<%BANNER%>

The Effect of Buyer-Supplier Relationships on Supplier Financing and Investment Policy

Permanent Link: http://ufdc.ufl.edu/UFE0024839/00001

Material Information

Title: The Effect of Buyer-Supplier Relationships on Supplier Financing and Investment Policy
Physical Description: 1 online resource (113 p.)
Language: english
Creator: Itzkowitz, Jennifer
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: buyer, cash, corporate, finance, relationships, supplier
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: Over the past few decades, the prevalence and importance of significant buyer-supplier relationships have increased dramatically. For suppliers, being in a relationship can be a double edged sword with both costs and benefits. To accentuate the benefits and minimize the costs, suppliers may take a variety of actions to change their financing and investment policy. Despite this, little extant literature addresses how major business-to-business relationships influence firm functioning. This dissertation contributes to the corporate finance literature by examining the effect of relationships in two distinct areas namely financial constraints and cash management. Specifically, I find that firms in relationships have lower sensitivity of investment to cash flow indicating that they have lower financing constraints. And, firms in relationships maintain higher levels of cash, all else equal, relative to firms which are not in relationships.
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 Jennifer Itzkowitz.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Houston, Joel F.
Local: Co-adviser: Naranjo, Andy.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024839:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024839/00001

Material Information

Title: The Effect of Buyer-Supplier Relationships on Supplier Financing and Investment Policy
Physical Description: 1 online resource (113 p.)
Language: english
Creator: Itzkowitz, Jennifer
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: buyer, cash, corporate, finance, relationships, supplier
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: Over the past few decades, the prevalence and importance of significant buyer-supplier relationships have increased dramatically. For suppliers, being in a relationship can be a double edged sword with both costs and benefits. To accentuate the benefits and minimize the costs, suppliers may take a variety of actions to change their financing and investment policy. Despite this, little extant literature addresses how major business-to-business relationships influence firm functioning. This dissertation contributes to the corporate finance literature by examining the effect of relationships in two distinct areas namely financial constraints and cash management. Specifically, I find that firms in relationships have lower sensitivity of investment to cash flow indicating that they have lower financing constraints. And, firms in relationships maintain higher levels of cash, all else equal, relative to firms which are not in relationships.
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 Jennifer Itzkowitz.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Houston, Joel F.
Local: Co-adviser: Naranjo, Andy.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024839:00001


This item has the following downloads:


Full Text

PAGE 1

1 THE EFFECT OF BUYER SUPPLIER RELATIONSHIPS ON SUPPLIER FINANCING AND INVESTMENT POLICY By JENNIFER ROSE ITZKOWITZ 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 Jennifer Rose Itzkowitz

PAGE 3

3 To my husband, Jesse, the lov e of my life and my best friend

PAGE 4

4 ACKNOWLEDGMENTS It is my pl easure to acknowledge some of my mentors, friends, and family that have contributed to this dissertation. I know that finishing my PhD would not have been possible without the wide base of support that I had. I would like to thank my supervisory committee members Joel Houston (committee chair), Andy Naranjo (committee co -chair), Mike Ryngaert, and Bart Weitz who made possible the completion of this dissertation. I am grateful for the time and energy they took to review my doctoral research and to provide me with constructive comments. Bart, it has been a pleasure getting to know you, and I am excited to continue our collaborations in the future Mike, I always enjoyed our discussions, both academic and otherwise. I am appreciative of Andy Naranjos coll aboration and advice. Our personal talks got me through the job search process, and your academic advice improved my dissertation substantially. Finally, to Joel, I am especially thankful for your excellent supervision, guidance, patience and friendship through all stages of my doctoral work including other research projects, finding a job, and writing this dissertation. From you I learned what an academic should be, and for this I am very grateful. I am also thankful for all the people in the Univer sity of Floridas Finance Department including the professors who guided me through coursework and supported my research and my peers who were instrumental in my growth as a scholar. Getting to know all of my peers has been one of the best parts of my doc toral experience. To my friends Alice and Lindsey I simply cannot thank you enough for collaborating with me, listen ing to me, and helping me to maintain my sanity Finally, I want to thank my family who has provided me with so much moral support during my education. I am very appreciative of my Mom and Dad for always encouraging, understanding, and being there for me. I love you both very much, and the pride you have for

PAGE 5

5 me got me through the toughest parts of the program. Thank you also to my sister J aime for being there to listen to me and always having the right thing to say. I would like to thank my machatunim, Doris and Gary, for reminding me to relax and enjoy each day. A ll of my grandparents deserve recognition. Bubby and Zadie instilled in me the importance and value of an education. And, Grandma and Poppy believed that I could accomplish anything and are the best cheerleaders I will ever have. And, last but not least, I would like to thank my husband, Jesse, for being my sounding board, edi tor, teammate and best friend.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 4 LIST OF TABLES ................................................................................................................................ 8 LI ST OF FIGURES .............................................................................................................................. 9 ABSTRACT ........................................................................................................................................ 10 CHAPTER 1 INTRODUCTION ....................................................................................................................... 12 Related Li terature ........................................................................................................................ 13 Data .............................................................................................................................................. 14 Influence of Relationships on Financing and Investment Policy ............................................. 15 Main Results ................................................................................................................................ 16 2 SUPPLIER RELATIONSHIPS AND INVESTMENT CASH FLOW SENSITIVITY ....... 20 Firm Relationships and Int ernal Capital Markets ..................................................................... 24 Inter -Firm Relationships ..................................................................................................... 24 Relationship costs ......................................................................................................... 24 Relationship benefits .................................................................................................... 25 Cross -sectional differences in the nature of relationships ......................................... 29 Liquidity, Investment, and Int ernal Capital Markets ......................................................... 32 Data Set ........................................................................................................................................ 34 Data Set Construction .......................................................................................................... 34 Fi rm Matching ..................................................................................................................... 37 Results .......................................................................................................................................... 38 Univariate Evidence ............................................................................................................ 38 Sensitivity of Investment to Cash Flow ............................................................................. 42 Conclusion ................................................................................................................................... 48 3 CASH HOLDINGS AND THE CHARACTERISTICS OF SUPPLIERS ............................. 61 Related Literature and Testable Hypotheses ............................................................................. 64 Cash Holding ........................................................................................................................ 64 Buyer -Supplier Relationships ............................................................................................. 65 The Effect of Buyer -Supplier Relationships on Cash Holdings ....................................... 66 Data .............................................................................................................................................. 70 Sources ................................................................................................................................. 70 Measure of Cash Holdings .................................................................................................. 73 Dependent variable ....................................................................................................... 73

PAGE 7

7 Independent variables of interest ................................................................................. 74 Control variables .......................................................................................................... 74 Empirical Analysis ...................................................................................................................... 75 Relation between Buyer -Supplier Relationships and Cash Holdings .............................. 75 Changes in cash holding by year ................................................................................. 75 Descrip tive findings ..................................................................................................... 76 Multivariate evidence ................................................................................................... 78 Determinants of the Effect of Relationships on Cash Holdings ....................................... 81 Relationship specific assets ......................................................................................... 82 Market position............................................................................................................. 84 The government as a buyer .......................................................................................... 86 Robustness ............................................................................................................................ 87 Conclusion ................................................................................................................................... 89 4 CONCLUSION AND FUTURE WORK ................................................................................ 101 Investment Cash Flow Sensitivity ........................................................................................... 101 Transmission Mechanism .................................................................................................. 103 Euler Met hod ...................................................................................................................... 103 Cash Holdings ........................................................................................................................... 103 Value of Cash Holdings for Firms in Relationships ........................................................ 105 Relationships and Risk ...................................................................................................... 105 The Link Between Investment -Cash Flow Sensitivity and Cash Holdings ........................... 106 LIST OF REFER ENCES ................................................................................................................. 108 BIOGRAPHICAL SKETCH ........................................................................................................... 113

PAGE 8

8 LIST OF TABLES Table page 2 1 Characteristics of f irms in relationships compared to a matched sample ........................... 51 2 2 Characteristics of durable goods manufacturers in relationships compared to a matched sample ...................................................................................................................... 52 2 3 Characteristics of non -durable goods manufacturers in relationships compared to a matched sample ...................................................................................................................... 53 2 4 Cross -sectional differences of firms in relationships by c ustomer sales ............................ 54 2 5 Cross -sectional differences of firms in relationships by relationship length ..................... 55 2 6 Investment Cash Flow Se nsitivity ........................................................................................ 56 2 7 Investment Cash Flow Sensitivity of Committed Firms ..................................................... 57 2 8 Investment Cash Flow Sensitivity by Buyers Stake in the Supplier ................................. 58 2 9 Differences in Investment Cash Flow Sensitivity during Recessions ................................ 59 2 10 Investment Cash Flow Sensitivity over Macro -Economic Cycles ..................................... 60 3 1 Relationships and Cash Holdings by Year ........................................................................... 92 3 2 Comparison of Suppliers with Major Customers to Suppliers without Major Customers ............................................................................................................................... 93 3 3 Summary Statistics ................................................................................................................. 94 3 4 All Manufacturing Firms ....................................................................................................... 95 3 5 Firms which Report Relationships ........................................................................................ 96 3 6 Descriptive Statistics by Industry .......................................................................................... 97 3 7 Determinant s of Relation between Cash Holdings and Buyer -Supplier Relationships ..... 98 3 8 Robustness ............................................................................................................................ 100

PAGE 9

9 LIST OF FIGURES Figure page 1 1 Trends in firm relationships over time ....................................................................................... 19

PAGE 10

10 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requir ements for the Degree of Doctor of Philosophy THE EFFECT OF BUYER SUPPLIER RELATIONSHIPS ON SUPPLIER FINANCING AND INVESTMENT POLICY By Jennifer Rose Itzkowitz August 2009 Chair: Joel Houston Cochair: Andy Naranjo Major: Business Administration Over t he past few decades there has been in increase in both the prevalence and importance of buyer -supplier relationships. Despite this, little extant literature examines the impact of buyer -supplier relationships on firm financing decisions. This dissertatio n examines whether and how buyer -supplier relationships influence different aspects of supplier financing and investment policy. In particular, I focus on the sensitivity of investment to cash flow and cash holdings. Heightened sensitivity of investment to cash flow is one indication that the supplier faces high external financial constraints. I find that the more committed a buyer is to its supplier and the greater a buyers stake in its supplier, the less sensitive the suppliers investment is to cash flow. Further, the investment -cash flow sensitivity of suppliers with principal customers is less affected by macroeconomic cycles than their peers. These effects are more pronounced for firms in industries with a high degree of relationship -specific in vestments. The findings suggest that firm relationships can help to ease capital market frictions which affect a firms investment decisions. With regard to cash holdings, I find that even after controlling for profitability, firm size, leverage, the eas e with which firms can access capital markets, growth opportunities, the

PAGE 11

11 availability of natural hedges, and a number of other control variables, firms which are in relationships hold more cash on average than firms not in relationships. Further, I find t hat as the importance of the relationship increases, measured by the percent of sales and sales concentrations to the reported customers, so does a suppliers cash holdings. Using the suppliers industry and R&D spending as proxies for the uniqueness of t he goods sold, I find that supplying unique goods mitigates the relation between customer importance and cash holdings. There is evidence consistent with the both the commitment and precautionary motivations for holding cash.

PAGE 12

12 CHAPTER 1 INTRODUCTION Ov er the past few decades both the prevalence and importance of buyer -supplier relationships have increased among American companies. Clearly, relationships provide both costs and benefits to the firms involved. But, the extent of these is largely unknown. Whether and how firms manage the costs or exploit the benefits of relationships remains an under researched area of corporate finance. This paper contributes to the literature by addressing the influence of buyer -supplier relationships on two important a reas of corporate finance research: financial constraints and cash management. Two distinct trends among American companies have contributed to the increase in the number and significance of buyer -supplier relationships over the last few decades. First, we have witnessed a decline in the vertical integration and industrial diversification of firms (Comment and Jarrell, 1995, Denis, Denis, and Yost, 2002, and Dimitrov and Tice, 2006). In general, there has been a trend for manufacturers to relinquish dir ect control of their suppliers and establish looser forms of collaboration (Zingales, 2000). As a result, the decline of vertical integration has given rise to a host of new buyer -supplier relationships. Simultaneously, many American companies have shif ted their overall business to -business strategy. In the hope of establishing enduring partnerships which encourage teamwork, foster innovation, and enhance product quality, companies have engaged fewer suppliers, but have increased the level of commitment with these suppliers (Emshwiller, 1991). This has increased the importance of each relationship for both the buyer and supplier involved. The data supports these anecdotes. As seen in Figure 1, in 1979 less than 30% of manufacturers report at least on e buyer that accounts for more than 10% of their sales. By 2006, more than 50% of manufacturers report such relationships. This demonstrates that over the past

PAGE 13

13 thirty years, the number of major buyer -supplier relationships increased. In a regression of the average sales to each major buyer (after adjusting for inflation) on a constant and time, time has a significantly positive coefficient, implying that the average sales per buyer has increased 6.39% per year. Additionally, the average number of major buyers reported by each supplier increases over time. The net effect of these two trends is that the median percent of total sales to all major customers increased to 39% from 26% between 1979 and 2006. Finally, customer concentration (measured as a Herf indahl -Hirshman Index of sales to each major customer) also increased during the same time period. Together, these demonstrate that the importance of relationships to each supplier has also increased. Related Literature To date, researchers have examine d the effect of inter -firm collaboration and firm dependence on a number of factors. Relative to suppliers which are not in substantial relationships, suppliers in relationships have similar growth rates (Kalwani and Narayandas, 1995) and lower information transmission costs (Gomes Casseres, Hagedoorn, and Jaffe, 2008). Buyer -supplier relationships also affect firm valuation. Hertzel, Li, Officer and Rodgers (2008) measure the average return of supplier during the periods prior to and around the time of the bankruptcy filing of buyers. They find negative abnormal return for suppliers of the distressed and bankruptcy filing -buyers. Relationships also effect firm functioning. Cohen and Frazzini (2008) observe buyer supplier pairs and show a correlation between the real activities of the firms by exploiting the time -series variation in the firms entering and leaving relationships over their sample period. Specifically, they observe a significantly greater correlation between the sales and operating inco mes of buyers and suppliers in relationships during the years that they engage in a relationship relative to the other years.

PAGE 14

14 In addition, researchers have examined firms responses to being in a relationship. Titman (1984) describes one way in which buyer -supplier relationships affect financing choices. He argues that buyers suffer costs if the supplier is liquidated. These costs are exacerbated when the supplying firms products are unique, because buyers have made specific investments to utilize the se products, which lose value upon liquidation of the supplier (Titman and Wessels, 1988). While this would normally lead to reluctance of the buyer to engage in specialized investments in relationship specific assets, a supplier can alter its capital s tructure, ex ante, such that they are pre positioned against liquidation, in order to elicit the maximum cooperation and investment of the buyer. Both Kale and Shahrur (2007) and Banerjee, Dasgupta, and Kim (2008) find empirical evidence consistent with t his theory. Kale and Shahrur (2007) measure the R&D intensity of the buyers industry as a measure of the uniqueness of the goods required. They find that the suppliers leverage is inversely proportional to the R&D intensity of the buyer. Similarly, du rable goods are thought to be more unique than nondurable goods. Banerjee, Dasgupta, and Kim (2008) find that the leverage ratio of durable goods manufacturers are inversely related to the proportion of purchases from a major buyer. The results are weak er for non -durable goods manufacturer and non-existent for non -manufacturers. Together, this set of results supports the hypothesis that suppliers of unique goods maintain a conservative capital structure in order to encourage the buyers to invest in rel ationship specific assets. Data To determine the effect of relationships on suppliers financing and investment policies, I construct an extensive buyer -supplier dataset that spans from 1979 to 2006. The Statement of Financial Accounting Standards No. 1 4 (later supplanted by SFAS No. 131) requires that firms report information for business relevant segments in their SEC 10K reports. This includes

PAGE 15

15 disclosure of the name of and amount of sales to a customer, if the revenue generated from the particular cu stomer exceeds 10% of revenue of the firm, or if the firm considers the sales important to its business. When a supplier sells more than 10% of its goods to one particular buyer I refer the buyer and supplier as engaging in a relationship. By relying of the buyers name and industry given to match the buyers to the Compustat data, I generate a sample of buyer -supplier pairs for which I know both partners financial information. Influence of Relationships on Financing and Investment Policy Relationships h ave the potential to affect the financing and investment polices of a supplier in variety ways. First, the presence of a major buyer may induce operating risk for the supplier. Most buyer -supplier relationships are governed by short term purchase orders Thus, there are no long term contracts or explicit guarantees and, in most cases, buyers can end the relationship at any time. Dependence on one customer for a large portion of sales increases a suppliers operating risk. Confirming this intuition, al most all suppliers which acknowledge the presence of a major customer also report in their 10 -K that the loss of this buyer would have a material adverse effect on the company. In response to this additional operating risk, suppliers may try to reduc e financial risk by either maintaining a lower leverage ratio or holding more cash to hedge against the uncertainty of expected future cash flows. Second, suppliers with concentrated sales to a particular buyer provide the buyer with additional bargaining power. Buyers may try to utilize this advantage by demanding concessions such as price reductions or the extension of favorable trade credit terms. If the buyer is successful, then these concessions can reduce profit margin and/or working capital of the supplier. Additional operating risk and a loss of bargaining power both represent costs of being in a relationship. However, relationships also have advantages for the supplier. One advantage of

PAGE 16

16 being in a major relationship is it provides the buyer and supplier with both the ability and the incentive to gain knowledge about one another. Suppliers can anticipate order placement and size and coordinate payment schedules helping suppliers to better manage inventory and account receivables which, in turn, reduces transactions costs and financing requirements. A second advantage to being in a relationship is that the buyer becomes a stakeholder in the supplier. While the buyer may not be a stakeholder in the traditional financial sense, the buyer has a vested interest in the ability of the supplier to deliver a quality product on time. Just as a relationship allows suppliers to learn about their buyers, buyers learn about suppliers by observing inventory turnover time, inventory quality, and prici ng. Together, the buyers stake in the supplier and this additional knowledge about the supplier provides the buyer with the ability and incentive to aid the supplier financially, if necessary. Finally, being in a relationship may indirectly reduce a s uppliers financing constraints. Knowing that a supplier is in a stable relationship may encourage a bank to lend to a supplier that is may not have lent to otherwise. Using a dataset of buyer -supplier relationships which spans from 1979 to 2006, I emp irically observe the effect of a relationship on a suppliers financial behavior. Overall, I find that being in a relationship affects the both investment -cash flow sensitivity and the cash holdings of suppliers. These findings are robust after controlli ng for a variety of different factors. Main Results In Chapter 2, I investigate when and how relationships affect suppliers financial constraints. To test whether a relationship affects firms financial constraints, I turn to a large literature docume nting the link between liquidity and investment. Here, liquidity, defined as the availability of internal funds, is an important determinant of investment when there are frictions

PAGE 17

17 in the capital market. On average, I find that suppliers with major buyers demonstrate lower investment cash flow sensitivity, suggesting that relationships ease suppliers liquidity constraints. However, not all relationships are the same. A variety of factors contribute to cross sectional differences in the extent to which relationships affect firm financing and investment. In particular, I find that the buyers commitment to the supplier and the buyers stake in the supplier lessen the sensitivity of the suppliers investment to cash flow. Buyers are considered to be mor e committed to a supplier if they are not in relationships with many other suppliers. And, the stake that a buyer has in its supplier is determined by the proportion of the suppliers total sales which they account for and the duration of the relationship. This is consistent with the theory that the buyer is the residual claimant on the suppliers investments. As a result the buyers incentive to aid the supplier increases in its stake in the supplier. Further, suppliers in relationships have lower fin ancing constraints at times when it matters the most. Using the corporate bond spread and a recession index as proxies for the accessibility of external financing, I find that the investment-cash flow sensitivity of suppliers with major customers is less affected by macroeconomic cycles than their peers. I also show that these effects are more pronounced for firms in industries with a high degree of relationshipspecific investments, in line with previous work. Together, the findings of this chapter suggest that firm relationships can help to ease capital market frictions which affect a firms investment decisions. Chapter 3 evaluates how relationships affect supplier cash holdings. Theoretically, a relationship may affect a suppliers cash holdings in a variety of ways. First, suppliers may maintain higher cash holdings as a means of committing to a buyer that they will not enter

PAGE 18

18 financial distress. Such a commitment may be necessary to induce buyers to make specialized investments in relationship speci fic assets. Next, suppliers in relationships may hold additional cash as a precautionary against the risk of the loss of a major customer. Alternatively, engaging in an important relationship may generate switching costs for the buyer. Moreover, these s witching costs may reduce the risk that the buyer changes suppliers abruptly, reducing the operating risk of the supplier. As a result, the supplier will not need to hold as much precautionary cash. I find that compared to their peers, suppliers in relat ionships hold more cash than suppliers which are not in relationships, ceteris paribus Suppliers cash holdings are positively correlated with both the ratio of sales to major customers to total sales and major customer sales concentration, consistent wi th the commitment and precautionary motives for holding cash. Two additional pieces of evidence make the case that of these two possibilities, the precautionary motive is the dominant rationale. First, I find a weaker relation between customer concentrat ion and cash holdings for manufacturers of unique goods relative to manufacturers of nonunique goods. This is consistent with the theory that the buyer faces higher switching costs when purchasing unique goods. And, switching costs reduce the risk of the buyer changing suppliers mitigating the need for holding precautionary cash. Second, I find that a suppliers cash holdings are affected when their customer is a corporation but are unaffected when the customer is a government agency. The U.S. government is not likely to go out of business or to abruptly end a relationship without warning. Therefore, this result lends support to the precautionary motivation for cash holdings. These additional results are consistent with the precautionary motivation for holding cash.

