Title: Convertible bond issues and equity price impacts
CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00102755/00001
 Material Information
Title: Convertible bond issues and equity price impacts
Physical Description: Book
Language: English
Creator: Kuhlman, Bruce Robert, 1949-
Copyright Date: 1988
 Record Information
Bibliographic ID: UF00102755
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: ltuf - AFM0874
oclc - 20063449

Full Text

CONVERTIBLE BOND ISSUES AND
EQUITY PRICE IMPACTS












By

BRUCE ROBERT KUHLMAN


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

1988


~I
~-~T~4B: ~ El I, ae~;~~ it S





To my wife, my parents, and Carolyn and Allyn





ACKNOWLEDGMENTS


Special thanks are due Professor Robert C. Radcliffe

who chaired my dissertation. His guidance and support

ensured the successful completion of this dissertation. I

also thank Professor David T. Brown, and Professor James T.

McClave for their assistance. Thanks are not enough for my

wife, Jacalyn, for her loving support over these years.


iii





TABLE OF CONTENTS


P..age


iii

vi


ACKNOWLEDGMENTS ...................

ABSTRACT . . . . . . . . . . . .


CHAPTER


ONE

TWO


INTRODUCTION ................

LITERATURE REVIEW ..............


Common Stock and Straight Bonds.....
Convertible Bonds ............
Theoretical Valuation of Convertible Bonds

HYPOTHESES . . . . . . . . .

Announcement Date ............
Issue Date . . . . . . . .
Notes . . . . . . . . .

DATA . . . . . . . . . .

Discussion of the Data..........
Notes . . . . . . . . .


THREE


FOUR


FIVE


EMPIRICAL DESIGN ............


Event Study Methodology.......
Announcement Date Tests of HIypotheses
Issue Date Tests of Hypotheses....

RESULTS . . . . . . . . .

Announcement Date Test Results ....
Issue Date Test Results .......
Summary of Results ..........
Additional Tests ...........
Single vs. Multi ..........
Announcement Date Tests .....
Issue Date Tests .........
CP Ratios . . . . . . .
Notes . . . . . . . .


.


SIX


.. 80
.. 88
. 99
.. 104
.. 105
.. 106
.. 113
.. 118
.. 125





SEVEN SUMMARY, CONCLUSIONS,
AND FUTURE RESEARCH ..........160

Summary and Conclusions . .. .. 160
Announcing Convertible Bond Issues .. 160
Selling Convertible Bond Issues .. .. 163
Additional Tests . .. .. .. ... 165
Future Research . . .. .. 167

REFERENCES ** ** ** ** ** ** ** *. . . 170

BIOGRAPHICAL SKETCH . ... . ** *. . .. 172





Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



CONVERTIBLE BOND ISSUES AND EQUITY PRICE IMPACTS


By


Bruce Robert Kuhlman


December 1988


Chairman: Dr. Robert C. Radcliffe
Major Department: Finance, Insurance, and Real Estate



That the announcement and sale of convertible bonds

elicit negative equity price impacts is an established

fact. The primary objective of this dissertation is to

determine those factors, characteristics of the firm and the

issue, which contribute to this impact.

Equity price impacts are studied at both the announce-

ment and issue dates. It is hypothesized that the degree to

which the market values the convertible issue as equity will

determine the extent of the equity impact. As a proxy for

the equity component of the convertible, the CP ratio is





calculated. This ratio, the conversion price divided by the

current stock price, measures the sensitivity of the market

value of the convertible to the underlying stock.

At both the announcement and issue dates, firms that set

the lowest CP ratios experience the greatest (most negative)

equity impacts. Although the results seem to support the

null hypothesis, tests show that the CP ratio might not be

capturing the equity component of the bonds.

To determine whether the CP ratio is actually masking

the effects of other variables, several firm and issue char-

acteristics are tested. When the firms are divided into

three strata by CP ratio, two variables, issue size and days

between announcement and issue dates, are found to be sig-

nificantly different across the strata.

A significant difference is observed between the issue

date equity price impacts of firms selling their only issue

of convertible bonds and those with outstanding previously

issued convertible bonds. These are called single and multi

firms, respectively. There is also evidence of a possible

relationship between firm size and relative issue size and

equity price impacts for multi firms.

The findings of this dissertation provide evidence that

equity price impacts at the announcement and subsequent sale

of convertible bonds may be related to characteristics of

the convertible issue and/or the issuing firm.


vii





CHAPTER ONE
INTRODUCTION


This dissertation examines abnormal stock price behav-

ior associated with the announcement and subsequent sale of

convertible bonds. Empirical research dealing with the im-

pacts which new securities have on the price of existing

shares has documented three common findings:


(1) The announcement and sale of a straight debt securi-
ty has little or no impact on firm equity value. For exam-
ple, Smith (1986), in a survey of research published
through 1986, reported an insignificant average two-day ab-
normal return of -0.26% at the announcement of an issue of
straight debt. Barclay and Litzenberger (1986) report nega-
tive but insignificant abnormal returns during the announce-
ment day. These examples are typical of the empirical re-
search associated with straight debt securities. Studies
have shown consistently that the announcement of straight
debt causes negative but insignificant effects on outstand-
ing share prices.

(2) The announcement and sale of common equity causes a
relatively sizable negative impact on firm equity value.
Smith (1986) reported an average two-day abnormal return of
-3.14%. This is confirmed by Barclay and Litzenberger who
find that stock prices fall, on average, 1.5% in the first
15 minutes following an equity sale announcement and be-
tween 2.7% and 3.0% during the three hour period surround-
ing the announcement. The consistent results of studies
dealing with common equity are significant negative abnor-
mal stock returns at both the announcement and issue dates.

(3) The announcement and sale of convertible bonds
causes an impact between those of straight debt and common
equity. Smith (1986) reports a statistically significant
average two-day abnormal return of -2.07%.





Theories which have been developed to explain such

price reactions can be broken into two general categories:

(1) leverage change, market inefficiency, and price pres-

sure hypotheses based on symmetric information between man-

agement and investors, and (2) hypotheses based on asymmet-

ric information.



Symmetric Information



Leverage Hypotheses

The various leverage hypotheses are restricted solely

to changes in capital structure. Investment decisions are

given and the offering conveys no information about profit-

ability of existing or future assets. Empirical tests of

leverage hypotheses, then, would be limited to security of-

ferings which are classified as refundings of outstanding

issues.

If a new security offering has the potential to change

financial leverage and is as least partially unexpected by

investors, its announcement can cause various effects on

firm equity value depending upon the theory to which one

subscribes. These theories can be classified as: (1) Tax

Advantage of Debt, (2) Optimal Capital Structure, and (3)

Redistribution of Wealth.





Tax advantage of debt

The tax advantage of debt hypothesis is based upon the

analyses of Modigliani and Miller (1963). According to

Modigliani and Miller, the effects of new debt or equity is-
sues will depend upon the application of the proceeds. An

issue which increases total interest tax shield will in-

crease share prices, while the opposite is true for a tax

shield decreasing issue. For example, an issue of common

equity sold to retire straight debt has the effect of reduc-

ing the firm's interest tax shield. Refunding outstanding
debt by selling a comparable issue of debt has no impact on

tax shields. And, repurchasing equity shares with funds

acquired through the sale of bonds increases tax shields.

The effect of selling convertible bonds, then, should

also depend upon the net effect on interest payments.

Since convertible bonds carry reduced coupons and are rare-

ly used as refinancing tools, their announcement date im-

pact is difficult to predict according to the Tax Advantage

of Debt Hypothesis. However, Mikkelson (1981 and 1985) has
examined stock returns associated with the call of an issue

of convertible bonds. Since the motivation for calling con-
vertible bonds is most often to force their conversion to

equity, the action will usually decrease the net interest

tax shield. Although the direction of stock price move-

ments was as would be predicted, Mikkelson was unable to

attribute the negative call date equity impact to loss of
tax shield.





Optimal capital structure

This hypothesis suggests that all unanticipated securi-

ty offerings should have a positive effect on share

prices. If (1) an optimal capital structure exists, (2)

management always acts in the interest of shareholders, and

(3) there is symmetry of information between managers and

shareholders, any new security offering could be viewed as

movement towards the optimal structure. Since sale of the

new issue is assumed to contain no information, this hypoth-

esis would imply that only issue size, not type, would af-

fect share prices.



Redistribution of wealth

The redistribution hypothesis is based upon option pric-

ing theory and suggests that new equity issues reduce de-

fault risk and transfer a portion of firm asset value from

stockholders to debtholders. Issuing debt increases de-

fault risk and shifts wealth from bondholders to stockhold-

ers. It is implied, then, that new equity issues will de-

crease stock prices, while new debt issues will have the op-

posite effect.

