Dummy Variable (L SDVC) of the following model, run separately for each country, along with the number of firms and mean number of observations per firm available in the estimation sample: = ( ) 1+ ( 1 ) 1+ + ( 33)

PAGE 117

117 Table 3 4 The Frequency and Size of the Capital Structure Adjustments Panel A. Slow Adjusters Country Frequency of Adjustment (%) Size of Adjustment (%) Access Debt D. Issue D. Retire Equity E. Issue E. Retire Debt D. Issue D. Retire Equity E. Issue E. Retire Argentina 82.47 77.92 51.30 26.62 53.25 29.87 43.51 20.20 17.19 25.04 14.69 19.81 10.19 Austria 83.43 72.24 38.37 33.87 58.87 28.05 42.30 23.76 17.67 30.44 19.50 19.22 19.79 Belgium 83.49 71.78 38.52 33.26 56.67 31.15 38.76 22.86 16.34 30.22 17.99 15.49 20.97 Brazil 87.50 75.36 51.32 24.04 61.90 34.13 42.07 21.95 15.53 34.60 16.28 11.41 21.39 Chile 76.59 63.64 42.95 20.68 45.23 28.64 30.00 12.36 12.04 12.90 10.93 11.28 10.35 Colombia 84.11 55.14 41.12 14.02 67.29 54.21 19.63 9.52 11.08 5.86 11.92 13.23 7.84 Greece 89.06 78.57 49.11 29.46 68.30 29.24 50.22 27.61 18.52 41.53 22.65 15.31 28.84 Ireland 77.09 68.35 48.16 20.19 43.50 28.54 31.07 16.40 17.56 14.28 10.52 12.68 7.47 Italy 82.50 72.91 39.08 33.83 57.03 26.18 44.72 23.02 15.08 32.55 18.03 12.86 23.13 Japan 62.59 56.13 24.58 31.55 32.33 20.20 29.52 9.73 9.08 10.29 5.60 6.26 4.72 Malaysia 73.18 65.17 41.96 23.22 45.05 26.40 35.97 15.87 16.79 14.48 10.67 13.93 7.11 Mexico 78.88 65.42 49.06 16.35 55.20 28.11 41.40 16.57 16.42 16.94 12.97 15.10 10.77 Pakistan 83.04 73.91 48.26 25.65 53.48 34.78 31.74 17.11 19.99 12.52 11.80 13.99 8.55 Peru 80.88 70.59 40.44 30.15 52.21 26.47 36.03 15.04 16.01 13.78 10.37 12.07 8.40 Philippines 77.12 66.47 39.84 26.63 49.31 28.40 36.09 14.96 13.34 17.18 10.68 10.30 11.18 Portugal 87.00 75.23 43.96 31.27 57.59 30.96 43.03 22.47 17.00 29.63 16.98 12.14 22.36 Spain 82.62 71.77 43.63 28.14 53.18 28.66 38.04 22.29 15.86 31.74 17.24 10.47 25.83 Switzerland 77.20 61.46 32.05 29.40 55.98 30.76 37.42 13.40 13.35 13.45 13.55 13.71 13.38 Turkey 96.36 85.43 63.97 21.46 80.57 55.47 29.55 30.96 24.62 47.64 26.70 19.94 40.82 Panel B. Fast Adjusters Country Frequency of Adjustment (%) Size of Adjustment (%) Access Debt D. Issue D. Retire Equity E. Issue E. Retire Debt D. Issue D. Retire Equity E. Issue E. Retire Australia 81.58 66.82 40.53 26.28 54.26 40.37 23.65 16.48 15.23 18.23 17.63 20.97 10.99 Canada 89.19 70.81 51.22 19.59 72.16 61.49 17.16 21.30 22.17 19.39 29.83 32.95 18.17 Denmark 76.30 67.58 38.58 29.00 45.85 23.52 37.64 15.33 15.17 15.52 10.55 10.50 10.69 Finland 80.92 69.44 32.82 36.62 53.85 25.03 44.31 19.26 14.50 23.65 14.15 11.31 16.77 France 82.63 70.74 39.14 31.59 56.79 31.97 35.85 23.76 18.08 30.63 19.02 15.48 24.00 Germany 81.17 68.31 34.45 33.86 56.68 28.31 40.66 19.32 14.91 24.05 15.45 15.24 15.74 Hong Kong 76.85 62.31 36.48 25.83 50.46 32.04 31.48 15.26 13.67 17.27 12.80 13.49 11.85 India 82.98 68.51 50.96 17.55 50.67 35.79 25.92 12.19 12.74 11.02 8.39 8.82 7.52 Indonesia 82.62 75.88 47.65 28.22 57.11 28.42 42.50 23.07 21.06 26.29 14.97 18.12 12.13 Israel 71.72 63.64 43.94 19.70 44.44 27.78 24.75 16.20 14.57 19.69 9.40 10.86 6.84 New Zealand 78.78 67.71 37.45 30.26 52.21 30.44 32.29 15.71 16.16 15.20 12.38 13.08 11.56 Norway 86.02 73.48 43.99 29.50 57.45 34.12 36.28 19.64 18.37 21.36 15.58 15.89 15.21 S. Korea 84.89 67.49 39.64 27.84 62.90 39.56 33.70 15.45 14.85 16.28 12.51 12.72 12.24 Singapore 78.79 69.79 40.76 29.03 50.09 28.36 36.43 18.14 18.18 18.09 12.43 15.07 9.41 South Africa 80.47 70.32 50.39 19.92 44.58 24.36 37.87 15.90 14.80 18.35 9.67 10.21 9.13 Thailand 79.80 72.63 39.18 33.45 54.58 29.14 38.89 17.69 16.55 18.99 11.12 12.23 9.88 United Kingdom 76.44 68.72 43.60 25.11 44.62 27.62 30.84 16.25 16.40 16.00 12.52 14.87 9.39 United States 73.35 62.83 41.52 21.31 43.80 30.84 27.55 15.09 15.90 13.72 13.41 17.47 6.71 Panel C. Slow versus Fast Adjusters Sample Frequency of Adjustment (%) Size of Adjustment (%) Access Debt D. Issue D. Retire Equity E. Issue E. Retire Debt D. Issue D. Retire Equity E. Issue E. Retire Slow Adjusters 72.32 62.73 38.39 26.69 44.16 28.31 31.44 14.68 14.65 14.83 12.05 14.11 8.74 Fast Adjusters 80.15 69.27 42.10 28.24 52.73 32.36 34.49 17.99 16.42 20.37 14.65 15.89 12.99 Difference 7.83 6.54 3.71 1.55 8.57 4.05 3.05 3.31 1.77 5.54 2.60 1.78 4.25 Significance *** *** *** *** *** *** *** *** *** *** *** *** *** *The table presents summary information on the frequency and magnitude of the capital structure adjustments for slow and fast adjusters Slow (fast) adjusters are the group of countries with adjustment speed smaller (larger) than the sample median. Access is defined as a change in absolute value of outstanding equity or debt exceeding 5% of total assets. An issuance or retirement is define d as having occurred in a given year if the net change in equity (E.) or debt (D.), normalized by the book value of assets at the end of the previous period, is greater than a certain cut off. Debt (Equity) is defined as a change in absolute value of outst anding debt (equity) exceeding 5% of total assets. Debt Issue, Debt Retirement, and Equity Issue each are defined as a net security issuance or repurchase of at least 5% of book assets. Equity Retirement is defined as a net securit y repurchase of at least 1.25% of book assets. Each year, the median value of the size of capital structure adjustment is calculated for each country. The reported statistic is the mean of these timeseries medians. *, **, *** indicate significant difference between the groups at the 10%, 5%, and 1% significance levels, respectively.

PAGE 118

118 Table 3 5 The Impact of the Legal and Financial Traditions on the Adjustment Speed Panel A. Legal Traditions SEPARATE POOLED Group Number of Mean Difference Significance Mean Difference Significance Countries (%) (%) (%) (%) Common 14 26.66 8.93 *** 18.91 14.12 *** Civil 23 17.73 4.79 English 14 26.66 11.73 *** 18.91 12.96 *** French 15 14.94 5.95 English 14 26.66 1.50 *** 18.91 4.13 *** Scandinavian 3 25.16 14.78 English 14 26.66 5.00 *** 18.91 12.54 *** German 5 21.66 6.37 French 15 14.94 10.23 *** 5.95 8.83 *** Scandinavian 3 25.16 14.78 French 15 14.94 6.73 *** 5.95 0.43 *** German 5 21.66 6.37 Scandinavian 3 25.16 3.50 *** 14.78 8.40 *** German 5 21.66 6.37 Panel B. Financial System Structure SEPARATE POOLED Group Number of Mean Difference Significance Mean Difference Significance Countries (%) (%) (%) Bank 22 19.94 2.87 *** 2.74 16.19 *** Market 15 22.82 18.93 Panel C. Financial System Development SEPARATE POOLED Group Number of Mean Difference Significance Mean Difference Significance Countries (%) (%) Low Size 16 15.56 9.36 *** 0.66 15.63 *** High Size 17 24.92 16.30 Low Efficiency 16 14.89 10.74 *** 3.58 12.83 *** High Efficiency 17 25.63 16.41 Low Aggregate Quality 16 15.08 10.34 *** 2.10 14.09 *** High Aggregate Quality 17 25.42 16.19 *The definitions of the variables are provided in Table 31. I allocate the countries in the sample into a legal origin or financial system based portfolio and compare the adjustment speeds reported in Table 33, column 1 using the SEPARATE and POOLED method. indicate significant difference between gro ups at the 10%, 5%, and 1% significance levels, respectively.