PAGE 19

19 Figure 1 1 Trends in firm relationships over time

PAGE 20

20 CHAPTER 2 SUPPLIER RELATIONSHI PS AND INVESTMENT CASH FLOW SENSITIVIT Y The nature of the firm is changing. Over the past few decades, there has been a noted decline in the vertical i ntegration and industrial diversification of firms (Comment and Jarrell, 1995, Denis, Denis, and Yost, 2002, and Dimitrov and Tice 2006). Overall, there has been a trend for manufacturers to relinquish direct control of their suppliers and establish loose r forms of collaboration (Zingales, 2000). For example, on Nov. 30, 2004 KimberlyClark, a maker of health and hygiene products, completed the spin -off of Neenah Paper, a premium uncoated paper mill. Despite the change of corporate control, KimberleyCla rk continued to purchase the same amount of paper goods from the mill creating a substantial buyer -supplier relationship. As a result, the decline of vertical integration has given rise to a greater number of important inter -firm relationships. Simultan eously, many American companies, including Xerox, Motorola, GM, and Ford, which once used many suppliers, changed strategies. These companies reduced the number of suppliers in hope of finding a few choice partners. They hoped to cement partnerships base d on teamwork and create enduring relationships to enhance product quality (Emshwiller, 1991). This trend increased the importance of each relationship to both the buyer and supplier. Despite the prevalence and importance of inter -firm relationships, little research has examined the costs and benefits of inter -firm relationships on a firms form or functioning. This paper focuses on when and how buyer -supplier relationships affect firm financing. Overall, I find evidence which suggests that relationshi ps ease capital market frictions which affect a firms investment decisions. Several aspects of buyer -supplier relationships can affect the suppliers financing policy. For a supplier, entering into a business relationship with a major customer can be a do uble -edged sword. A major customer accounts for a large portion of the suppliers total sales. But, the buyer does not does not provide an explicit guarantee that it will continue to purchase these goods in the future. As a

PAGE 21

21 result, the relationship indu ces additional operating risk for the supplier which may prompt the supplier to minimize financial risks. After developing a relationship, the threat to suppliers financial heal does not come from just losing the customer. In their 2004 SEC 10 -K report, Doane Pet Care acknowledged that relying on a customer for a large portion of their sales might allow the customer to make greater demands of us. The report enumerates several possible demands that could be made, including the exertion of price pressur e or requiring trade credit. Despite these risks, there are benefits to being in a relationship. Specific benefits include gaining valuable information about their buyers and the potential for joint investment made by the firms. Information acquired over the course of business may result in better management of inventory and accounts receivable. Joint investments and investments in relationship specific assets may increase the ability to pass price increases on to the buyer. In addition, buyer -suppli er relationships may create inter -firm capital markets which can ameliorate the impact of external capital market frictions. Further, as the result of implicit future commitments, an indirect benefit of a relationship may be the ability to acquire bank de bt with greater ease. This paper hypothesizes that relationships benefit the supplier by easing the suppliers financial constraints. To determine the effect of relationships on suppliers, I construct a buyer -supplier dataset which spans from 1979 to 2006. The Statement of Financial Accounting Standards No. 14 (later supplanted by SFAS No. 131) requires that firms report information for segments that represent 10% or more of consolidated sales. This includes disclosure of the name of and amount of sale s to a customer, if the revenue generated from the particular customer exceeds 10% of revenue of the firm, or if the firm considers the sales important to its business. When a supplier sells more than 10% of its goods to one particular buyer I refer to this as a relationship. Using the firm name given and industry, I match buyers to the Compustat database to generate a sample of relationships which

PAGE 22

22 includes financial data for both the buyer and suppliers. I then compare the suppliers to a matched sample of firms which do not report being in relationships. I begin by comparing the investment cash flow sensitivity of firms in relationships to a matched sample of firms which do not report any material relationship. The results of the tests provide eviden ce consistent with the hypothesis that suppliers in relationships have lower financing constraints than suppliers which are not in relationships. Further, this paper hypothesizes that certain characteristics of suppliers, buyers and their relationship enh ance the benefits of being in a relationship. First, the more committed the buyer is to the supplier the greater the suppliers incentive to aid the supplier and the lower the risk that the supplier will end the relationship. Buyers are considered to be m ore committed to a supplier if they are not in relationships with many other suppliers. I find that suppliers which are in relationships with committed buyers have lower investment cash flow sensitivity than suppliers in relationships with less committed buyers. Second, buyers are the residual claimant on the suppliers investments. As such, the benefits accrued to the suppliers will vary proportionately with the buyers stake in the supplier. The stake that a buyer has in its supplier is measured as t he proportion of the suppliers total sales which they account for and the duration of the relationship. The results indicate that the suppliers sales to the major customer and length of the relationship are inversely proportional to the investment cash flow sensitivity of the supplier. This is consistent with the theory that as the stake that the buyer has in the supplier increases so does their incentive to aid the supplier. Further, suppliers in relationships have lower financing constraints at time s when it matters the most. Using the corporate bond spread and a recession index as proxies for the accessibility of external financing, I find that the investment-cash flow sensitivity of suppliers with major customers is less affected by macroeconomic cycles than their peers. Finally, these effects are more pronounced

PAGE 23

23 for firms in industries typified by a high degree of relationship specific assets. The findings suggest that firm relationships can help to ease capital market frictions which affect a f irms investment decisions. The analysis in this paper is related to a number of distinct literatures in both corporate finance and financial intermediation. First, this paper is related to the small but growing literature with in corporate finance that explores the impact of buyer -supplier relationships on firms financing and investment policy. Existing research shows that relationships have a variety of both positive and negative effects. And, depending on the influence of the relationship on the fir m, some firms actively change their financial management policy to mitigate the costs and accentuate the benefits (Titman, 1984, Kale and Shahrur, 2007, and Banerjee, Dasgupta, and Kim 2008). This analysis also contributes to the financial intermediati on literature by suggesting that like banks, buyers have an ability to collect information and monitor firms, mitigating capital market frictions.1 Due to their routine interactions, buyers and suppliers in relationships transmit soft information to each other. While banks may have superior ability in this regard, interdependent firms should also have a similar ability to collect information at low cost. Finally, I build on the investment cash flow sensitivity literature pioneered by Fazzari, Hubbard an d Petersen (1988) by acknowledging another distinction between financially constrained and unconstrained firms. This paper demonstrates that interdependence reduces a firms sensitivity of investment to cash flow. This finding is consistent with theory t hat firm relationships can help to ease capital market frictions which affect a firms investment decisions. Beginning with a review of the firm relationship literature, Section I establishes hypotheses regarding how and when firm relationships affect fir m financing and investment. Following this, I 1 See Diamond (1984), James (1987), and Billet t Flannery and Garfinkel (1995) for evidence of banks ability to collect information.

PAGE 24

24 provide a brief synopsis of work investigating the sensitivity of investment to cash flow with a focus on studies that examine the influence of capital market imperfections on real investment decisions. Secti on II describes the data set used in my study. Section III presents evidence supporting the hypotheses. Section IV, confirms the robustness of the results to differences in types of buyer and supplier firms. Section V concludes. Firm Relationships and Internal Capital Markets Inter -Firm Relationships To date, researchers have examined the effect of inter -firm collaboration and firm dependence on a number of factors that influence firms form and functioning, including sales, profit margin, transactio n costs (Kalwani and Narayandas, 1995), distress costs (Hertzel, Li, Officer, and Rodgers, 2006), information transmission costs (Gomes -Casseres, Hagedoorn, and Jaffe, 2008), and capital structure (Banerjee, Dasgupta, and Kim, 2008, and Kale and Shahrur, 2007). I extend this line of research by documenting the effect of inter -firm relationships on firm financing and investment. The effect of a relationship on a firms ability to access the external capital market is not clear. On one hand, the risk inher ent in being in a significant relationship may exacerbate financing frictions. On the other hand, being in a significant relationship may be able to directly and/or indirectly ease financing frictions. If firms in relationships have a vested interest in helping one another, then they can create an implicit inter -firm capital market to facilitate investment financing. In addition to internal capital market creation, implicit future commitments resulting from a relationship may enable suppliers to more eas ily acquire bank debt. Relationship c osts Maintaining a relationship with a major buyer can pose a number of risks to the suppliers financial health. By depending on one or more customers for a significant part of the companys sales, suppliers open th emselves up to the risk of losing a large portion of sales at once which would

PAGE 25

25 cripple their financial health. The more goods that the supplier sells to the principal customer the more that the supplier loses if the buyer chooses to end the relationship. SEC 10 -K reports include statements detailing the additional firm risk partnerships pose. For example, in their June 28, 2003 annual report, Salton Inc. reported, If we were to lose one or more of our major customers, our financial results would suffe r. The statement continues: We do not have long -term agreements with our major customers, and purchases are generally made through the use of individual purchase orders. A significant reduction in purchases by any of these major customers could h ave a material adverse effect on our business, financial condition and results of operations. Further, a reduction in sales may not be the customers choice but the result of financial distress or bankruptcy. In particular, Hertzel, Li, Officer and Rod gers (2008) find that the contagion effects of buyer bankruptcy filings significantly affect the suppliers of filing firms. Moreover, because these relationships rely on short term and implicit contracting, their future remains uncertain, generating busin ess risk which can exacerbate financial market frictions. In addition to the risk of losing sales, when a supplier relies on one customer for a large portion of their sales, the buyer may gain bargaining power. This power could enable the buyer to put pressure on the supplier to reduce prices and/or to extend trade credit. Prominent examples of price pressure include Wal -Mart pressuring manufacturers to accept pricing terms or GM demanding price reductions from suppliers each year. Price reductions red uce suppliers profit margins and free cash flows. Similarly, the extension of trade credit reduces working capital. Uncertainty about output prices and costs can generate cash constraints and increase business risk. Relationship b enefits While being in a relationship may be costly, the benefits may outweigh the costs. Potential benefits from being in a relationship include reductions in business risk, financial market frictions,

PAGE 26

26 and/or transaction costs. These may result in higher operating profit, lower required working capital, and a greater ease with which to access capital. Each of these benefits enhances firm value. One specific way in which a relationship benefits a supplier is by generating information which is beneficial to the supplier. It is believed that banks have a unique ability to collect information and monitor firms, mitigating capital market frictions.2 While banks may have expertise in this area, it is certainly plausible that the ability to collect information is not limited t o them. Due to their routine interactions, buyers and suppliers in relationships may transmit soft information to each other. Similar to the manner in which banks acquire information about borrowers, suppliers may be able to acquire private information about their customers as the result of a continuing business relationship (Lummer and McConnell, 1989). Through the course of business, buyers and suppliers have access to information unavailable to outsiders (Smith, 1987, Biais and Gollier, 1997). Buye rs and suppliers can observe each others inventory levels, inventory turnover, ability to pay, actual growth, and potential trouble. Regular interaction can provide insights into management and also result in direct access to each others employees. Fur ther, buyers and suppliers have market/industry specific information. Gomes Casseres, Hagedoorn, and Jaffe (2006) find that the flow of knowledge between pairs of firms working together is greater than of the flow of knowledge between pairs of non allied firms. Although firms may not enter into explicit contracts specifying future interactions, over the course of the relationship the supplier can gain information about the buyers order schedule and financial position reducing transactions costs. Success ful reductions in transaction costs affect firm finances, improving profit margins. Additionally, because the information is gained over the routine course of business it comes at a low cost. Lower information costs may help to mitigate 2 For a theoretical explanation see Diamond (1984). James (1987), and Billet, Flan nery and Garfinkel (1995) provide empirical evidence.

PAGE 27

27 incentive problem s in capital markets which can limit the ability of cash constrained firms to make valuable investments.3 Another reason that relationships benefit suppliers is that buyers have an incentive to help their suppliers. Although buyers are not stakeholders in the traditional sense, buyers have vested interest in the suppliers financial well -being. Buyers can benefit from suppliers successes and can be hurt when suppliers are in distress. In effect, buyers in relationships become the residual claimants o n their suppliers projects, and vice versa. That is, in addition to its traditional stakeholders, a suppliers surplus may be allocated among its buyers. For example, if a manufacturer upgrades its factory, resulting in the ability to generate goods at a lower cost, then part of the savings may be passed onto the buyer. Therefore, if the supplier recognizes a positive NPV investment but lacks the liquidity to take advantage of the situation, their partner may be willing to support the investment because they are aware of the circumstances surrounding the event, and because they are also invested in the outcome. Buyers also act like stakeholders because they may be harmed by supplier distress. Supplier distress may cause buyers to worry about product quality, the value of warranties, and the continuity of supply and serviceability. For example, in August 2008, Boeing announced it would miss its second quarter earnings projections in part because otherwise completed jetliners could not be delivered due to lack of suppliers inventory. To resolve the issue, Boeing is helping its suppliers improve their operations (Michaels and Lunsford, 2008). As stakeholders, buyers have a vested interest in ensuring the continued financial health of their suppliers. As a result, buyers may aid suppliers by creating an inter -firm capital market to 3 The theoretical model of Myers and Majluf (1984) ; and the empirical evidence documented by Fazzari, Hubbard and Petersen (1988) and Hoshi, Kashyap, and Scharfstein (1991) speak to this point.

PAGE 28

28 mitigate external capital market frictions.4 First, reducing the amount of trade credit extended by the supplier can ease credit constraints. It is recognized that suppli ers can provide buyers with liquidity by providing trade credit. Biais and Gollier (1997) make the assumption that suppliers have private information about their buyers. They show that a suppliers private information helps to determine whether or not to grant trade credit and simultaneously helps to alleviate information asymmetries that would otherwise have precluded the financing of its buyers positive NPV projects. Further, Nilsen (1994) offers evidence that when small firms are financially constrai ned as the result of monetary contractions, they react by using higher levels of trade credit from their suppliers. While this shows that a supplier has private information about its buyers and incentives to help them the reciprocal arrangement may also e xist. Having discussed why a buyer has an incentive to assist a supplier, it remains to be determined how a buyer has the ability to assist its supplier. A buyer can provide its supplier with additional liquidity by paying off trade debts and by accepting pricing terms favorable to the supplier. The majority of the trade credit literature assumes that the extension of trade credit is a one-sided decision made exclusively by the supplier. However, the amount of trade credit outstanding is jointly determi ned by the amount offered by the supplier and the amount that the buyer chooses to accept. By paying on time, the buyer can reduce the suppliers trade credit outstanding, lessening the suppliers liquidity constraint. Reducing trade credit is not the on ly means by which a buyer can support its supplier. Individual buyers and their suppliers negotiate the price of the goods or services to be purchased. While examples of Wal Mart and GM are conspicuous, not all buyer -supplier relationships are one sided or adversarial. A buyer can help to provide its supplier with financial liquidity by agreeing to 4 Ther e are indeed other ways in which the supplier can help the buyer but they are outside the scope of this paper.

PAGE 29

29 price increases when necessary. For example, in 2005, the Tractor Supply Company reported in their SEC 10 -K that the Company is subject to market risk with respect to the pricing of certain products. The report continues by explaining that the ability (or inability) to pass the price increases on to their customers would materially affect the firms gross margins. In another example, the worlds largest pr inter, RR Donnelley, recognized that they relied heavily on their paper and ink suppliers and that if the quality and reliability of supplies improved, the company could reduce waste and improve their bottom line. Like many other firms, RR Donnelley encou raged suppliers to research potential improvements in both production and products. The most striking feature of this arrangement is not that the buyer asked their supplier to improve its performance, but that RR Donnelley also offered to split any result ant savings with their suppliers (Lee, 2004). Finally, being in a significant relationship can help relieve supplier credit constraints indirectly. Being in a relationship with a financially stable buyer can reduce the risks of future cash flows, allevia ting business risk. As a result, banks may be more willing to lend to firms in relationships. Therefore, the result of being in a relationship indirectly provides liquidity by enabling the supplier access to bank credit. Cross-s ectional d ifferences in the n ature of r elationships While there may be some disadvantages to entering into a significant relationship, overall, the hypotheses presented imply that suppliers benefit from being in material relationships. A variety of factors may contribute to cr oss-sectional differences in the nature and extent to which relationships affect financing and investment. In addition to the general hypothesis that firms benefit from relationships, this paper proposes several specific hypotheses regarding the buyer, supplier, and relationship characteristics which influence the benefits of relationships. First, the more committed a buyer is to the relationship, the greater its incentive to aid the supplier. On November 16th, 2004 the Senior Producer and Correspondent Hedrick Smith stated on

PAGE 30

30 Frontline that, By figuring out how to exploit two powerful forces that converged in the '90s, the rise of information technology and the explosion of the global economy, Wal Mart has changed the balance of power in the world of bu siness. While manufacturers previously had the power in relationships, that may no longer be the case. Now a few extremely large companies account for the sales of many dramatically smaller manufacturers. Although the relationship is material to the su pplier, it may be immaterial to the buyer. The largest buyers in the sample account for a disproportionate number of the relationships studied. The top 20 buyers in the sample account for nearly 50% of the firm -year relationship observations. Wal -Mart is the most frequently reported buyer and accounts for the highest number of unique relationships and the greatest amount of sales.5 In contrast, buyers which are reported to be in relationships with exactly one supplier account for a mere 5% of the sample. While the largest firms may have the greatest means to help their suppliers if they choose, they also have a market advantage and an ability to switch suppliers. Therefore, firms in fewer relationships will be more committed to their relationships. Second, suppliers will benefit more from relationships in which the buyer has a larger stake in the supplier. One measure of the buyers stake in the supplier is the percentage of the suppliers sales purchased by the buyer. The higher the percentage of sales the buyer is responsible for the more of the benefit the buyer will reap from helping the supplier. For example, when Ford renovated its Five Hundred series assembly plant, it asked a dozen of its critical suppliers to build factories within a few miles of the plant in order to implement a relatively new system of manufacturing referred to as a modular manufacturing. Following this, Ford began outsourcing many of its components, rather 5 The other 19 buyers in order of the number of buyer supplier relationship years are General Motors, Ford, IBM, Chrysler, Sears, Amerisource B ergen, Cardinal Healthcare, Ingram Micro, J.C Penney, General Electric, Motorola, Boeing, AT&T, Lucent Technologies, McKesson Drugs, Hewlett Packard, Nortel Networks, Home Depot, and Target

PAGE 31

31 than just individual parts, to these suppliers. Due to their location, the factories were dedicated to Ford, and Ford made up the great majority of company wide sales for most of the suppliers. As a result, investments made at these factories translated directly into higher quality and/or lower prices for Ford alo ne. Another measure of the buyers stake is the length of the relationship. At the beginning of a relationship, firms may not have fully committed to the relationship. Over time, firms gain knowledge about each other, build trust, and increase interde pendence. Therefore, as the length of a relationship increases, so does the buyers incentive to support the supplier. Third, suppliers in relationships that sell specialized products ought to be less liquidity constrained than their peers. Financial constraint may force the supplier to scale back its operations, resulting in lower inventory, which has the potential to harm the buyer. More seriously, illiquidity may risk firm liquidation. Titman (1984) and Titman and Wessels (1988) argue that buyers face switching costs when their supplier is liquidated. Further, these costs are especially high when the suppliers products are unique. Producers of unique products invest more heavily in relationship specific assets and therefore impose higher liquida tion costs on their buyers. Therefore, to avoid the costs imposed by supplier liquidation, buyers may be more willing to provide liquidity to their suppliers. Alternatively, suppliers in relationships may choose to stay more liquid. Both Kale and Shahrur (2007) and Banerjee, Dasgupta, and Kim (2008) provide evidence consistent with the disparity between manufacturers of unique and nonunique goods. They show that for firms which produce unique goods, the strength of the buyer -supplier relationship is inversely proportional to the suppliers leverage ratio. Further, the effect of a relationship on leverage is weaker for manufacturers of nonunique goods.

PAGE 32

32 Firms in the durable goods industry are thought to produce more unique goods. I therefore compare the investment cash flow sensitivities of firms in durable goods industries to those in nondurable goods industries. Because many goods in the durable goods industry cannot be easily substituted, this effect of being in a relationship should be stronger for suppliers of durable goods relative to other industries. Additionally, while capital market information problems arise at the firm level, financial constraints have a clear macroeconomic dimension because fluctuations in a firms cash flow and liqu idity are correlated with the aggregate economy over the business cycle. If buyers are harmed by supplier illiquidity, they should provide additional liquidity when market conditions demand it. In other words, buyers should provide assistance when suppli ers have difficulty accessing external capital. One measure of the cost of accessing external capital is the spread between BBB and AAA rated bonds. Widening of the spread indicates a decline in the external financial markets willingness to fund risky i nvestment. Another measure of the cost of external capital is whether or not the country is in a recession because the cost of external financing rises during recessions. Kayshap, Lamont and Stein (1994) study the 1981 1982 recession and present eviden ce that liquidity constrained firms contributed substantially to the overall inventory decline. Following a tightening in monetary policy, Gertler and Gilchrist (1994) find that small firms account for a disproportionate share of the decline in manufactur ing. This is especially relevant since the average firm in the dataset is significantly smaller than the average manufacturing firm.6 Liquidity, Investment, and Internal Capital Markets Models of simple perfect capital markets imply that, when cash flow s fall due to market conditions, all firms should be similarly affected. However, if capital markets are imperfect, then 6 For a discussion of firms size and other sample characteris tic see Section II.

PAGE 33

33 different firms will face different consequences as a result of changing economic environments. Specifically, firms without access to capital will have different behaviors than those which do. Thus, if a buyer -supplier relationship reduces capital market frictions, then the capital expenditures of buyer and supplier will be less affected by price shocks than their independent peers. To test whether being in a relationship is related to a firms financial constraints, I turn to a large literature that documents the link between liquidity and investment. I focus on a common theme of this work: that liquidity, defined as the availability of internal funds, should be an important determinant of investment when there are frictions in the capital market. Work by Fazzari, Hubbard and Petersen (1988), Hoshi, Kayshap and Sharfstein (1991), Houston, James, and Marcus (1997), Lamont (1997), Shin and Stulz (1998), Rauh (2006), and Fee, Hadlock, and Pierce (2008) document a strong correlation between cash and investment. Although this correlation is well -documented, the causation is more difficult to establish because both investment and cash flow are driven by underlying shocks to profitability. Since exogenous instruments for cash that are uncorrelated with the profitability of investment are difficult to find, researchers have instead focused on examining the differences in cash investment corre lations between groups of firms hypothesized to have different dependencies on internal financing. This paper follows this tradition. Studies typically use panel data on firms to estimate: where the dependent variable is the ratio of capital expenditures to assets, Ai,t1 is a measure of book assets or fixed capital at the beginning of the period, Qi,t 1 is Tobins Q at the beginning of the period, and CFit is a measure of cash flow or cash stock. To test the hypotheses that two groups of firms face different constraints, the coefficient 2 is compared across different firms with firms categorized according to different measures of financial constraints.

PAGE 34

34 This methodology has faced much criticism. In particular, this method relies on average Q to m easure marginal Q which may generate measurement error. However, for the mismeasurment of Q to affect the results in this paper, the error would have to differ systematically between firms which are in relationships and firms which are not. Moreover, the error would have to differentially affect distinct sub -groups of firms and affect the firms differentially at over time for the results to obtain. Data Set Data Set Construction The Statement of Financial Accounting Standards No. 14 Financial Report ing for Segments of Business Enterprise (SFAS No. 14) of the Financial Accounting Standards Board (FASB) requires that firms report information for segments that represent 10% or more of consolidated sales, for fiscal years ending after 1977. This include s disclosure of sales to principal customers, if the revenue generated from sales to a particular customer exceeds 10% of revenue of the firm, or if the firm considers the sales important to its business. Prior to 1997, suppliers were also required to rep ort the customers name. In 1997, the FASB issued SFAS 131, revising SFAS no. 14, which permitted firms to optionally report the customers name. Customers and revenue from each customer is collected in the Compustat Segments Data. My sample includes al l U.S. manufacturing firms (primary SIC code within 2000 3990) which have assets exceeding $10 million included in Compustat with non -missing values of sales and total assets which have sales to a customer identifiable in Compustat between 1979 and 2006.7 The minimum size requirement reduces the noise created when scaling the variables. For each firm, I identify whether its principal customer is listed in Compustat and assign it the corresponding firm identifier. Identification of firm customers requ ires individual verification as the 7 Total assets are converted to 1984 dollars using the consumer price index.