Since, by definition, convertible bonds are (usually)

convertible into common equity, they have characteristics

of both debt and equity. If the bonds increase (decrease)

the firm's debt-equity ratio, their sale will be met with





increased (decreased) share prices. Without the ability to

accurately measure the firm's debt ratio before and after

the sale, the exact effect of selling convertible bonds is

very difficult to predict.



Testing the application of the leverage hypotheses to

the sale of convertible bonds is very difficult. First,

they are rarely used solely for realigning capital struc-

ture, so both investment and capital structure impacts

might be present at the announcement of a proposed issue of

convertible bonds. Also, convertible issues have both im-

mediate and future impacts on firm debt structure which can

not be predicted at the announcement or sale of the issue.

For example, the future date at which the issue will be con-

verted and financing decisions at that time are unknown.



Market Inefficiency

One view of the negative abnormal returns associated

with common equity and convertible issues is that investors

consider a new issue as a statement by management that the

security is overvalued. By issuing securities at inflated

values, management effectively transfers wealth from new

shareholders to existing shareholders.

There are two reasons why investors might believe that

management thinks the security is overvalued: (1) manage-

ment has privileged information and, thus, knows better





what the security is worth, and (2) management and share-

holders have symmetric information, but management is some-

how better able to value the firm's debt and equity. It is

the latter that is considered market inefficiency, but it

is difficult to accept that managers can consistently as-

sess firm values more accurately than the market when they

hold the same body of information.



Price Pressure

The sale of new securities is often said to have a neg-

ative price impact due to "price pressure" associated with

the issue. Price pressure can arise from either a downward

sloping demand curve or from transactions costs.

If a security has a downward sloping demand curve, ad-

ditional securities can only clear at increasingly lower

prices. A downward sloping demand curve can be the result

of differences in information among investors in a complete

market or an incomplete market with symmetry of informa-

tion. Regardless of the cause, if perfect substitutes for

a firm's securities do not exist, the issuing firm will

face a downward sloping demand curve for its securities.

High grade (straight) debt securities are generally

thought to have the greatest number of substitutes and,

thus, the least price pressure, while common stock is said

to have fewer substitutes and more price pressure. Convert-

ible bonds are difficult to assess, but the relatively few





convertibles outstanding and the high costs of replicating

convertible payoffs with the firm's common stock and debt

suggest the number of substitutes is small. This would sug-

gest that price pressure associated with convertible bonds

is considerable. The convertible bond is a hybrid securi-

ty, however, and might add to the "completeness" of the mar-

ket. This characteristic might be viewed as attractive by

investors who would be willing to pay a premium for the con-

vertible. The degree to which this would offset the price

pressure associated with the issue makes it difficult to as-

sess the net effect of selling convertible bonds using the

downward sloping demand curve hypothesis.

Price pressure can also arise from price inducements

which might be given to initial purchasers. These induce-

ments are considered transactions costs associated with the

sale of securities and are commonly thought of as an offset

to the costs which buyers incur in adjusting their portfo-

lios to optimal positions.

The distinguishing difference between these two forms

of price pressure is that negative equity impacts caused by

downward sloping demand curves are permanent while price in-

ducements cause only temporary price declines. It is impor-

tant to note that price pressure impacts should affect only

the security being sold. Convertible bonds, with their

debt and equity components, are difficult to assess using a

price pressure argument.





Asymmetric Information


The leverage, market inefficiency, and price pressure

hypotheses for equity price impacts of new security offer-

ings are all based on forces other than asymmetric informa-

tion between investors and firm managers. That is, no new

information is conveyed by the unanticipated sale of new se-

curities. Recent theoretical work, however, has investigat-

ed the possible role of management signalling in security

offerings. This literature is not yet well developed, but

three information hypotheses have become popular: (1) cash

flow signalling, (2) asset value signalling, and (3) good-

poor firm signaling.



Cash Flow Signalling

The cash flow hypothesis was first suggested by Miller

and Rock (1985). They assume symmetric information about

the level of the firm's planned investments and the firm's

value (conditional on the firm's internal cash flow), but

asymmetric information about the firm's cash flows. Their

analysis suggests that any unanticipated security offering

conveys unfavorable information about the firm's internal

cash flow and will result in decreased equity prices. Un-

anticipated announcements of new security sales signal that

internal cash flows are insufficient to finance a known





investment budget. Thus, announcements of debt, equity,

and convertible securities all provide negative information

and cause negative stock price reactions. An implication

is that the larger the issue size, regardless of type, the

greater (more negative) the equity impact.



Asset Value Signalling

This hypothesis is based on the work of Myers and

Majluf (1984). They suggest that management and investors

have asymmetric information about the value of the firm's

assets in place and potential new investments. A principal

result of their model is that management might not to ac-

cept all positive NPV projects if new securities need to be

issued to do so. This can occur if benefits which existing

stockholders would receive from investing in the projects

are insufficient to offset losses in stock values caused by

negative information released when a new security offering

is disclosed to the public. An implication of their model,

however, is that equity sales would cause a larger negative

equity impact than a straight debt issue. This would imply

that the larger the equity component inherent in the con-

vertible bond, the greater the negative stock price reac-

tion.



Good-Poor Firm Signalling

In a recent working paper Narayanan and Mishra (1987)





present a different manner in which information asymmetries

might affect stock prices. This will be referred to as the

Good-Poor Firm signalling hypothesis. Their model is ad-

dressed directly to the issue of convertible securities and

is driven by the option which management has to decide the

call date of the security.

They assume that all convertibles are issued with a

call provision. (Although not strictly true, virtually all

convertible bonds are callable at the firm's option with

very little or no call protection.) Managers are assumed

to operate either good or bad firms. A good firm is one in

which managers have relatively good knowledge about the po-

tential outcomes from current and future investments. A

poor firm, on the other hand, is one in which managers have

more diffuse knowledge about investment outcomes. As a re-

sult, managers of poor firms prefer financing alternatives

which provide them with a device by which they can signal

information to investors in the future. In their model, a

call conveys negative information and a decision not to

call conveys either no information or good information.

The critical aspect of their paper is their suggestion that

the issue of a convertible is in itself a signal that man-

agement desires the ability to provide future signals via

the call feature. That is, the sale of a convertible is a

signal that the firm is a "poor firm" and the stock price

will fall at the time of issue.





11

The intent of this study is to examine the empirical

validity of the above information asymmetry hypotheses.

Both the Asset Value Signalling and the Good-Poor Firm

Signalling hypotheses would suggest that abnormal returns

will be positively related to the size of the equity compo-

nent in a new convertible bond issue. In addition, these

hypotheses would predict larger (more negative) returns for

larger issues. In contrast, the Cash Flow Signalling hy-

pothesis would predict only a relationship between negative

abnormal returns and issue size. The relative size of the

equity component is irrelevant.

The measurement of the equity and debt components inher-

ent in a typical convertible bond is a theoretically and em-

pirically difficult problem. Conceptually the investor pur-

chases a portfolio of the following securities: (1) a

straight debt security with fixed coupons and maturity

date, plus (2) an option to purchase shares by tendering

the straight debt security, less (3) a call option sold

back to the firm which allows management to determine the

exercise date of the conversion option.

The straight debt value would be the present value of

the coupons expected between the issue date and expected

call date and the present value of the expected call

price. The discount rate would reflect yields on bonds of

similar default risk having a maturity equal to the

expected call date. Of course, determining the market's





12

perception of the expected call date is an empirically dif-

ficult task.

In short, an explicit valuation for the debt and equity

components of a convertible bond is presently not possible

with any degree of accuracy. As a proxy for the relative

amount of equity inherent in the convertible bond, the con-

version price is divided by the market price of the issuing

firm's common equity. This will be referred to as the CP

ratio throughout the dissertation. The ratio is basically

a measure of how far out of the money the convertible's

call option is at the announcement or issue date. As such,

it is an approximation of how sensitive the convertible's

price is to the underlying stock. All else equal, the high-

er the CP ratio, the less sensitive the call's value to

changes in stock prices.

The CP ratio can be interpreted in a slightly different

manner in the case of the Good-Poor Firm Signalling hypoth-

esis. In that case, the lower the CP ratio, the earlier

management's ability to use the information value of a

call/no call decision. As such, the lower the CP ratio,

the more likely the firm is to be a Poor firm. With either

of the interpretations, the suggestion is that the lower

the CP ratio, the greater (more negative) the equity price

impact.

For each convertible in the respective sample, a CP

ratio will be calculated at both the announcement and issue





13

dates. In each case, the closing stock price two days pre-

ceding the event and the conversion price at issue are

used. It is anticipated that those firms setting low CP

ratios will experience the greatest equity price impact.