PAGE 119

119 Table 3 6 Institutional Determinants of the Adjustment Speed Individual Indices Panel A. Adjustment Costs SEPARATE POOLED Panel A 1.Ease of Access Group # Countries Mean Difference Significance Mean Difference Significance EQUITY CONTRACTS RIGHTS Poor 20 18.48 5.73 *** 5.68 11.41 *** Strong 17 24.21 17.10 ENFORCEMENT Poor 19 18.00 6.39 *** 7.25 9.74 *** Strong 18 24.39 16.99 DEBT CONTRACTS RIGHTS Poor 22 17.30 11.74 *** 13.58 1.06 *** Strong 15 29.04 14.64 ENFORCEMENT Strong 19 23.78 5.49 *** 16.71 12.65 *** Poor 18 18.29 4.07 Panel A 2.Asymmetric Info Group # Countries Mean Difference Significance Mean Difference Significance TRANSPARENCY Poor 19 17.83 8.14 *** 5.96 11.04 *** Strong 15 25.97 17.01 EQUITY MARKETS DISCLOSURE Poor 23 18.18 7.75 *** 5.66 11.48 *** Strong 14 25.93 17.14 LIABILITY Poor 19 18.64 5.08 *** 8.98 7.59 *** Strong 18 23.72 16.57 ENFORCEMENT Poor 21 18.74 5.48 *** 5.21 12.74 *** Strong 16 24.22 17.95 INSIDER TRADING Poor 19 18.72 5.52 *** 10.03 6.49 *** Strong 17 24.24 16.52 DEBT MARKETS INFORMATION SHARING Poor 7 21.24 0.16 10.46 5.09 *** Strong 30 21.08 15.55 Panel A 3.Cash Constraints Group # Countries Mean Difference Significance Mean Difference Significance EQUITY CONTRACTS DIVIDENDS Low 33 22.15 9.63 *** 15.22 2.76 High 4 12.52 17.99 DEBT CONTRACTS RESERVE REQUIREMENTS Low 21 24.99 8.96 *** 17.44 13.39 *** High 16 16.02 4.05 Panel B. Adjustment Benefits SEPARATE POOLED Panel B 1. Distress Costs Group # Countries Mean Difference Significance Mean Difference Significance TIME TO REPAY Short 1 7 25.79 9.23 *** 13.03 2.09 Long 1 8 16.56 15.12 BANKRUPTCY COSTS Low 22 22.36 2.82 *** 16.93 12.39 *** High 13 19.53 4.54 BANKRUPTCY EFFICIENCY Low 18 17.83 7.17 *** 6.29 11.56 *** High 17 25.00 17.85 Panel B 2. Tax Shields Group # Countries Mean Difference Significance Mean Difference Significance TAX SHIELDS Low 19 20.95 0.33 *** 16.25 6.22 High 18 21.28 10.03 Panel B 3. Deviation Costs Group # Countries Mean Difference Significance Mean Difference Significance EXECUTIVE Low 18 16.53 7.87 *** 9.27 7.18 *** High 17 24.41 16.45 ENFORCE Low 18 15.92 8.68 *** 7.87 8.90 *** High 17 24.60 16.76 LAW & ORDER Low 23 18.42 6.55 *** 6.44 12.37 *** High 13 24.97 18.81 GOVERNANCE Low 19 17.18 8.09 *** 6.89 9.71 *** High 18 25.26 16.60 *The definitions of the variables are provided in Table 3 1. I allocate the sample countries into portfolios based on the sample median of the individual institutional indices and compare the adjustment speeds reported in Table 3 3, column 1 using the SEPARATE and POOLED method. Countries with a strong (poor) leve l of an institutional attribute are countries with an index above (below) the sample median. indicate significant difference between the groups at the 10%, 5%, and 1% significance levels, respectively.

PAGE 120

120 Table 3 7 Institutional Determinants of the Adjustment Speed -Principal Components Panel A. Adjustment Costs SEPARATE POOLED Panel A 1. Ease of Access Group # Countries Mean Difference Significance Mean Difference Significance SHAREHOLDER Poor 19 17.97 6.46 *** 5.4 2 12.22 *** PROTECTION Strong 18 24.43 17.64 CREDITOR Poor 19 16.41 9.67 *** 13.24 2.43 *** PROTECTION Strong 18 26.07 15.67 INVESTOR Poor 19 17.29 7.86 *** 5.616 11.75 *** PROTECTION Strong 18 25.15 17.37 Panel A 2. Asymmetric Info Group # Countries Mean Difference Significance Mean Difference Significance EQUITY MARKETS Poor 17 16.71 9.43 *** 12.61 3.82 *** Strong 17 26.14 16.43 DEBT MARKETS Poor 17 17.85 8.10 *** 8.778 8.11 *** Strong 17 25.95 16.89 ALL MARKETS Poor 17 16.07 10.71 *** 9.485 7.28 *** Strong 17 26.78 16.76 Panel A 3. External Financing Costs Group # Countries Mean Difference Significance Mean Difference Significance EQUITY MARKETS Poor 17 16.19 10.46 *** 5.7 8 11.51 *** Strong 17 26.65 17.29 DEBT MARKETS Poor 17 16.26 10.33 *** 6.597 10.62 *** Strong 17 26.59 17.22 ALL MARKETS Poor 17 15.65 11.56 *** 6 .60 10.56 *** Strong 17 27.20 17.16 Panel A 4. Cash Constraints Group # Countries Mean Difference Significance Mean Difference Significance CASH CONSTRAINTS High 21 17.54 8.26 *** 4.2 2 13.74 *** Low 16 25.80 17.95 Panel A 5. Combined Adjustment Costs Group # Countries Mean Difference Significance Mean Difference Significance EQUITY COSTS Low 17 16.19 10.46 *** 5.7 8 11.52 *** High 17 26.65 17.29 DEBT COSTS Low 17 15.64 11.56 *** 3.6 5 14.83 *** High 17 27.20 18.48 COMBINED COSTS Low 17 15.65 11.56 *** 6. 60 10.56 *** High 17 27.20 17.16 Panel B. Adjustment Benefits SEPARATE POOLED Group # Countries Mean Difference Significance Mean Difference Significance DISTRESS COSTS Strong 1 7 23.82 5.18 *** 17.60 11.35 *** Poor 1 8 18.65 6.24 TAX SHIELDS Poor 19 20.95 0.33 *** 16.25 6.22 Strong 18 21.28 10.03 DEVIATION COSTS Poor 17 15.96 7.60 *** 7.05 9.05 *** Strong 17 23.56 16.10 COMBINED BENEFITS Poor 16 16.14 7.50 *** 7.16 9.19 *** Strong 16 23.64 16.35 The definitions of the variables are provided in Table 31. I allocate the sample countries into portfolios based on the sample median of the composite institutional indices and then compare the adjustment speeds reported in Table 3 3, column 1 using the SEPARATE and POOLED method. Countries with a strong (poor) level of an institutional attribute are countries with an index abo ve (below) the sample median. *, **, *** indicate significant difference between the groups at the 10%, 5%, and 1% significance levels, respectively.

PAGE 121

121 Table 3 8 Alternative Definition of Leverage and Estimation Method BB LSDVC Panel A. Adjustment Costs MARKET BOOK MARKET Panel A 1.Ease of Access Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance EQUITY CONTRACTS RIGHTS Poor 20 23.81 5.39 *** 21.09 5.14 *** 27.91 5.40 *** Strong 17 29.20 26.23 33.30 ENFORCEMENT Poor 19 22.61 7.57 *** 20.59 5.87 *** 28.11 4.68 *** Strong 18 30.18 26.46 32.79 DEBT CONTRACTS RIGHTS Poor 22 22.83 10.66 *** 20.61 8.74 *** 28.81 4.87 *** Strong 15 33.49 29.36 33.68 ENFORCEMENT Strong 18 29.13 5.83 *** 26.24 5.73 *** 31.29 1.85 *** Poor 19 23.29 20.51 29.44 Panel A 2.Asymmetric Info Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance TRANSPARENCY Poor 19 22.83 9.05 *** 21.19 6.06 *** 28.61 4.96 *** Strong 15 31.88 27.25 33.57 EQUITY MARKETS DISCLOSURE Poor 23 22.94 8.84 *** 20.63 7.44 *** 28.16 5.88 *** Strong 14 31.78 28.08 34.04 LIABILITY Poor 19 23.21 6.32 *** 21.78 3.44 *** 28.23 4.44 *** Strong 18 29.53 25.21 32.67 ENFORCEMENT Poor 21 23.92 5.47 *** 21.97 3.43 *** 29.64 1.73 *** Strong 16 29.39 25.39 31.37 INSIDER TRADING Poor 19 23.65 6.32 *** 22.25 3.25 *** 28.57 4.65 *** Strong 17 29.97 25.51 33.22 DEBT MARKETS INFORMATION SHARING Poor 7 24.75 1.89 *** 21.39 2.54 *** 25.34 6.23 Strong 30 26.65 23.93 31.57 Panel A 3.Cash Constraints Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance EQUITY CONTRACTS DIVIDENDS Low 33 27.07 7.27 *** 23.69 2.22 *** 30.70 2.93 *** High 4 19.81 21.47 27.78 DEBT CONTRACTS RESERVE REQUIREMENTS Low 21 29.44 7.29 *** 26.16 6.26 *** 31.53 2.64 *** High 16 22.15 19.89 28.89