PAGE 35

35 names are not entered in the Segment data in a uniform way.8 I use a string matching algorithm to generate a list of potential matches to the customers name and then hand -match customers by inspecting the firms name s egment and industry information.9 I am deliberately conservative in assigning customer names and firm identifiers in order to ensure that the customers are matched to the appropriate financial information. After using the matching algorithm, I sort the dataset by firm name and year to confirm continuity over time. In several cases it is clear from inspection that the buyer -supplier relationship has remained constant over time although the matching algorithm missed a section of years.10 Customers for whic h I could not identify a unique match are excluded from the sample. Each year, a supplier can list multiple significant buyers in the Compustat sample. To compare changes within each company, each firm should only have one observation per year. I keep the buyer that accounts for the largest amount of firm revenue each year. Including multiple borrowers per supplier as distinct firm -year observation produces erroneously small standard errors. The sample includes 28 years (19792006) and covers a numbe r of macroeconomic cycles. To account for changes over time, I supplement the Compustat data with additional data from three sources. From the Bureau of Labor Statistics, I use the Consumer Price Index to adjust firm assets for inflation to 1984 dollars. To evaluate the effects of recessions and monetary tightening, I include data on economic recessions from the National Bureau of Economics and corporate bond yields from Federal Reserves website. 8 For example, throughout Shiloh Industries 14 year relationship with General Motors Compustat reported their principal customer as GEN MTR, GEN MOTORS, and General Motors Corp. 9 The SAS function Spedis is used to generate spelling distances 10 For example, in one year the matching algorithm had no problem matching the buyer PENNEY (J.C.) CORP. INC. to an existing Compustat firm. The following yea r for the same supplier the matching algorithm was unable to match the buyer JCPenneys to an existing Compustat firm

PAGE 36

36 Finally, for several reasons, before analyzing the dat a firms in the 3 digit SIC Code (283)for drug manufacturers are omitted from the sample. In the context of product market relationships, there is a continuum of organizational forms that range from distinct firms conducting arms -length transactions to com plete vertical integration. Fee, Hadlock, and Thomas (2006) investigate partial equity ownership which represents one form of this type of relationship. Anecdotal evidence suggests that equity stakes are not coincidental in trading relationships, but rat her they serve a role in maintaining such a relationship. They find that equity stakes are more common when the supplier is a n R&D intensive firm and when the companies have formal alliance agreements and conclude that the equity stake frequently serves t wo purposes. First, the equity stake helps to align incentives and to help provide contractual completeness. Second, many stakes represent newly issued shares indicating that it could serve to solve a financing constraint. If the relationship is fully d escribed and contracted, these firms should be omitted from the sample to avoid biasing the results. To determine which firms should be omitted from the sample, I choose a random sub-sample of 250 firms. For each of these firms, I downloaded their SEC 10-K reports from EDGAR and searched through the report for the name of their main trading partner to determine the nature of their relationship. Of the 250 firms, I identified explicit block equity ownership or partial firm ownership as the result of a joi nt venture in 16 cases or 6.4% of the subsample. This is consistent with Fee, Hadlock, and Thomas (2006) who find that equity stakes exist in 3.31% of all buyer -supplier relationships. Of the 16 cases found 25% fall under the SIC sub-category 283 which represents the creation and manufacture of drugs. In no other 3 digit SIC code are there more than 2 cases. In general, these firms are highly R&D intensive and require large amounts of capital up front. Additionally, of the 40 firms in this industry mo re than 75% had explicitly outlined and contracted strategic alliances. While the majority of manufacturers operate based on purchase

PAGE 37

37 orders or cancelable contracts, biotechnology firms are dependent on large upfront investments and milestone payments from their customers. Firms in the SIC sub -category 283 are omitted from the sample. This is further justified by the stream of literature devoted solely to characterizing and understanding the firms in the biotechnology field (i.e. Lerner, Shane, and Tsa i, 2003). Firm Matching To provide the best possible test of the data, a matched sample is created to control for firm differences. The reporting process requires that if sales to one particular buyer account for 10% or more of a suppliers profit th en the supplier must report the sales. This biases the sample of suppliers towards the smaller firms in Compustat.11 Additionally, firm financing practices differ by industry. To control for the obvious difference in size and different industry practices I create a oneto -one matched sample of firms which do not report relationships. First, matched firms are chosen from a sample of firms that never report sales to a single customer as exceeding 10% of their revenues during the life of the firm. Then, t he matched firm is chosen from the subsample based on the year, 3 digit SIC code, and net sales. In the first year that a supplier identifies a buyer the independent firm in the same industry with the smallest absolute difference in net sales is chosen as the relationship firms match. These firms remain matched over time. In the event that the independent matched firm is de listed from Compustat before the supplier stops listing a substantial buyer, another firm is substituted in its place using the sam e criteria from that year forward. During the first year of each relationship, the median difference in firm size, measured by total assets, is a mere $1.59 million and is not statistically different from zero. However, after the first year, the firms in relationships grow at a somewhat faster rate than their matched peers, resulting in an average difference of $63.37 million dollars. This finding indicates two things. First, the match 11 The average inflation adjusted firm size in Compustat over the sample period is $704 million dollars which is considerably larger tha n the supplier sample average of $45 1 million dollars

PAGE 38

38 achieves the goal of controlling for firm size. Second, the growth o f suppliers in relationships outpaces their matched peers. This suggests that firm relationships are beneficial to the supplier. Results Univariate Evidence Table 2 1 reports the mean, median and standard deviation of firm characteristics for suppliers in relationships, matched suppliers which do not report relationships, and pairwise differences between each supplier and its match. The average difference is found by first taking the difference of each firm and its specific intra industry size -matched f irm each year and then finding the average of the differences. This method controls for differences in industry and to a lesser extent firm size. The results in Table 2 1 provide conflicting evidence regarding the benefits and costs of being in a relati onship. Despite having customers which account for a larger percent of their total sales, suppliers in relationships have lower accounts receivable relative to net sales than their matched peers. This frees up working capital and is a benefit to being in a relationship. Additionally, firms in relationships have greater sales relative to total assets tha n their matched peers. Dependent suppliers have higher relative amounts of investment and R&D congruent with the findings of Banerjee, Dasgupta and Kim ( 2008). On the other hand, in contrast to the findings of Kalwani and Narayandas (1995), dependent suppliers appear to be less profitable as measured by operating profit margin and return on assets. The influence of relationships on firm value is influenc ed by the nature of the firm. Table s 2 2 and 2 3 divide the sample of firms into two groups, durable and non-durable goods manufacturers, respectively. Table 2 3 reveals that among durable goods manufacturers, the benefits of relationships include lower accounts receivable and larger potential growth (as measured by Q). Relationship costs include larger cash holdings and lower return on assets and operating profit margin. It is interesting

PAGE 39

39 to note that variability of both cash flow and investment is lar ger for firms in relationships than their matched peers as measured by the standard deviation. Turning to non -durable goods manufacturers, Table 2 3 indicates that among non -durable goods manufacturers, suppliers in relationships have lower accounts rec eivables than their matched peers. Non durable goods manufacturers in relationships also have higher sales as a ratio of total assets indicating that they are more efficient with their capital and have lower transaction costs. Further, non-durable goods manufacturers in relationships hold lower cash reserves. This may indicate one of two things. One possibility is that holding less cash is a symptom of financial distress, because these firms cannot raise enough cash. Another possibility is that the pre sence of a large customer reduces business risk and allows a supplier in a relationship to safely hold less cash allowing firms to earn greater returns on their capital. Comparing the statistics in Tables 2 2 and 2 3 offers some insights into the differ ences between durable and non -durable goods manufacturers. Supporting the theory that durable goods manufacturers offer unique goods, durable goods manufacturers spend more money on both capital expenditures and research and development as a percent of to tal assets than non -durable goods manufacturers. Additionally, durable goods manufacturers have larger growth potential than non durable goods manufacturers. However, durable goods manufacturers have considerably lower returns on assets and operating profit margins than non -durable goods manufacturers. Further, Tables 2 2 and 2 3 reveal interesting differences regarding how relationships affect the different types of firms. Among both types of firms, suppliers in relationships have lower accounts recei vable than their matched peers. Further, both types of firms benefit from a higher sales to asset ratio.

PAGE 40

40 While there are some similarities between the influence of a relationship on durable and non durable manufacturers, there are also many differences Durable goods manufacturers hold more cash than their matched peers while non -durable goods manufacturers hold less cash. Relationships also influence cash holdings and investment differently for durable and non -durable goods. Further, the variability of cash flow and investment is much larger among durable goods manufacturers than it is among nondurable goods manufacturers. As the sensitivity of investment to cash flow is the primary effect studied in this paper, the influence of a relationship on t he sensitivity of investment to cash flow is investigated separately for these different groups of firms. After evaluating the differences between suppliers which are in relationships and independent firms, I turn to differences among the group of suppl iers in relationships. The firms that are in relationships are split into groups based on the principal buyers stake in the supplier. First, firms are divided into groups based on the strength of each relationship which is measured by the percent of sal es to the buyer. Then, suppliers are divided into groups based on the length of their relationship. Table 2 4 compares the mean and median values for suppliers which are in relationships grouped by the strength (percent of sales to the buyer) of their relationship each year. Surprisingly, there does not appear to be a high correlation between the number of years in a relationship nor firm age and the percent of revenue from a customer. Firm size is affected by the grouping. Smaller suppliers, as measured by both total assets and net sales, tend to be most dependent on one buyer. Smaller firms are also the ones that find it hardest to acquire external financing. So, it is for these firms that access to capital via a relationship may be most important. As the strength of the relationship increases, accounts receivable to total assets decreases. While the result appears to be economically significant, it is not statistically significant. Surprisingly, when comparing group 1 to group 2 the leverage ra tio does not decrease significantly. When comparing group 2 to group 3 we

PAGE 41

41 see that the market leverage ratio decreases as relationship strength increases, consistent with the results of Banerjee, Dasgupta, and Kim (2007) and Kale and Sharur (2008).12 Prof itability as measured by return on assets, cash flow to assets, and operating profit margin, declines as supplier dependence increases. Table 2 5 compares the mean and median values for dependent suppliers grouped by the length of their relationship ea ch year. As a result, every firm in the second group is in the first group for all 4 years. Indeed, firm age is in part mechanically related to the number of years that firms are in relationships. A firm cannot be in a relationship for more than 4 years if it is not more than 4 years old. As firms tend to grow in size over time, the average firm in a longer relationship is significantly larger, as measured by total assets, than firms in shorter relationships. By this measure, suppliers in longer relati onships are larger, as opposed to the previous specification in which smaller firms accounted for a larger percentage of sales to one buyer. The majority of firm differences follow the opposite pattern of the previous table which compared firms based on the percent of sales to the primary buyer. Both total assets and net sales increase with relationship length and decrease as percent of sales to the primary buyer increases. Further, as the length of the relationship increases, both leverage and profita bility increase. Simultaneously, overall investment and investment in R&D decrease as the length of the relationship increases. But, these all increase as the percent of sales to the buyer increases. If the sensitivity of investment to cash flow decline s as each of the measures of relationship dependence increases then we can conclude that it is due to the influence of the relationship not other firm characteristics. To confirm that the difference across groups is not the result of differences in firm a ge and size, I use a difference -in -difference approach. First, for each firm characteristic I find the difference in 12 A replication of the multivariate regression in Banerjee, Dasgupta, and Kims (2008) Table VII confirms the efficacy of the dataset.

PAGE 42

42 value between each firm in a relationship and its match. Then I average these to find the mean difference after controlling for firm size, age and industry. Then, I compare the differences of the average value between the two groups according to length of the relationship. The Satterthwaite pvalue for the difference between the groups of the average difference is shown. The results conf irm that for (almost) all of the firm characteristics there is a statistically significant difference between firms in relationships for 1 4 years and firms in relationships for 5 or more years. The one difference that remains constant across groups is the most critical to this study. As relationship length increases, accounts receivable decreases. Paying off trade credit in a timely fashion is one way by which buyers can influence the financial liquidity of their suppliers. While admittedly weak, this provides one of the first pieces of evidence that relationships can ease financial frictions. Sensitivity of Investment to Cash Flow As shown above, in general, firms in relationships have less accounts receivable outstanding. This demonstrates that buyers have the means to ease their suppliers financing constraints. As discussed in Section II, this is not the only way a buyer can influence its suppliers financial liquidity. They may also do so both directly, by agreeing to price concessions, or i ndirectly, by reducing firm risk and thus improving the chance that a supplier can acquire a bank loan. Alternatively, suppliers in relationships may actively pursue higher levels of cash or lower levels of debt in an effort to be less liquidity constrain ed. To determine if the presence of a material buyer eases a suppliers financial constraints, I utilize the methodology pioneered by Fazzari, Hubbard and Petersen (1988). The regressions include investment, cash flow and Q. To eliminate the effects of scale, I normalize investment and cash flow by the firms total assets at the beginning of the year. Firm fixed effects are included to absorb sources of variation which may be undesirable due to their correlation with investment opportunities. Because t he sample corresponds to years in which stages of the business

PAGE 43

43 cycle differ, the regressions include yearly dummies for each year except 1980. The coefficients of the dummy variables are not reported in the tables. Capital expenditure taken from the fi rms statement of cash flow (data128) is the measure of investment used. Cash flow is defined as income plus depreciation and amortization (data18 + data14). Because depreciation and amortization are noncash charges they should be added back to determin e the true cash flow. The theory predicts that Tobins Q should be the only determinant of investment. Tobins Q is defined as the market value of equity plus the book value of assets minus the book value of equity all divided by the book value of assets A standard criticism of this literature is that liquidity proxies for an important omitted variable, namely the profitability of investment. When a firms liquidity is high, it is likely to be doing well and it should have good investment opportunities Thus, it is not surprising that they invest more. To control for this, most researchers try to control for the value of investment opportunities by including Tobins Q in the equation. However, it is a distinct possibility that Q is mismeasured. Ther efore, instead of estimating the effect of liquidity on investment for all firms, I separate the firms according to a priori beliefs about how liquidity should affect investment. Table 2 6 compares firms in relationships to the matched sample of firms w hich do not report being in a relationship. Firm cash flow is interacted with a dummy variable which indicates whether the firm is in a relationship. If a buyer either directly or indirectly helps to provide liquidity for the supplier, then the investmen t cash flow sensitivity of suppliers in relationships should be smaller than their matched peers. The results in Table 2 6 support this conclusion. At the most aggregate level, being in a relationship lowers a firms sensitivity of investment to cash flo w. However, the result is not economically or statistically significant. This is expected as there are a variety of types of firms and relationships pooled together.

PAGE 44

44 Columns 2 and 3 of Table 2 6 divide the sample into groups of firms which manufacture d urable and nondurable goods, respectively. Durable goods have primary SIC codes from 3400 to 3990. Durable goods are unique and require relationship specific investments. Among these firms, being in a relationship may be more valuable to the buyer when the relationship is functioning properly and more risky and costly when it is not. In column 2, among durable goods manufacturers, there is not a significant difference between the investment cash flow sensitivity of firms in relationships relative to in dependent firms. In column 3, the results for non-durable goods are consistent with the prediction. Firms in relationships have significantly lower sensitivity of investment to cash flow. Together these results imply that relationships affect firm finan cing and investment and that being in a relationship affects firms in durable goods industries differently than firms in nondurable goods industries. Table 2 7 shows the effect of a buyers commitment to the supplier. In this table, cash flow is interact ed with a dummy variable indicating whether the firm is in a relationship with one of the top 20 buyers based on the number of relationships that are reported for the buyer. Here, cash flow is interacted with a dummy variable which indicates whether the f irm is in a relationship with a buyer which is not one of the top 20 buyers. Therefore, the results indicate the marginal effect of being in each one of these types of relationships relative to not being in a relationship at all. Presumably, the fewer re lationships a buyer is in, the more the buyer will be committed to each individual relationship. The results in Table 2 7 are consistent with the hypothesis that buyers of easily substitutable goods, which have the ability to easily switch suppliers, do not alter the cash flow sensitivity of the suppliers which sell to them. Among non -durable goods manufacturers, there is no difference between being in a relationship with one of the top 20 buyers and not being in a relationship at all.

PAGE 45

45 However, for buye rs with an incentive to aid their suppliers, that is buyers which engage in only a few important relationships, non -durable goods suppliers display lower sensitivity of investment to cash flow.13 Table 2 8 shows the effect of the size of the buyers stake in the supplier on investment cash flow sensitivity. In columns 1, 2 and 3, the buyers stake in the supplier is proxied for by the percent of sales to the buyer. Unlike papers that categorize firms once for all of the years of the firms life, here fir ms can change groups as their relationship strength changes. The first group of firms report that a buyer accounts for less than 25% of their sales. The second group reports that a buyer accounts for 25% 50% of their sales. Firms in the final group ha ve a buyer which accounts for more than 50% of their sales. I then divide the firms into groups based on the current length of the relationship. Cash flow is interacted with an indicator for whether the relationship has lasted for more or less than 4 yea rs. The table reports the marginal difference between being in each group relative to not being in a relationship at all. Table 2 8 column 2 reports results for the set of durable goods manufacturers while column 3 reports non -durable goods manufacture rs. The results in Table 2 8 demonstrate that inter -firm relationships help to ease a suppliers liquidity constraints, at least for firms in the strongest relationships. However, firms in the weakest relationships (those with less than 25% of their sale s to their largest buyer) are indistinguishable from firms which are not in relationships. As the strength of the relationship increases, investment cash flow sensitivity falls. 13 To test the robustness of these results I investigate the ma rginal difference between the sensitivity of not being in a relationship to being in a relationship with different types of buyers. The results are not tabulated but available upon request. First, cash flow is interacted with the buyer being a retailer or wholesaler (SIC 5000 5999). In a second set of tests, cash flow is interacted with the buyer being a manufacturer (SIC 2000 3990). Both provide results consistent with those found in Table 3, but neither provide any further insight into the specific types of relationships which are beneficial to suppliers.

PAGE 46

46 The latter columns in Table 2 8 report the estimates of investment equatio ns for manufacturers grouped by their relationship length. Firms must be in the first group in order to appear in the second group. The results are reported for the full sample from 1980 2006 and then results are reported again for the subsample from 1 984 2006. During the years 1980 1983, any firm that reports a relationship is determined to have been in the relationship for less than 4 years due sample construction. Regression results from subsample of firms from 1984 2006 are reported to elimi nate the bias that could come from placing firms in longer relationships in the first sub -category. Among durable goods manufacturers in the full sample from 1980 2006, the cash flow coefficient is negative and significant when firms have been in rela tionships for more than 4 years. This suggests that when firms are in relationships for longer periods of time, they become less liquidity constrained. This is not the case for non -durable goods manufacturers. While the cash flow coefficient is negative for firms in longer relationships, it is not statistically significant. Among durable goods manufacturers in the sub-sample from 1984 2006, the cash flow coefficient decreases as relationship length increases. This further supports the above conclusi on that inter -firm dependence eases capital market frictions by mitigating liquidity constraints. This additional evidence is a critical piece of the puzzle. Without this finding, one could claim that other firm characteristics such as size cause spuriou s results. However, while firm size is inversely related to relationship strength, it is positively related to relationship length. This confirms that the results are driven by relationship characteristics rather than from omitted firm characteristics. Among non -durable goods manufacturers, the results are not statistically significant. Although being in a relationship reduces the sensitivity of investment to cash flow, length does not affect the strength of the results. This is consistent with how d urable and non durable goods

PAGE 47

47 manufacturers differ. Since they invest in relationship specific assets, durable goods manufacturers stand to gain more from relationships over time compared to non -durable goods manufacturers. Tables 6 and 7 provide evidenc e that macroeconomic policy affects both firm financing and investment. By studying an exogenous shock to credit and economic cycles we can determine the influence of relationships on liquidity constraints. If a buyer is harmed by supplier illiquidity, t hen they should provide additional liquidity when demanded by market conditions. Recessions are characterized by an inaccessibility of external finance. The economy is defined to be in a recession for a year if the economy is classified by the NBER as be ing in a recession for at least 6 months of that year. I construct a recession indicator variable which is equal to one if the U.S. economy is in a recession for the year and zero if not. Table 2 9 column 1 confirms that firms have greater liquidity co nstraints during recessions. In the first 3 columns, firm cash flows are interacted with indicators for whether the country is in a recession. Across all specifications, investment is more sensitive to cash flow during recession years. In Table 2 9 co lumns 4, 5, and 6, both suppliers in relationships and independent suppliers cash flows are independently interacted with the recession indicator. This allows multiple comparisons. First, for suppliers in relationships there is not a significant differe nce in cash flows between recession years and non recession years. Interestingly, these firms do not show more liquidity constraints during a recession. This is consistent with the hypothesis that additional liquidity is available to firms in relationshi ps during times when accessing external capital is most difficult. However, for independent firms there is a significant difference between recession and nonrecession years, consistent with the proposed theory that relationships provide liquidity (or in this case that lack of a relationship fails to provide liquidity).

PAGE 48

48 Initially, Table 2 7 reported that independent firms were less liquidity constrained than firms in relationships. However, comparing firms during recessions offers a different picture. Comparing the investment cash flow sensitivity of durable goods manufacturers in relationships to those not in relationships during recessions indicates that when suppliers need it most, liquidity is provided. To confirm the robustness of these results I use an alternative measure of the cost of accessing external capital, namely the spread between BBB and AAA rated bonds, by Moodys. Widening of the spread indicates a decline in the external financial markets willingness to fund risky investment. A gain, I create an indicator variable for the difficulty of accessing external capital. The indicator is equal to 1 when the spread is above the 75th percentile of the years in the sample. Table 2 10 reports the results of the regressions in which cash flow is interacted with the spread dummy. The main results are the same as in Table 2 9 First, on average, firms are significantly more liquidity constrained when accessibility of external capital is limited. Second, among durable -goods manufacturers t hese results show no differences. That is, the F test cannot reject the hypothesis that the coefficients are the same. So, when the credit spread widens, dependent durable goods suppliers are no more liquidity constrained than otherwise. Finally, when t he credit spread widens, firms in relationships are less liquidity constrained than their independent counterparts. Table 2 10 also provides additional evidence of the difference between durable and non -durable goods manufacturers. The results are consis tent with relationships influencing durable goods manufacturers but not nondurable goods manufacturers. Conclusion Over the past few decades, there has been a surge in the prevalence and importance of significant buyer -supplier relationships. Clearly, relationships have both costs and benefits. Despite this, little extant literature addresses how major relationships influence firm financing and investment

PAGE 49

49 policy. This paper addresses one facet of this complex issue, namely the effects of being in a fi rm relationship on supplier investment cash flow sensitivity. This paper provides evidence that suppliers in relationships are less affected by capital market frictions. At the most aggregate level, suppliers in significant relationships demonstrate low er investment cash flow sensitivity than their independent peers, indicating that suppliers in relationships are less liquidity constrained than their peers. All of the tests are based on an industry and size matched sample to ensure differences across in dustries do not influence the results. In theory, the more committed a buyer is to a relationship, the greater their incentive to help the supplier. The empirical results indicate that this is the case. Buyers reported to be in only a few relationships affect the sensitivity of investment to cash flow of their suppliers. In contrast, suppliers in relationships with the largest buyers (who can easily switch suppliers) are no better off than their independent peers. This is consistent with the hypothesis that suppliers in relationships with more committed buyers face lower financing frictions. Using the percent of suppliers sales from the principal buyer and the length of the relationship as proxies for the buyers stake in a relationship, I find tha t as the buyers stake in its supplier increases the suppliers sensitivity of investment to cash flow decreases. This is consistent with the idea that, in effect, the buyer is the residual claimant on its suppliers investments. Therefore, as the buyer s stake in the supplier increases, the buyers incentive to aid the supplier increases as well. Further, using the corporate bond spread and a recession index as proxies for the accessibility of external financing, I find that the investment to cash flo w sensitivity of suppliers with principal customers is less affected by macroeconomic cycles than the sensitivity of their peers. This result is consistent with the theory that, when necessary, firms in relationships face lower financing frictions.

PAGE 50

50 Final ly, firms in the durable goods industry are thought to produce more unique products and invest more heavily in relationship specific assets and therefore impose higher liquidation costs on buyers. Consistent with this idea, I find that the effects are mor e pronounced for firms in industries with a high degree of relationship specific investments. Consequently, I find evidence consistent with the idea that being in a relationship helps to ease capital market frictions which affect suppliers investments.