CHAPTER TWO
LITERATURE REVIEW


This study deals with issuing convertible bonds and the

accompanying impact on the issuing firm's stock price. To

understand this impact one must study not only the litera-

ture dealing with convertible bonds, but also the litera-

ture dealing with other forms of financing. This chapter

is divided into a brief review of the literature dealing

with the impacts of issuing common stock and straight debt

and a somewhat more detailed review of the literature deal-

ing with the impacts of issuing convertible bonds. These

sections are followed by a theoretical discussion of the

valuation of convertible bonds.



Common Stock and Straight Bonds



The reaction of equity prices to the issuance of new

securities is an area of Finance that has received consid-

erable attention. Masulis (1980) looked at the equity

price reactions to security exchanges, the issuance of one

security to retire another. Of particular interest to this

study are his findings dealing with debt for equity










and equity for debt exchanges. Masulis found that when

debt was issued to retire equity, stockholders benefited

with positive residuals of over 10.5%. However, where

equity was exchanged for debt, the resulting residuals were
negative and significant at around -7.5%. Masulis inter-

preted the results as support for the theory that equity
value is a positive function of leverage. Studies of lev-

erage increasing transactions which lend support to
Masulis' 1978 findings are studies by Dann (1981), and

Vermaelen (1981) who found residuals of 15.41%, and 14.14%,

respectively, for stock repurchases, and Dann and Mikkelson

(1984) and Eckbo (1986) who find insignificant residuals of
-.37% and -.06%, respectively, with the issuance of

(straight) debt. Korwar (1982), Asquith and Mullins

(1983), and Asquith and Mullins (1986) find negative
residuals of about -2.5%, -3%, and -3.14%, respectively,

with the sale of equity which is a leverage decreasing
transaction. This abundance of evidence would seem to

indicate that Masulis (1983) was correct in suggesting that

residuals experienced by the issuing firm's stockholders

are a positive function of the change in leverage.

Several studies, including some of the above, offer

results that seem to dispute the conclusion that equity

impacts are due to changes in leverage. Eckbo (1986) looks
at offering size, tax shield change, rating, method of issu-

ance, and possible abnormal earnings experienced before





16

and after an issue of debt is sold in an attempt to explain

price behavior witnessed with debt offerings, but finds

none significant. He interprets this as a direct contra-

diction to Masulis (1983). Mikkelson and Partch (1986)

test issue rating and size, type of security issued, and

relative amount of new offering for refundings and find

only security type to be significant. Rogers and Owers

(1985) study equity for debt exchanges and five possible

causes for the negative response noted in equity returns.

After ruling out wealth transfer, tax shield loss, etc.,

they suggest that the only remaining possible cause is

information contained in the transaction. Vermaelen

(1981), as noted before, found positive residuals with the

repurchase of equity. However, since the firm usually ten-

dered for shares in the market at a premium, and a large

percentage of the premium remained in the stock price long

after the tender expired, he concluded that the tender

offer actually signalled to the market that the firm's

shares were undervalued. This signalling of "inside" infor-

mation was offered as reason for the positive residuals

with the repurchase. This is in agreement with the sugges-

tions of Myers and Majluf (1984) who suggest that firms

will offer securities for sale when management feels the

securities are overvalued in the market and purchase the

securities if they are undervalued.

According to Miller and Rock (1985), as the issuance of

securities becomes more regular, the reaction in the market





will become less severe. That is, if financing is unantic-

ipated, equity prices will drop with its announcement.
Even if the security issue is anticipated but the actual

amount issued is greater than expected, equity prices will

drop. In essence Miller and Rock suggest that a larger
than expected issue signals lower than expected earnings.

The findings of Eckbo (1986) and Mikkelson and Partch

(1986) that type of security offered, not offering size,
seems to affect the reaction in equity returns would,
however, tend to contradict Miller and Rock.

From the above studies it would seem that capital struc-

ture changes, size of issue, loss of tax shield, method of

issue, quality rating of issue, and transactions costs all

fall short of explaining the reaction in equity returns

when the firm announces the need for external funds through

the issuance of new securities. Only the type of security

which is issued remains a significant variable. And, it

has been demonstrated repeatedly that issuing equity causes

significant negative residuals while the residuals at the

issuance of debt are not statistically differently from
zero.



Convertible Bonds


Whether the negative reaction in equity returns from

the issue of common stock is caused by leverage change or









18

signalling, the fact remains that the issuance of new

equity is usually met with negative equity price reac-

tions. Issuing straight debt, as many studies have shown,

seems not to be an economically important event. It would

seem reasonable that convertible bonds, seen as some combi-

nation of debt and equity, could cause a reaction somewhere

in between those of debt and equity, and studies have shown

this to be the case.

Dann and Mikkelson (1984) study the impact on equity

returns for 132 convertible debt issues between 1970 and

1979. Treating the convertible bonds as 100% debt and

using market value for equity and book value for debt, they

measure the two-day abnormal returns around announcement

date and issue date to find the correlation between lever-

age change caused by the convertible debt issue and the

sign of any residuals. They reason that in every case the

convertible must have increased leverage, but in every case

the residuals are negative. The average two-day announce-

ment date residual for convertible debt issues was -2.31%

and significant at the 0.01 level. Of their sample of 132,

29 had positive residuals, and the remaining 103 had nega-

tive residuals. With the straight debt offerings from 150

firms during the same time period, they found insignificant

residuals of -0.37% at announcement date. At the issue

date they find two-day abnormal returns of -1.54% for con-

vertible debt, and, again, reject the null hypothesis at





19
the 0.01 significance level. The abnormal returns for

straight debt at issue date are insignificant. Also, for

the issues of convertible debt, 32 had positive abnormal

returns and 97 had negative, while the distribution for

straight debt was about half and half.

Dann and Mikkelson (1984) tested three possible expla-

nations for the response to convertible debt issuance: (1)

change in leverage, defined as book value of long term debt
divided by market value of equity, (2) new, unfavorable

information, defined as the degree to which the proceeds

are used for new financing rather than refunding existing

debt, and (3) underpricing of the issue. Their finding of
negative residuals, even though the convertible bonds

increased leverage, was interpreted as contrary to the

change in leverage hypothesis that would predict a positive
relationship between the direction of the leverage change

and the sign of the residual. Recall, however, that, in
measuring debt ratios, the authors treated the convertibles

as 100% long term debt. They argue that "the equity

component of the new convertible would have to comprise
almost two-thirds of the new financing for the average firm

in order for the issuance to reduce leverage." Whether or

not convertible bonds increase the firm's financial

leverage may not be the appropriate question, however.

The term "leverage" can take on two meanings: (1) the

tax subsidy connected with the deductibility of interest





20

payments, and (2) the signalling connotation of Myers and

Majluf (1984). If the authors are referring to the firm's

use of tax shields, the assumption is questionable. It is

obvious that convertible bonds, even with their reduced cou-

pons, have the potential to increase the firm's fixed obli-

gations. However, since the proceeds are often partially

used to repay other short and long term borrowings with, no

doubt, higher interest costs, the net change in tax shield

is uncertain without close scrutiny. It is unclear whether

the authors actually measured interest costs before and

after the convertible issue to determine any changes in tax

shield.

If the authors use debt-equity ratios to measure lever-

age in the Myers and Majluf (1984) signalling context,

again their conclusion is questionable. They ignore the

potential for complete conversion to equity at a future

date and the uncertainty about actions the firm might take

at that time if funds are needed. Also, the proceeds of

the convertibles send mixed signals of refunding existing

long or short term debt, financing capital expansion or

working capital needs, or redeeming common or preferred

stock. Perhaps it would be better to measure the relative

sizes of the debt and equity components of the convertible,

the issue's debt-equity ratio, than to look at that of the

issuing firm.

To test whether the convertible bond issues signalled

new, unfavorable information, they broke their sample into





21

those classified as debt refundings and those representing

new financing. They argue that debt refundings send no sig-

nal, while using convertibles to finance new projects,

etc., represents the unexpected need for external financ-

ing; a negative signal in the Miller and Rock (1985)

context. When the two samples experienced almost identical

announcement period average prediction errors, the authors

interpreted it as evidence that new information was not nec-

essarily responsible for the entire reaction. It might be

incorrect to assume that issuing convertible bonds to

refund debt is not a negative signal, however. Since con-

vertible bonds are recognized as containing an equity com-

ponent, this action must be seen as issuing both debt and

equity to replace the outstanding debt.

The signal contained in issuing convertible bonds to

refund existing debt, then, is mixed but probably negative

as Myers and Majluf (1984) would predict. That the authors

found no significant difference between the impacts from

issuing convertibles for new financing or to refund exist-

ing debt is not surprising.