PAGE 122

122 Table 3 8 Continued BB LSDVC Panel B. Adjustment Benefits MARKET BOOK MARKET Panel B 1. Distress Costs Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance TIME TO REPAY Short 18 31.36 9.47 *** 27.22 7.28 *** 33.98 6.44 *** Long 17 21.90 19.95 27.54 BANKRUPTCY COSTS Low 13 27.49 1.96 *** 24.82 3.05 *** 30.21 1.74 High 22 25.53 21.77 31.94 BANKRUPTCY EFFICIENCY Low 18 23.51 6.70 *** 20.41 6.74 *** 28.41 5.02 *** High 17 30.21 27.16 33.43 Panel B 2. Tax Shields Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance TAX SHIELDS Short 19 25.71 1.19 *** 23.89 0.92 30.49 0.22 Long 18 26.90 22.98 30.28 Panel B 3. Deviation Costs Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance EXECUTIVE Short 18 21.62 8.45 *** 19.76 6.36 *** 27.39 5.74 *** Long 17 30.07 26.11 33.13 ENFORCE Low 18 21.68 7.88 *** 19.13 7.54 *** 27.93 4.35 *** High 17 29.56 26.66 32.28 LAW & ORDER Low 23 22.74 8.78 *** 21.26 5.24 *** 28.70 4.01 *** High 13 31.53 26.50 32.72 GOVERNANCE Low 19 22.93 6.90 *** 20.15 6.79 *** 28.70 3.47 *** High 18 29.83 26.93 32.17

PAGE 123

123 Table 3 8 Continued BB LSDVC Panel A. Adjustment Costs MARKET BOOK MARKET Panel A 1. Ease of Access Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance SHAREHOLDER Poor 19 22.95 6.87 *** 20.89 5.26 *** 28.83 3.20 *** PROTECTION Strong 18 29.82 26.15 32.03 CREDITOR Poor 19 21.66 9.51 *** 19.88 7.33 *** 28.88 3.09 *** PROTECTION Strong 18 31.17 27.21 31.97 INVESTOR Poor 19 22.86 7.06 *** 20.50 6.06 *** 29.15 2.55 *** PROTECTION Strong 18 29.91 26.56 31.70 Panel A 2. Asymmetric Info Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance EQUITY MARKETS Poor 17 21.95 9.73 *** 20.65 6.43 *** 28.19 5.21 *** Strong 17 31.69 27.08 33.40 DEBT MARKETS Poor 17 23.05 8.56 *** 21.77 4.75 *** 28.58 5.04 *** Strong 17 31.60 26.52 33.61 ALL MARKETS Poor 17 20.92 11.81 *** 20.24 7.24 *** 27.54 6.51 *** Strong 17 32.72 27.48 34.05 Panel A 3. External Financing Costs Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance EQUITY MARKETS Poor 17 21.71 10.21 *** 19.56 8.60 *** 28.00 5.59 *** Strong 17 31.93 28.16 33.59 DEBT MARKETS Poor 17 21.27 11.11 *** 19.46 8.81 *** 27.82 5.95 *** Strong 17 32.38 28.27 33.77 ALL MARKETS Poor 17 21.42 10.80 *** 18.77 10.19 *** 28.19 5.21 *** Strong 17 32.22 28.96 33.40 Panel A 4. Cash Constraints Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance CASH CONSTRAINTS High 21 22.78 8.12 *** 21.03 5.59 *** 28.76 3.76 *** Low 16 30.90 26.62 32.52 Panel A 5. Combined Adjustment Costs Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance EQUITY COSTS Low 17 21.71 10.21 *** 19.56 8.60 *** 28.00 5.59 *** High 17 31.93 28.16 33.59 DEBT COSTS Low 17 21.69 10.26 *** 18.59 10.55 *** 28.67 4.26 *** High 17 31.95 29.14 32.93 COMBINED COSTS Low 17 21.42 10.80 *** 18.77 10.19 *** 28.19 5.21 *** High 17 32.22 28.96 33.40 BB LS DVC Panel B. Adjustment Benefits MARKET BOOK MARKET Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance DISTRESS COSTS Strong 18 29.18 4.98 *** 25.51 3.76 *** 31.86 2.07 *** Poor 17 24.20 21.76 29.79 TAX SHIELDS Poor 19 25.71 1.19 *** 23.89 0.92 30.49 0.22 Strong 18 26.90 22.98 30.28 DEVIATION COSTS Poor 17 20.81 8.56 *** 19.31 6.28 *** 27.08 5.41 *** Strong 17 29.37 25.59 32.49 COMBINED BENEFITS Poor 16 22.50 6.06 *** 19.38 6.54 *** 28.87 2.77 *** Strong 16 28.57 25.92 31.64 *The defini tions of the variables are provided in Table 31. I allocate the sample countries into portfolios based on the sample median of the simple and composite institutional indices and then compare the adjustment speeds reported in Table 3 3, column 24 using the SEPARATE method. Countries with a strong (poor) level of an institutional attribute are countries with an index above (below) the sample median. indicate significant difference between the groups at the 10%, 5%, and 1% significance levels, res pectively.

PAGE 124

124 Table 3 9 Variable Adjustment Speed Estimates The definitions of the variables in the reg ressions are provided in Table 31. I report the adjustment speed estimates (%) obtained from the Blundell and Bonds (1998) twostep system GMM estimations of the following model, run separately for each country: 1= 1 1 1 + 1+ (3 7) Uniform Adjustment Variable Adjustment Mean Median Argentina 12.81 16.41 16.03 Australia 38.29 44.48 45.00 Austria 19.10 28.57 29.02 Belgium 15.08 24.05 24.01 Brazil 13.29 20.13 20.78 Canada 29.31 36.11 37.41 Switzerland 15.47 16.73 15.83 Chile 21.34 30.17 29.09 Columbia 4.03 8.95 7.39 Germany 25.87 18.62 19.02 Denmark 26.01 13.87 13.46 Spain 20.11 33.46 32.80 Finland 23.43 30.86 32.19 France 27.66 19.34 18.77 United Kingdom 31.45 40.05 39.72 Greece 11.40 37.60 36.31 Hong Kong 35.82 33.58 33.01 Indonesia 23.63 32.92 33.68 India 21.90 27.70 28.22 Ireland 15.65 39.07 38.86 Israel 32.86 41.01 37.85 Italy 9.46 13.83 13.91 Japan 15.07 14.09 14.62 S. Korea 32.79 49.60 47.83 Mexico 20.43 40.36 39.02 Malaysia 12.96 14.59 14.60 Norway 26.04 16.09 16.37 New Zealand 40.61 27.61 26.07 Pakistan 13.37 20.13 20.79 Peru 9.42 14.15 14.06 Philippines 16.68 21.72 21.96 Portugal 6.40 17.97 18.75 Singapore 25.64 31.78 31.46 Thailand 24.20 32.29 32.10 Turkey 12.27 25.16 23.47 United States 24.10 27.82 28.03 South Africa 27.07 34.76 36.13

PAGE 125

125 Table 3 10. Institutional Determinants of the Variable Adjustment Speed Individual Indices Panel A. Adjustment Costs SEPARATE POOLED Panel A 1.Ease of Access Group # Countries Mean Difference Significance Mean Difference Significance EQUITY CONTRACTS RIGHTS Poor 20 23.62 3.50 *** 10.78 13.51 *** Strong 17 27.13 24.29 ENFORCEMENT Poor 19 22.09 5.34 *** 12.39 11.83 *** Strong 18 27.43 24.22 DEBT CONTRACTS RIGHTS Poor 22 24.29 5.82 *** 22.57 3.74 Strong 15 30.11 18.83 ENFORCEMENT Poor 18 24.51 2.22 *** 9.41 14.32 *** Strong 19 26.73 23.73 Panel A 2.Asymmetric Info Group # Countries Mean Difference Significance Mean Difference Significance TRANSPARENCY Poor 19 25.37 0.94 *** 11.56 12.65 *** Strong 15 26.32 24.21 EQUITY MARKETS DISCLOSURE Poor 23 22.87 4.25 *** 10.53 13.79 *** Strong 14 27.11 24.32 LIABILITY Poor 19 21.55 6.08 *** 14.83 8.89 *** Strong 18 27.63 23.72 ENFORCEMENT Poor 21 19.31 10.85 *** 9.81 16.15 *** Strong 16 30.16 25.96 INSIDER TRADING Poor 19 25.94 0.46 *** 15.04 8.64 *** Strong 17 26.40 23.68 DEBT MARKETS INFORMATION SHARING No 7 26.73 0.05 17.28 5.24 *** Yes 30 26.23 22.52 Panel A 3.Cash Constraints Group # Countries Mean Difference Significance Mean Difference Significance EQUITY CONTRACTS DIVIDENDS Low 33 26.30 0.44 22.22 3.11 High 4 25.86 25.33 DEBT CONTRACTS RESERVE REQUIREMENTS Low 21 29.42 11.50 *** 25.29 16.93 *** High 16 17.92 8.36 Panel B. Adjustment Benefits SEPARATE POOLED Panel B 1. Distress Costs Group # Countries Mean Difference Significance Mean Difference Significance TIME TO REPAY Long 18 25.99 0.52 *** 22.74 3.42 *** Short 17 26.52 19.32 BANKRUPTCY COSTS High 13 25.27 2.37 *** 8.70 15.42 *** Low 22 26.65 24.12 BANKRUPTCY Low 18 21.31 6.67 *** 10.51 14.62 *** High 17 27.99 25.13 Panel B 2. Tax Shields Group # Countries Mean Difference Significance Mean Difference Significance TAX SHIELDS Low 19 28.77 6.61 22 .00 6.44 High 18 22.17 15.56 Panel B 3. Deviation Costs Group # Countries Mean Difference Significance Mean Difference Significance EXECUTIVE Poor 18 24.33 2.08 *** 15.39 7.91 *** Strong 17 26.41 23.30 ENFORCE Poor 18 25.33 1.08 *** 12.13 11.80 *** Strong 17 26.41 23.93 LAW & ORDER Poor 23 19.79 11.17 *** 11.09 15.81 *** Strong 13 30.97 26.90 GOVERNANCE Poor 19 25.49 0.97 *** 11.22 12.47 *** Strong 18 26.46 23.69 I allocate the sample countries into portfolios based on the sample median of the individual institutional indices and take t he variable adjustment speed estimates of E quation 3 7 to compare those using SEPARATE and POOLED methods. Countries with a strong (poor) level of an institutional attribute are countries with an index above (below) the sample median. indicate significant difference between the groups at the 10%, 5%, and 1% significance levels, respectively.