PAGE 51

51 T able 2 1 Characteristics of firms in relationships compared to a matched sample This table shows the mean, median, and standard deviations of firm characteristics for the sample of U.S. manufacturing firms from 1980 2006 which are in identifiable relationships. For each supplier a matched firm is chosen based on the suppliers industry, firm sales and year. For each firm characteristic, I find the difference between the supplier and the matched independent firm and then take the average. Total Assets is firm assets (data6) adjusted for inflation to 1984 dollars using the Consumer Price Index. AR/S is total firm receivables (data2) divided by net firm sales (data12). The mark et leverage ratio is defined as long term debt d ivided by the book assets minus book equity plus market equity ( data9 / (data6 data60 + (data199 data25)). Book leverage is defined as total long term debt (data9) divided by total assets. ROA is income before extraordinary items (data18) divided by total assets. Cash flow is the sum of income before extraordinary items and depreciation and amortization (data18 + data14) divided by total assets. Cash holdings / TA is cash and short term receivables (data1) divided by total assets. Sales/TA is net f irm sales (data12) divided by total assets. Operating Profit Margin is the difference between operating income before depreciation and depreciation and amortization (data13data14) divided by net firm sales. Q is defined as market equity plus book debt divided by the book value of firm assets (((data199*data25) + data6 data216)/(data181 + data216)). Investment/TA is determined by capital expenditures (data128) scaled by total assets. R&D is the research and development expenses of a firm (data46) divi ded by total assets. When Compustat reports this value as missing, it is assumed to be zero. All variables are winsorized at the 1st and 99th percentiles. The students T test is used to determine if the means are statistically different from zero. The Wilcoxon RankSum test is used to determine if the median is statistically different from zero. The symbols ***, **, and indicate statis tical significance at the 1%, 5%, and 10% levels, respectively. Mean Median Std. Dev. Mean Median Std. Dev. Mean Median Total Assets 451.58 82.25 1132.01 393.48 78.06 1048.54 63.37 *** 3.79 *** Net Sales 509.92 102.20 1252.74 405.09 93.04 1038.28 118.99 *** 2.21 *** AR/S 0.16 4 0.153 0.070 0.172 0.160 0.081 0.011 *** 0.005 *** Leverage Ratio (Market) 0.137 0.094 0.147 0.136 0.086 0.150 0.000 0.000 Leverage Ratio (Book) 0.170 0.131 0.172 0.165 0.121 0.173 0.003 0.000 ROA 0.003 0.039 0.169 0. 011 0.041 0.141 0.015 *** 0.005 *** Cash Flow/ TA 0.044 0.080 0.159 0.055 0.080 0.133 0.011 *** 0.002 *** Cash Holdings / TA 0.155 0.073 0.185 0.144 0.075 0.167 0.011 *** 0.000 *** Sales / TA 1.273 1.229 0.559 1.230 1. 168 0.542 0.044 *** 0.034 *** Operating Profit Margin 0.011 0.062 0.262 0.035 0.066 0.213 0.027 *** 0.005 *** Q 1.636 1.311 1.020 1.614 1.289 1.026 0.022 0.021 ** Investment / TA 0.059 0.046 0.049 0.052 0.040 0.042 0.008 *** 0.004 *** R&D / TA 0.055 0.025 0.074 0.040 0.017 0.055 0.017 *** 0.001 *** All firms Pairwise differences Matched Sample Firms in Relationships

PAGE 52

52 Table 2 2. Characteristics of durable goods manuf acturers in relationships compared to a matched sample This table shows the mean, median, and standard deviations of firm characteristics for the sample of U.S. manufacturing firms from 1980 2006 which are in identifiable re lationships. For each supplier a matched firm is chosen based on the suppliers industry, firm sales and year. For each firm characteristic, I find the difference between the supplier and the matched independent firm and then take the average. Total Asse ts is firm assets (data6) adjusted for inflation to 1984 dollars using the Consumer Price Index. AR/S is total firm receivables (data2) divided by net firm sales (data12). The mark et leverage ratio is defined as long term debt divided by the book assets minus book equity plus market equity ( data9 / (data6 data60 + (data199 data25)). Book leverage is defined as total long term debt (data9) divided by total assets. ROA is income before extraordinary items (data18) divided by total assets. Cash flow i s the sum of income before extraordinary items and depreciation and amortization (data18 + data14) divided by total assets. Cash holdings / TA is cash and short term receivables (data1) divided by total assets. Sales/TA is net firm sales (data12) divided by total assets. Operating Profit Margin is the difference between operating income before depreciation and depreciation and amortization (data13data14) divided by net firm sales. Q is defined as market equity plus book debt divided by the book value o f firm assets (((data199*data25) + data6 data216)/(data181 + data216)). Investment/TA is determined by capital expenditures (data128) scaled by total assets. R&D is the research and development expenses of a firm (data46) divided by total assets. When Compustat reports this value as missing, it is assumed to be zero. All variables are winsorized at the 1st and 99th percentiles. The students T test is used to determine if the means are statistically different from zero. The Wilcoxon RankSum test is used to determine if the median is statistically different from zero. The symbols ***, **, and indicate statistical signif icance at the 1%, 5%, and 10% levels, respectively. Mean Median Std. Dev. Mean Median Std. Dev. Mean Median Total Assets 412.09 73 .03 1078.29 392.52 68.71 1131.34 37.76 *** 3.61 *** Net Sales 453.82 81.52 1195.91 395.17 75.85 1114.98 89.53 *** 1.18 *** AR/S 0.176 0.162 0.071 0.183 0.169 0.083 0.010 *** 0.005 *** Leverage Ratio (Market) 0.116 0.068 0.136 0.119 0.065 0.143 0.003 0.000 Leverage Ratio (Book) 0.147 0.101 0.160 0.146 0.088 0.171 0.001 0.000 ROA 0.013 0.037 0.183 0.002 0.038 0.152 0.016 *** 0.004 *** Cash Flow/ TA 0.036 0.080 0.171 0.045 0.075 0.143 0 .010 *** 0.001 Cash Holdings / TA 0.189 0.114 0.200 0.167 0.099 0.177 0.022 *** 0.002 *** Sales / TA 1.176 1.141 0.519 1.173 1.099 0.521 0.004 0.013 Operating Profit Margin 0.008 0.061 0.299 0.021 0.062 0.240 0.032 *** 0. 004 *** Q 1.719 1.369 1.090 1.632 1.315 1.012 0.087 *** 0.044 *** Investment / TA 0.060 0.046 0.050 0.048 0.037 0.041 0.012 *** 0.007 *** R&D / TA 0.072 0.047 0.079 0.053 0.034 0.060 0.022 *** 0.009 *** Pairwise differenc es Firms in Relationships Matched Sample Durable Goods

PAGE 53

53 Table 2 3. Characteristics of non -durable goods manufacturers in relationsh ips compared to a matched sample This table shows the mean, median, and standard deviations of firm characteristics for the sample of U.S. manufacturing firms from 1980 2006 which are in identifiable relationships. For each supplier a matched firm is chosen based on the suppliers industry, firm sales and year. For each firm characteristic, I find the difference between the supplier and the matched independent firm and then take the average. Total Assets is firm assets (dat a6) adjusted for inflation to 1984 dollars using the Consumer Price Index. AR/S is total firm receivables (data2) divided by net firm sales (data12). The mark et leverage ratio is defined as long term debt divided by the book assets minus book equity plus market equity ( data9 / (data6 data60 + (data199 data25)). Book leverage is defined as total long term debt (data9) divided by total assets. ROA is income before extraordinary items (data18) divided by total assets. Cash flow is the sum of income be fore extraordinary items and depreciation and amortization (data18 + data14) divided by total assets. Cash holdings / TA is cash and short term receivables (data1) divided by total assets. Sales/TA is net firm sales (data12) divided by total assets. Ope rating Profit Margin is the difference between operating income before depreciation and depreciation and amortization (data13data14) divided by net firm sales. Q is defined as market equity plus book debt divided by the book value of firm assets (((data1 99*data25) + data6 data216)/(data181 + data216)). Investment/TA is determined by capital expenditures (data128) scaled by total assets. R&D is the research and development expenses of a firm (data46) divided by total assets. When Compustat reports thi s value as missing, it is assumed to be zero. All variables are winsorized at the 1st and 99th percentiles. The students T test is used to determine if the means are statistically different from zero. The Wilcoxon RankSum test is used to determine if the median is statistically different from zero. The symbols ***, **, and indicate statistical significance at the 1%, 5%, and 10% levels, respectively. Mean Median Std. Dev. Mean Median Std. Dev. Mean Median Total Assets 541.50 111.47 1241.27 395.67 108.28 829.84 121.70 *** 4.69 *** Net Sales 637.69 161.70 1365.08 427.68 13 6.27 837.5 186.08 *** 10.57 *** AR/S 0.136 0.132 0.057 0.145 0.137 0.069 0.012 *** 0.006 *** Leverage Ratio (Market) 0.183 0.149 0.159 0.176 0.139 0.157 0.006 0.003 Leverage Ratio (Book) 0.221 0.193 0.186 0.209 0.189 0.168 0.013 *** 0.002 ROA 0.019 0.041 0.130 0.033 0.047 0.106 0.013 *** 0.006 *** Cash Flow/ TA 0.062 0.082 0.125 0.078 0.090 0.104 0.016 *** 0.008 *** Cash Holdings / TA 0.079 0.032 0.111 0.092 0.036 0.125 0.012 *** 0.002 *** Sales / TA 1.493 1.453 0.585 1.361 1.307 0.565 0.136 *** 0.090 *** Operating Profit Margin 0.054 0.064 0.137 0.066 0.075 0.126 0.016 *** 0.005 *** Q 1.446 1.199 0.806 1.575 1.219 1.056 0.125 *** 0.032 *** In vestment / TA 0.056 0.044 0.046 0.059 0.049 0.044 0.002 0.004 *** R&D / TA 0.012 0.000 0.028 0.007 0.000 0.016 0.004 *** 0.000 *** Pairwise differences Firms in Relationships Matched Sample Non Durable Goods

PAGE 54

54 Table 2 4. Cross -sectional differences of firms in relationships by customer sales Mean Median Std. Dev. Mean Median Std. Dev. Mean Median Std. Dev.Revenue from Buyer 0.144 0.140 0.052 0.342 0.330 0.069 0.649 0.618 0.124 -30.39 *** -19.10 *** Years in Relat 4.383 3 3.884 5.298 4 4.266 4.787 4 3.413 -17.50 *** 5.60 *** Firm Age 10.890 9 6.808 10.330 9 6.792 8.785 7 5.984 3.43 *** 6.05 *** Total Assets 538.06 98.53 1253.04 231.59 60.78 650.62 162.71 39.30 641.8 42.69 *** 56.41 *** 0.000 *** 0.346 Net Sales 599.37 118.77 1362.57 283.08 74.88 823.13 208.78 42.89 902.14 40.64 *** 51.90 *** 0.000 *** 0.683 AR/S 0.165 0.155 0.067 0.161 0.148 0.074 0.154 0.137 0.086 0.61 0.49 0.561 0.005 *** Leverage Ratio (Market) 0.139 0.096 0.147 0.139 0.098 0.150 0.100 0.046 0.123 0.02 2.25 ** 0.260 0.026 ** Leverage Ratio (Book) 0.174 0.135 0.173 0.166 0.125 0.172 0.128 0.068 0.153 0.73 2.00 ** 0.083 0.301 ROA 0.001 0.039 0.163 -0.012 0.036 0.177 -0.026 0.033 0.201 1.26 0.63 0.334 0.729 Cash Flow/ TA 0.048 0.081 0.153 0.036 0.077 0.167 0.024 0.081 0.188 1.13 0.57 0.664 0.663 Cash Holdings / TA 0.147 0.070 0.174 0.160 0.064 0.198 0.235 0.156 0.232 -1.15 -3.26 *** 0.298 0.000 *** Sales / TA 1.264 1.219 0.532 1.317 1.286 0.613 1.220 1.142 0.670 -2.68 *** 2.48 ** 0.076 0.933 Operating Profit Margin 0.027 0.064 0.218 -0.018 0.056 0.316 -0.087 0.051 0.454 3.25 *** 2.19 ** 0.004 ** 0.134 Q 1.634 1.319 0.995 1.572 1.268 1.000 1.883 1.385 1.317 2.42 ** -5.78 *** 0.385 0.000 *** Investment / TA 0.058 0.045 0.047 0.060 0.046 0.052 0.066 0.052 0.057 -0.48 -0.53 0.000 *** 0.061 R&D / TA 0.053 0.025 0.070 0.057 0.021 0.080 0.070 0.032 0.094 -0.58 -0.86 0.004 *** 0.069 Group 1 2 Group 2 3 P-Value for diff. in differences > 50% 25% 50% 10% < 25% T-stat for diff. between means Group 2 3 Group 1 2 This table shows the mean, median and standard deviation of firm characteristics for the sample of U.S. manufacturing firms from 1980 2006 which are in identifiable relationships. The sample is divided into groups based on the degree of supplier dependence. In Pa nel A firms are divided according to the strength of the relationship defined by the percent of sales derived from the firms largest customer. In Panel B firms are divided ac cording to the length of the relationship where firms that are in the first 4 yea rs of their relationships are in the first columns and firms in their fifth or greater year of relationship are in the latter columns. The null hypothesis tested is that the there is no difference between the means of the different groups. The t value co rresponding to the difference of means is shown. P value states the Satterthwaite value for the probability that the difference between the two groups (where the value used for each group is the average of the difference between the firm and its match) is statistically significant. Revenue from the buyer is defined as the amount of sales from the largest buyer divided by net sales. Years in relationship is the total number of years prior to the reporting year that t he supplier lists the buyer as a material customer. Firm age is the number of years that the firm is listed in Compustat beginning in 1980. For definitions of the rema ining variables see Table 21

PAGE 55

55 Table 2 5. Cross -sectional differences of firms in relationships by relationship length Mean Median Std. Dev. Mean Median Std. Dev. Revenue from Buyer 0.205 0.160 0.146 0.233 0.190 0.151 -3.43 *** Years in Relat 2.290 2 1.059 8.56 7 3.975 -171.94 *** Firm Age 9.350 7 6.679 12.92 12 6.309 -26.97 *** 0.000 *** Total Assets 370.58 70.82 1011.84 590.60 110.56 1301.32 -29.24 *** 0.011 ** Net Sales 403.63 83.47 1083.40 692.36 143.70 1481.95 -36.97 *** 0.000 *** AR/S 0.167 0.155 0.074 0.159 0.150 0.062 1.45 0.027 ** Leverage Ratio (Market) 0.130 0.078 0.147 0.149 0.115 0.145 -2.34 ** 0.009 *** Leverage Ratio (Book) 0.163 0.116 0.173 0.182 0.152 0.170 -2.13 ** 0.144 ROA -0.013 0.037 0.186 0.014 0.042 0.133 -3.19 *** 0.000 *** Cash Flow/ TA 0.034 0.077 0.174 0.061 0.084 0.126 -3.36 *** 0.000 *** Cash Holdings / TA 0.177 0.092 0.198 0.117 0.049 0.151 6.85 *** 0.000 *** Sales / TA 1.225 1.173 0.569 1.354 1.308 0.532 -8.18 *** 0.000 *** Operating Profit Margin -0.009 0.061 0.310 0.045 0.063 0.141 -5.61 *** 0.000 *** Q 1.732 1.355 1.12985 1.470 1.258 0.76843 12.91 *** 0.000 *** Investment / TA 0.060 0.045 0.051662 0.057 0.046 0.044263 0.71 0.745 R&D / TA 0.062 0.029 0.07962 0.043 0.020 0.06053 3.31 *** 0.000 *** 1 4 Years P-value for difference in differences T-stat for difference between means 5+ Years This tab le shows the mean, median and standard deviation of firm characteristics for the sample of U.S. manufacturing firms from 1980 2006 which are in identifiable relationships. The sample is divided into groups based on the degree of supplier dependence. In Panel A firms are divided according to the strength of the relationship defined by the percent of sales derived from the firms largest customer. In Panel B firms are divided ac cording to the length of the relationship where firms that are in the first 4 years of their relationships are in the first columns and firms in their fifth or greater year of relationship are in the latter columns. The null hypothesis tested is that the there is no difference between the means of the different groups. The t value corresponding to the difference of means is shown. P value states the Satterthwaite value for the probability that the difference between the two groups (where the value used for each group is the average of the difference between the firm and its match) is statistically significant. Revenue from the buyer is defined as the amount of sales from the largest buyer divided by net sales. Years in relationship is the total number of years prior to the reporting year that t he supplier lists the buyer as a mater ial customer. Firm age is the number of years that the firm is listed in Compustat beginning in 1980. For definitions of the rema ining variables see Table 21.

PAGE 56

56 Table 2 6. Investment -Cash Flow S ensitivity Cash Flow 0.076 *** 0.067 *** 0.115 *** (0.000) (0.000) (0.000) Cash Flow Firm in Relat Dum -0.006 0.012 -0.026 ** (0.366) (0.184) (0.017) Q 0.014 *** 0.013 *** 0.017 *** (0.000) (0.000) (0.000) Intercept 0.012 *** 0.013 *** 0.010 *** (0.000) (0.000) (0.008) Number of Obs. 18489 12824 5665 Number of Groups 2850 1925 925 R20.19 0.21 0.17 All Firms Durable Non-Dur. Table 23 reports panel regressions in which the dependent variable is investment divided by beginning of the period total assets. All models include unreported year dummies and firm fixed effects. The sample includes only U.S. manufacturing firms from 1980 2006 which are in identifiable relationships and their matched peers. Cash flow is scaled by beginning of period total assets and interacted with a dummy which takes the value of one when the firm is in the given sample. P values are shown in parentheses. Q is a beginning of period value. The symbols ***, **, and indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

PAGE 57

57 Table 2 7. Investment -Cash Flow Sensitivity of Committed Firms Cash Flow 0.076 *** 0.067 *** 0.115 *** (0.000) (0.000) (0.000) Cash Flow Top 20 Buyer 0.007 0.011 -0.013 (0.358) (0.240) (0.456) Cash Flow Buyer not in Top 20 0.005 0.013 -0.039 ** (0.498) (0.113) (0.034) Q 0.014 *** 0.013 *** 0.017 *** (0.000) (0.000) (0.000) Intercept 0.012 *** 0.013 *** 0.010 *** (0.000) (0.000) (0.009) Number of Obs. 18489 12824 5665 Number of Groups 2850 1925 925 R20.19 0.21 0.17 All Firms Durable Non-Dur. Table 24 reports panel regressions in which the dependent variable is investment di vided by beginning of the period total assets. All models include unreported year dummies and firm fixed effects. The sample includes only U.S. manufacturing firms from 1980 2006 which are in identifiable relationships and their matched peers. Cash fl ow is scaled by beginning of period total assets and interacted with a dummy which takes the value of one when the firm is in the given sample. P values are shown in parentheses. Q is a beginning of period value. The symbols ***, **, and indicate stat istical significance at the 1%, 5%, and 10% levels, respectively.

PAGE 58

58 Table 2 8 Investment -Cash Flow Sensitivity by Buyers Stake in the Supplier Cash Flow 0.075 *** 0.067 *** 0.115 *** 0.076 *** 0.067 *** 0.114 *** 0.065 *** 0.068 *** 0.098 *** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Cash Flow Buyer accounts for 10% 25% of Sales 0.004 -0.008 -0.021 (0.606) (0.333) (0.195) Cash Flow Buyer accounts for 25% 50% of Sales 0.013 -0.023 ** -0.038 (0.149) (0.023) (0.077) Cash Flow Buyer accounts for > 50% of Sales 0.007 -0.028 ** -0.046 ** (0.586) (0.022) (0.024) Cash Flow Relationship is 1 4 years old 0.009 0.018 ** -0.034 ** -0.015 ** -0.013 -0.024 (0.200) (0.020) (0.041) (0.040) (0.082) (0.173) Cash Flow Relationship > 4 years old -0.001 -0.015 -0.011 0.005 -0.023 *** 0.007 (0.875) (0.062) (0.561) (0.543) (0.004) (0.72) Q 0.014 *** 0.013 *** 0.017 *** 0.014 *** 0.013 *** 0.017 *** 0.013 *** 0.012 *** 0.016 *** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Intercept 0.012 *** 0.013 *** 0.010 *** 0.013 *** 0.013 *** 0.009 *** 0.014 *** 0.015 *** 0.011 *** (0.000) (0.000) (0.008) (0.000) (0.000) (0.015) (0.000) (0.000) (0.003) Number of Obs. 18489 12824 5665 18489 12824 566516818 11761 5057 Number of Groups 2850 1925 925 2850 1925 9252679 1823 856 R20.19 0.21 0.170.19 0.21 0.170.20 0.21 0.17 All Firms Durable Non-Dur. 1984 2006 1980 2006 1980 2006 All Firms Durable Non-Dur. All Firms Durable Non-Dur. Table 25 reports panel regressions in which the dependent variable is investment divided by beginning of the period total assets. All models include unreported year dummies and firm fixed effects. The sample includes only U.S. manufacturing firms which are in identifiable relationshi ps and their matched peers. The first 6 columns include observations from 1980 2006 while the last 3 columns only include observations from 1984 2006. Cash flow is scaled by beginning of period total assets and interacted with a dummy which takes the value of one when the firm is in the given sample. The sampl e is divided into groups based on the degree of supplier dependence. Strength groups are determined by the percent of sales derived from the firms largest cus tomer. Suppliers in Group 1 sell less than 25%, Group 2 sells 25 50%, and group 3 sells more than 50% of t heir goods to their main buyer. Length groups are determined by the length of the relationship where firms in the first 4 years of their relationships are in the first group and firms in their 5 or greater y ear of relationship are in group 2. P values ar e shown in parentheses. Q is a beginning of period value. The symbols ***, **, and indicate statistical significance at t he 1%, 5%, and 10% levels, respectively.

PAGE 59

59 Table 2 9. Differences in Investment -Cash Flow Sensitivity during Recessions Cash Flow 0.077 *** 0.072 *** 0.095 *** (0.000) (0.000) (0.000) Cash Flow Recession 0.015 0.020 ** 0.030 (0.053) (0.027) (0.098) 0.094 *** 0.083 *** 0.128 *** (0.000) (0.000) (0.000) 2 0.080 *** 0.078 *** 0.085 *** (0.000) (0.000) (0.000) 3 0.090 *** 0.096 *** 0.128 *** (0.000) (0.000) (0.000) 4 0.073 *** 0.062 *** 0.113 *** (0.000) (0.000) (0.000) Q 0.014 *** 0.013 *** 0.017 *** 0.014 *** 0.013 *** 0.017 *** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Intercept 0.013 *** 0.013 *** 0.010 *** 0.013 *** 0.013 *** 0.010 *** (0.000) (0.000) (0.006) (0.000) (0.000) (0.007) P-Value for from the F-Test H012 0.163 0.659 0.057 H01 3 0.791 0.089 0.997 H0240.333 0.007 *** 0.085 H0340.119 0.019 ** 0.051 Number of Obs. 18489 12824 5665 18489 12824 5665 Number of Groups 2850 1925 925 2850 1925 925 R20.19 0.21 0.17 0.19 0.21 0.17 Durable Non-Dur. All Firms All Firms Durable Non-Dur. Table 26 reports panel regressions in which the dependent variable is investment divided by beginning of the period total assets. All models include unreported year dummies and firm fixed effects. The sample includes only U.S. manufacturing firms from 1980 2006 which are in identifiable relationships and their matched peers. Cash flow is also scaled by beginning of period total assets and interacted with a dummy which takes the value of one when the firm is in the given sample. P values are shown in parenthese s. Q is a beginning of period value. The symbols ***, **, and indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

PAGE 60

60 Table 2 10. Investment Cash Flow Sensitivity over Macro-Economic Cycles Cash Flow 0.076 *** 0.074 *** 0.081 *** (0.000) (0.000) (0.000) Cash Flow Wide Spread 0.031 *** 0.022 *** 0.106 *** (0.005) (0.01) (0.000) 0.106 *** 0.079 *** 0.182 *** (0.000) (0.000) (0.000) 2 0.076 *** 0.077 *** 0.071 *** (0.000) (0.000) (0.000) 3 0.082 *** 0.082 *** 0.198 *** (0.000) (0.000) (0.000) 4 0.075 *** 0.066 *** 0.099 *** (0.000) (0.000) (0.000) Q 0.014 *** 0.013 *** 0.017 *** 0.014 *** 0.013 *** 0.016 *** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Intercept 0.013 *** 0.013 *** 0.012 *** 0.013 *** 0.013 *** 0.012 *** (0.000) (0.000) (0.002) (0.000) (0.000) (0.002) P-Value for from the F-Test H012 0.001 *** 0.291 0.000 *** H01 3 0.048 ** 0.013 ** 0.547 H024 0.813 0.038 ** 0.097 H034 0.518 0.015 ** 0.000 *** Number of Obs. 18489 12824 5665 18489 12824 5665 Number of Groups 2850 1925 925 2850 1925 925 R20.19 0.21 0.18 0.19 0.21 0.18 All Firms Durable Non-Dur. All Firms Durable Non-Dur. Table 27 reports panel regressions in which the dependent variable is investment divided by beginning of the period total assets. All models include unrepor ted year dummies and firm fixed effects. The sample includes only U.S. manufacturing firms from 1980 2006 which are in identifia ble relationships and their matched peers. Cash flow is also scaled by beginning of period total assets and interacted with a dummy which takes the value of one when the firm is in the given sample. P values are shown in parentheses. Q is a beginning of period value. The symbols ***, **, and indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

PAGE 61

61 CHAPTER 3 CASH HOLDINGS AND THE CHARACTERISTICS OF SUPPLIERS Since Keynes (1936) proposed the transaction, precaution, and speculat ion motivations for holding cash, financial economists have sought to more fully understand the motivations of a firms cash holdings. Research on this topic has provided important insight into the effect of agency issues, information asymmetries, bankrup tcy costs, taxes, and other financial frictions on corporate cash holdings.1 I contribute to the cash holding literature by investigating how inter firm relationships affect the suppliers motivation for holding cash. Being in a relationship with a buyer may affect a suppliers motivation for holding cash in a number of distinct ways. First, building on Keynes original proposition, suppliers in relationships may hold additional cash as a precautionary measure. If one customer comprises a large portion o f the suppliers sales, then the loss of that one customer will result in a large adverse cash flow shock. That is, the sale of a large portion of goods to one particular customer induces operating risk. The greater the risk, the more cash firms will hol d as a precaution. In contrast to this theory, as the relationship becomes more important to the buyer and supplier, it becomes less likely that relationship will end abruptly. This may reduce operating risk encouraging the suppliers to hold less precauti onary cash. Prior work in this area is restricted to the belief that firm boundaries are determined strictly through explicit contracts. However, buyer -supplier relationships rely heavily on implicit contracting and over time the buyer and supplier may become stakeholders in one another. Recognizing that the firm may be comprised of both implicit and explicit contracts provides an 1Extant research finds that a a variety of factors influence firm cash holdings including the following; the e ffect of agency issues and information asymmetries Jensen ( 1986 ), Mikkleson and Partch (2003) and Harford, Mansi, Maxwell ( 2009) bankruptcy costs Opler Pinkowitz, Stulz, Williamson (1999) Bates, Kahle, Stulz (2009), Almeida, Campello, Weisbach (2004), Acharaya, Almeida, Campello (2007) taxes Foley, Hartzell, Titman and Twite (2007), r elative product market position Haushalter, Klasa, and Maxwell (2007) and relative strength of labor unions Klasa, Maxwell and Ortiz Molina (2009).