To test the underpricing hypothesis, Dann and Mikkelson

tested for equity price reactions to the issuance of con-

vertible bonds sold through rights offerings. Their

reasoning was that underpricing of new debt securities

unnecessarily strips wealth from the stockholders and is

accompanied by negative residuals, while an issue through a





22

rights offering should not shift wealth even if the (debt)

security is underpriced. Their findings do not support

underpricing as the cause for the equity reactions, how-

ever. Further, they calculated that convertible debt would

have to be underpriced by an implausible 15% for underpric-

ing to account for the entire reaction in equity returns.

Mikkelson and Partch (1986) took a random sample of 360

firms from the Center for Research in Security Prices Daily

Returns File (CRSP) for 1972. They studied equity price

reactions to announcements by these firms of issues of com-

mon stock, straight debt, convertible debt, and preferred

stock over the period from 1972 to 1982. They found signif-

icant residuals for common stock and convertible debt of

-3.56% and -1.97%, respectively, and insignificant resid-

uals for straight debt and preferred stock. As with the

results of Dann and Mikkelson (1984), about three-fourths

of the observations for both common stock and convertible

debt were negative, while about half of the straight debt

observations were negative. The authors apply the results

to the hypotheses of Miller and Rock (1985) and Myers and

Majluf (1984). Miller and Rock (1985) suggest that unan-

ticipated financing conveys unfavorable future earnings po-

tential, and, hence, is followed by a negative reaction in

equity returns. A necessary implication of this hypothesis

is that the size of the issue will affect the size of the

equity impact. Miller and Rock do not distinguish between





23

types of securities, however. The model of Myers and

Majluf (1984), on the other hand, suggests that new financ-

ing conveys information to the market about the firm's

assets in place. Specifically, the issuance of equity con-

veys less favorable information than does the issuance of

debt. The findings of Mikkelson and Partch (1986) do not

support Miller and Rock in that offering size and net

amount of new financing are insignificant in explaining the

abnormal returns experienced at the announcement of a con-

vertible bond issue. On the other hand, they find security

type to be significant, as Myers and Majluf would predict.

Mikkelson and Partch conclude by stating that their evi-

dence is consistent with the prediction of Myers and Majluf

(1984) in that issues of common stock and convertible debt

convey less than favorable information to the market about

the firm's assets in place. Market participants infer

equity and convertible debt to be overpriced whenever the

firm issues them.

Vu (1986) studied the effects of the announcement of

issues of straight (non-convertible) and convertible debt.

He ran cross-sectional regressions to test for the effects

of size and riskiness of the issue, tax shield increase,

and use of the funds. Again, the author's results agree

with the predictions of Myers and Majluf (1984) and dis-

agree with Miller and Rock (1985). While he found no sig-

nificant relationship between any of the variables and





announcement date abnormal returns, he did find a signif-

icant difference in the reactions to straight debt and

convertible debt. Although straight debt caused no sig-

nificant reaction in equity returns, convertible debt

caused a significant abnormal return of -1.25%. Since Vu

found no relationship between increase in tax shield and

equity impacts, he interprets this as evidence that lever-

age-increasing issues do not necessarily increase stock-

holder wealth.

It has been demonstrated in many of the above tests

that the equity price impact associated with the announce-

ment of a security issue is not a function of increased or

decreased tax-shield. In each case the only significant

variable in determining announcement date equity price

impacts is the type of security offered. And, the impact

from announcing convertible bonds consistently falls

between those from announcing common equity or straight

debt.





Theoretical Valuation of Convertible Bonds



That convertible bonds contain components of both debt

and equity is widely accepted, and the above empirical

results would tend to support that belief. Further, it is

reasonable to assume that it is this mixture of debt and





25

equity that causes the announcement date reaction for con-

vertible bonds to be less than that for common equity but

more than the reaction to announcing straight debt.

However, this is only conjecture without a theoretical

basis for the assumption that convertible bonds are indeed

valued as part debt and part equity. Ingersoll (1977) pre-

sented a thorough model for the valuation of convertible

bonds and demonstrated the source of their debt and equity

components. He began by stating the value of a straight,

non-callable bond to be a function of firm value, V, matur-

ity, t, coupon, c, and final payment, B. Letting F denote

the value of a straight, non-callable bond, then



(4) F = F(V7, t; B, c).



If we let G be a non-callable bond equivalent to F in

all respects except that it is convertible into fraction,

a, of the firm's post-conversion equity, Ingersoll showed

that G can be represented as



(5) G = G(V, t; B, C, a).



Note that if G is convertible into a fraction a = 0 of

the firm's equity, G is identical to F.

If we now allow the same bond to be callable, and let H

denote the value of this callable convertible, the relation

ship can be expressed as follows:





(6) H = H(V, t; k(t), B, a)



where k(t) is the call price at time t. If k(t > 0) is

infinite, H( )is equivalent to G( ). In words, the value

of a callable convertible with infinite call price prior to

maturity is the same as a non-callable convertible with

otherwise identical characteristics. In fact, the differ-

ence in value for callable and non-callable debt arises

purely out the firm's right to call the issue. The dif-

ference in the value of straight, non-callable debt and

that of callable, convertible debt is then due to the com-

bination of the call feature and the conversion feature.

Ingersoll shows that the value of convertible debt must

be a non-decreasing function of a, the fraction of equity

into which the bond is convertible, and k(t), the price at

which the convertible is callable. Therefore, a convert-

ible issue, G( ), must be at least as valuable as a noncon-

vertible issue, F( ). And, since a non-callable issue can

be thought of as a callable with infinite call price, a

callable convertible cannot be more valuable than a noncall-

able one. That is, G( ) H( ).

Ingersoll continues by showing the value of a callable,

convertible discount bond to be





27

(7) H(V, t) = F( ) + W( ) + [Fl( ) F"2 )1
where


F( ) = The value of a non-callable, non-
convertible discount bond of matur-
ity, t, face value, B, and issued
by firm with value V. (See equa-
tion 4 for the equivalent coupon
bond.)

W( ) = The value of a warrant issued on
the same firm convertible into a
portion, a, of the firm's value.
(Ingersoll considers the firm to
have only common stock and one
issue of convertible debt outstand-
ing.)


Fl( ) and F2( ) represent the reduction in value
due to the call feature. This call feature reduction in

value implicitly considers the reduction in the value of

the bond, F( ) as well as reduction in the value of the

warrant, W( ) In other words, the value of the noncall-

able straight debt portion of the convertible is reduced by

the addition of the call feature, and the value of the war-

rant, the investor's option to convert to common equity, is

reduced by the firm's ability to determine its maturity

through calling the issue. That is, there are two compo-

nents to the security which bondholders "sell back" to man-

agement which allow management to determine both the matu-

rity of the bond and that of the option to convert.

Ingersoll shows that the value of the equity portion of

the convertible is a function of the post-conversion value





28

of the percentage of firm held if the bonds are converted

and the maturity of the "attached" warrant. The value of

this warrant is derived through the Black-Scholes model

(1973), and is shown to be a positive function of stock

price and maturity, and a negative function of exercise

price (conversion price).

Although Ingersoll does not present a closed form model

for pricing the various components of a convertible bond,

he has demonstrated the theoretical basis for valuing them

as both debt and equity. The exact valuation of the compo-

nents of the convertible bond is a topic for future re-

search.

It is the purpose of this study to demonstrate that con-

vertible bonds vary in the extent to which they are valued

as equity and, further, that announcement date equity price

impacts are a function of this equity valuation. It is hy-

pothesized that those convertibles perceived to have a rela-

tively large equity component will impact much the same as

an issue of common equity. If the market feels the conver-

sion terms will never be attractive enough to warrant con-

version, from the standpoint of the firm or the investor,

the convertible will be perceived as mostly debt and will

impact much the same as an issue of straight debt.





CHAPTER THREE
HYPOTHESES


The hypotheses tested in this study are divided into

two categories: (1) those dealing with equity price

impacts at the announcement of a firm's intent to issue con-

vertible bonds, and (2) those dealing with equity price

impacts at the date of issue. As such this chapter is

divided into two sections dealing with the respective

hypotheses.



Announcement Date

The Announcement Date Event

Investors continually gather and analyze information about

publicly held firms, and, at any point in time, some consen-

sus expectation will exist as to the probability that a giv-

en firm will issue convertible bonds. Whether changes in

this probability have an effect on equity share prices is

the subject of this dissertation.

The announcement by management of the intent to issue

convertible bonds will have no effect on equity prices if

the announcement provides no information of value to market

participants. In an efficiently priced market this has two





30

possible explanations. First, the issuance of convertible

bonds might be an economically important event, but the

announcement of such is fully anticipated by the market.

That is, knowledge of a convertible bond issue is valuable,

but the public announcement by management is not. Second,

the issue announcement might be unexpected but economically

unimportant.