PAGE 126

126 Table 3 11. Institutional Determ inants of the Variable Adjustment Speed-Principal Components Panel A. Adjustment Costs SEPARATE POOLED Panel A 1.Ease of Access Group # Countries Mean Difference Significance Mean Difference Significance SHAREHOLDER Poor 19 22.40 10.72 *** 9.93 14.98 *** PROTECTION Strong 18 33.12 24.91 CREDITOR Poor 19 23.13 4.09 *** 19.72 3.73 *** PROTECTION Strong 18 27.21 23.45 INVESTOR Poor 19 18.82 11.91 *** 10.22 14.37 *** PROTECTION Strong 18 30.72 24.59 Panel A 2. Asymmetric Info Group # Countries Mean Difference Significance Mean Difference Significance EQUITY MARKETS Poor 17 21.04 6.79 *** 19.13 4.43 *** Strong 17 27.83 23.56 DEBT MARKETS Poor 17 25.22 1.11 *** 14.95 9.07 *** Strong 17 26.33 24.02 ALL MARKETS Poor 17 22.42 4.65 *** 15.27 8.64 *** Strong 17 27.07 23.91 Panel A 3.External Financing Costs Group # Countries Mean Difference Significance Mean Difference Significance EQUITY MARKETS Poor 17 22.69 4.30 *** 10.35 14.17 *** Strong 17 26.99 24.52 DEBT MARKETS Poor 17 21.00 6.34 *** 11.79 12.60 *** Strong 17 27.34 24.39 ALL MARKETS Poor 17 21.00 6.34 *** 11.38 13.00 *** Strong 17 27.34 24.38 Panel A 4. Cash Constraints Group # Countries Mean Difference Significance Mean Difference Significance CASH CONSTRAINTS High 21 19.19 11.85 *** 8.93 17.21 *** Low 16 31.04 26.14 Panel A 5. Combined Adjustment Costs Group # Countries Mean Difference Significance Mean Difference Significance EQUITY COSTS Low 17 23.67 2.92 *** 10.35 14.17 *** High 17 26.60 24.52 DEBT COSTS Low 17 22.69 4.30 *** 8.18 18.22 *** High 17 26.99 26.4 COMBINED COSTS Low 17 22.69 4.30 *** 11.38 13.00 *** High 17 26.99 24.38 Panel B. Adjustment Benefits SEPARATE POOLED Group # Countries Mean Difference Significance Mean Difference Significance DISTRESS COSTS Low 18 21.68 5.89 *** 9.96 14.88 *** High 17 27.57 24.84 TAX SHIELDS Low 19 28.77 6.61 22.00 6.44 High 18 22.17 15.56 DEVIATION COSTS Low 17 25.40 0.66 *** 11.44 11.73 *** High 17 26.05 23.17 COMBINED BENEFITS Low 16 24.15 2.12 *** 12.38 10.75 *** High 16 26.28 23.13 I allocate the sample countries into portfolios based on the sample median of the composite institutional indices and take the variable adjustment speed estimates of E quation 37 to compare those using SEPARATE and POOLED methods. Countries with a strong (poor) level of an institutional attribute are countries with an index above (below) the sample median. indicate significant difference between the groups at the 10%, 5% and 1% significance levels, respectively.

PAGE 127

127 Table 3 12. Asymmetric Response to Adjustment Costs and BenefitsIndividual Indices Panel A. Adjustment Costs OVER LEVERAGED UNDER LEVERAGED SMALL LARGE (1) (2) (3) (4) Panel A1. Ease of access Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance Mean Difference Significance EQUITY CONTRACTS RIGHTS Poor 20 15.14 11.14 *** 9.68 14.03 *** 9.03 16.34 *** 10.45 15.71 *** Strong 17 26.28 23.71 25.37 26.16 ENFORCEMENT Poor 19 16.06 10.58 *** 11.29 12.7 *** 10.99 13.80 *** 11.73 13.43 *** Strong 18 26.64 23.99 24.79 25.16 DEBT CONTRACTS RIGHTS Poor 22 23.91 0.74 *** 22.34 4.56 24.80 10.45 23.71 5.01 Strong 15 24.65 17.78 14.35 18.7 0 ENFORCEMENT Poor 18 14.07 11.52 *** 8.11 15.47 *** 24.59 16.64 *** 25.31 16.94 *** Strong 19 25.59 23.58 7.95 8.37 Panel A2. Asymmetric Info Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance Mean Difference Significance TRANSPARENCY Poor 19 15.75 10.45 *** 10.2 0 13.74 *** 11.64 13.38 *** 10.63 15.32 *** Strong 15 26.2 0 23.94 25.02 25.95 EQUITY MARKETS DISCLOSURE Poor 23 14.99 11.58 *** 9.38 14.55 *** 8.88 6.23 *** 9.88 15.9 *** Strong 14 26.57 23.93 15.11 25.78 LIABILITY Poor 19 17.83 8.52 *** 13.77 9.62 *** 12.48 11.7 *** 15.43 9.61 *** Strong 18 26.35 23.39 24.18 25.04 ENFORCEMENT Poor 21 12.57 16.22 *** 8.73 16.28 *** 10.68 12.82 *** 10.10 15.32 *** Strong 16 28.79 25.01 23.5 0 25.42 INSIDER TRADING Poor 19 19.55 6.04 *** 14.31 8.99 *** 14.76 9.72 *** 15.27 10.75 *** Strong 17 25.59 23.30 24.48 26.02 DEBT MARKETS INFORMATION SHARING No 7 20.43 4.48 *** 17.53 4.29 *** 17.53 4.96 *** 14.93 9.37 *** Yes 30 24.91 21.82 22.49 24.3 0 Panel A3. Financial Constraints Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance Mean Difference Significance EQUITY CONTRACTS DIVIDENDS No 33 24.92 1.11 21.77 1.74 21.96 5.83 *** 23.23 5.06 Yes 4 26.03 23.51 16.13 28.29 DEBT CONTRACTS RESERVE REQUIREMENTS Low 21 28.20 17.25 *** 24.46 17.19 *** 22.62 13.59 *** 25.13 16.88 *** High 16 10.95 7.27 9.03 8.25

PAGE 128

128 Table 3 12. Continued Panel B. Adjustment Benefits OVER LEVERAGED UNDER LEVERAGED SMALL LARGE (1) (2) (3) (4) Panel B1. Distress Costs Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance Mean Difference Significance TIME TO REPAY Long 18 23.29 1.22 *** 18.37 4.00 *** 17.9 0 3.62 22.56 0.37 *** Short 17 24.51 22.37 21.52 22.19 BANKRUPTCY COSTS High 13 13.64 12.59 *** 7.95 15.77 *** 25.32 17.3 *** 25.41 16.87 *** Low 22 26.23 23.72 8.02 8.54 BANKRUPTCY EFFICIENCY Low 18 14.61 12.49 *** 9.74 14.97 *** 8.59 18.73 *** 10.44 16.75 *** High 17 27.10 24.71 27.32 27.19 Panel B2. Tax Shields Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance Mean Difference Significance TAX SHIELDS Low 19 26.21 6.88 23.00 8.16 22.12 6.35 24.68 9.55 High 18 19.33 14.84 15.77 15.13 Panel B3. Deviation Costs Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance Mean Difference Significance EXECUTIVE Poor 18 17.54 8.03 *** 13.61 9.43 *** 12.04 12.85 *** 13.92 11.62 *** Strong 17 25.57 23.04 24.89 25.54 ENFORCE Poor 18 16.36 9.56 *** 11.33 12.33 *** 11.31 13.73 *** 12.32 13.52 *** Strong 17 25.92 23.66 25.04 25.84 LAW & ORDER Poor 23 14.09 14.57 *** 10.36 15.45 *** 11.12 13.74 *** 12.29 12.9 0 *** Strong 13 28.66 25.81 24.86 25.19 GOVERNMENT Poor 19 15.86 9.83 *** 10.36 12.94 *** 10.23 14.62 *** 11.34 14.14 *** Strong 18 25.69 23.30 24.85 25.48 I allocate the sample countries into portfolios based on the sample median of the individual institutional indices and take t he variable adjustment speed estimates of Equation 37 to compare those using the POOLED method. Countries with a strong (poor) level of an institutional attribute are countries wi th an index above (below) the sample median. A firm is over levered (under levered) if the actual leverage is 10% above (below) th e calculated target leverage. indicate significant difference between the groups at the 10%, 5%, and 1% significance levels, respectively.