PAGE 62

62 additional motivation for cash holdings. I hypothesize that as a commitment to their customers, suppliers will maintain higher cash holdings because a supplier can use cash holdings to induce buyers to make specialized investments in relationship specific assets. This hypothesis is based on the work of Titman (1984) which suggests that as part of their relationship a supplier may require its customers to undertake investments that lose value if the supplier encounters financial distress. A suppliers higher cash holdings provide a valuable signal to the buyer that the supplier is and will remain financially stable. The empir ical implication of this hypothesis is that suppliers which expect their buyers to engage in specialized investments will hold higher levels of cash. In order to test these hypotheses, using Compustat data, I construct a buyer -supplier relationship datas et. The Statement of Financial Accounting Standards No. 14 (which was later supplanted by SFAS No. 131) requires that firms report information for segments that represent 10% or more of consolidated sales. This includes disclosure of sales to principal c ustomers, if the revenue generated from sales to a particular customer exceeds 10% of revenue of the firm, or if the firm considers the sales important to its business. Based on this dataset, I investigate when and how buyer -supplier relationships affe ct supplier cash holdings. I find that even after controlling for profitability, firm size, leverage, the ease with which firms can access capital markets, growth opportunities, the availability of natural hedges, and a number of other control variables, firms which are in relationships hold more cash on average than firms not in relationships. Further, I find that as the importance of the relationship increases, measured by the percent of sales and sales concentrations to the reported customers, so does a suppliers cash holdings. These results are consistent with both the commitment and precautionary motivations for holding cash.

PAGE 63

63 Further, I investigate how differences in buyer and supplier characteristics affect supplier cash holdings. According to T itman (1984) and Titman and Wessels (1988) manufacturers of unique goods depend more on relationship specific assets than manufacturers of nonunique goods. Therefore, buyers of unique goods have a greater stake in the financial welfare of their suppliers If the primary motivation for cash holdings is as a commitment device, then unique goods manufacturers in relationships should hold more cash than nonunique goods manufacturers. Using the suppliers industry and R&D spending as proxies for the uniquene ss of the goods sold, I find that supplying unique goods mitigates the relation between customer importance and cash holdings. This evidence does not support the commitment hypothesis. However, the evidence is consistent with the precautionary motivation. According to this hypothesis, greater switching costs make it less likely that a buyer will end the relationship abruptly. As a result, supplying unique goods lessens the operating risk of being in a relationship. Next, I study the influence of market position on the association between the importance of the relationship and cash holdings. If the supplier produces an output for which there are few alternatives, then the buyer will require a greater commitment from the supplier. In contrast with few al ternatives, the buyer will be less likely to end the relationship abruptly. This decreases the risk to the supplier and reduces the need to hold cash as a precaution. Using the industryadjusted price cost margin and the industry Herfindahl Hirschman Index as proxies for market position, I find that the association between the importance of the customers and cash holdings is more important for suppliers in which there are few alternative suppliers This is consistent with the commitment motivation for cash holdings. Finally, certain buyers may help to provide stability. Government agencies are not likely to go out of business or to become unable to make payments. So, being in a relationship with a

PAGE 64

64 government agency does not induce the same operating risk as being in a relationship with a private firm. As a result, if suppliers hold cash as a precaution against the risk of an adverse cash flow shock then firms in relationships with non-government firms should hold more cash then firms in relationships with government agencies (such as the military). This is indeed what the evidence shows. Overall, the evidence in this paper is consistent with both the precautionary and commitment motive s for holding cash. This study contributes to the literature on the determinants of corporate cash holdings. In particular, I provide new evidence that buyer -supplier relationships effect both the motivation for and the amount of cash holding. Prior work suggests alternative motivations for cash holdings. Additionally, t his study contributes to the small but growing body of literature that examines how corporate financial policy decisions are affected by strategic interactions between buyers and suppliers. Understanding the link between corporate cash holdings and a firm relationship is important since the ability of a firm to compete in the market depends on its relationships with its suppliers and customers. This paper proceeds as follows. Following a brief review of the literature, Section 2 discusses three possible hypothesis of how and why inter -firm relationships affect corporate cash holdings. Section 3 describes the data set used in my study. Section 4 presents evidence used to distinguish between the hypotheses. Section 5 concludes. Related Literature a nd Testable Hypotheses Cash Holding Holding cash provides firms with benefits and costs. Among the benefits, corporate cash holdings may serve as a risk management tool (Bates, Kahle, and Stulz 2009). Indeed, Acharaya, Almeida, and Campello (2007) provide evidence that cash holdings are distinct from maintaining low leverage and benefit companies by serving as a risk management tool. Cash

PAGE 65

65 used as a risk management tool can be especially valuable to firms with high external financing costs and large growth opportunities by reducing under investment problems (Opler, Pinkowitz, Stulz, and Williamson 1999 and Mikkleson and Parch 2003). Used in this way, cash holdings provide benefits similar to the use of derivatives (as described by Froot, Scharfstein and St ein 1993). Further, Haushalter, Klasa and Maxwell (2007) find that holding cash enables firms to successfully compete in product markets by allowing firms to fully invest in growth opportunities. On the other hand, holding cash may be disadvantageous. Holding cash may result in reductions in valuable investments. In firms with agency problems, cash holdings provide managers with the opportunity to invest in value -decreasing projects (Jensen 1986). Consistent with this hypothesis, Dittmar and Mahrt -S mith (2007) and Harford, Mansi, and Maxwell (2009) provide evidence suggesting that entrenched managers are more likely to build excess cash balances, but then spend this excess cash quickly. Buyer Supplier R elationships If the firm is merely a nexus of explicit contracts, each piece can be evaluated independently and the value of the firm is simply the sum of its parts. According to this theory of the firm, inter -firm relationships are merely contracts that can be replicated at market determined prices As a result, relationships between firms should not affect firm value, financing policy nor investment policy. But, if the firm is considered to be a nexus of both implicit and explicit contracts (e.g., Baker, Gibbons, and Murphy, 2002) then each fi rm is unique and its value becomes greater than the sum of components which are readily available in the market place. Inter -firm relationships, which rely heavily on implicit contracting (Macaulay, 1963), therefore may affect firm functioning and the va lue of the firm (Zingales, 2000).

PAGE 66

66 Suppliers which sell more than 10% of their goods or services to one particular buyer are required by the FASB to report both the level of sales and the name of the buyer. For the remainder of the paper, I will refer to suppliers which sell more than 10% of their goods or services to one or more buyers as being in a relationship with the buyer. These relationships affect financing policy and investment policy. To date, researchers have examined the effect of inter -firm relationships on a number of factors that influence firms form and functioning, including sales, profit margin, transaction costs (Kalwani and Narayandas, 1995), distress costs (Hertzel, Li, Officer, and Rodgers, 2008), information transmission costs (Gomes -Casseres, Hagedoorn, and Jaffe, 2008), bargaining power (Fee and Thomas, 2004 and Brown, Fee, and Thomas, 2009) and capital structure (Banerjee, Dasgupta, and Kim 2008, and Kale and Shahrur 2007). I hypothesize that relationships affect the supp liers cash holdings in a number of distinct ways. The Effect o f Buyer -Supplier Relationships o n Cash Holdings Relationships between buyer and supplier sometimes necessitate specialized investments. For these suppliers, the ability to induce buyers t o undertake investments in these specialized investments is necessary for successful firm operation. Further, these specialized investments have lower value outside of the particular business relationship because the assets are specific to the particular c ontext. As a result, these relationship specific assets lose value if the supplier enters into financial distress or liquidation. Take for example the case of the supplier, Ducommun Incorporated, and their primary buyer, Boeing. Ducommun Incorporated designs, engineers, and manufactures complex

PAGE 67

67 aerostructure components.2 An important feature of this relationship is that completion of Boeings 777 relies on delivery of Ducommuns components. Financial distress may render Ducommun unable to deliver its product in a timely fashion. Without Ducommuns parts, Boeing cannot deliver the plane. Each additional day that the plane is not delivered increases Boeings labor, storage and financing costs. Value loss results from the fact that Boeing has made spec ialized investments. Further, it is more costly to redesign the 787 based on inputs from another supplier.3 For this reason, one should expect buyers to limit their relationships with suppliers that are expected to face future financial difficulties. O ne way for suppliers to avoid buyer underinvestment is to signal the firms long term financial viability to its customers. In other words, the supplier can commit to a financial policy that takes into consideration the effect of financial distress on the buyer by choosing to hold high levels of cash. Suppliers can maintain financial viability by holding a large cash reserve as a hedge against financial instability. If a supplier is selling to a large number of customers then the product is not likely to require specialized investments in relationship specific assets. Therefore, if commitment to their buyer is a suppliers main motivation for holding cash, then suppliers in relationships should hold more cash than suppliers with many customers. Titman (1984) and Titman and Wessels (1988) argue that buyers face switching costs when their supplier is liquidated. Further, these costs are greater when the suppliers products are unique. Therefore, buyers of unique goods have a greater stake in the financi al welfare of their 2In the SEC 10K report dated January 31, 2006 the management of Ducummon stated, The Company Is Dependen t on Boeing Commercial Aircraft. The Companys sales for Boeing commercial aircraft and the C 17 aircraft are principally for n ew aircraft production; and the Companys sales for the Apache helicopter are principally for replacement rotor blades. Any s ignificant change in production rates for these programs would have a material effect on the Companys results of operations and cash flows. In addition, there is no guarantee that the Companys current significant customers will continue to buy products f rom the Company at current levels. The loss of a key customer could have a material adverse effect on the Company. 3 Recently, Boeing acknowledged that a slowdown of less technologically advanced parts such as seats and customized gallies was slowing dow n production was taking a toll on its bottom line (Michaels and Lunsford 2008).

PAGE 68

68 suppliers and require a greater commitment from the supplier. If commitment to the buyer is a suppliers motivation for holding cash then suppliers of unique goods should hold more cash than suppliers of nonunique goods. Both Kale and Shahrur (2007) and Banerjee, Dasgupta, and Kim (2008) provide evidence consistent with the disparity between unique and non unique goods manufacturers. They show that for firms which produce unique goods, the strength of the buyer -supplier relationsh ip is inversely proportional to the suppliers leverage ratio. Further, the effect of a relationship on leverage is weaker for manufacturers of non unique goods. As pointed out by Opler, Pinkowitz, Stulz and Willliamson (1999), most of the variables tha t are associated with higher cash levels are related to variables that are also known to be associated with low debt. Therefore, we might expect that in addition to motivating suppliers to maintain lower leverage ratios, commitment s to their buyers also m otivate suppliers to maintain larger cash reserves. While one motivation for a supplier to hold additional cash may be to protect the buyer, another compelling reason for a supplier to hold additional cash is to protect itself. In this case, holding cas h benefits the supplier by reducing the likelihood of incurring financial distress costs if the firms operations do not generate sufficient cash flow. Selling a large portion of goods to one particular buyer induces operating risk due to the potential for highly variable future cash flows. By depending on one or more customers for a significant part of the companys sales, suppliers can potentially lose a large portion of sales at once for unexpected and idiosyncratic reasons which could cripple their fi nancial health. This is not the case for firms which do not rely on one significant customer but instead have a diversified range of customers. For suppliers that sell their goods to many buyers, the loss of one customer for idiosyncratic reasons does not impact the operating performance of the firm.

PAGE 69

69 The more goods that the supplier sells to a principal customer, the more that supplier loses if the buyer ends the relationship. Indeed, SEC 10-K reports include statements detailing the additional opera tional risk partnerships pose. For example, in their June 28, 2003 annual report, Salton Inc. reported, If we were to lose one or more of our major customers, our financial results would suffer. The statement continues: We do not have long -term agree ments with our major customers, and purchases are generally made through the use of individual purchase orders. A significant reduction in purchases by any of these major customers could have a material adverse effect on our business, financial con dition and results of operations. As a result, firms which rely on one or a few customers for a large proportion of their sales may have a higher expected probability of financial distress than firm with a large diversified sales base. Further, a redu ction in sales may not be the customers choice but the result of financial distress or bankruptcy of the customer. In particular, Hertzel, Li, Officer and Rodgers (2008) study bankruptcy contagion and find that buyer bankruptcy filings significantly affect the suppliers which are in relationships with the filing firms. Firms in relationships have higher expected financial distress costs because they have a higher risk of adverse cash flow shocks and as a result a higher risk of distress. Opler, Pink owitz, Stulz, Williamson (1999) argue that holding higher levels of cash is one way in which firms hedge against adverse cash flow shocks. Thus firms in relationships may hold higher amounts of cash to protect themselves against the risk of financial dist ress. Further, if precaution against financial distress is the primary motivation for firms to hold cash, then both the strength and concentration of the relationship should be proportionately related to the suppliers cash holdings. While the commitment motivation for holding larger cash reserves is primarily to entice the buyer to invest in relationship specific assets, the precautionary motive for holding cash is to protect the firm itself from financial distress.

PAGE 70

70 The theory that firms in relationsh ips hold more cash for precautionary reasons is based on the assumption that being in a relationship increases operating risk and as a result raises the cost of financial distress. However, not all relationships increase the suppliers operating risk. Onc e in a relationship, the buyer may not have the desire or the ability to end the relationship. A supplier which supplies goods for which there are few alternatives leaves the buyer with no other options. Additionally, a supplier which has made relationship specific investments may generate switching cost large enough that it becomes prohibitively expensive for the buyer to change suppliers. In both of these cases, a large buyer may provide the supplier with a stable sales base mitigating some of the supp liers operating risk. As a result, supplier will need to hold less cash as a hedge against future uncertainty. In addition, certain buyers may help to provide stability for the suppliers. Many American suppliers sell their goods to the U.S. government The government is likely to represent a more stable source of demand than other principal customers due to its high creditworthiness. Therefore, a supplier which sells a large portion of its goods directly to the government should not suffer from highe r expected risk of financial distress relative to a supplier which sells its goods to non-government companies. As a result, if the precautionary motive is the primary reason for firms to hold cash then firms which sell to non -government companies should hold more cash than suppliers in relationships with the government. Data Sources The Statement of Financial Accounting Standards No. 14 Financial Reporting for Segments of Business Enterprise (SFAS No. 14) of the Financial Accounting Standards Board (FASB) requires that firms report information for segments that represent 10% or more of consolidated sales, for fiscal years ending after 1977. This includes disclosure of sales to

PAGE 71

71 principal customers, if the revenue generated from sales to a particular customer exc eeds 10% of revenue of the firm Prior to 1997, in addition to the 10% cutoff, suppliers had to report buyers if the sales to the buyer were considered important for business operations and suppliers were required to report the customers nam e In 1997, the FASB issued SFAS 131, revising SFAS no. 14, which permitted firms to optionally report the customers name. To maintain consistency throughout time, firms which report customers which account for less than 10% of sales are omitted from th e relationship sample. Robustness tests (reported later) confirm that the policy change does not affect the results. Customers and revenue from each customer is collected in the Compustat Segments Data. My sample includes all U.S. manufacturing firms (pr imary SIC code within 2000 3990) included in Compustat with non-missing values of sales and total assets in Compustat between 1979 and 2006.4 For each firm customer reported, I identify whether the customer is listed in Compustat and assign it the cor responding firm identifier. Identification of firm customers requires individual verification as the names are not entered in the Segment data in a uniform way.5 I use a string matching algorithm to generate a list of potential matches to the customers name and then hand -match customers by inspecting the firms name segment and industry information.6 I am deliberately conservative in assigning customer names and firm identifiers in order to ensure that the customers are matched to the appropriate financial information. After using the matching algorithm, I sort the dataset by firm name and year to confirm continuity over time. In several cases it is clear from inspection that the buyer -supplier relationship has remained 4 Total assets are converted to 1984 dollars using the consumer price index. 5 For example, throughout Shiloh Industries 14 year relationship with General Motors Compustat rep orted their principal customer as GEN MTR, GEN MOTORS, and General Motors Corp. 6 The SAS function Spedis is used to generate spelling distances.

PAGE 72

72 constant over time although the matching algorithm missed a section of years.7 I correct these errors by hand. The sample includes 28 years (19792006) and covers a number of macroeconomic cycles. To account for changes over time, I supplement the Compustat data with additional data fr om three sources. From the Bureau of Labor Statistics, I use the Consumer Price Index to adjust firm assets for inflation to 1984 dollars. Before analyzing the data, firms in the 3 digit SIC Code (283) for drug manufacturers are omitted from the sample. In the context of product market relationships, there is a continuum of organizational forms that range from distinct firms conducting arms -length transactions to complete vertical integration. Fee, Hadlock, and Thomas (2006) investigate partial equity ownership which represents one form of this type of relationship. Anecdotal evidence suggests that equity stakes are not coincidental in trading relationships, but rather they serve a role in maintaining such a relationship. They find that equity stakes are more common when the supplier is an R&D intensive firm and when the companies have formal alliance agreements and conclude that the equity stake frequently serves two purposes. First, the equity stake helps to align incentives and to help provide con tractual completeness. Second, many stakes represent newly issued shares indicating that it could serve to solve a financing constraint. If the relationship is fully described and contracted, these firms should be omitted from the sample to avoid biasing the results. To determine which firms should be omitted from the sample, I choose a random sub-sample of 250 firms. For each of these firms, I downloaded their SEC 10 -K reports from EDGAR and searched through the report for the name of their main trading partner to determine the nature of their relationship. Of the 250 firms, I identified 7 For example, in one year the matching algorithm had no problem matching the buyer PENNEY (J.C.) COR P. INC.to an existing Compustat firm. The following year for the same supplier the matching algorithm was unable to match the buyer JCPenneys to an existing Compustat firm

PAGE 73

73 explicit block equity ownership or partial firm ownership as the result of a joint venture in 16 cases or 6.4% of the sub-sample. This is consistent with Fee, Hadlock, and Thomas (2006) who find that equity stakes exist in 3.31% of all buyer -supplier relationships. Of the 16 cases found 25% fall under the SIC sub-category 283 which represents the creation and manufacture of drugs. In no other 3 digit SIC code are there more than 2 cases. In general, these firms are highly R&D intensive and require large amounts of capital up front. Additionally, of the 40 firms in this industry more than 75% had explicitly outlined and contracted strategic alliances. While the ma jority of manufacturers operate based on purchase orders or cancelable contracts, biotechnology firms are dependent on large upfront investments and milestone payments from their customers. Firms in the SIC sub -category 283 are omitted from the sample. This is further justified by the stream of literature devoted solely to characterizing and understanding the firms in the biotechnology field (i.e. Lerner, Shane, and Tsai 2003). Measure of Cash Holdings Dependent v ariable I examine whether cash holdi ngs can be explained by firm relationships. The primary measure of cash holdings used is the ratio of cash and marketable securities to total assets minus cash and marketable securities, based on Opler, Pinkowits Stulz and Williamson (1999). I deflate li quid asset holdings by the book value of total assets, net of liquid assets, under the assumption that a firms ability to generate future profits is a function of its assets in place. Because this creates some extreme outliers for firms holding most of t heir assets in cash, I use the logarithm of one plus the ratio of cash to net assets as the dependent variable.8 The literature employs several alternative definitions of the cash ratio Although not tabulated, I reproduce all 8 This method is also employed by Foley et al. (2007) and Bates, Kahle, and Stulz (2009)

PAGE 74

74 of the regressions using ea ch of the following as the dependent variable: c ash to assets, cash to net assets, log of cash to net assets, cash minus debt to net assets, and cash to sales Despite generating extreme outliers, t he qualitative results are not altered by the alternative specifications in a material way. Independent v ariables of i nterest The primary independent variables of interest are measures of firm relationships. I create an indicator for whether the firm reports being in a relationship during at least one year.9 Then, I create the two additional continuous measures of the importance of the relationships which are conditional on reporting a relationship. The next relationship measure, percent of sales to major customers, is the sum of a firms sales to its major cu stomers divided by total sales. The final relationship measure, major customer concentration, is the concentration of a firms sales to the major customers it reports. As a measure of concentration, I use a Herfindahl Hirschman Index determined as the sum of the squares of the share of sales to each reported customers where the shares are expressed as a percentage of total sales. The result is proportional to the average share of sales, weighted by share. As such, it can range from 0 to 1, where zero rep resents no reported customers and one represents exactly one customer which purchases 100% of the suppliers goods. One benefit of this index is that it gives more weight to buyers which purchase a larger share of the suppliers goods. Control v ariables The goal of this study is to determine if and when firm relationships alter a firms cash holdings independent of known firm characteristics. A number of other factors are known to influence firm cash holdings. So as control variables, I include the firm characteristics suggested 9 For consistency, for every relationship measure, even if they are reported I omit reported relationships that account for less than 10% of the firms total sales.

PAGE 75

75 by Opler, Pinkowitz, Stulz and Williamson (1999) and Bates, Kahle, and Stulz (2009). These variables control for profitability, size, leverage and other firm characteristics. I follow their methodology for variable construction including winsorizing outliers in firm -year observations. Empirical Analysis Relation between Buyer -Supplier Relationships a nd Cash Holdings The first question I investigate is whether the presence and importance of a buyer supplier relationship affect firm cash holdings. I measure the importance of the relationship using the percent of sales to major customer and customer concentration, the two proxies described previously. Changes in cash holding by year Table 3 1 displays univariate statistics of the primary variables of interest by year. The first column reports the percent of manufacturers which report being in a relationship. T he number of firms which report relationships increases consistently over time. In other words, the prevalence of rel ationships in this sample increases over time. Further, the results confirm that the regulation change in 1997 does not alter the amount of firms reporting customers. The second column reports the percent of sales to all major customers. Not only does t he number of firms which report major customers increase, but the percent of sales to major customers increases as well. So, on average, for each firm reporting the presence of a relationship, the importance of that relationship to the operating performan ce of the supplier increases as well. In the third column, I report major customer concentration. Though it is not linear, customer concentration has increased over time as well. Taken together, the evidence suggests that relationships have increased in prevalence and importance over time. The right panel of Table 3 1 displays firm cash holdings over time. Consistent with Bates, Kahle, Stulz (2009), cash holding as a percent of total sales for all firms increases just

PAGE 76

76 about every year. The next two c olumns divide the sample according to whether the firm has reported a relationship in that year. These columns provide support for the proposition that firms in relationships hold more cash than firms which are not in relationships. For each year (except 1988), the difference between the cash to assets ratio of firms in relationships and firms not reporting relationships is statistically significant. Importantly, this difference is economically significant. At an aggregate level, firms in relationships hold approximately 30% more cash as a proportion to total assets than firms which are not in a major relationship. This is consistent with both the precautionary and commitment motivations for cash holding. Additionally, the difference between the cash to assets ratio of firms in relationships and firms not in relationships widens over time as well. In a regression of the average cash to assets ratio on time, time has a significantly positive coefficient. Regressing average cash to assets on time separ ately for each of the two subgroups, firms with customers and firms without customer, implies that the average cash to assets ratio has increased by .27% and .44%, respectively. Therefore, firms with relationships account for more of the overall change in cash holding over time. As a result, the puzzle of the increase in overall cash holdings over time reported by Bates, Kahle, and Stulz (2009) may be partially explained by firm relationships. Descriptive findings Table 3 2 presents a summary of the diff erences in firm characteristics between firm years in which a relationship is reported and firm -years in which firms do not report being in a relationship. The difference between the cash holdings of firms in relationships relative to firms which are not in a relationship is both economically and statistically significant. However, a number of other firm characteristics differ between the two groups of firms as well. Firms which report customers tend to be smaller, younger firms.