While it is theoretically conceivable that a convert-

ible bond announcement could be accurately predicted by the

market, as a practical matter it is doubtful. Thus, we

will assume that the announcement does change the public's

expectations that the firm will issue convertibles. Given

this assumption, the nature of equity price movements at

the date of announcement can be used to infer the type of

information contained in the announcement. For example, if

share prices remain unchanged at the date of public

announcement, we can infer that the announcement contains

no economically important information. However, if on

average prices fall by a statistically significant amount,

the announcement contains negative information.

Investors become aware of the announcement at the actu-

al announcement date through various wire services, compu-

ter networks, news services, etc. However, the announce-

ment might be made during or after market trading hours,so

any response to the announcement can take place on that day

or on the following day when the announcement appears in





31

the Wall Street Journal or other financial news media. To

capture any equity price impact on either of these days,

this study utilizes a two-day window consisting of the date

the issue is announced in the Wall Street Journal and the

preceding day.



The Announcement

The type of information disclosed at the announcement

date is important since this could have implications about

possible price impacts at the date of issue or withdrawal.

As part of this study over 200 convertible bond announce-

ments were read in the Wall Street Journal. Virtually all

of these announcements contained only the following type of

information: (1) that the firm intended to sell convert-

ible bonds, (2) the approximate size of the issue, (3) the

name of the underwriterss, and (4) a period during which

the issue might be sold. The proposed use of the proceeds

was occasionally mentioned, but only in very broad terms

such as "general corporate needs."

Certain potentially important information is not reveal-

ed in a convertible bond announcement, however. In partic-

ular, the issue's conversion terms, coupon rate, maturity,

and call features are not available until the issue date.

Expectations about such terms can be formed at the announce-

ment date given the market's knowledge of past convertible

offerings, current capital market conditions, the financial





32

status of the issuing firm, etc., They will not be known

with certainty, however, until the issue is sold.



Announcement Date Impact

One purpose of this study is to investigate why convert-

ible bond sales have a negative impact on the stock price

of the issuing firm. However, as noted above, certain in-

formation is not disclosed until the actual issue date.

Thus, any price impacts observed at the announcement date

might be due to the fact that a firm intends to issue a

generic convertible bond. Evidence pointing to the true

source of the negative returns, however, might not be ob-

served until the actual issue date when all terms are pub-

licly disclosed for the first time.

For example, if a larger equity component inherent in a

convertible bond leads to a larger negative equity return,

we might not be able to observe such a relationship until

the issue date...the date when the public is told the terms

of the issue. Announcement date equity price impacts might

be due to the equity component inherent in any convertible

bond issue but be unrelated to the, as yet, undisclosed

terms of the specific issue in question.

Even the probability that the firm will actually sell

the convertible is not 1.0 at the announcement date since

the firm retains an option to withdraw the issue. Due to

the research design of this study, precise statistics on





33

announced convertible issues that are subsequently with-

drawn from sale are not available. However, security

offerings are frequently withdrawn. Mikkelson and Partch

(1986), in a randomly drawn sample of security offerings

from 1972 to 1982, find 23% (18 of 80) common stock offer-

ings cancelled, 20% (35 of 172) straight debt offerings can-

celled, and 24% (8 of 23) convertible bond offerings can-

celled. Therefore, investors must not only estimate any

potential economic impact from the firm's proposed issue of

convertible bonds, they must estimate the probability that

the issue will actually be sold.

In short, equity price impacts might be observed at

both the announcement and issue dates, but the nature of

such impacts might be different due to different infor-

mation available at each date.





Announcement Date Hypotheses

Hol: The announcement of an issue of convertible
bonds has no impact on the announcing firm's equity value.

This test will examine whether the announcement of an

issue of convertible bonds is an economically important

event. If the announcement is fully anticipated by the mar-

ket and the market efficiently interprets the information,

there should be no effect on the firm's stock price. If,

on the other hand, the announcement of convertible bonds





34

imparts new information to the market, the effect on the

firm's equity value should be reflected in the firm's price

per share.

It is anticipated that the announcement of an issue of

convertible bonds cannot be accurately predicted by the mar-

ket, and that it is an economically important event as

Myers and Majluf (1984) or Miller and Rock (1985) would

predict. As such, a significant negative average equity

price response is expected at the announcement date.



Ho2: The announcement of an issue of convertible
bonds causes no increase in the variance of daily stock

returns.

On any given day excess returns for a sample of firms

will be expected to have a zero mean and some non-zero vari-

ance due to insistently movements in stock prices. That

is, on any day the movement of a firm's stock may result in

a positive or negative excess return, but over a large sam-

ple of firms the average excess return on that day should

be zero. The release of information, however, could affect

the distribution of excess returns, and the type of infor-

mation can be a factor in the effect. For example, there

might be a consensus that on average the type of informa-

tion released is a negative signal about true equity

values. Then, most firms announcing the same type of infor-

mation will experience a negative equity price impact on





35

that date, and the average excess return for a sample of

firms making the same type of announcement will be non-

zero.

If the announcement made by each firm does not contain

exactly the same information, however, there might be vary-

ing equity price impacts. For example, consider the follow-

ing which demonstrates that the return for any given firm

on any given day depends on some normal flow of information

as well as any information that might be contained in an

unexpected convertible bond announcement:



R = Rn + Rc
where

R = the total return.

Rn= the return due to some "normal" flow of infor-
mation.

Ro= the return due to information contained in an
unexpected convertible bond announcement.
and
(1) (2) (3) (4)
whereVAR(R) = VAR(R,) + VAR(Rc) + 2COV(R,,Rc)

(1) = the variance of the return on any given day.

(2) = the normal variance caused by the usual flow
of information.

(3) = the variance caused by information contained
in the convertible bond announcement.

(4) = the covariance of the returns due to normal
information flow with returns caused by any
information in the convertible bond
announcement.





36

On a typical non-event day, terms (3) and (4) are zero,

and the variance of the firm's return is simply VAR(R,).

However, with the release of the convertible bond announce-

ment ,term (2) might become non-zero, and the value of

term (3) depends upon the correlation, r, of R, and

Rc. That is, if managers only make convertible bond

announcements on "good" days, rRn,Rc = -1.0. If managers

pay no attention to the normal information flow when making

these "bad" public announcements, rRn,Rc is more or less
zero, and if convertible bond announcements are released

only on "bad" days, rRn,IR = +1.0.
VAR(R) can be summed up as follows:



if rRn,Rc 0, VAR(R) > VAR(R,)

if rRn,Rc < 0, VAR(R,) < [VAR(R ) + VAR(Rc)I

In either case, as long as VAR(Rc) > 0, the variance
of the daily returns will probably increase with the an-

nouncement of a convertible bond issue. If rRn,Ro is neg-
ative, as a practical matter it is probably not -1.0. And,

even if the correlation of "normal" returns with these

"abnormal" returns was perfectly negative, term (4) will

not necessarily make VAR(R) < VAR(R,).

Firms announcing convertible bonds come from a variety

of different industries, vary greatly in size, and have var-

ying capital structures and degrees of financial soundness.





37

In that case, even if firms issued identical convertible

bonds, the information contained in the convertible bond

announcement might be valued differently from firm to

firm. And, the equity price impacts might vary across the

firms, even if they are mostly negative, causing VAR(Rc)

to be firm-specific.

When a firm announces a convertible bond issue, it

could be signalling information about its assets in place,

Myers and Majluf (1984), or its projected cash flows,

Miller and Rock (1985). Investors must analyze each firm

and issue to determine the type and value of any signal.

On average the announcement date equity impact could be neg-

ative. But, since the signal can vary greatly across a

large sample of convertible bond announcements, it is ex-

pected that the announcement of convertible bond issues

will be met with not only a negative shift in the distri-

bution of daily excess returns but also an increase in the

variance of those returns, indicating that the announcement

of an issue of convertible bonds has a significantly differ-

ent impact across firms.



Ho3: Variation in the relative values of the debt
and equity components of the convertible bond has no effect

on any impact on the announcing firm's equity value.

Using the reasoning of Myers and Majluf (1984) the type

of security issued signals something about the condition of





38

the issuing firm. Specifically, issuing equity signals

weakness. The empirical results of studies discussed in

the literature survey support Myers and Majluf. In all

cases the announcement of an equity issue is met with sig-

nificant negative abnormal returns. The announcement date

equity price impacts for firms selling straight debt are

insignificant.

The straight debt component of convertible bonds, fixed

coupons and maturity, should cause no announcement date

equity price impact. However, convertible bonds may be con-

verted into the common equity of the issuing firm at the

discretion of the holder or may be forced into equity by

the issuing firm.2 Therefore, they must be considered

part equity, and this equity component may affect the value

of the firm's outstanding common equity. In that case the

announcement date equity price impact would be directly

related to the size of the equity component.