PAGE 129

129 Table 3 13. Asymmetric Response to Adjustment Factors Principal Components Panel A. Adjustment Costs OVER LEVERAGED UNDER LEVERAGED SMALL LARGE (1) (2) (3) (4) Panel A1. Ease of access Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance Mean Difference Significance SHAREHOLDER Poor 19 14.50 12.32 *** 8.85 15.68 *** 8.45 17.29 *** 9.29 17.27 *** PROTECTION Strong 18 26.82 24.53 25.74 26.56 CREDITOR Poor 19 21.01 8.78 *** 19.26 2.70 *** 21.73 3.31 21.22 2.14 *** PROTECTION Strong 18 29.79 21.96 18.42 23.36 INVESTOR Poor 19 14.72 11.82 *** 9.04 15.29 *** 8.23 17.16 *** 9.88 16.49 *** PROTECTION Strong 18 26.54 24.33 25.39 26.37 Panel A2. Asymmetric Info Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance Mean Difference Significance EQUITY MARKETS Poor 17 22.72 2.95 *** 17.88 5.33 *** 16.96 7.40 *** 19.84 5.40 *** Strong 17 25.67 23.21 24.36 25.24 DEBT MARKETS Poor 17 18.06 7.98 *** 13.87 9.71 *** 13.14 11.69 *** 15.39 10.55 *** Strong 17 26.04 23.58 24.83 25.94 ALL MARKETS Poor 17 18.61 7.58 *** 13.89 9.64 *** 14.13 10.49 *** 15.95 9.45 *** Strong 17 26.19 23.53 24.62 25.40 Panel A3. External Financing Costs Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance Mean Difference Significance EQUITY MARKETS Poor 17 14.56 12.00 *** 9.04 15.21 *** 9.02 16.09 *** 10.31 16.09 *** Strong 17 26.56 24.25 25.11 26.40 DEBT MARKETS Poor 17 15.19 11.35 *** 10.47 13.60 *** 10.07 15.20 *** 12.66 13.55 *** Strong 17 26.54 24.07 25.27 26.21 ALL MARKETS Poor 17 15.62 10.78 *** 10.00 14.13 *** 9.61 15.28 *** 11.68 14.70 *** Stron g 17 26.40 24.13 24.89 26.38 Panel A4. Financial Constraints Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance Mean Difference Significance CASH CONSTRAINTS High 21 11.80 17.15 *** 7.87 17.34 *** 9.86 13.57 *** 9.05 16.32 *** Low 16 28.95 25.21 23.43 25.37 Panel A5. Combined Adjustment Costs Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance Mean Difference Significance EQUITY COSTS Low 17 14.56 12.00 *** 9.04 15.21 *** 9.02 16.09 *** 10.31 16.09 *** High 17 26.56 24.25 25.11 26.40 DEBT COSTS Low 17 11.07 17.59 *** 7.06 18.42 *** 8.97 15.41 *** 7.99 19.18 *** High 17 28.66 25.48 24.38 27.17 COMBINED COSTS Low 17 15.62 10.78 *** 10.00 14.13 *** 9.61 15.28 *** 11.68 14.70 *** High 17 26.40 24.13 24.89 26.38

PAGE 130

130 Table 3 13. Continued Panel B. Adjustment Benefits OVER LEVERAGED UNDER LEVERAGED SMALL LARGE (1) (2) (3) (4) Group # Countries Mean Difference Significance Mean Difference Significance Mean Difference Significance Mean Difference Significance DISTRESS COSTS Low 18 14.22 12.66 *** 9.35 15.00 *** 26.17 17.66 *** 26.63 16.51 *** High 17 26.88 24.35 8.51 10.12 TAX Low 19 26.21 6.88 23.00 8.16 22.12 6.35 24.68 9.55 High 18 19.33 14.84 15.77 15.13 DEVIATION COSTS Low 17 15.77 9.34 *** 10.74 11.94 *** 11.47 13.03 *** 12.05 12.95 *** H igh 17 25.11 22.68 24.50 25.00 COMBINED BENEFITS Low 16 16.85 8.19 *** 11.82 10.83 *** 9.42 15.63 *** 9.67 16.42 *** H igh 16 25.04 22.65 25.05 26.09 I allocate the sample countries into portfolios based on the sample median of the composite institutional indices and take the variable adjustmen t speed estimates of Equation 37 to compare those using the POOLED method. Countries with a strong (poor) level of an institutional attribute are countries wi th an index above (below) the sample median. A firm is over levered (under levered) if the actual leverage is 10% above (below) the calculated target leverage. indicate significant difference between the groups at the 10%, 5%, and 1% significance levels, respectively.

PAGE 131

131 Figure 3 1 .Adjustment Speed Estimates. The figure ranks the adjustment speed estimates (%) reported in Table 3 3 from smallest to largest based on column 1 and plot s columns 1 through 4 for the countries in the sample. BB and LSDVC refer to Blundell and Bonds (1998) Two-step System GMM and Brunos (2005) Corrected Least Squares Dummy Variable Approach respectively. 0 10 20 30 40Adjustment Speed Colombia Portugal Peru Italy Greece Turkey Argentina Malaysia Brazil Pakistan Japan Belgium Switzerland Ireland Philippines Austria Spain Mexico Chile India Finland Indonesia United States Thailand Singapore Germany Denmark Norway South Africa France Canada United Kingdom S. Korea Israel Hong Kong Australia New ZealandAdjustment Speeds-Book Leverage (BB) 0 10 20 30 40Adjustment Speed Colombia Portugal Peru Italy Greece Turkey Argentina Malaysia Brazil Pakistan Japan Belgium Switzerland Ireland Philippines Austria Spain Mexico Chile India Finland Indonesia United States Thailand Singapore Germany Denmark Norway South Africa France Canada United Kingdom S. Korea Israel Hong Kong Australia New ZealandAdjustment Speeds-Book Leverage (LSDVC) 0 10 20 30 40 50Adjustment Speed Colombia Portugal Peru Italy Greece Turkey Argentina Malaysia Brazil Pakistan Japan Belgium Switzerland Ireland Philippines Austria Spain Mexico Chile India Finland Indonesia United States Thailand Singapore Germany Denmark Norway South Africa France Canada United Kingdom S. Korea Israel Hong Kong Australia New ZealandAdjustment Speeds-Market Leverage (BB) 0 10 20 30 40 50Adjustment Speed Colombia Portugal Peru Italy Greece Turkey Argentina Malaysia Brazil Pakistan Japan Belgium Switzerland Ireland Philippines Austria Spain Mexico Chile India Finland Indonesia United States Thailand Singapore Germany Denmark Norway South Africa France Canada United Kingdom S. Korea Israel Hong Kong Australia New ZealandAdjustment Speeds-Market Leverage (LSDVC)

PAGE 132

132 Figure 3 2 Capital Structure Adjustments The figure illustrates the frequency and size of capital structure adjustments for slow and fast adjusters in the sample. Slow (fast) adjusters are the group of countries with adjustment speed smaller (larger) than the sample median. An issuance or retirement is defined as having occurred in a given year if the net change in equity (E.) or debt (D.), normalized by the book va lue of assets at the end of the previous period, is greater than a certain cut -off. Debt (Equity) is defined as a change in absolute value of outstanding debt (equity) exceeding 5% of total assets. Debt Issue, Debt Retirement, and Equity Issue each are def ined as a net security issuance or repurchase of at least 5% of book assets. Equity Retirement is defined as a net security repurchase of at least 1.25% of book assets. Each year, the median value of the size of capital structure adjustment is calculated f or each country. The reported statistic is the mean of these time -series medians. 62.73 69.27 38.39 42.1 26.69 28.24 44.16 52.73 28.31 32.36 31.44 34.49 0 20 40 60 80 Debt Debt Issue Debt Retire Equity Equity Issue Equity RetireFrequency of Adjustments Slow Adjusters Fast Adjusters 14.68 17.99 14.65 16.42 14.83 20.37 12.05 14.65 14.11 15.89 8.74 12.99 0 5 10 15 20 Debt DebtIssue DebtRetire Equity EquityIssue EquityRetireSize of Adjustments Slow Adjusters Fast Adjusters

PAGE 133

133 Figure 3 3 Adjustment Speeds and Adjustment Cost Factors. The figure ranks countries according to various composite indices of adjustment costs and plots their adjustment speed estimates reported in Table 3 3 column 1. The solid line illustrates the fitted values of the adjustment speed estimates. Table 3 1 provides detailed information on the formulation of the composite indices of adjustment costs. Mexico Philippines Peru Argentina Greece Spain Italy Germany Austria Switzerland Colombia Portugal France Brazil Belgium South Korea Finland Indonesia Chile Turkey Denmark Norway Pakistan Japan Thailand Israel India US Ireland Canada Australia Malaysia Singapore UK South Africa New Zealand HongKong 0 .1 .2 .3 .4Adjustment Speed Access to capital markets Adjustment Speed Fitted US Canada UK Singapore Australia Malaysia HongKong Finland Philippines South Africa New Zealand Israel France Switzerland Japan Denmark South Korea Norway India Thailand Spain Italy Belgium Chile Mexico Germany Portugal Brazil Austria Turkey Greece Colombia Peru Argentina 0 .1 .2 .3 .4Adjustment Speed Information Asymmetry Adjustment Speed Fitted Indonesia Australia Philippines US Pakistan New Zealand Finland HongKong Ireland UK Canada Malaysia South Africa Israel Singapore India Thailand Austria France Belgium Germany Spain Norway Portugal Turkey Argentina Mexico Peru Italy Japan Denmark South Korea Switzerland Chile Greece Brazil Colombia 0 .1 .2 .3 .4Adjustment Speed Cash Constraints Adjustment Speed Fitted HongKong UK Singapore Canada US Australia Malaysia South Africa New Zealand Israel Finland India Japan France Thailand Norway Denmark Philippines Belgium Switzerland South Korea Spain Chile Italy Turkey Germany Portugal Austria Argentina Mexico Brazil Peru Greece Colombia 0 .1 .2 .3 .4Adjustment Speed Adjustment Costs Adjustment Speed Fitted