PAGE 77

77 Additionally, there is a significant difference in the coefficient of variation of the cash flows of firms which report customers and firms which do not. The coefficient of variation of the firms cash flows is calculated as the standard deviation of a firms cash flows divid e d by the firms mean cash flow using as many years as are available for each firm over the prior 20 year period.10 It indicates which firms experience larger idiosyncratic risk. Moreover, I compare the coefficient of variation of cash flow for firms whic h have been in a relationship in any prior year to firms which have never previously been in a relationship. After adjusting each firms coefficient of variation by the industry mean, I find an even larger and statistically significant difference between t he industry adjusted idiosyncratic risk of firms previously in relationships compared to those which have never previously been in a relationship. This piece of evidence is necessary (but not sufficient) for precautionary motivation to be a reasonable exp lanation of cash holdings because one of the underlying assumptions of the precautionary motivation is that being in a relationship induces operating risk. Table 3 3 presents a summary of firm characteristics. For each firm characteristic, the mean, me dian, and standard deviation are shown. I break the sample up into subsamples based on a measure of relationship strength. Specifically, the sample of firms which report relationships is divided into quartiles based on the percentage of sales to the majo r customers. This table provides evidence that firms in relationships hold more cash than firms which are not in relationships. Further, as the strength of the relationship increases, suppliers hold proportionately more cash. This evidence suggests that firms may hold cash as either a precaution against financial distress or as a commitment to their buyers. 10 To compute this statistic I utilize on Compustat data dating back to 1959.

PAGE 78

78 When the sample is broken up into subsamples based on the percentage of sales to the major customers, I find that in addition to differences in cas h holdings, firm relationships affect a number of other firm characteristics. Firms which do not report a relationship are larger than firms which report a relationship. This is in part an artifact of the reporting process which biases the sample of supp liers towards the smaller firms in Compustat. Consistent with Kale and Shahrur (2007) and Banerjee, Dasgupta, and Kim (2008), firm leverage, measured as either the market leverage ratio or the long term book debt ratio, is inversely correlated with the pe rcent of sales to the major buyer. Overall, the univariate findings support the hypothesis that firms in relationships strategically maintain higher cash reserves. However, the differences in firm characteristics across the group of firms indicate that tests which control for other firm effects will be far more powerful than univariate tests. The next section reports the results of such tests. Multivariate evidence The univariate tests provide a description of the data set. In addition they provide preliminary evidence that firms which are in relationships hold more cash than firms which are not. Additionally, as the strength of the relationship increases so does the amount of cash held. This initial evidence is consistent with both the precautiona ry and commitment motivations for holding cash. Next, in a multivariate setting, I examine the effect of firm relationships on corporate cash holdings. Both the precautionary and commitment motivations for holding cash imply that suppliers in relationships should hold more cash then suppliers which are not in relationships. Further, if a firms primary motivation to hold cash is as a precaution against risk, then cash holdings should increase as relationships induce greater operating risk. As sales to one (or a

PAGE 79

79 few) particular customer(s) increases or the concentration of sales to customers increases, so should cash holdings. Table 3 4 presents regressions predicting the cash holdings between 1979 and 2006 using the independent variables described ea rlier. The natural logarithm of one plus the ratio of cash to net assets is the measure of liquidity used as the dependent variable. In each panel, the first column reports time -series cross-sectional OLS regressions which include year dummies. The stan dard errors are clustered by firm. The second column of each panel reports estimates based on the method presented in Fama and MacBeth (1972) (which I will refer to as the Fama MacBeth method). Using this approach, for each year, a cross -sectional regress ion is estimated. This method effectively treats each year as an independent cross -section eliminating the problem of serial correlation in the residuals of time -series cross -sectional regressions. Additionally, the Fama -MacBeth method is robust to the c hanges in regulation in 1997. The coefficients and standard errors based on the Fama -MacBeth method are estimated using the Stata code associated with Petersen (2009). The third column reports the results of regression which includes firm fixedeffects. If the coefficients of the relationship measurements occur as the result of an omitted firm specific factor, then including firm fixed effects should mitigate the impact of the relationship variables. In the first set of regressions, the primary varia ble of interest is an indicator variable which equals one if the firm reports being in a relationship during at least one year between 1979 and 2006. The coefficient indicates that firms which have been or are in relationships, all else equal, hold more c ash than firms which are not in relationships and have never been in a relationship. The coefficient of interest in the first column indicates that the difference in not being in a relationship and being in a relationship, all else equal, corresponds to a n increase in

PAGE 80

80 the cash to net assets ratio of 0.029 which is economically significant when considering tha the median cash to net assets ratio is 0.070 In the second set of regressions, the primary variable of interest the percent of sales to all major c ustomers. Not only do firms which are in relationships hold more cash, as the percent of sales to their major customers increases so does the amount of cash that they hold. In the third set of regressions, the primary variable of interest is customer con centration. The regressions indicate that customer concentration and cash holdings are positively correlated. Overall, the evidence in this table is consistent with both the precautionary and commitment motivations for holding cash. The sign and magnitu de of the control variables in Table 3 3 are consistent with the prior work that studies the determinants of cash holdings including Opler, Pinkowitz, Stulz, and Williamson (1999). Further, the control variables are generally statistically significant. T his indicates that the relationship measures broaden the scope of variables known to be related to cash holding. The inclusion of firm fixed effects confirms that this there is not a spurious relation between cash holdings and firm relationships. Finally there are not material differences between the OLS, Fama -MacBeth, and Fixed -Effects models. This indicates that the affect of being in a relationship is robust to model specification. In Table 3 4, all manufacturing firms in Compustat were included. This constrains firms in relationships and firms not in reported relationships to act in the same way. Table 3 5 tests a sub -sample of firms which were present in the previous tests. This sample includes all firms years of suppliers which report sales to a primary customer during at least one year during 1979 2006.11 In each panel, the first column reports the results of cross -sectional time -series regressions which include year dummies. The second column reports the results of the Fama 11 Using this subsample is consistent with the work of Banerjee Dasgupta, and Kim (2009).

PAGE 81

81 MacBeth meth od. The third column reports the results of regression which includes firm fixedeffects. The results in Table 3 5 support the hypothesis that suppliers selling proportionately more to their primary customers and suppliers with more concentrated sales maintain higher cash ratios. Additionally, as the amount of sales and the concentration of sales to major customers increases each supplier increases their cash holdings. The results are very similar to those based on the whole sample of firms. Further, the coefficients of the control variables are similar to that of the whole sample. To determine the economic implications, I multiplied the interquartile range of the percent of sales to the primary customers by the coefficient of percent of sales to th e primary customers in the first column of Table 3 5. Next, taking the exponent of the resulting value indicates that the difference between a firm at the 25th and 75th percentiles of percent of sales to their primary customers shows a change of 0 .0 29 in the cash to net assets ratio. Given that the median cash to net sales ratio for firms in this sample is 0 .085 this change is economically significant. It is not surprising that the coefficients of amount of sales to major customers and major customer concentration are similar. These two variables have a pair wise correlation coefficient of .823 which is highly statistically significant. One of the benefits of customer concentration as a measure of relationship importance to the supplier is that it gi ves more weight to buyers which purchase a larger share of the suppliers goods. Therefore, for the remainder of the paper, I will focus on only the concentration measure. Determinants of the Effect o f Relationships o n Cash Holdings There is strong evidence that, all else equal, firms in relationships hold more cash. However, not all firms and relationships are the same. I now study how the positive relation

PAGE 82

82 between cash holdings and buyer -supplier relationships is affected by firm and relationship cha racteristics. I therefore augment the models provided with additional explanatory variables. The interaction of the additional explanatory variables with customer concentration helps to understand different factors which affect the importance of cash hol dings for firms in relationships. Relationship s pecific a ssets The first characteristic I investigate is the presence of relationship specific assets. The presence of specialized investments in relationship specific assets suggests that switching costs are higher for buyers of unique goods and may influence a suppliers cash holdings differently depending on their primary motivation for holding cash. If a suppliers primary motivation for holding cash is as a commitment to the buyer, then suppliers whic h rely on specialized investments to produce relationship specific assets should have a stronger relation between cash holdings and customer concentration. On the other hand, the presence of relationship specific assets binds the buyer and supplier togeth er and makes it less likely that the buyer will choose to end the relationship. As a result, the precautionary motive predicts a weaker relation between cash and customer concentration for firms with a high degree of relationship specific assets. I use two different proxies for the likelihood of specialized investments resulting in relationship specific assets. First, I categorize firms according to their primary industry. According to Titman (1984) and Titman and Wessels (1988) durable goods manufacturers depend more on relationship specific assets than non-durable goods manufacturers. Second, I use research and development (R&D) scaled by sales as a proxy for asset specificity as is prevalent in the empirical literature on transactions cost economics (see Boerner and Macher 2001 for a review). Further, Allen and Phillips (2000) suggest that R&D intensive industries are more likely to create relationship specific assets.

PAGE 83

83 Table 3 6 presents the primary variables of interest and the additional variab les discussed sorted according to their two -digit SIC code. The sample of manufacturing firms is divided into those producing durable goods (primary SIC codes from 3400 to 3990) and those producing nondurable goods (primary SIC codes from 2000 to 3390). Durable goods manufacturers report being in a relationship more frequently and also hold more cash on average. Further, durable goods manufacturers spend more on research and development than non-durable goods manufacturers. Table 3 7 presents the resu lts of the primary regression model augmented to include the additional explanatory variables and their interaction with the firms customer concentration. The first column includes the interaction of customer concentration with an indicator for whether t he manufacturer produces durable goods. The insignificant coefficient on the interaction term indicates that the relation between cash holding and customer concentration is not significantly different for manufacturers of durable goods than non-durable go ods. This provides a contrast with the results of Banerjee, Dasgupta, and Kim (2008 ) who provide evidence that durable goods manufacturers maintain lower leverage ratios than non -durable goods manufacturers and conclude that the evidence is consistent wit h the commitment motivation. The second column includes the interaction between the ratio of R&D to sales and customer concentration. The negative coefficient on the interaction term indicates that the relation between cash holding and customer concent ration is less pronounced for firms which invest more heavily in research and development. However, the sum of the coefficients indicates that the relation between customer concentration and cash holdings remains positive for suppliers investing in R&D. This is consistent with the theory that suppliers hold cash for precautionary reasons. Further, this provides a contrast with Kale and Shahrur (2007) who provide evidence

PAGE 84

84 that manufacturers which invest in R&D maintain lower leverage ratios than manufactur ers which do not. Taken together the results indicate that producers of unique goods hold less cash than producers of nonunique goods which is consistent with the precautionary motivation for holding cash. The difference between manufacturers that inv est in R&D and those that do not provides additional evidence of the efficacy of the results. For the relation between cash holdings and customer concentration to be spurious, certain firm characteristics would have to simultaneously cause firms to enter into relationships and also cause them to maintain higher cash ratios. Additionally, these same firm characteristics would have to differentially affect these firms Market p osition The next determinant of the affect of buyer -supplier relationships of cash holdings which I study is the effect of a firms market position. Buyers which transact with suppliers with a dominant market position or suppliers in a concentrated industry may have few viable alternatives. If the supplier produces an output for w hich there are few alternative suppliers then the buyer may be even more concerned of the event of supplier financial distress. In addition, in the presence of many suppliers, a buyer may be able to find one who can salvage some of the value of the special ized investments. As a result, if there are few alternative suppliers and the suppliers primary motivation for holding cash is as a commitment to the buyer, then there will be a stronger relationship between cash holdings and sales to the buyer. Alter natively, if the supplier produces a good for which there are few alternatives then risk of a buyer ending a relationship will be less pronounced. In the presence of numerous alternative suppliers, buyers can make a credible threat to withhold future busi ness from the existing supplier (e.g. Holstrom and Roberts 1998). Therefore, if the supplier holds cash as a precaution against risk, then effect of buyer sales on cash holdings will be weaker.

PAGE 85

85 I use two proxies for the suppliers market position. Firs t, I consider the excess price cost margin, based on the Lerner Index (Lerner 1934), as a measure of asset specificity within an industry. The excess price -cost margin determines the relative rent which the supplier can extract. If goods can be easily substituted, then competition will drive this margin down. The excess price-cost margin is defined as the difference between a firms operating profit margin and the average operating profit margin of its industry. This measure is used prevalently as a pro xy for market position in the empirical industrial organization literature because it measures a firms ability to extract rents relative to its peers (Gaspar and Massa 2006). The price -cost margin is industry adjusted because different industries have st ructurally different profit margins for reasons unrelated to market position. Therefore, this measure captures the intra industry difference in market position. The second proxy that I use for market position is the supplier industry concentration measu red by the Herfindahl Hirschman index (HHI) which is frequently used to measure market position in antitrust policy. The HHI is equal to the sum of the squared market shares of firms in the industry. I collect the Herfindahl Hirshman Index (HHI) of indus try concentration from the Census of Manufacturers publication, which is published as part of the US Census in 1982, 1987, 1992, 1997, and 2002. I assume that the HHI is valid for the 5 years prior to the survey.12 Each firm in the sample is assigned the HHI of the industry. Therefore, this measure captures the inter industry differences in market position. The third and fourth columns of Table 3 7 presents the results of model including the interaction between the excess price-cost margin and custome r concentration and the interaction between HHI and customer concentration, respectively. The positive coefficient on the 12 I also use the 2002 survey for the years from 2003 2006. The 2007 HHIs are due to be released this year. As soon as they are published, I will update the dataset

PAGE 86

86 interaction term indicates that as the excess -price cost margin increases the relation between cash holding and customer concentratio n becomes more pronounced. This is consistent with the commitment motivation for cash holdings. The g overnment as a buyer Finally, not all customers pose equal risk to the supplier. The U.S. government represents a stable source of demand which is not l ikely to go out of business or become unable to make payments. So, being in a relationship with a government agency is not as risky as being in a relationship with a private firm. As a result, if the precautionary motive is the primary motivation for hol ding cash then firms in relationships with nongovernment firms should hold more cash then firms in relationships with government agencies. However, as a customer, the government still has an interest in protecting its investments. So, if commitment is the primary motivation for holding cash then suppliers to the government should be indistinguishable from suppliers which report non-government primary customers. Suppliers report the names of their primary customers. I categorize customers reported as Domestic Government or U.S. Navy etc as government agencies. The fifth column of Table 3 7 presents the results after distinguishing between sales to non -government firms and to government agencies. The customer concentration of nongovernment firms is defined as the Herfindahl Hirshman Index of only non-government firms. As a result, the coefficient of this variable represents the marginal difference between sales to all customers and sales to non-government firms only. First, I cannot reject the hypothesis that manufacturers which sell goods to a diverse set of customers have the same cash holding policy as manufacturers which sell their goods to the government. Second, the correlation between customer relationship and firm cash holding is clearl y dominated by sales to non-government entities. The coefficient of the concentration of sales to nongovernment companies is

PAGE 87

87 statistically significant. This result presented is consistent with the precautionary motive for holding cash and inconsistent w ith the commitment motive. Robustness To confirm the robustness of the results, I try a variety of other tests. Although not tabulated, I reproduce each of the regressions presented using c ash to assets, cash to net assets, log of cash to net assets, cas h minus debt to net assets, and cash to sales as the dependent variable. Although some of the dependent ratios mentioned generate extreme outliers, t he qualitative results are not altered by the alternative specifications in a material way. Next, I inclu d e additional control variables. I then parse the data into alternative subgroups to confirm that there is support for the primary results among all subgroups. The results are robust to alternative specification and provide evidence consistent with both the precautionary and commitment motive s for holding cash. Because I argue that firms are likely to use cash holdings as a risk management tool, I control for two potential determinants of corporate hedging decisions including the amount of foreign sales relative to total sales and the number of business segments in which the firm operates. The number of business segments in which a firm operates controls for the extent to which a firms could be diversified and whether firms have non-core assets that coul d be liquidated in periods of economic distress. The amount of foreign sales could be a natural hedge again changes in the domestic economy. Alternatively, as shown by foreign sales may generate higher cash holdings because cash repatriation incurs high taxes (Foley, Hartzell, Titman, and Twite (2007)). These additional variables are included in the regression specification in the second column in Table 3 8. The results are similar to those of Haushalter, Klasa and Maxwell (2007). Despite the inclusion of these two variables, the main result still holds.

PAGE 88

88 It may be the case that (for several reasons unrelated to being in a relationship) smaller firms hold more cash than larger firms. First, smaller firms may have a more difficult time accessing extern al capital. As a result, smaller firms hold more cash. Second, if there are economies of scale in holding cash, then smaller firms will hold marginally more cash than larger firms. Third, suppliers reporting large buyers are smaller firms by constructi on of the dataset. As a result, the tests may pick up a spurious relationship between cash holding and sales to customers. In all of the previous tests, the natural log of inflation adjusted assets is included in the regressions to control for differenc es in firm size. In case this is not enough to control for the spurious relationship, I divided the sample up into subgroups based on firm size. Table 3 8 reports the results of regressions based on firm size. The third column shows only firms that with less than $39 million in inflation adjusted total assets (which is the median inflation adjusted firm size among suppliers which report relationships). The fourth column shows firms with greater than $39 million in total assets. Several interesting pa tters emerge which illuminate the differences in cash holdings related to firm size. None of the small firms have an investment grade rating. So, either they do have public debt and it is unrated or they do not currently have a rating at all. Consistent with there being economies of scale to holding cash (Mulligan 1997), firm size and cash flow affect the cash holdings of small and large firms differently. Additionally, larger manufacturers are more likely to substitute cash holdings with net working ca pital than smaller firms. Despite these differences, the main results hold regardless of firm size. The cash holdings of small and large firms are positively correlated with the sales concentration to their major customers. Although not shown, the resul ts are qualitatively the same if percentage of sales to major customers is used in place of major customer concentration.

PAGE 89

89 Prior to 1997, the Federal Accounting Standards Board governed reporting of segments according to SFAS no. 14. According to this reg ulation, suppliers were obligated to report if any buyer accounted for more than 10% of sales. In addition to this requirement, suppliers had to report any customer which was material to their business operation. In 1997 the regulation was supplanted by SFAS 131, this regulation was more lenient and permitted report the company name only on an optional basis. To maintain uniformity throughout time, I omitted all buyer supplier relationships in which the buyer accounted for less than 10% of sales. Howeve r, as a result of the regulation change, the results may be dominated by the time when the reporting standards were stricter. To determine if this is the case, I divided the sample by time period. The last two columns of Table 3 8 reports the regression s based on time dependent subsamples. Regardless of the time period, the primary variable of interest is positive and significant. Although not shown, the same results obtain if percent of sales to the major customer is used as the primary variable of i nterest. Conclusion I find evidence that both the number of firms reporting relationships and the strength of these reported relationships measured by both percent of sales to the customers and sales concentration increases over time. Despite the increas e in both the prevalence and importance of buyer -supplier relationships, little extant research examines the effect of buyer -supplier relationships on suppliers financing or investment policy. I contribute to the small but growing literature about buyer -supplier relationships by providing evidence that suppliers which are in relationships hold more cash, all else equal, than firms which are not in relationships. Further, I provide evidence that firms cash holding policies are affected by the operating risk generated by the presence of a relationship. Specifically, I find that firms in relationships hold more cash than firms not in relationships. This result suggests that firms in relationships

PAGE 90

90 hold cash as a precaution against the adverse cash flow sho ck due to the loss of a major customer. Additionally, I find that as the percentage of sales to the major customers and as sales concentration to the major customers increase, so does cash holdings. This result suggests that as relationships induce great er risk, firms hold even more cash to offset that risk. These results also suggest that as the buyers stake in the supplier increases, the supplier holds more cash as a commitment to the buyer. These results hold even after I control for profitability, size, leverage, debt capacity, product market position, availability of natural hedges and other firm characteristics. Supporting the notion that greater cash holdings are motivated by precautionary measures, I present additional findings. First, in contr ast to the capital structure papers of Kale and Shahrur (2007) and Banerjee, Dasgupta and Kim (2008), who find manufacturers of unique goods maintain lower leverage ratios than manufacturers of non unique goods as a means of commitment to their buyers. I find that manufacturers of unique goods actually hold less cash than manufacturers of nonunique goods. Supp orting the hypothesis that cash holdings is motivated by commitment to the buyer, I find that manufacturers in relationships with a dominant market position hold more cash than suppliers with a weak market position. This is consistent with the theory that when buyer have few alternative suppliers they require their suppliers to hold higher levels of cash. Additionally, I find that firms in relatio nships with non-government corporations manage their cash holdings differently from firms in relationships with the government. The government is distinct from non -government corporations in that it is more stable and less likely to go out of business. As a result, it is not necessary for firms which sell to the government to fear an abrupt

PAGE 91

91 loss of sales. This result provides further support for the precautionary motivation for holding cash.

PAGE 92

92 Table 3 1. Relationships and Cash Holdings by Year All Firms Firms without Customers Firms with Customers All Firms Firms without Customers Firms with Customers Mean Median Mean Median Mean Mean Mean Median Median Median 1979 1986 28.0% 32.2% 26.8% 0.109 0.048 0.080 0.077 0.089 0.012 ** 0.043 0.043 0.045 0.002 1980 2030 29.5% 33.1% 25.4% 0.118 0.053 0.092 0.087 0.103 0.015 *** 0.048 0.048 0.048 0.000 1981 2237 33.2% 34.7% 27.0% 0.133 0.053 0.109 0.102 0.122 0.020 *** 0.053 0.051 0.061 0.010 ** 1982 2178 37.8% 36.3% 28.0% 0.145 0.058 0.112 0.103 0.126 0.023 *** 0.058 0.055 0.066 0.011 ** 1983 2294 40.6% 37.8% 30.2% 0.154 0.063 0.152 0.133 0.179 0.046 *** 0.084 0.071 0.102 0.032 *** 1984 2301 42.6% 38.6% 31.0% 0.161 0.063 0.130 0.119 0.146 0.027 *** 0.064 0.057 0.074 0.017 *** 1985 2253 43.8% 39.0% 31.3% 0.167 0.066 0.135 0.123 0.150 0.027 *** 0.065 0.060 0.078 0.019 *** 1986 2264 44.9% 38.2% 30.0% 0.163 0.061 0.141 0.134 0.150 0.016 ** 0.071 0.068 0.076 0.009 1987 2313 47.7% 38.3% 29.5% 0.158 0.058 0.142 0.136 0.149 0.013 ** 0.071 0.065 0.078 0.013 ** 1988 2193 47.1% 38.2% 29.9% 0.154 0.063 0.127 0.123 0.131 0.008 0.059 0.057 0.060 0.004 1989 2112 47.2% 38.9% 31.0% 0.154 0.066 0.125 0.117 0.134 0.017 *** 0.058 0.051 0.065 0.014 ** 1990 2048 48.2% 38.9% 32.0% 0.150 0.061 0.123 0.113 0.133 0.019 *** 0.054 0.052 0.060 0.009 ** 1991 2057 48.0% 39.6% 32.8% 0.153 0.068 0.134 0.123 0.147 0.024 *** 0.064 0.058 0.072 0.015 *** 1992 2129 50.0% 39.6% 32.0% 0.147 0.064 0.144 0.127 0.161 0.034 *** 0.068 0.060 0.079 0.019 *** 1993 2266 52.0% 39.8% 33.1% 0.140 0.071 0.154 0.138 0.168 0.030 *** 0.072 0.061 0.088 0.026 *** 1994 2361 52.9% 38.2% 32.8% 0.130 0.063 0.139 0.124 0.152 0.028 *** 0.065 0.055 0.078 0.023 *** 1995 2601 53.4% 39.9% 35.0% 0.134 0.068 0.150 0.127 0.170 0.042 *** 0.063 0.051 0.085 0.033 *** 1996 2749 54.9% 39.9% 34.0% 0.139 0.070 0.161 0.132 0.185 0.053 *** 0.074 0.058 0.093 0.035 *** 1997 2768 54.0% 40.8% 36.0% 0.142 0.078 0.164 0.136 0.187 0.050 *** 0.074 0.056 0.098 0.042 *** 1998 2649 53.1% 40.8% 36.0% 0.141 0.078 0.150 0.126 0.171 0.045 *** 0.060 0.046 0.078 0.032 *** 1999 2529 42.6% 40.9% 37.0% 0.143 0.081 0.161 0.141 0.188 0.047 *** 0.060 0.050 0.079 0.029 *** 2000 2437 53.7% 42.3% 38.5% 0.145 0.084 0.171 0.139 0.198 0.059 *** 0.064 0.046 0.088 0.042 *** 2001 2263 54.2% 42.8% 39.0% 0.149 0.086 0.176 0.143 0.204 0.061 *** 0.082 0.058 0.108 0.050 *** 2002 2121 56.6% 43.5% 40.0% 0.152 0.093 0.183 0.149 0.208 0.059 *** 0.095 0.072 0.121 0.049 *** 2003 2002 56.4% 43.6% 40.0% 0.158 0.088 0.201 0.174 0.221 0.047 *** 0.119 0.093 0.144 0.051 *** 2004 1938 56.5% 44.4% 40.0% 0.165 0.091 0.214 0.183 0.238 0.055 *** 0.129 0.103 0.165 0.062 *** 2005 1867 56.8% 44.1% 40.0% 0.166 0.090 0.216 0.181 0.242 0.061 *** 0.140 0.108 0.170 0.061 *** 2006 1517 54.0% 43.5% 38.8% 0.156 0.084 0.217 0.192 0.238 0.046 *** 0.125 0.098 0.160 0.062 *** Total 62463 47.9% 40.0% 34.0% 0.148 0.072 0.149 0.129 0.172 0.043 *** 0.070 0.058 0.086 0.028 *** Cash / Assets Difference of Means Difference of Medians Firms Reporting Customers Number of Firms Year Percent of Sales to Major Customers Major Customer Concentration This ta ble reports the mean and median values of firm characteristics of manufacturing between 1979 and 2006. Percent of sales to m ajor customers is the sum of sales to all major customer divided by total sales. Major Customer Concentration is a Herfindahl Hir schman Index of sales determined as the sum of the squares of the share of sales to each reported customers where the shares are expressed as a percentage of total sales. Cast /Assets is cash and marketable securities divided by total assets minus cash and marketable securities The students T test is used to determine if the differences between means are statistically different from zero. The Wilcoxon RankSum test is used to determine if the differences between medians are statistically different from z ero. The symbols ***, **, and indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