If market participants can reasonably predict the

equity component inherent in a given announcement, then the

announcement date equity price impact will be directly re-

lated to the relative size of the underlying equity compon-

ent. Hence, those convertibles perceived as having a rel-

atively large equity component will cause a greater (more

negative) equity price reaction than issues with relatively

smaller equity components.





39

Ho4: Issue size has no effect on any equity price
impact at the announcement date.

This hypothesis tests whether issue size, measured rela-

tive to the market value of the issuing firm's equity, is

important in determining announcement date equity price

impacts. Miller and Rock (1985) maintain that, by going to

the financial markets for external funds, the firm signals

an unexpected short fall in cash flows. Any announced need

for external financing will be interpreted as negative

information by the market and should result in a negative

reaction in stock price. Myers and Majluf (1984) agree but

state further that weaker firms issue equity.

Convertible bonds contain characteristics of both debt

and equity. If we assume that the size of the equity com-

ponent is not a function of issue size, a sample of large

convertible bond issues would be expected to have a larger

average equity valuation than a sample of small issues.

On average, then, we would expect relatively larger issues

of convertibles to cause a greater announcement date equity

price impact.



Ho5: Firm size has no effect on any equity price
impact at the announcement date.

According to Miller and Rock (1985), any unexpected

need for external funds signals a short fall in internally

generated funds. Since this will be read as a negative





40

signal, the firm's equity will be analyzed and reduced in

value accordingly. Myers and Majluf (1984) maintain that a

firm's need for external financing signals that its assets

are over valued By definition, the announcement is unex-

pected and should cause reevaluation of the firm's equity.

If we assume that the unexpected component of any re-

lease of information is the cause of stock price adjust-

ments, then the size of the announcing firm could influence

the equity price impact at the convertible bond announce-

ment date. For example, information on a larger firm may

be more readily available than that on a smaller firm,

especially if the larger firm is older and/or has gone to

the markets in the past. Or the larger firm may be listed

on one of the larger exchanges. In any case, expectations

would be that actions of larger firms should be more easily

monitored.

It has already been established that the announcement

of an issue of convertible bonds is an economically impor-

tant event. By definition, then, the announcements are

unexpected signals of weakness. By the above argument, in-

formation on larger firms should be more complete, and a

signal of weakness might be more of a surprise when issued

by a larger firm. If a relationship exists between firm

size and announcement date equity price impacts, it would

be expected that the largest firms would experience the

greatest (most negative) impacts.













Issue Date



The Issue Date Event

In an efficient capital market, any impact from issuing

securities should be experienced when the issue is first

announced. Any economically important information con-

tained in the announcement should be analyzed immediately

by market participants, and the firm's stock price should

reflect any necessary adjustments instantaneously. The

date the issue is actually sold should be of no economic

importance, since selling the issue is simply the trans-

action announced previously and has already been priced by

the market. However, this makes two important assump-

tions: (1) all necessary information is contained in the

announcement, and (2) there is no uncertainty that the

announced issue will be sold.

Violation of either of these assumptions can cause an

issue date equity price impact. As noted earlier, the

announcement of a convertible bond issue might not contain

all the information necessary to price the impact of the

sale at the announcement date. The typical announcement

contains information only that the firm plans to sell con-

vertibles, the underwriterss, an estimate of when the sale

may take place, and some probable uses of the proceeds.









42

A potentially crucial item, the conversion ratio, is only

released immediately prior to the sale and, therefore, can

only be estimated at the announcement date.3 If the con-

version terms have economic importance, the equity price

impact can only be estimated at the announcement date.

Then, release of the terms at the issue date may constitute

new information and cause a reevaluation of the issue and

an adjustment in stock prices.

The second implied assumption is that investors know

with probability one that the announced issue will not be

withdrawn. If any doubt remains that the issue will be

sold, the full equity impact might not be felt at the an-

nouncement date. The market will continually revise esti-

mates of if and when the issue will be sold, but the proba-

bility of issue can never be one until the sale actually

takes place.



Issue Date Impacts

At the sale or cancellation date, then, all uncertainty

is resolved. If the issue is to be sold, the terms of the

issue are made public, and the market can properly assess

the economic importance of the particular convertible. The

firm's stock price will then be adjusted accordingly. If

the issue is withdrawn, however, the announcing firm should

experience a positive stock price reaction as found by

Mikkelson and Partch (1986).









43

Although not tested empirically, it seems reasonable to

assume that the average issue date impact for a mixed sam-

ple of firms selling and firms withdrawing convertible bond

issues might be zero. Since this study includes no issues

that were withdrawn, the average issue date impact will

only be some combination of the economic impact of releas-

ing the issue terms and resolution of uncertainty about

whether the issue would be sold. Since there will be no

firms experiencing the positive impact from withdrawing the

announced issue, it is expected that the average issue date

impact will be negative.



Issue Date Hypotheses

Hol: The sale of an issue of convertible bonds has
no impact on the issuing firm's equity value.

Any issue date equity price impact will depend upon

whether or not economically important information is releas-

ed when convertible bonds are sold. Anything that has the

capacity to affect the market's estimate of the firm's equi-

ty value can be construed as information. For example,

many of the convertible bond's terms, such as the conver-

sion rights, coupon, and maturity, are not released until

the issue is actually sold. If these terms are critical in

determining the equity price impact of issuing the convert-

ible, and if the market is not capable of precisely estimat-

ing them at the announcement date, it may be reasonable to

expect their release to constitute information.









44

Even if the market can accurately predict the terms of

the issue and the issue's impact on the market value of the

firm's equity, there remains uncertainty that the issue

will be sold. In that case, only a portion of the total

equity impact will be felt at the announcement date as the

market waits to see the final disposition of the issue. If

the issue is withdrawn, the firm might recoup all or part

of its announcement date loss as documented by Mikkelson

and Partch (1986). If, however, the issue is sold, and the

exact sale date has not been predicted by the market, the

firm could experience a further reduction in equity value.

Whether it is the release of the terms of the issue or

simply the resolution of uncertainty about the sale or with-

drawal of the issue, it is expected that economically impor-

tant information is released at the issue date, and the

average issue date equity price impact will be negative.



Ho2: The sale of an issue of convertible bonds
causes no increase in the variance of daily stock returns.

That firms on average experience a negative equity

price impact on the date convertible bonds are sold has

been established. If the noted equity impact represents

only a shift in the mean issue date excess return, we would

say that it is due to resolution of uncertainty. That is,

if firms sell convertible bonds, there is a more or less

constant equity cost that will be assessed. If, however,





45

along with a shift in the mean comes an increase in the

cross-sectional variance of excess returns, we would say

that issue date equity price impacts vary significantly

across firms. Selling an announced issue of convertible

bonds conveys different information about different firms.

There are four scenarios that might explain the nega-

tive average equity price impact at the sale of convertible

bonds: (1) There is a common equity price impact from issu-

ing convertible bonds, and the probability of issue is con-

sidered about equal across firms. The sale of the issue

only constitutes resolution of uncertainty about whether

the issue will be sold, and any issue date impact is more

or less equal across firms. The issue date reaction is a

shift in the distribution of daily excess returns. (2) The

probability of sale or withdrawal is considered constant

across firms. The equity price impact from issuing convert-

ible bonds is firm specific but accurately estimated at the

announcement date. Any impact at the issue date is the

probability of withdrawal times the true equity impact,

and, therefore, it can vary across firms. The issue date

reaction is a shift in mean and increased cross-sectional

variance. (3) The probability of selling an announced

issue of convertible bonds is considered constant across

firms, but firm specific equity price impacts cannot be

accurately predicted at the announcement date. If the

terms of a convertible bond issue are necessary in





46

calculating equity price impacts, and they can not be ac-

curately estimated, their release at the issue date is ec-

onomically important. Each firm will experience an issue

date equity price impact according to how management sets

the terms. Then, since terms vary from issue to issue, it

is reasonable to expect the issue date equity impacts to

vary in size even though they are mostly negative. The is-

sue date reaction is an increase in cross-sectional vari-

ance of daily excess returns and a negative shift in the av-

erage daily excess return. (4) The probability of selling

an announced issue of convertible bonds varies across

firms, and the firm specific equity price impacts cannot be

accurately predicted at the announcement date. There is

much uncertainty surrounding the signal contained in a con-

vertible bond announcement. The market tends to penalize

firms about the same amount at the announcement date and

wait for the release of the terms at the sale date to de-

termine how much, if any, further adjustment in equity val-

ue is needed. Much of the equity valuation is left for the

issue date, and the effect on the equity value of the sell-

ing firms can vary. The result is a shift in the average

daily excess return and an increase in cross-sectional

variance.



Ho3: Variation in the relative values of the debt
and equity components of the convertible bond has no effect





47

on any issue date impact on the issuing firm's equity

value.