PAGE 134

134 Figure 3 4 Adjustment Speeds and Adjustment Benefit Factors. The figure ranks countries according to various composite indices of adjustment benefits and plots their adjustment speed estimates reported in Table 3 3 column 1. The solid line illustrates the fitted val ues of the adjustment speed estimates. Table 3 1 provides detailed information on the formulation of the composite indices of adjustment benefits Singapore Japan Norway Canada New Zealand UK Finland Belgium Ireland Australia HongKong South Korea US Spain Portugal Germany Austria Colombia Denmark France Switzerland Greece Mexico Israel Peru Malaysia Italy South Africa Argentina Brazil Chile Thailand Turkey Indonesia Philippines 0 .1 .2 .3 .4Adjustment Speed Distress Costs Adjustment Speed Fitted HongKong Ireland Singapore Malaysia Switzerland France South Korea Chile Brazil Portugal Finland Belgium Turkey South Africa US Norway Spain UK Greece India Indonesia Austria Canada Denmark Australia Peru Thailand Philippines Mexico Germany Argentina Italy Colombia Israel New Zealand Japan Pakistan 0 .1 .2 .3 .4Adjustment Speed Tax Shields Adjustment Speed Fitted Philippines Pakistan Indonesia Peru Mexico Argentina Colombia Brazil Turkey India Chile Thailand Greece Malaysia South Africa Portugal Spain Israel Italy Singapore France Australia Japan Ireland Belgium Germany Finland Austria US Canada UK Denmark Norway Switzerland 0 .1 .2 .3 .4Adjustment Speed Deviation Costs Adjustment Speed Fitted Philippines Indonesia Peru Brazil Turkey Argentina Mexico Chile Thailand Colombia Malaysia Greece South Africa Portugal Italy Israel Spain France Singapore Austria Australia Ireland Germany Denmark US Belgium Switzerland Japan Finland UK Canada Norway 0 .1 .2 .3 .4Adjustment Speed Adjustment Benefits Adjustment Speed Fitted

PAGE 135

135 Figure 3 5 Variable Adjustment Speed Estimates. The figure compares the estimates of uniform adjustment to mean variable adjustment in each country by plotting the adjustment speed estimates (%) reported in Table 3 9 columns 1 and 2. 0 10 20 30 40 50 Argentina Australia Austria Belgium Brazil Canada Chile Columbia Denmark Finland France Germany Greece Hong Kong India Indonesia Ireland Israel Italy Japan Malaysia Mexico New Zealand Norway Pakistan Peru Philippines Portugal S. Korea Singapore South Africa Spain Switzerland Thailand Turkey United Kingdom United States Uniform Adjustment Variable Adjustment

PAGE 136

136 CH APTER 4 DIRECTIONS FOR FUTURE RESEARCH Motivation Results reported in Chapter 2 and 3 suggest that the effectiveness of a countrys legal, financial and political institutions systematically affects cross -country differences in the firms choices of capit al structure via influencing the bankruptcy costs, agency costs and information asymmetry costs imposed on firms. F urther research is clearly needed to establish an institutional framework that would enable systematic evaluations on the differences in det erminants of capital structure across countries. A comprehensive model simultaneously using information on firm, industry, and country attributes needs to be developed in order to further explore if there are major differences in the capital structure of f irms across different institutional settings, as well as how institutional arrangements affect corporate financial leverage, and h ow the firm industry, and macroeconomic l evel covariates relate to capital structure in different institutional environments Empirical Methodology First, such a model should allow the deviations from target leverage not to be offset immediately. Chapter 2 and 3 illustrate that i t would be unreasonable to impose the assumption that firms always operate at their optimal leverage ratios. 1= 1 + (4 1 ) where is firm is debt ratio in year t and in country j; is firm is desired debt ratio in country j and in year t; In Equation 4 1, is the adjustment parameter representing the ma gnitude of the adjustment If equals one, then all the gap between the observed and optimal leverage is closed for that firm leading the actual leverage to equal the optimal leverage. In the presence of

PAGE 137

137 market frictions, will be less than one in which case the firm would only be able to eliminate a proportion of the gap between its actual and optima l leverage. Second, e ach firms optimal leverage should be modeled as a function of both the observed firm industry, and macroeconomic characteristics, the obs erved country characteristics, and the unobserved firm heterogeneity, To allow for the impact of institutions on the within country correlations of leverage and its factors as demonstrated in Chapter 2 and 3 the interaction between firm and country characteristics should be added to the model as well. T he refore, the target equation shou ld permit for the possibility that optimal leverage might differ across firms countries and over time as follows : = 1+ + 1+ (4 2 ) where and are coefficient vectors to be estimated. 1 and are vector s of firm (industry, macroeconomic ) and country characteristics related to the costs and benefits of operating with various leverage ratios. 1 is the interaction between the firm (industry, macroeconomic) and country determinan ts of optimal capital structure Under the trade off theory, 0 and the variation in is non trivial. Substituting Equa tion 4 2 into the par tial adjust ment specification in Equation 4 1 and re arranging yields: = 1+ ( 1 ) 1+ + 1+ + (4 3 ) Equation 4 3 constitutes a typical partial adjustment model of capital structure where th e base adjustment speed is the coefficient on the lagged dependent variable 1 subtracted from one (1-(1 )) In Equation 4 3, denotes the effect of firm (industry or macroeconomic) features on firms capital structure, which I refer to as the firm (industry or macroeconomic) effect. To

PAGE 138

138 quantify the impact of the institutions in explaining international d ifferences in the capital structure choice, one can focus on two different components. The first piece is the direct impact of the institutional determinants on leverage. denotes the effect of country level institutional characteristics on firms ca pital structure, which I refer to as the direct institutional effect. The second piece is the indirect impact that occurs due to the heterogeneity of the firm specific determinants captured by the interaction term between the firm and country factors. Ac cordingly, denotes the effect of the country level institutional attributes on the impact of the firm level attributes, which I call indirect institutional effect. Total firm and institution effects are + and + 1, respectively. The dynamic panel model in Equation 4 3 requires instruments for the endogenous transformed lagged dependent variable (Baltagi (2001)) and a correction for the short panel bias (Bruno (2005)). Accordingly, Blundell -Bond (1998) gener alized method of moments estimation (BB) can be used to estimate Equation 4 3 Test Design Direct Effect of the Firm, Industry, and Macroeconomic Determinants of Leverage In Chapter 2, I evaluate which capital structure theories are able to account for the observed patterns in the international data. To do so, I compare the direction and the significance of the observed relationships between leverage and its determinants with the theories predictions. To test which leverage determinants have robust impact on capital structure according to Equation 4 3, one can carry out the following test: = 0 If the leverage factor in question is reliable, 0 should hold. One then may evaluate the merits of the capital structure theories by comparing its sign with the theories predictions. For instance according to the tradeoff theory, for tangibility, > 0 should hold.

PAGE 139

139 Effects of the Institutional Determinants on Leverage Country characteristics may influence firms costs and benefits for determining their capi tal structure Countries differ in the quality of institutions that may potentially affect the trade off between the bankruptcy costs, tax benefits, agency costs, and information asymmetry costs imposed on firms. The institutional setting may influence the capital structure of a firm in two different ways: either by affecting its long run leverage target, hence its leverage choice, a direct institution effect; and/or may attenuate or intensify the relationship between firm specific determinants and capita l structure, an indirect effect. Direct institutional effect T o test whether the country characteristics matter for the leverage determination, that is if the direct institution effect is significant, one should test: = 0 If the institutional feature in question is relevant for the capital structure choice 0 should hold. Furthermore, the direction of the impact of the institutional variables should be consistent with the predictions of the capital structure theories. For instance, if the institution that is considered is creditor protection, I would expect > 0 under the tradeoff theory, as higher creditor protection would lower bankruptcy and agency costs of debt resulting in higher leverage Indirect institutional effect In Chapter 2, I analyze how institutions affect the within country correlations between leverage and its determinants. In Chapter 3, I investigate this issue in detail for the most outstanding leverage factor, lagged leverage. Overall, I find that the reliable core fact ors of leverage have similar impact across firms from diverse institutional environments when the direction of the relationship is considered. However, some differences emerge as well. Firm, industry, and macroeconomic features illustrate a relationship th at is different in its magnitude

PAGE 140

140 with leverage across different institutional settings, indicating that country characteristics are at work. To test whether the influence of the firm level, in dustry -level, and macroeconomic -level determinants of leverage depends on the institutional features of the country the firm operates in, one can test whether the indirect institutional effect is non trivial, i.e. = 0 If > 0 t he institutional characteristic can be thought to strengthen the effect of the fi rm industry, or macroeconomic feature on leverage. On the other hand, if < 0 the institutional characteristic can be thought to moderate the effect of the firm industry, or macroeconomic feature on leverage. For example, if the positive effect of t angibility on leverage is more pronounced in countries with weaker creditor rights, I would expect to find < 0

PAGE 141

141 CHAPTER 5 CONCLUSION This study examines international differences in the determinants of capital structure and firms adjustment to optimal capital structure across 37 countries ove r the period from 1991 to 2006. My intuition is that the capital structure choice of a firm is not only the outcome of its own characteristics, but is also the result of the environment and traditions in which it operates. I address the econometric challenges involved in modeling the capital structure decisions of firms by using panel data, employ ing Blundell and B onds (1998) dynamic system GMM, and simultaneously taking into account the heterogeneous nature of the data In Chapter 2, I examine which leverage factors are consistently important for the capital structure decisions of firms around the world. I also assess why each determinant has th e observed relationship with leverage based on the main capital structure theories. Finally, I explore the effect of conditioning on firm circumstances on the relative reliability of leverage determinants. My findings reveal that not all leverage factors a re equally reliable because firms are exposed to different circumstances as a consequence of the intrinsic characteristics of their countrys legal and economic structure. Institutional factors change how selected determinants affect leverage. In C hapter 3 I suggest a theory for the adjustment to optimal capital structure and conjecture that the effectiveness of a countrys legal, financial and political institutions is systematically related to cross -country differences in the adjustment speeds. The avera ge adjustment speed estimates I find for a large cross section of countries support the burgeoning dynamic tradeoff literature that has documented relatively quick adjustment to optimal leverage The variation in adjustment speeds is driven by the adjustme nt costs and benefits that determine the worldwide frequency and size of the capital structure rebalancing, establishing the relevance