PAGE 93

93 Table 3 2 Comparison of Suppliers with Major Customers to Suppliers without Major Customers Mean Median Std Deviation N Mean Median Std Deviation N Cash / Net Assets 0.247 0.062 0.592 32503 0.355 0.094 0.694 29946 -0.107 *** -0.032 *** Cash / Total Assets 0.129 0.058 0.168 32507 0.172 0.086 0.199 29947 -0.043 *** -0.028 *** Total Assets 1107.32 83.99 6628.14 32512 330.94 32.90 1506.61 29951 776.38 *** 51.09 *** Age 16.232 13 12.804 32512 12.848 9 11.310 29951 3.384 *** 4.000 *** R&D / Sales 0.101 0.009 0.443 32512 0.121 0.022 0.389 29951 -0.020 *** -0.013 *** CAPX / Assets 0.070 0.053 0.064 32010 0.074 0.052 0.072 29696 -0.004 *** 0.001 Debt / Assets 0.284 0.239 0.283 32457 0.274 0.217 0.300 29896 0.010 *** 0.022 *** Cash Flow / Assets -0.057 0.070 0.552 31222 -0.079 0.065 0.537 28179 0.022 *** 0.005 *** Market to Book 1.972 1.331 2.182 32437 2.172 1.421 2.304 29933 -0.200 *** -0.090 *** Net Working Capital / Assets 0.161 0.206 0.355 31902 0.170 0.222 0.363 29790 -0.009 *** -0.016 *** CV of Cash Flow 2.201 2.238 0.726 32512 2.367 2.441 0.633 29951 -0.166 *** -0.203 *** Difference between Medians Means Firms with no Major Customers Firms with Major Customers This table reports the mean, median, and sta ndard deviations of the values of firm characteristics of manufacturing between 1979 and 2006. Cash/ Net Assets is cash and marketable securities divided by total assets minus cash and marketable securities. Cast/Assets is cash and marketable securities divided by total assets. Total Assets is the book value of assets in 1984 dollars. M/B is the book value of assets minus the book value of equity plus the market value of equity all divided by the book value of assets. R&D/Sales is R&D expense divided b y sales. CAPX/Assets is capital expenditures divided by assets. Debt/Assets is long term debt plus debt in short term liabilities divided by assets. Cash Flow/Assets is earnings after interest, dividends and taxes but before depreciation divided by ne t assets. NWC/Assets is net working capital divided by assets. The coefficient of variation of cash flow is the industry a verage of standard deviation of cash flow divided by the mean cash flow over the previous 20 years. The students T test is used to determine if the differences between means are statistically different from zero. The Wilcoxon RankSum test is used to determine if the differences between medians are statistically different from zero. The symbols ***, **, and indicate statistical significance at the 1%, 5%, and 10% levels, respectively

PAGE 94

94 Table 3 3 Summary Statistics Cash / Assets Total Assets M/B Age logiata Market Leverage R&D / Sales CAPX / Assets Debt / Assets NWC / Assets CV of Cash Flow Firms Years in which no single customer accounts for more than 10% of salesMean 0.143 582.006 2.077 15.312 4.109 0.134 0.120 0.070 0.288 0.163 2.284 Median 0.066 56.470 1.356 12 4.034 0.089 0.012 0.051 0.236 0.214 2.355 N 18087 18091 18035 18091 18091 18016 18091 17779 18063 17796 18091Firm Years in which the sum of sales to all major customers between 10% and 18%Mean 0.137 466.234 1.825 14.577 4.054 0.131 0.063 0.070 0.266 0.219 2.305 Median 0.066 49.128 1.344 11 3.894 0.083 0.016 0.053 0.226 0.243 2.365 N 7444 7444 7440 7444 7444 7432 7444 7370 7432 7385 7444Firm Years in which the sum of sales to all major customers between 18% and 34%Mean 0.152 331.392 1.973 13.725 3.758 0.125 0.083 0.070 0.274 0.199 2.358 Median 0.072 37.226 1.369 10 3.617 0.075 0.018 0.051 0.232 0.231 2.437 N 7507 7507 7503 7507 7507 7494 7507 7453 7496 7467 7507Firm Years in which the sum of sales to all major customers between 34% and 56%Mean 0.173 282.926 2.132 12.265 3.518 0.114 0.115 0.071 0.276 0.171 2.367 Median 0.087 30.808 1.411 9 3.428 0.057 0.024 0.049 0.214 0.225 2.478 N 7368 7369 7366 7369 7369 7349 7369 7304 7350 7336 7369Firm Years in which the sum of sales to all major customers is greater than 56%Mean 0.224 244.886 2.746 10.861 3.160 0.095 0.220 0.084 0.283 0.093 2.437 Median 0.137 21.076 1.608 7 3.048 0.032 0.033 0.056 0.193 0.181 2.519 N 7628 7631 7624 7631 7631 7614 7631 7569 7618 7602 7631 This table reports the mean and median values of firm characteristics of manufacturing between 1979 and 2006. With the exception of total assets, assets refers to net assets defined as total assets minus cash and marketable securities. Cash/Assets is cash and marketable securities divided by assets. Total Assets is the book value of assets in 1984 dollars. M/B is the book value of assets minus the book value of e quity plus the market value of equity all divided by the book value of assets. R&D/Sales is R&D expense divided by sales. CAPX/Assets is capital expenditures divided by assets. Debt/Assets is long term debt plus debt in short term liabilities divided by assets. Cash Flow/Assets is earnings after interest, dividends and taxes but before depreciation divided by net assets. NWC/Assets is net working capital divided by assets. The coefficient of variation of cash flow is the industry average of standa rd deviation of cash flow divided by the mean cash flow over the previous 20 years.

PAGE 95

95 Table 3 4 All Manufacturing Firms Ever in a Relationship 0.029 *** 0.031 *** (0.000) (0.000) Percent of Sales 0.070 *** 0.063 *** 0.070 *** (0.000) (0.000) (0.000) Customer Concentration 0.104 *** 0.107 *** 0.119 *** (0.000) (0.000) (0.000) Log(Assets) 0.003 ** 0.001 0.004 ** 0.001 0.029 *** 0.003 ** 0.001 0.029 *** (0.038) (0.863) (0.015) (0.747) (0.000) (0.024) (0.766) (0.000) R&D/Sales 0.121 *** 0.115 *** 0.120 *** 0.115 *** 0.054 *** 0.119 *** 0.113 *** 0.053 *** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) CAPX/Assets 0.654 *** 0.651 *** 0.643 *** 0.639 *** 0.498 *** 0.652 *** 0.645 *** 0.500 *** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Debt/Assets -0.302 *** -0.290 *** -0.300 *** -0.289 *** -0.173 *** -0.300 *** -0.289 *** -0.173 *** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Investment Grade Dum -0.069 *** -0.058 *** -0.071 *** -0.059 *** -0.008 -0.072 *** -0.061 *** -0.008 (0.000) (0.000) (0.000) (0.000) (0.126) (0.000) (0.000) (0.118) Dividend Dummy -0.059 *** -0.049 *** -0.057 *** -0.048 *** -0.011 *** -0.059 *** -0.050 *** -0.011 *** (0.000) (0.000) (0.000) (0.000) (0.006) (0.000) (0.000) (0.006) Cash Flow/Assets -0.187 *** -0.191 *** -0.187 *** -0.191 *** -0.194 *** -0.187 *** -0.192 *** -0.194 *** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) M/B 0.001 0.001 0.000 0.000 0.001 0.000 0.000 0.001 (0.841) (0.796) (0.897) (0.873) (0.081) (0.970) (0.983) (0.148) NWC/Assets -0.061 ** -0.090 *** -0.058 -0.087 *** -0.062 *** -0.057 -0.086 *** -0.062 *** (0.047) (0.000) (0.059) (0.000) (0.000) (0.062) (0.000) (0.000) Coefficient of Variation 0.037 *** 0.039 *** 0.037 *** 0.039 *** 0.005 ** 0.037 *** 0.039 *** 0.005 ** of Cash Flow (0.000) (0.000) (0.000) (0.000) (0.011) (0.000) (0.000) (0.014) Intercept 0.204 *** 0.127 *** 0.208 *** 0.139 *** 0.054 *** 0.218 *** 0.145 *** 0.063 *** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Year Dummies Yes No Yes No Yes Yes No Yes Number of Observations 58028 58028 58028 58028 58028 58028 58028 58028 Number of Groups 28 28 6007 28 6007 R20.29 0.29 0.29 0.30 0.21 0.29 0.30 0.21 Fixed Effects Fama-MacBeth OLS OLS Fixed Effects OLS Fama-MacBeth Fama-MacBeth This table reports OLS and Fama MacBeth regressions of the determinants of cash holding for firms between 1979 and 2006. OLS regre ssions include year dummies and the standard errors are clustered by firm. The dependent variable is the log of one plus the ratio of cash and marketable securities to total assets minus cash and marketable securities. Ever in Relationship is an indicato r for whether the firm reports being in a relationship during at least one year. Percent of Sales is the sum of the sales to each reported customer divided by total sales. Customer Concentration is a Herfindahl Hirschman Index of sales determined as the sum of the squares of the share of sales to each reported customers where the shares are expressed as a percentage of total s ales. Log(Assets) is the natural logarithm of the book value of assets in 1984 dollars. M/B is the book value of assets minus the book value of equity plus the market value of equity all divided by the book value of assets. R&D/Sales is R&D expense divided by sales. CAPX/Assets is capital expenditures divided by ass ets. Debt/Assets is long term debt plus debt in short term liabilities divided by assets. Investment Grade Dum is an indicator variable equal to one if the firm has an investment grade d ebt rating. Cash Flow/Assets is earnings after interest, dividends and taxes but before depreciation divided by net assets. NW C/ Assets is net working capital divided by assets. The coefficient of variation of cash flow is the industry average of standard deviation of cash flow div ided by the mean cash flow over the previous 20 years. P values are shown in parentheses. The symbols ***, **, and indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

PAGE 96

96 Table 3 5 Firms which Report Relationships Percent of Sales 0.057 *** 0.052 *** 0.072 *** (0.000) (0.000) (0.000) Customer Concentration 0.086 *** 0.091 *** 0.119 *** (0.000) (0.000) (0.000) Log(Assets) 0.007 *** 0.002 0.032 *** 0.006 *** 0.002 0.032 *** (0.001) (0.432) (0.000) (0.001) (0.402) (0.000) R&D/Sales 0.125 *** 0.135 *** 0.055 *** 0.124 *** 0.134 *** 0.054 *** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) CAPX/Assets 0.645 *** 0.629 *** 0.526 *** 0.652 *** 0.634 *** 0.528 *** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Debt/Assets -0.316 *** -0.303 *** -0.183 *** -0.316 *** -0.303 *** -0.183 *** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Investment Grade Dum -0.081 *** -0.065 *** -0.017 ** -0.082 *** -0.066 *** -0.017 ** (0.000) (0.000) (0.017) (0.000) (0.000) (0.018) Dividend Dummy -0.061 *** -0.050 *** -0.008 -0.062 *** -0.051 *** -0.008 (0.000) (0.000) (0.072) (0.000) (0.000) (0.067) Cash Flow/Assets -0.189 *** -0.177 *** -0.195 *** -0.189 *** -0.178 *** -0.195 *** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) M/B 0.002 0.002 0.001 0.002 0.001 0.001 (0.467) (0.476) (0.132) (0.516) (0.576) (0.224) NWC/Assets -0.073 ** -0.101 *** -0.071 *** -0.073 ** -0.100 *** -0.070 *** (0.048) (0.000) (0.000) (0.05) (0.000) (0.000) Coefficient of Variation 0.044 *** 0.045 *** 0.007 *** 0.045 *** 0.046 *** 0.007 *** of Cash Flow (0.000) (0.000) (0.006) (0.000) (0.000) (0.007) Intercept 0.188 *** 0.126 *** 0.056 *** 0.196 *** 0.131 *** 0.067 *** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Year Dummies Yes No Yes Yes No Yes Number of Observations 44521 44521 44521 44521 44521 44521 Number of Groups 28 4238 28 4238 R20.29 0.29 0.21 0.29 0.30 0.21 OLS Fama-MacBeth Fixed Effects OLS Fama-MacBeth Fixed Effects This table reports OLS, Fama MacBeth, and Firm Fixed Effect regressions of the determinants of cash holding for f irms which report relationships between 1979 and 2006. OLS regressions include year dummies and the standard errors are clustered by firm. Firm Fixed effect regressions include firm and year fixed effects. The dependent variable is the log of one plus the ratio of cash and marketable securities to total assets minus cash and marketable securities. Percent of Sales is the sum of the sales to each reported customer divided by total sales. Customer Concentration is a Herfindahl Hirschman Index of sales determined as the sum of the squares of the share of sales to each reported customers where the shares are expressed as a percentage of total sales. Log(Assets) is the natural logarithm of the book value of assets in 1984 dollars. M/B is the book value of assets minus the book value of equity plus the market value of equity all divided by the book value of assets. R&D/Sales is R&D expense divided by sales. CAPX/Assets is capital expenditures divided by assets. Debt/Assets is long term debt plus debt in short term liabilities divided by assets. Investment Grade Dum is an indicator variable equal to one if the firm has an investment grade debt rating. Cash Flow/Assets is earnings after interest, dividends and taxes but before depreciation divided by net assets. NWC/Assets is net working capital divided by assets. The coefficient of variation of cash flow is the industry average of standard deviation of cash flow divided by the mean cash flow over the previous 20 years. P values are shown in paren theses. The symbols ***, **, and indicate statistical significance at the 1%, 5%, and 10% levels, respectively .

PAGE 97

97 Table 3 6 Descriptive Statistics by Industry Mean Std Dev Mean Std Dev Mean Std Dev Mean Std Dev Mean Std Dev Mean Std DevNon -Durable Goods Manufacturers 20 Food and Kindred Products 33% 0.172 0.448 0.319 0.199 0.098 0.131 0.016 0.192 0.073 0.049 0.048 0.095 21 Tobacco Products 33% 0.140 0.289 0.202 0.084 0.040 0.027 0.009 0.044 0.062 0.107 0.157 0.246 22 Textile Mill Products 44% 0.077 0.174 0.302 0.171 0.076 0.080 0.005 0.011 0.056 0.036 0.053 0.055 23 Apparel and Finished Products Made From Fabrics 54% 0.166 0.420 0.395 0.217 0.122 0.154 0.002 0.018 0.036 0.037 0.047 0.081 24 Lumber and Wood Products 25% 0.196 0.315 0.317 0.215 0.111 0.162 0.012 0.172 0.028 0.021 0.043 0.077 25 Furniture and Fixtures 34% 0.099 0.183 0.293 0.234 0.082 0.128 0.007 0.014 0.047 0.036 0.064 0.054 26 Paper and Allied Products 29% 0.102 0.320 0.241 0.155 0.050 0.062 0.014 0.117 0.069 0.038 0.080 0.060 27 Printing, Publishing, and Allied Industries 23% 0.217 0.540 0.292 0.165 0.074 0.077 0.013 0.137 0.034 0.043 0.068 0.127 28 Chemicals and Allied Products 37% 0.262 0.657 0.378 0.257 0.139 0.190 0.147 0.566 0.076 0.059 0.172 0.756 29 Petroleum Refining and Related Industries 25% 0.078 0.174 0.314 0.217 0.091 0.132 0.037 0.315 0.058 0.017 0.021 0.173 30 Rubber and Miscellaneous Plastics Products 48% 0.149 0.345 0.318 0.211 0.095 0.130 0.040 0.278 0.031 0.046 0.030 0.131 31 Leather and Leather Products 47% 0.223 0.378 0.382 0.260 0.170 0.282 0.003 0.007 0.085 0.041 0.062 0.057 32 Stone, Clay, Glass, and Concrete Products 30% 0.134 0.368 0.391 0.258 0.127 0.155 0.026 0.182 0.072 0.056 0.051 0.116 33 Primary Metal Industries 44% 0.113 0.360 0.308 0.193 0.083 0.101 0.017 0.162 0.068 0.048 0.040 0.082 Durable Goods Manufacturers 34 Fabricated Metal Products 45% 0.144 0.373 0.396 0.264 0.161 0.246 0.020 0.157 0.039 0.042 0.047 0.104 35 Machinery and Computer Equipment 50% 0.362 0.656 0.394 0.248 0.143 0.198 0.139 0.426 0.070 0.059 0.100 0.428 36 Electronic and Electrical Equipment 65% 0.413 0.745 0.435 0.252 0.160 0.205 0.156 0.452 0.088 0.057 0.093 0.418 37 Transportation Equipment 64% 0.137 0.348 0.483 0.267 0.197 0.228 0.049 0.283 0.110 0.075 0.042 0.101 38 Measuring, Analyzing, and Controlling Instruments 49% 0.523 0.916 0.437 0.275 0.192 0.251 0.254 0.641 0.048 0.039 0.363 1.068 39 Miscellaneous Manufacturing Industries 44% 0.215 0.471 0.380 0.234 0.117 0.139 0.034 0.199 0.046 0.041 0.015 0.259 Firms in Relationships Industry SIC Code Cash Ratio Customer Concentration Percent of Sales to the Customer R&D / Sales Price-Cost Margin HerfindahlHirshman Index This table reports the mean standard deviation of firm characteristics of manufacturing betwe en 1979 and 2006 separated by two digit SIC code. Cash ratio is the ratio of cash and marketable securities to total assets minus cash and marketable securities. Percent of sales to major cust omers is the sum of sales to all major customer divided by tot al sales. Major Customer Concentration is a Herfindahl Hirschman Index of sales determined as the sum of the squares of the share of sales to each reported customers where the shares are expressed as a percentage of total sales. Cast Ratio is cash and m arketable securities divided by total assets. R&D/Sales is R&D expense divided by sales. Excess price cost margin is operating profit margin before depreciation divided by sales minus the industry average of this ratio. Industry HHI is a measure of indus try concentration provided by the U.S. The students T test is used to determine if the differences between means are statistically different from zero. The Wilcoxon Rank Sum test is used to determine if the differences between medians are statistically different from zero. The symbols ***, **, and indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

PAGE 98

98 Table 3 7 Determinants of Relation between Cash Holdings and Buyer -Supplier Relationships Customer Concentration 0.136 *** 0.167 *** 0.145 *** 0.184 *** 0.046 *** (0.000) (0.000) (0.000) (0.000) (0.004) Customer Concentration Durable Goods Dummy 0.027 (0.282) Customer Concentration R&D/Sales -0.019 (0.061) Customer Concentration Excess Price-Cost Margin 0.034 *** (0.001) Customer Concentration Industry HHI 0.010 (0.486) Excess Price-Cost Margin 0.031 *** (0.000) Industry HHI 0.017 *** (0.000) Government Customer Concentration -0.115 *** -0.118 *** -0.119 *** -0.136 *** (0.000) (0.000) (0.000) (0.000) Non-Government Customer Concentration 0.111 *** (0.000) Log(Assets) 0.032 *** 0.032 *** 0.033 *** 0.032 *** 0.032 *** (0.000) (0.000) (0.000) (0.000) (0.000) R&D/Sales 0.054 *** 0.057 *** 0.051 *** 0.071 *** 0.054 *** (0.000) (0.000) (0.000) (0.000) (0.000) CAPX/Assets 0.526 *** 0.525 *** 0.546 *** 0.529 *** 0.526 *** (0.000) (0.000) (0.000) (0.000) (0.000) Debt/Assets -0.183 *** -0.183 *** -0.180 *** -0.183 *** -0.183 *** (0.000) (0.000) (0.000) (0.000) (0.000) Investment Grade Dum -0.017 ** -0.017 ** -0.019 ** -0.017 ** -0.017 ** (0.022) (0.022) (0.011) (0.020) (0.022) Dividend Dummy -0.009 -0.009 -0.011 ** -0.010 ** -0.009 (0.06) (0.061) (0.023) (0.036) (0.06) Cash Flow/Assets -0.195 *** -0.195 *** -0.203 *** -0.210 *** -0.195 *** (0.000) (0.000) (0.000) (0.000) (0.000) M/B 0.001 0.001 0.000 0.001 0.001 (0.241) (0.239) (0.656) (0.293) (0.235) NWC/Assets -0.070 *** -0.070 *** -0.066 *** -0.071 *** -0.070 *** (0.000) (0.000) (0.000) (0.000) (0.000) Coefficient of Variation 0.007 *** 0.007 *** 0.006 ** 0.006 ** 0.007 *** of Cash Flow (0.006) (0.006) (0.022) (0.012) (0.006) Intercept 0.065 *** 0.063 *** 0.053 *** 0.069 *** 0.065 *** (0.000) (0.000) (0.000) (0.000) (0.000) Year Dummies Yes Yes Yes Yes Yes Number of Observations 44521 44521 41941 44378 44521 Number of Groups 4238 4238 4131 4226 4238 R20.21 0.21 0.22 0.21 0.21 (5) (4) (3) (2) (1) This table reports fixed effects r egressions of the determinants of cash holding for firms which report relationships between 1979 and 2006. The dependent variable is the log of one plus the ratio of cash and marketable securities to total assets minus cash and marketable securities. Cus tomer Concentration is a Herfindahl Hirschman Index of sales determined as the sum of the squares of the share of sales to each reported customers where the shares are expressed as a percentage of total sales. Durable goods dummy is an indicator variable which equals 1 if the firms has a primary SIC code from 3400 to 3990. Excess price cost margin is operating profit margin before depreciation divided by sales minus the industry average of this ratio. Industry HHI is a measure of industry concentration

PAGE 99

99 pr ovided by the U.S. Census. Nongovernment Customer Concentration is a Herfindahl Hirschman Index of sales to nongovernment companies. Log(Assets) is the natural logarithm of the book value of assets in 1984 dollars. M/B is the book value of assets minus the book value of equity plus the market value of equity all divided by the book value of assets. R&D/Sales is R&D expense divided by sales. CAPX/Assets is capital expenditures divided by assets. Debt/Assets is long term debt plus debt in short term liabilities divided by assets. Investment Grade Dum is an indicator variable equal to one if the firm has an investment grade debt rating. Cash Flow/Assets is earnings after interest, dividends and taxes but before depreciation divided by net assets. NWC/Assets is net working capital divided by assets. The coefficient of variation of cash flow is the industry average of standard deviation of cash flow divided by the mean cash flow over the previous 20 years. P values are shown in parentheses. The s ymbols ***, **, and indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

PAGE 100

100 Table 3 8 Robustness Customer Concentration 0.117 *** 0.137 *** 0.066 *** 0.138 *** 0.074 *** (0.000) (0.000) (0.000) (0.000) (0.000) Foreign Sales -0.004 (0.351) Number of Business Segments -0.008 *** (0.000) Log(Assets) 0.033 *** 0.066 *** -0.01 *** 0.020 *** 0.069 *** (0.000) (0.000) (0.000) (0.000) (0.000) R&D/Sales 0.054 *** 0.039 *** 0.141 *** 0.073 *** -0.005 (0.000) (0.000) (0.000) (0.000) (0.488) CAPX/Assets 0.527 *** 0.542 *** 0.253 *** 0.399 *** 0.673 *** (0.000) (0.000) (0.000) (0.000) (0.000) Debt/Assets -0.183 *** -0.239 *** -0.005 -0.179 *** -0.131 *** (0.000) (0.000) (0.399) (0.000) (0.000) Investment Grade Dum -0.013 -0.027 *** -0.004 -0.051 ** (0.069) (0.000) (0.640) (0.015) Dividend Dummy -0.008 0.001 0.011 *** -0.007 -0.005 (0.067) (0.911) (0.005) (0.139) (0.660) Cash Flow/Assets -0.195 *** -0.227 *** -0.049 *** -0.182 *** -0.185 *** (0.000) (0.000) (0.000) (0.000) (0.000) M/B 0.001 -0.006 *** 0.029 *** -0.004 *** 0.003 *** (0.218) (0.000) (0.000) (0.000) (0.010) NWC/Assets -0.071 *** -0.097 *** -0.102 *** -0.102 *** -0.060 *** (0.000) (0.000) (0.000) (0.000) (0.000) Coefficient of Variation 0.007 *** 0.010 ** 0.002 0.002 0.013 ** of Cash Flow (0.008) (0.045) (0.274) (0.489) (0.019) Intercept 0.063 *** 0.080 *** 0.176 *** 0.123 *** -0.074 *** (0.000) (0.000) (0.000) (0.000) (0.005) Year Dummies Yes Yes Yes Yes Yes Number of Observations 44521 22008 22513 30411 14110 Number of Groups 4238 2940 2333 3689 2526 R20.21 0.27 0.33 0.20 0.12 Additional Variables Post-1997 Pre-1997 Large Firms Small Firms This table reports fixed effects regressions of the determinants of cash holding for firms which report relationships between 1979 and 2006. The dependent variable is the log of one plus the ratio of cash and marketable securities to total assets minus cash and marketable securities. Percent of Sales is the sum of the sales to each reported customer divided by total sales. Percent of sales to nongovernment corporations is the sum of sales to each customer which is not a government agency divided by total sales. Customer Concentration is a Herfindahl Hirschman Index of sales determined as the sum of the squares of the share of sal es to each reported customers where the shares are expressed as a percentage of total sales. Log(Assets) is the natural logarithm of the book value of assets in 1984 dollars. M/B is the book value of assets minus the book value of equity plus the market value of equity all divided by the book value of assets. R&D/Sales is R&D expense divided by sales. CAPX/Assets is capital expenditures divided by assets. Debt/Assets is long term debt plus debt in short term liabilities divided by assets. Investmen t Grade Dum is an indicator variable equal to one if the firm has an investment grade debt rating. Cash Flow/Assets is earnings after interest, dividends and taxes but before depreciation divided by net assets. NWC/Assets is net working capital divided by assets. The coefficient of variation of cash flow is the industry average of standard deviation of cash flow divided by the mean cash flow over the previous 20 years. P values are shown in parentheses. The symbols ***, **, and indicate statistical significance at the 1%, 5%, and 10% levels, respectively.