The probability of conversion is a potentially impor-

tant variable in the market's assessment of the signal con-

tained in a convertible bond announcement, and where manage-

ment sets the conversion terms could have an effect on this

probability. For example, if the conversion price is set

very close to the firm's current stock price, conversion in

the near future might be highly probable. In that case,

the market might consider the convertibles no more than a

delayed equity issue. Setting the conversion price very

high with respect to the current stock price could cause

the market to view conversion as improbable. The issue

could be considered non-convertible from a practical stand-

point.

Since the terms of a convertible bond issue are not re-

leased until the issue date, it is reasonable to expect in-

vestors to estimate announcement date equity price impacts

based on some average conversion terms or CP ratio. For

this study, the CP ratio is defined as the ratio of con-

version price to current stock price. The CP ratio could

be thought of as a measure of the value of the issue's call

option to the firm. A low CP ratio would indicate that the

call option is "near the money." That is, management might

not have to wait long before the stock price is greater

than the conversion price and they can call the bonds and





48

force conversion to equity. A higher CP ratio for the same

firm could indicate less probability of forced conversion

in the near future.

If, at the issue date, the terms of the issue vary from

the perceived average, there could be a reaction by market

participants to further reevaluate the issuing firm's equi-

ty. It is expected that issues with lower than average CP

ratios will be viewed more as equity than the announcement

date impact would indicate. The issue date equity price

impact for those firms should more negative than for firms

setting terms at or above the average.



Ho4: Issue size has no effect on any equity price im-
pact at the issue date.

Any issue date equity price impact must be a result of

the release of information or the resolution of uncertainty

that the announced issue would be sold. Rejecting the null

hypothesis, then, suggests a relationship between issue

size, measured as the size of the convertible bond issue

relative to the market value of the firm's equity, and the

market's ability to accurately predict equity price impacts

and/or the probability that the announced issue will be

sold.

The arguments of Miller and Rock (1985) and Myers and

Majluf (1984) suggest a reaction to the announcement of a

proposed issue of convertible bonds. Both arguments





49

suggest that the size of the issue may affect the market's

announcement date reaction. But even if the size of the

issue has an effect on announcement date equity price

impacts, there is no theoretical argument which suggests

that issue size can cause systematic under or overesti-

mation of equity price impacts or probability of issue.

There should be no recognizable relationship between rel-

ative size and any issue date equity price impact.



Ho5: Firm size has no effect on any equity price
impact at the issue date.

Issue date equity price impacts must be the result of

the release of economically important information. As dis-

cussed above, this information can take the form of above

or below average conversion terms or other information con-

tained in the issue terms, or it can be simply the resolu-

tion of uncertainty that the issue would be sold.

There is no theory that would support the hypothesis

that any of these would be more or less predictable accord-

ing to the size of the issuing firm. That is, there should

be no systematic over or underestimation of the terms of

the issue associated with the size of the issuing firm, and

the probability of sale has not been shown empirically to

be related to the size of the announcing firm. There might

be an increase in the cross sectional variance of daily ex-

cess returns and even a shift in the average excess return.





However, there should be no discernible relationship be-

tween any observed issue date impacts and the size of the

issuing firm.




Notes


1. For announcements to be classified as information, they
must be unexpected or at least not fully anticipated.

2. If the conversion price is below the market price of
the underlying stock, the issuing firm can force bond-
holders to convert to equity by calling the issue.

3. The conversion ratio is the number of shares into which
each bond is convertible. The lowest conversion ratio
in the sample is 6.58 and the highest is 142.86. Al-
though this might seem quite variable, it is somewhat
misleading. To better compare conversion ratios, con-
version prices are computed ($1000/conversion ratio)
and compared to the firm's current stock price. This
is called the CP ratio in the study and ranged from
1.00 to 1.91 with the majority falling between 1.13 and
2.00. Looked at from this angle, there is not a great
difference in the terms of most convertible bonds.





CHAPTER FOUR
DATA


An initial sample of 558 outstanding convertible bond

issues was obtained from Moody's Bond Guide.1 As stock

price information is critical to the study, the issuing

firm must be listed on the Center for Research in Security

Prices (CRSP) tapes. 259 issues were deleted from the sam-

ple because they were (1) issued prior to July 2, 1962, (2)

the issuing firm was never listed on the CRSP tapes, or (3)

the firm was listed after issuing the convertibles.2

The date of sale or issue date was identified in

Moody's Industrial Manual. Twenty convertibles that were

issued along with other securities were removed from the

issue date sample. Of the 20, 15 were issued with common

stock, four with straight debt, and one with preferred

stock.

Another 25 were deleted from the sample for other rea-

sons. Six of the issues were convertible into the common

stock of a parent company, two were convertible into pre-

ferred stock or preferred plus common, and one issue was

convertible into straight debt. Seven issues were deleted





because they were issued outside the United States, and

four were deleted because detailed information was unavail-

able. One issue was deleted because the exact issue date

was unclear. This leaves a total of 298 issues in the

issue date sample.3

Original announcements of the proposed convertible bond

issues were located in the Wall Street Journal.4 Of the

298 convertibles in the issue date sample, fourteen issues

were deleted from announcement date testing for release of

confounding information along with the announcement of the

proposed issue of convertible bonds. This confounding

information was the announcement of earnings (9 issues), a

dividend change (1 issue), a change in bond ratings (1

issue), or any other information with potential for

economic impact on the firm's stock price (3 issues).

Another 53 issues were deleted from announcement date

testing because no information was released in the Wall

Street Journal prior to the issue date.5 This left a

final announcement date sample of 232 convertible bond

issues. Table 1 contains a list of the data gathered, when

available, for each issue in both samples.

The convertibles in the final sample were issued over

the years 1963 to 1986. Table 4-2 shows the frequency of

issue per year along with the maturities of the issues.

The typical convertible bond is issued with a maturity of

20 or 25 years. Only seven have maturities less than 20





years, and only two of those are less than 15 years. This

is not unexpected since virtually all convertible bonds are

callable, and the value of the call option to the firm is a

positive function of the length of maturity of the option.

A survey by Brigham (1966) revealed that managers might

look at convertible bonds, with their generally reduced cou-

pons, as "cheap" debt. Table 4-2 shows average yields on

long term (> 10 years) treasury issues and AAA rated cor-

porate bonds for each year in the test sample. If managers

are simply trying to avoid high borrowing costs, convert-

ible bonds should be particularly attractive during periods

of high or climbing interest rates. Over 61%, 184 of 300,

of the bonds in the sample were issued between 1965 and

1971.6 During this time, interest rates on AAA corpor-

ates climbed steadily from 4.57% to 7.28%, and those on

government issues also rose. It appears that, during the

earlier years in the sample, cheap borrowing might have

played a part in convertible bond sales. However, about

30% of the issues in the sample were sold from 1980 to

1986, and interest rates were generally falling over this

period. Also, during the eight years beginning with 1972,

very few convertible bonds were issued, and interest rates

were rising over this period. This would seem to indicate

that now there are motives other than reduced coupons for

selling convertible bonds.

Tables 4-3, 4-4, and 4-5 show the distribution of

issue, firm, and relative sizes, respectively. The average





54

convertible bond issue in the sample is about $51.9 mil-

lion. Ninety percent of the issues are $100 million or

less. The smallest issue sold was $5.5 million, and the

largest was just over $250 million. The smallest firm sell-

ing convertible bonds had total equity value of just over

$6.6 million. That for the largest was almost $7 billion,

and the average firm was valued at about $550 million.

Since not all firms in the issue date sample are included

in the announcement date sample, Tables 4-4 and 4-5 show

the figures for the announcement date firms at both the

announcement date and the issue date, and the total issue

date sample of firms. Comparing the issue date distri-

bution of the firms in the announcement date sample and the

distribution of "All Firms Issue Date" shows the distribu-

tion of the issue date firms that were not included in the

announcement date tests.

Table 4-5 shows that the average issue of convertible

bonds is about 19% of the value of the issuing firm's

equity. Over 30% of the issues are 10% or less of the val-

ue of the firm's equity, and over 89% are less than 35%.

Table 4-6 gives the distributions of CP and conversion

ratios. The conversion ratio is the number of shares of

common stock into which an individual bond is convertible.

Dividing the face value of the convertible by the conver-

sion ratio yields the conversion price, which, divided by

the current stock price, gives the CP ratio. The





55

conversion ratio and CP ratio are not necessarily related,

however. Much depends on the firm's current stock price.