PAGE 142

142 of the dynamic tradeoff theory. I empirically test and quantify the impact of the hypothesized institutional characterist ics on the adjustment speed to optimal leverage T he institutional proxies I employ to account for the ease of access to capital, the information asymmetry, the financial constraints, the tax benefits of debt, the costs of distress and the deviation costs all affect the speed of adjustment in a manner consistent with my theory. Specifically, the adjustment is significantly faster in countries with better protection and enforcement of investor rights, stronger legal, judicial, financial systems, and better c apital, corporate, and political governance. My findings suggest that the effectiveness of a countrys legal, financial and political institutions systematically affects cross -country differences in the firms choices of capital structure by influencing t he bankruptcy costs, agency costs, and information asymmetry costs imposed on firms. In Chapter 4, I suggest directions for future research. I propose a comprehensive model using information on firm, industry, macroeconomic and country -level characteristic s to explore how firm, industry, macroeconomic, and institutional arrangements simultaneously affect corporate financial leverage of firms around the world. An increasing amount of research is being conducted to investigate the channels through which a cou ntrys institutional heritage affects its financing patterns. I offer a new channel through which countries legal and institutional structure matter for capital structure decisions, by its impact on the leverage determinants and the speed of convergence t o the optimal capital structure. The empirical evidence that I present provides substantial support for the dynamic tradeoff theory.

PAGE 143

143 LIST OF REFERENCES Allen, Franklin, and Gale Douglas, 1999, Diversity of op inion and financing of new technologies, Journal of Financi al Intermediation 8, 1 22. Amihud, Yakov, and Baruch Lev, 1981, Risk reduction as a managerial motive for conglomerate mergers, The Bell Journal of Economics 12, 605617. Amihud, Yakov, and Haim Mendelson, 1986, Asset pricing and bid ask spread, Journal of Financial Economics 17, 223249. Antoniou, Antonios, Yilmaz Guney, and Krishna Paudyal, 2008, The determinants of capital structure: Capital Market Oriented versus Bank Oriented Institutions, Journal of Financial and Quantitative Analysis 43, 5992. Bae, KeeHong, and V idhan K. Goyal, 2004, Property rights protection and bank loan pricing, Working paper. Baker, M alcolm Wurgler, J effry 2002, Mark et timing and capital structure, Journal of Finance 57, 132. Baltagi, Badi H., 2001, Econometric Analysis of Panel Data John Wiley and Sons, West Sussex. Bancel, Franck, a nd Usha R. Mittoo, 2004, Cross -country determinants of capital structure choice: A survey of European f irms, Financial Manage ment 33, 103132. Barry, Christopher B., and Stephen J. Brown, 1985, Differential information and security market equilibrium, Journal of Financial and Quantitative Analysis 20, 407422. Bencivenga, Valerie, R., and Bruce D. Smith, 1991, Financial interme diation and endogenous growth, Review of Economics Studies 58, 195209. Berger, N. Allen, Robert DeYoung, Mark J. Flannery, David K. Lee, and zde ztekin, 2008, How do large banking organizations manage their capital r atios?, Journal of Financial Services Research 34,123149. Berkowitz, Daniel, Katharina Pistor, and Jean Francois Richard, 2003, Economic development, legality and the transplant effect, European Economic Review 47, 165195. Bhattacharya, Utpal, and Hazem Daouk, 2002, The world price of insider t rading, Journal of Finance 57, 75108. Blundell, Richard and Stephen Bond, 1998, Initial conditions and moment r estrictions in dynamic panel data m odels, Journal of Econometrics 87, 115143. Boot, Arnoud W. A. and An jan V. Thakor, 1997, Financi al system a rchitecture, Review of Financial Studies 10, 693733.

PAGE 144

144 Booth, Laurence, Varouj Aivazian, Asli Demirgunc -Kunt, and Voji slav Maksomovic, 2001, Capital s tructure in developing countries, Journal of Finance 56, 87130. Bruno, S. F. Gio vanni, 2005, A pproximating the bias of the LSDV estimator for d ynamic u nbalanced panel data m odel, Economic Letters 87, 361366. Campbell Y. John and Tuomo Vuolteenaho 2004 Inflation illusion and stock prices The American Economic Review 94, 19 23. Chang, Xin, and Sudipto Dasgupta, 2009, Target behavior and financing: How conclusive is the evidence?, Journal of Finance, forthcoming. Claessens, Stijn, Djankov Simeon, and Lang Larry, 2000, The separation of ownership and control in East Asian corporat ions, Journal of Financial Economics 58, 81 112. Clark, Brian, Bill Francis, and Hasan Iftekhar, 2009, Do firms adjust toward target capital structures? Some international evidence, Working Paper. Coles, Jeffrey L., Uri Loewenstein, and Jose Suay, 1995, On equilibrium pricing under parameter uncertainty, Journal of Financial and Quantitative Analysis 30, 347364. Daouk, Hazem, Charles M.C. Le e, and David Ng, 20 06, Capital market governance: H ow do security laws affect market performance?, Journal of Corpor ate Finance 12, 560593. DeAngelo, Harry, and Ronald W. Masulis, 1980, Optimal capital structure under corporate and personal taxation, Journal of Financial Economics 8, 3 29. Demirguc -Kunt, Asli, and Vajislav Maksimovic, 1996, Stock market d evelopment, a nd firm financing c hoices, World Bank Economic Review 10, 341371. Demirg -Kunt, Asli and Vajislav Maksimovic, 1 998, Law, finance, and firm g rowth, Journal of Finance 53, 210737. Demirguc -Kunt, Asli, and Vajislav Maksimovic, 1999, Institutions, financia l markets, and firm debt maturity, Journal of Financial Economics 54, 295 336. Demirguc -Kunt, Asli, and Vajislav Maksimovic, 2002, ''Funding growth in bank -based and market based financial systems: E vidence from firm level data, Journal of Financial Economics 65, 337363. Diamond, Douglas, 1984, Financial i ntermedia ion and delegated m onitoring, Review of Economic Studies 51, 393414. Djankov, Simeon, Rafael La Porta, Florencio Lopez de Silanes, and Andrei Shleifer, 2002, The regulation of entry, The Quar terly Journal of Economics 1, 1 37. Djankov, Simeon, La Porta, Rafael, Florencio Lopez de Silanes, and Andrei Shleifer, 2003, Courts, The Quarterly Journal of Economics 118, 453517.

PAGE 145

145 Djankov, Simeon, Caralee McLie sh, and Andrei Shleifer, 2007, Private cred it in 129 countries, Journal of Financial Economics 84, 299329. Djankov, Simeon, Rafael La Porta, Florencio Lopez -de Silan es, and Andrei Shleifer, 2008. The law and economics of self -dealing, Journal of Financial Economics 88, 430 465. Djankov, Simeon, T im Ganser, Caralee McLiesh, Rita Ramalho, and Andrei Shleifer, 2008, The effect of corporate taxes on i nvest ment and e ntrepreneurship, Working paper. Easterbrook, Frank H., 1984, Two agency-cost explanations of dividends, American Economic Review 74, 650659. Fama, F. Eugene, and Ken neth R. French, 1997, Industry c osts of e quity, Journal of Financial Economics 43, 153193. Fama, F. Eugene, and Ke nneth R. French, 2002, Testing tradeoff and pecking order p redictions about d ividends and d ebt, Review and Financial Studies 15, 134. Fan, P.H., Joseph, Sheridan Titman and Garry J. Twite, 2006, An i nternational comparison of capital s tructure and debt maturity c hoices, Working paper. Farhat Joseph, Carmen Cotei, and Benjamin Abugri, 2006, The pecking order hypothesis vs. the static trade -off theory under different institutional environments Working paper. Faulkender, Michael, Mark J. Flannery, Kristine W. Hankins, and Jason M. Smith, 2008, Do a djustment costs i mpede the realization of target capital s tructure?, Working paper. Faulkender, Michael, and Mitche ll A. Peterson, 2006, Does the source of capital affect capital s tructure?, Review of Financial Studies 19, 4579. Flannery, Mark, J., and Ka sturi P. Rangan, 2006, Partial adjustment toward target c a pital s tructures, Journal of Financial Economics 79, 469506. Flannery, Mark, J., and Kristine Hankins, 2007, Estimating dynamic panel models in corporate f inance Working paper. Frank, Murray Z., and Vidhan K. Goyal, 2005, Tradeoff and pecking order t heories of d ebt, working paper. Frank, Murray Z., and Vidhan K. Goyal, 2007, Capital structure d ecisions: W hich factors are reliably i mportant?, Working paper. Gi annetti, Mariassunta, 2003, Do better institutions mitigate agency p roblems? Evidence from corpor ate finance choices, Journal of Financial and Quantitative Analysis 38, 185212. Graha m, John R., 1996, Debt and the marginal tax r ate, Journal of Financial Economics 41, 4173.