PAGE 101

101 CHAPTER 4 CONCLUSION AND FUTUR E WORK Over the past few decades, the prevalence and importance of significant buyer -supplier relationships has increased dramatically. For suppliers, being in a rel ationship can be a double edged sword with both costs and benefits. To accentuate the benefits and minimize the costs, suppliers may take a variety of actions to change their financing and investment policy. Despite this, little extant literature addresse s how major relationships influence firm functioning. This dissertation contributes to the literature by examining the effect of relationships in two distinct areas of corporate finance namely financial constraints and cash management. Specifically, firms in relationships have lower financing constraints and they maintain higher levels of cash. However, not all relationships are the same. I further investigate the determinants of the influence of being in a relationship on financial constraints and cash holdings. In future work I hope to confirm the robustness of the results presented here and address when and how being in a relationship influences other aspects of firm financing and investment policy. Investment -Cash Flow Sensitivity I provide evidence that suppliers in relationships are less affected by capital market frictions. To do so, I turn to a large literature documenting the link between liquidity and investment. Here, liquidity, defined as the availability of internal funds, is an important determinant of investment when capital market frictions exist. On average, suppliers with major buyers demonstrate lower investment cash flow sensitivity than a matched set of their peers, suggesting that relationships ease suppliers liquidity constraint s. I choose a matched sample of peer firms to ensure that neither differences across industries nor firm size bias the results. The buyers commitment to the supplier and the buyers stake in the supplier both influence the suppliers financing constraints. In theory, the more committed a buyer is to a

PAGE 102

102 relationship, the greater their incentive to help the supplier. Further, as a partner and stakeholder, the buyer is the residual claimant on its suppliers investments. Therefore, the greater the buyers stake in the supplier, the greater the buyers incentive to aid the supplier. These claims are supported by the empirical results. A buyers stake in the supplier, as determined by how many other suppliers they are in a relationship with, affects the sen sitivity of investment to cash flow of their suppliers. In contrast, suppliers in relationships with the largest buyers (who can easily switch suppliers) are no better off than their independent peers. This is consistent with the hypothesis that supplier s in relationships with more committed buyers face lower financing frictions. Using the percent of suppliers sales from the principal buyer and the length of the relationship as proxies for the buyers stake in a relationship, I find that as the buyers stake in its supplier increases the suppliers sensitivity of investment to cash flow decreases. Providing additional evidence, I use the corporate bond spread and a recession index as proxies for the accessibility of external financing to show that investment to cash flow sensitivity of suppliers with principal customers is less affected by macroeconomic cycles than the sensitivity of their peers. This result is consistent with the theory that firms in relationships face lower financing frictions even in times of general economic downturn. Finally, I examine relationships in the durable goods industry. Firms in the durable goods industry, in general, produce unique products and invest more heavily in relationship specific assets therefore imposing high er liquidation costs on buyers. Consistent with this idea, I find that the effects are more pronounced for firms in industries with a high degree of relationshipspecific investments. Consequently, I find evidence consistent with the idea that being in a relationship helps to ease capital market frictions.

PAGE 103

103 Transmission Mechanism While Chapter 2 provides evidence consistent with the idea that being in a relationship eases capital market frictions, it does not detail the means through which this occurs. B uyers may provide relief to suppliers directly by paying off trade credit faster or by agreeing to timely price increases. Alternatively, by providing a steady stream of revenue and reducing business risk, buyers may indirectly provide suppliers with grea ter access to bank capital. Future work will investigate the channel by which buyers provide aid to their suppliers. First, comparing the accounts receivable of firms in relationships relative to their independent peers may provide some evidence of the direct credit channel of support. Additionally, testing the level of accounts receivable over the course of the business cycle can provide insights into the channel of transmission. Next, comparing the operating profit of firms in relationships to their independent peers may provide evidence of the presence of the pricing channel. Euler Method The methodology used in Chapter 2, pioneered by Fazzari, Hubbard, and Petersen (1988), of measuring investment cash flow sensitivity to determine the relative f inancial constraint of firms has faced sharp criticism. In addition to this methodology, to confirm the robustness of the results, I will employ the Euler equation method to model financial constraints similar to that of Whited (1992) and Whited and Wu (2 006). This structural approach has the advantage of avoiding the difficult problem of accurately measuring Tobins Q. Cash Holdings This dissertation further investigates how relationships affect supplier cash holdings. Compared to their peers, supplie rs in relationships hold more cash. Further, suppliers cash holdings are positively correlated with both the ratio of sales to major customers to total sales and major customer sales concentration. Two different theories of cash holdings can explain

PAGE 104

104 these results. First, suppliers may choose to hold more cash as a commitment to the buyer that they will remain financially stable in the future. Such a commitment may be necessary in order to induce buyers to make specialized investments in relationship specif ic assets. Second, suppliers may choose to hold more cash as a hedge against future risk. Generally, when buyers purchase goods from manufacturers they utilize individual purchase orders rather than entering into long -term contracts. By depending on one or more customers for a significant part of the companys sales, suppliers open themselves up to the risk of losing a large portion of sales at once which would cripple their financial health. As a result, maintaining additional cash on hand is necessary as a precaution against the operating risk. Therefore, the commitment motivation for holding cash reserves is primarily to provide the buyer with confidence in the in the suppliers financial position while the precautionary motive for holding cash is to protect the firm itself from financial distress. While, in general, suppliers in relationships hold more cash, not all relationships are the same and some relationships are not as risky as others. For some buyers, engaging in a relationship may generate p rohibitive switching costs. Such switching costs reduce the risk that the buyer change suppliers abruptly, reducing the operating risk of the supplier. As a result, the supplier will not need to hold as much precautionary cash. Consistent with this theor y, I find a weaker relation between customer concentration and cash holdings for manufacturers of unique goods relative to manufacturers of nonunique goods. I also find that suppliers cash holdings are affected when their customer is a corporation but a re unaffected when the customer is a government agency. The U.S. government is not likely to go out of business or become unable to make payments for goods. Together these results are consistent with the precautionary motivation for holding cash.

PAGE 105

105 Value of Cash Holdings for Firms in Relationships Firms in relationships hold more cash than firms which are not in relationships as a precaution against cash flow risk. Holding cash benefits the supplier by reducing the likelihood of incurring financial dist ress costs if their major buyer ended the relationship and the supplier can no longer generate sufficient cash flow. However, corporate liquidity comes at a cost, since cash may provide funds for managers to invest in value decreasing projects (Jensen 1986 ). It is therefore possible that while firms hold extra cash because of the risk induced by being in a relationship, the market does not perceive extra cash holdings as a benefit. Future work will examine when and how differences in firm and relationshi p characteristics affect the marginal value to stock holders of corporate cash holdings. Faulkender and Wang (2006) and Pinkowitz and Williamson (2006) document that cash holdings are more valuable for constrained firms rather than unconstrained firms. Us ing their methodology, I may be able to determine if the marginal value of holding cash is greater for firms in relationships than for firms which are not in relationships. Relationships and Risk A number of different metrics indicate that being in a re lationship increases a firms idiosyncratic cash flow volatility. Additionally, certain types of relationships increase the expected volatility of a firms cash flows. When the supplier accounts for a significant part of the suppliers current sales and th at customer has no future commitment to continue purchasing, buyer supplier relationships induce expected cash flow volatility. As a result of this risk, firms in relationships hold cash as a financial hedge. While this indicates that firm management recog nizes this risk, it is not clear whether the market perceives this source of risk as well. In future work, I will investigate the link between inter -firm relationships and the idiosyncratic volatility of the firms stock market returns.

PAGE 106

106 The Link Between I nvestment -Cash Flow Sensitivity a nd Cash Holdings This dissertation addressed two distinct means through which relationships influence firm behavior. Despite studying financial constraints and cash holdings separately, they are fundamentally linked. The importance of holding cash is affected by a firms access to external capital markets. Theoretically, if capital markets are perfect, then financial condition and investment decisions are unrelated. And, corporate liquidity becomes irrelevant. In contr ast, if financial frictions do exist, then managing cash holdings may be an important issue for firms. I find that firms in relationships simultaneously have lower financing frictions and higher cash holdings than firms which are not in relationships. While suppliers in relationships have lower financial constraints they are still subject to some financial constraints. Therefore, for all firms, liquidity management remains an important consideration. These findings are further related in two ways. First firms with lower financing frictions face a smaller wedge between internal and external funds. Therefore, when necessary, firms that face lower financing frictions should find it easier to access external capital markets which justifies holding less cash But, despite this, firms in relationships still hold more cash. This offers further evidence that the presence of a relationship drives firms to hold even more cash, all else equal, relative to firms which are not in relationships. Second, arguably, f irms in relationships may appear to have lower financial constraints because they invest in new projects out of their current cash holdings rather than cash flow resulting in lower investment -cash flow sensitivity. However, firms will only appear to have lower financial constraints if they are willing to spend the cash that they have on hand. Therefore, changes in cash holdings not the level of cash holdings contribute to lower investment -cash flow sensitivity. Further, the claim is not supported by the evidence. First,

PAGE 107

107 there is persistence in the difference between the cash holdings of firms in relationships relative to those not in relationships. If suppliers in relationships fund new investments with current cash holdings then cash holdings would be depleted over time, this is not the case. Next, the cash holdings regressions control for capital expenditures and the difference between supplier in relationships and those not in relationships remains. Therefore, we can conclude that firms in relat ionships have lower financing constraints and that this is the result of a feature of the relationship, not additional cash holdings. Future work will confirm the difference in financial constraints while controlling for the difference in cash holdings by evaluating suppliers propensity to save cash out of cash flows as described by Almeida, Campello, and Weisbach (2004).

PAGE 108

108 LIST OF REFERENCES Acharya, Viral V., Heitor Almeida, and Murillo Campello, 2007, Is cash negative debt? A hedging perspective on corporate financial policies, Journal of Financial Intermediation 16, 515 554. Allen, Jeffrey W. and Gordon Phillips, 2000, Corporate equity ownership, strategic alliances, and product market relationships, Journal of Finance 55, 2791 2815. Almeida, Heit or, Murillo Campello, and Michael Weisbach, 2004, The Cash Flow Sensitivity of Cash, Journal of Finance 54, 1777 1804. Biais, Bruno, and Christian. Gollier, 1997, Trade Credit and Credit Rationing, Review of Financial Studies 10, 903 937. Baker, George Robert Gibbons, and Kevin Murphy, 2002, Relational Contracts and the Theory of the Firm, Quarterly Journal of Economics 117, 39 84. Banerjee, Shantanu, Sudipto Dasgupta, and Yungsan Kim, 2008, Buyer -Supplier Relationships and the Stakeholder Theory of Capital Structure, Journal of Finance 58, 2507 2552. Bates, Thomas W., Kathleen M. Kahle, and Rene Stulz, 2009, Why do U.S. firms hold so much more cash than they used to? Journal of Finance forthcoming. Berger, Philip, and Eli Ofek, 1995, Diversifica tions Effect on Firm Value, Journal of Financial Economics 37, 39 65. Billett, Matthew, Mark Flannery, and Jon Garfinkel, 1995, The Effect of Lender Identity on a Borrowing Firms Equity Return, Journal of Finance 50, 699 718. Brown Fee and Thomas, 2009, Financial leverage and bargaining power with suppliers: Evidence from leveraged buyouts, Journal of Corporate Finance 15, 196211. Cleary, Sean, 1999, The Relationship between Firm Investment and Financial Status, Journal of Finance 54, 673 692. Com ment, Robert, and Gregg Jarrell, 1995, Corporate Focus and Stock Returns Journal of Financial Economics 37, 67 87. Diamond, Douglas, 1984, Financial Intermediation and Delegated Monitoring. Review of Economic Studies 51, 393 414. Denis, David, Diane D enis, and Keven Yost, 2002, Global Diversification, Industrial Diversification, and Firm Value Journal of Finance 57, 1951 1979. Dimitrov, Valentin, and Sheri Tice, 2006, corporate diversification and credit constraints: real effects across the business cycle, Review of Financial Studies 19, 14651498.

PAGE 109

109 Dittmar, Amy, and Mahrt -Smith, J., 2007, Corporate governance and the value of cash holdings, Journal of Financial Economics 83, 599 634. Emshwiller, John R., Aug 16, 1991, Suppliers Struggle to Improve Q uality as Big Firms Slash Their Vendor Rolls, Wall Street Journal p. B1. Faulkender, Michael and Rong Wang, 2006, Corporate financial policy and the value of cash, Journal of Finance 61, 1957 1990. Fazzari, Steven, Glenn Hubbard, and Bruce Petersen, 1988, Financing Constraints and Corporate Investment, Brookings Papers on Economic Activity 1, 141 195. Fama, E.F., MacBeth, J.D., 1973, Risk, return, and equilibrium: Empirical tests, Journal of Political Economy 81, 607 636. Fee, Edward, Charles Hadlock and Joshua Pierce, 2008, Investment, Financing Constraints, and Internal Capital Markets: Evidence from the Advertising Expenditures of Multinational Firms, Review of Financial Studies forthcoming. Fee, Edward, Charles Hadlock, and Shawn Thomas, 2006, C orporate equity ownership and the governance of product market relationships, Journal of Finance 61, 1217 1251. Fee, Edward and Shawn Thomas, 2004, Sources of gains in horizontal mergers: Evidence from customer, supplier and rival firms, Journal of Financial Economics 74, p. 423 460. Foley, C. Fritz, Jay C. Hartzell, Sheridan Titman, and Garry Twite, 2007, Why do firms hold so much cash? A tax -based explanation, Journal of Financial Economics 86, 579 607. Froot, Kenneth A, David A. Sharfstein and Jere my C. Stein, 1993, Risk management: coordinating corporate investment and financing policies The Journal of Finance 12.16291658. Gaspar, J.M. and M. Massa, 2006, Idiosyncratic volatility and product market competition, Journal of Business 79, 3125 315 2. Gertler, Mark, and Simon Gilchrist, 1994, Monetary Policy, Business Cycles, and the Behavior of Small Manufacturing Firms, Quarterly Journal of Economics 109, 309 340. Gertner, Robert, David Scharfstein, and Jeremy Stein, 1994, Internal versus Externa l Capital Markets, Quarterly Journal of Economics 109, 1211 1230. Gomes -Casseres, Benjamin, John Hagedoorn, and Adam Jaffe, 2006, Do Alliances Promote Knowledge Flows? Journal of Financial Economics 80, 5 33. Grossman Sanford, and Oliver Hart, 1986, T he Costs and Benefits of Ownership: A Theory of Vertical and Lateral Integration, Journal of Political Economy 94, 691 719.

PAGE 110

110 Harford, Jarrad, Sattar A. Mansib, and William F. Maxwell, 2009, Corporate governance and firm cash holdings in the U.S. Journal of Financial Economics forthcoming. Haushalter, David, Sandy Klasa, William F. Maxwell, 2007, The influence of product market dynamics on a firms cash holdings and hedging behavior, Journal of Financial Economics 84, 797 825. Hoshi, Takeo, Anil Kashyap, and David Scharfstein, 1991, Corporate Structure, Liquidity, and Investment: Evidence from Japanese Industrial Groups, Quarterly Journal of Economics 106, 33 60. Hertzel, Michael, Zhi Li, Micah Officer, and Kimberly Rodgers, 2008, Inter -Firm Linkages a nd the Wealth Effects of Financial Distress along the Supply Chain, Journal of Financial Economics 87, 374 387. Holmstrom, Bengt and John Roberts, 1998, The boundaries of the firm revisited, The Journal of Economic Perspectives 12, 7394. Houston, Joel Christopher James, and David Marcus, 1997, Capital Market Frictions and the Role of Internal Capital Markets, Journal of Financial Economics 46, 135 164. James, Christopher, 1987, Some Evidence on the Uniqueness of Bank Loans, Journal of Financial Economics 19, 217 235. Jensen, M., 1986, Agency costs of the free cash flow, corporate finance and takeovers, American Economic Review 76, 323 329. Kale, Jayant, and Husayn Shahrur, 2007, Corporate Capital Structure and the Characteristics of Suppliers and Cu stomers, Journal of Financial Economics 83, 321 365. Kalwani, Manohar, and Narakesari Narayandas, 1995, Long Term Manufacturer Supplier Relationships: Do They Pay Off for Supplier Firms? Journal of Marketing 59, 1 16. Kaplan, Steven, and Luigi Zingales 1997, Do Investment Cash -Flow Sensitivities Provide Useful Measures of Financing Constraints? Quarterly Journal of Economics 169 215. Kashyap, Anil, Owen Lamont, and Jeremy Stein, 1994, Credit Conditions and the Cyclical Behavior of Inventories, Quarte rly Journal of Economics 109, 565 592. Klasa, Sandy, William F. Maxwell, and Hernan Ortiz Molina, 2009, The strategic use of corporate cash holdings in collective bargaining with labor unions, Journal of Financial Economics forthcoming. Keynes, J.M., 19 36. The General Theory of Employment. In: Interest and Money. Harcourt Brace, London. Lamont, Owen, 1997, Cash Flow and Investment: Evidence from Internal Capital Markets, Journal of Finance 52, 83 109.

PAGE 111

111 Lang, Larry, and Rene Stulz, 1994, Tobins q, Corpo rate Diversification, and Firm Performance, Journal of Political Economy 102, p 1248 1280. Lee, Hau, Oct. 1, 2004, The Triple -A Supply Chain, Harvard Business Review 2 12. Lerner, Alfred, 1934, The concept of monopoly and the measurement of monopoly p ower, The Review of Economic Studies 1, 57 175. Lerner, Josh, Hilary Shane, and Alexander Tsai, 2003, Do Equity Financing Cycles Matter? Evidence from Biotechnology Alliances, Journal of Financial Economics 67, 411 446. Lummer, Scott, and John McConnel l, 1989, Further Evidence on the Bank Lending Process and the Capital Market Response to Bank Loan Agreements, Journal of Financial Economics 25, 99 122. Macaulay, Stewart, 1963, Non-Contractual Relations in Business: A Preliminary Study, American Sociol ogical Review 28, 55 67. Michaels, Daniel, and Lynn Lunsford, August 8, 2008, Lack of Seats, Galleys Delays Boeing, Airbus, Wall Street Journal p. B1. Mikkelson, W., Partch, M., 2003, Do persistent large cash reserves hinder performance? Journal of Fina ncial and Quantitative Analysis 38, 275 294. Mulligan, C.B., 1997, Scale economies, the value of time, and the demand for money: longitudinal evidence from Firms, Journal of Political Economy 105, 1061 1079. Myers, Stewart, and Nicholas Majluf, 1984, Corporate Financing and Investment Decisions when Firms Have Information that Investors Do Not Have, Journal of Financial Economics 13, 187 221. Nilsen, Jeffrey, 2002, Trade Credit and the Bank Lending Channel, Journal of Money, Credit and Banking 34, 226 253. Opler, Tim, Lee Pinkowitz, Rene Stulz, and Rohan Williamson, 1999, The determinants and implications of corporate cash holdings, Journal of Financial Economics 52, p 3 34. Petersen Mitchell, 2009, Estimating standard errors in finance panel data Set s: Comparing approaches, Review of Financial Studies 22, 435 80. Petersen, Mitchell, and Raghuram Rajan, 1997, Trade Credit: Theories and Evidence, Review of Financial Studies 10, 661 691. Pinkowitz, Lee Rohan Williamson, 2006, What is the market value of a dollar of corporate cash, Journal of Applied Corporate Finance 19, 7481 Rauh, Joshua, 2006, Investment and Financing Constraints: Evidence from the Funding of Corporate Pension Plans, Journal of Finance 61, 33 71.

PAGE 112

112 Shin, Hyun Han, and Rene Stulz, 1998, Are Internal Capital Markets Efficient? Quarterly Journal of Economics 113, 531 552. Smith, Janet, 1987, Trade Credit and Informational Asymmetry, Journal of Finance 42, 863 872. Stein, Jeremy 1997, Internal Capital Markets and the Competition for Corporate Resources, Journal of Finance 52, 111 133. Titman, Sheridan, 1984, The Effect of Capital Structure on a Firms Liquidation Decision, Journal of Financial Economics 13, 137 151. Titman, Sheridan, and Roberto Wessels, 1988, The Determinants of Capital Structure Choice, Journal of Finance 43, 1 19. Zingales, Luigi, 2000, In Search of New Foundations, Journal of Finance 55, 1623 1653.

PAGE 113

113 BIOGRAPHICAL SKETCH Since graduating from Nova High School in 1997, Jennifer Itzkowitz has spent her time to date studying at the University of Florida. Along the way, she earned a Bachelor of Science with high honors and a Master of Science with majors in mathematics Requirements for the degree of Doctor of Philosophy in finance were completed at the Unive rsity of Florida during the summer of 2009. Upon graduation from the University of Florida, she will join Seton Hall University in New Jersey as an assistant professor in finance. Jens research interests include corporate finance, financial intermediation international finance, and banking.