PAGE 146

146 Graham, John R., and Campbell Harvey, 2001, The theory and practice of corporate finance: E vidence from the field Journal of Financial Economics 60, 187243. Graham, John R. Campbell R. Harvey, and Manju Puri, 2008, Managerial attitudes and cor porate a ctions, working paper. Greene, H. William, 2002, Econometric Analysis 5th ed. Prentice Hall. Haas, Ralph, and Marga Peeters, 2006, The dynamic adjustment towards target capital structures of firms in transition economies, Economics of Transition 1 4, 133169. Hahn, Jinyong Jerry Hausman and Guido Kuersteiner, 2007, Long difference instrumental variables e stimation for dynamic panel models with fixed effects, Journal of Econometrics 140, 574617. Hail, Luzi, and Christian Leuz, 2006, International differences in the cost of equity capital: D o legal institutions and securities regulation m atter? Journal of Accounting Res earch 44, 485531. Healy, Paul M., Amy P. Hutton., Krishna G. Palepu, 1999, Stock performance and intermediation changes surrounding sustained increases in disclosure, Contemporary Accounting Research 16, 485 520. Heaton, James B. 2002, Managerial optimi sm and corporate finance, Financial Management 31, 3345. Hellwig, Martin, 1991, Banking, financial intermediation, and corporate finance, A. Giovanni and C. Mayer (eds.) European Financial Intermediation, Cambridge University Press, Cambridge, 35 63. Hennessy, A.Christopher, and Toni M. Whited, 2005, Debt dynamics, Journal of Finance 60, 11291165. Holmstrom, Bengt, and Jean Tirole, 1993, Market liquidity and performance monitoring, Journal of Political Economy 101, 678709. Hovakimian, Armen, Tim Ople r, and Sheridan Titman, 2001, The debt -equity c hoice, Journal of Financial and Quantitative Analysis 36, 1 24. Huang, Ron gbing, and Jay Ritter, 2009 Testing theories of capital structure and estimating the speed of a djustment, Journal of Financial and Quantitative Analysis, forthcoming. Je nsen, Michael C., 1986, Agency c osts of free cash flows, corporate finance and t akeovers, American Economic Review 76, 323339. Jensen, Michael C., and Murphy, Kevin J., 1990, Performance pay and top-management incentives Journal of Political Economy 98, 22564.

PAGE 147

147 Jensen, M ichael.C., and William H. Meckling, 1976, Theory of the firm: M anagerial behavior, agency costs and ownership structure, Journal of Financial Economics 3, 305 360. Jong, Abe de, Rezaul Kabir, and Thuy Thu Nguyen, 2007, Capital structure around the world: T he roles of firm and countryspecific determinants Working Paper. Korajczyk, Robert A. and Amnon Levy, 2003, Capital structure choice: Macroeconomic conditions and Financial Constraints, Journal of Financial Economics 68, 75109. La Porta, Rafael, Florencio Lopez -De -Silanes, and An drei Shleifer, 1999, Corporate ownership around the World, Journal of Finance 54, 471517. La Porta, Rafael, Florencio Lopez -de -Silan es, and Andrei Shleifer, 2006, Wha t works in securities laws? Journal of Finance 61, 1 32. La Porta, Rafael, Florencio Lopez -De -Silanes, Andrei Shleifer, and Robert W. Vishny, 1997, Legal d eterminan ts of external f inance, Journal of Finance 52, 11311150. La Porta, Rafael, Florencio Lopez -de -Silanes, Andrei Shleifer, and R obert W. Vishny, 1998, Law and f inance, Journal of Political Economy 106, 11131155. La Porta, Rafael, Florencio Lopez -de -Silanes, Andrei Shleifer, and Robert W. Vishny, 1999, The quality of government, Journal of Law, Economics, and Organization 15, 222 79. La Porta, Rafael, Florencio Lopez -de -Silanes, Andrei Shleifer, and Robert W. Vishny, 2000a, Agency problems and dividend polic ies a round the w orld, Journal of Finance 55, 133. La Porta, Rafael, Florencio Lopez -de -Sila nes, Andrei Shleifer, and Robert W. Vishny, 2000b, Investor Protection and corporate governance, Journal of Financial Economics 58, 327. La Porta, Rafael, Florencio Lopez -de -Silanes, Andrei Shleifer, and Robert W. Vishny, 2002, Investor protection and cor porate valuation, Journal of Finance 57, 1147 1170. La Porta, Rafael, Florencio Lopez -de -Silanes, Christian Pop -Eleches, and Andrei Shleifer, 2004, Judicial checks and balances, Journal of Political Economy 112, 44570. Lambert, Richard, A., Christian Leuz and Robert E. Verrecchia, 2007, Accounting Information, Disclosure, and the cost of c apital, Journal of Accounting Research 45, 385420. Leary, Mark, T., and Michael R. Roberts, 2005, Do firms rebalance their capital structures?, Journal of Finance 60, 25752619. Lemmon, Michael, L., Michael R. Roberts and Jaime F. Zender 2008, Back to the beginning: P ersistence and the cross -section of corporate capital structure, Journal of Finance 63, 15751608. Lemmon, Michael L. and Jaime F. Zender, 2004, Debt capacity and tests of capital structure theories, Working Paper, University of Utah.

PAGE 148

148 Levine, Ross, 1991, Stock markets, growth, and tax policy, Journal of Finance 46, 144565. Levine, Ross, 1999, Law, finance and economic growth? Journal of Financial Intermediation 8, 8 35. Levine, Ross, 2002, Bank -Based or m arke t -b ased financial systems: W hich is b etter? Journal of Financial Intermediation 11, 130. Lockhart, G. Brandon, 2009, Adjusting to target capital structure: The effect of credit lines, Working Paper. Lombardo, Davide, and Marco Pagano, 2002, Law and equity markets: A simple model, Corporate Governance Regimes: Convergence and Diversity, Oxford University Press, Oxford. Lf, Hans, 2003, Dynamic o ptimal capital s tructure and technical c hange, ZEW Discu ssion Paper No. 0306. Merton, Robert C., 1987, A simple model of capital market e quilibrium with i ncomplete i nformation, Journal of Finance 42, 483510. Murphy, Kevin J., 1985, Corporate p erformance and managerial remuneration: A n e mpirical a nalys is, Journal of Accounting and Economics 7 1142. Myers, Stewart C., 1977, Determinants of corporate borrowing, Journal of Financial Economics 5, 147 175. Myers, Stewart C., 1984, The capital structure p uzzle, Journal of Financ e 39 575592. Myers, Stewart C. and Nicho las S. Majluf, 1984, Corporate f inancing and investment decisions w hen firms have information that i nvestors do n ot h ave, Journal of Financial Economics 13, 187224. Myers, Stewart, 2003, Financing of corporations, Handbook of the Economics of Fi nance ed. George Constantinides and Milton Harris, Elsevier, Amsterdam, The Netherlands. Obstfeld, Maurice, 1994, Risk taking, global diversification, and growth, American Economic Review 84, 131029. Qian, Jun, and Philip E. Strahan, 2007, How laws and i nstitutions shape financial contracts: T he case of bank loans, Journal of Finance 62, 28032834. Rajan, Raghuram, G., 1992, Insiders and outsiders: The choice between informed and arms length Debt, Journal of Finance 47, 13671400. Rajan, G. Raghuram, and Luigi Zingales, 1995, What do we know about capital structure? Some evidence from international data, Journal of Finance 50, 1421 1460.

PAGE 149

149 Rajan, Raghuram and L uigi Zingales, 1998, Financial dependence and g rowth, American Economic Rev iew 88, 55986. Ramakrishnan, Ram, T.S. and Anjan V. Thakor, 1984, The v aluation of a s sets under moral h azard Journal of Finance 39, 229238. Ritter Jay and Richard S. Warr, 2002, The decline of inflation and the bull market of 19821999, Journal of F inancial and Quantitative Analysis 37, 2961. Shleifer, A ndrei and Robert W. Vishny, 1986, Large shareholders and corporate control, Journal of Political Economy 94, 461488. Sirri, Erik R.m and Tufano, Peter, 1995, The economics of pooling, in The Global Financial System: A Functional Approach (D.B. Crane, et al., Eds.), Harvard Business School Press, Cambridge, MA, 81 128. Smith, C lifford and J erold Warner, 1979, On financial contracting: A n analysis of bond covenants Journal of Financial Economics 7 pp. 1171 61. Song JoonYoung, and George C. Philippatos, 2004, Have we resolved some critical issues related to international capital structure? Empirical evidence from the 30 OECD countries, Working Paper. Stiglitz, Joseph, and Andrew Weiss, 1 981, Credit rationing in markets with imperfect i nformation, American Economic Review 71, 393410. Strebulaev, Ilya, A., 2007, Do tests of capital structure theory mean what they s ay? Journal of Finance 62, 17471787. Stulz, R ene, M., and Herb Johnson, 1985, An analysis of secured debt, Journal of Financial Economics 14, 501 521. Stulz, R ene, M., 1990, Managerial discretion and optimal financing policies, Journal of Financial Economics 26, 3 27. Verrecchia, Robert, E., 2001, Essays of disclosure, Journal of Acc ounting and Economics 32, 97 180. Welch, Ivo, 2004, Capital structure and stock returns, Journal of Political Economy 112, 106 131.

PAGE 150

150 BIOGRAPHICAL SKETCH zde ztekin was born in Istanbul, Turkey. She graduated from Lycee Saint -Michel High School in June 1994. She received her Bachelor of S cience in economics from Bogazici University in July 1999. On completing her bachelors degree, s he worked in banking as a treasury dealer in Garanti Bank for two years. Following her industry experience, she ear ned her M aster of A rts in economics from Bogazici University and University of Florida. She then started her doctoral studies in finance in Fall 2004 in Gainesville, Florida. The requirements for the degree of Doctor of Philosophy in finance were complete d in May 2009 at the University of Florida. She specialized in corporate finance, international finance, financial markets and institutions, investments, fixed income securities and banking. In August 2009, she will join the faculty of the Finance, Economi cs, and Decision Sciences at the University of Kansas Lawrence, Kansas.