For example, consider the following:







Firm A

Conversion Ratio 40.00 30.00 20.00

Conversion Price $25.00 $33.33 $50.00

Current Stock Price $25.00 $25.00 $25.00

CP Ratio 1.00 1.33 2.00



Firm B

Conversion Ratio 40.00 30.00 20.00

Conversion Price $25.00 $33.33 $50.00

Current Stock Price $15.00 $15.00 $15.00

CP Ratio 1.67 2.22 3.33



The average conversion ratio was approximately 28.4,

indicating that the average bond in the sample was convert-

ible into just over 28 shares of the issuing firm's common

stock. Most of the bonds were convertible into between 10

and 40 shares of common.

The call option on a convertible bond allows the firm

the opportunity to force conversion of the bonds to equity

once the firm's stock price is above the conversion price.





The CP ratio, therefore, measures how far "out of the mon-

ey" the option to force conversion is. The minimum CP

ratio was 1.0, and the maximum was 1.91. The average CP

ratio was 1.17, with over 95% falling between 1.00 and

1.3. Further, almost 60% of the CP ratios were between 1.1

and 1.2.

Table 4-7 shows the distribution of 2-day excess re-

turns at both the announcement and issue dates. The aver-

age announcement date return is -1.43%. The minimum and

maximum, respectively, were -10.55% and 10.72%. Approx-

imately 73% of the firms, 169 of 232, experienced negative

announcement date 2-day excess returns. The average issue

date 2-day excess return was -0.61%, and, again, about 73%,

219 of 298, of the returns were negative. The minimum 2-

day issue date excess return was -12.97%, and the maximum

was 15.13%.




NOTES


1. Issues were found outstanding in at least one of the
Moody's Bond Guides dated 1968, 1970, 1972, 1975, 1977,
1980, 1982, 1984, or 1986. Therefore, any issue called
before early 1968 or issued in the latter part of 1986
is not in the sample.

2. The CRSP tapes began recording stock price and return
information on July 2, 1962.

3. None of the 298 issues in the final sample was issued
to retire outstanding convertible bonds.

4. The typical announcement for a convertible bond issue
(as it appears in the Wall Street Journal) contains





information on the approximate size of the proposed
issue, a period of time during which the issue will
occur, and the underwriterss. In most cases the pro-
posed use of the funds is stated but only in very broad
terms such as general corporate needs, refund debt,
plant expansion, working capital needs, or some com-
bination of any or all the above. Only in a very small
number of cases was a specific use detailed in the
announcement, such as retirement of a specific issue of
straight debt or preferred stock.

5. In some cases announcement of the issue came the day
the issue was sold or only a day or two prior to the
issue date. Usually the announcement and issue dates
were separated by three to four weeks.

6. None of the issues in the final sample was sold during
1973, and very few were sold during the period 1972 to
1979. To test whether there were different motives for
issuing convertibles and, hence, potentially different
market reactions, the sample was divided into two
periods: those issues sold from 1963 through 1972, and
those sold from 1974 through 1986. The average
announcement and issue date two-day equity price
impacts for the two samples were virtually identical.
The announcement date two-day abnormal returns for
1963-1974 and 1974-1986 were -1.41 and -1.47,
respectively. The corresponding issue date returns
were -0.56 and -0.67.

To test whether equity valuation affected equity price
impacts, each sample was divided into three approxi-
mately equal strata by CP ratio. The three strata for
each sample are designated High, Mid, and Low. It was
assumed that, although the relative values of CP ratios
may vary across comparable strata (ie. high vs. high),
each stratum was indicative of a "high", "medium", or
"low" CP ratio for its time of issuance.

The issue date two-day residuals for the high, mid, and
low CP strata for the pre and post 1973 issues were
0.12, -0.82, -1.82 and 0.44, -1.15, -2.14,
respectively. Although the high and low strata for
each sample were significantly different, comparison of
the respective strata for the two samples showed no
significance. Hence, it will be assumed that the
impact of the relative equity valuation of convertibles
does not vary across time.









TABLE 4-1
Data gathered for each issue in the final sample.*

ISSUE DATE: The date the issue was sold.

ANNOUNCEMENT DATE: The first time the proposed issue is
mentioned in the Wall Street Journal.

MATURITY DATE: The date on which the issue matures.

COUPON: The annual coupon paid to holders of the bond.
This is usually paid in two semi-annual payments, the dates
for which were also gathered.

ISSUE SIZE: The original dollar amount registered with the
Securities and Exchange Commission.

QUALITY RATING: The original rating given by Moody's Bond
Rating service. This was obtained from Moody's Weekly News
Reports in the week following the issue of the bonds.
CONVERSION RATIO: The original number of shares into which
each bond is convertible. It is used to compute conversion
price, the price at which a bond may be converted into the
issuing firm's common. It was obtained from Moody's Weekly
News Reports in the week following the issue of the bonds.

CLOSING STOCK PRICE: The closing stock price two days
prior to both the announcement and issue dates. These were
obtained from the CRSP Daily Stock Master tape.

SHARES OUTSTANDING: The number of shares outstanding two
days prior to both the announcement and issue dates. These
were obtained from the CRSP Daily Stock Master tape.

ANNOUNCED USE OF FUNDS: The use of the proceeds from the
issue as announced at the time the bonds were sold.

ACTUAL CALL DATE: The date on which the issue was called
for redemption by the issuing firm. This was obtained from
Standard and Poor's Bond Guide.

ACTUAL CALL PRICE: The dollar amount offered to each
bondholder for redemption of a single bond. This was
obtained from Standard and Poor's Bond Guide.

CONVERSION PRICE AT CALL: The conversion price on the date
the issue was called. It was usually adjusted over the
life of the bonds for stock dividends and/or splits. This
and the date on which the conversion privilege expired were
obtained from Standard and Poor's Bond Guide.


Unless otherwise noted, data were gathered from Moody's
Industrial Manual.





Table 4-2
Frequency of issue and maturity of convertible bonds in the sample and
average yields on long term U.S. government and AAA corporate bonds.


Convertible Bonds


Maturity


Ave. Yield *


Number
Year Issues


25 Yrs


20 Yrs


< 20 Yrs


US Gov't


AAA Corp.


1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986


Total


4.07
4.14
4.32
4.65
5.11
5.46
6.47
6.28
5.62
5.54
6.23
6.78
7.06
6.58
7.62
8.61
9.67
11.75
13.30
11.61
11.49
11.94
10.29
7.93


4.30
4.46
4.57
5.33
6.00
6.55
8.03
8.02
7.28
7.17
7.68
8.77
8.99
8.25
8.12
9.00
10.15
12.86
15.36
13.91
12.30
12.04
11.45
9.24


300


204


k Source: Moody's Industrial Manual


















< IS

< IS

< IS

< IS

< IS

< IS

< IS

< IS

< IS

< IS

< IS

< IS

< IS

< IS

< IS

< IS

< IS

< IS

IS


60

Table 4-3
Convertible Bond Issue Sizes ($ Million)


Issue S


iize Issues




< 250 7

< 200 2

< 175 11

< 150 1

< 140 0

< 130 6

< 120 2

< 110 25

< 100 0

< 90 3

< 80 16

< 70 20

< 60 36

< 50 18

< 40 43

< 30 46

< 20 47

< 10 16

AVE. IS = 51.9


250

200

175

150

140

130

120

110

100

90

80

70

60

50

40

30

20

10


















Firm Equity Value
($Million)


61

Table 4-4
Distribution of Firm Sizes, Announcement and Issue Date.

Announcement
Date Sample


All Firms
Issue Date


Announ.
Date


Issue
Date


5000
4000
3000
2000
1000
900
800
700
600
500
400
300
200
100
90
80
70
60
50
40
30
20
10


FS
FS <
FS <
FS <
FS <
FS <
FS <
FS <
FS <
FS <
FS <
FS <
FS <
FS <
FS <
FS <
FS <
FS <
FS <
FS <
FS <
FS <
FS <
FS <


5000
4000
3000
2000
1000
900
800
700
600
500
400
300
200
100
90
80
70
60
50
40
30
20
10




















Announ. Issue
Date Date


2 3

2 1

0 0

3 3

1 0

0 0

5 4

6 7

13 15

16 15

23 20

35 38

47 40

44 46

16 16


Table 4-5
Relative Sizes of Convertible Bond Issues,
Announcement Date and Issue Date


Announcement
Date Sample


Relative
Size ($Million)


All Firms
Issue Date


4

1

0

4

2

2

6

9

16

17

28

44

49

61

18


1.00

0.90

0.80

0.70

0.60

0.50

0.40

0.35

0.30

0.25

0.20

0.15

0.10

0.05


1.00

0.90

0.80

0.70

0.60

0.50

0.40

0.35

0.30

0.25

0.20

0.15

0.10

0.05





63

Table 4-6
Distribution of Conversion and Issue Date CP Ratios


Conversion Ratios


CP Ratios


Ratio


Ratio


Issues


Issues


100 <

90 <

80 <

70 <

60 <

55 <

50 <

45 <

40 <

35 <

30 <

25 <

20 <

15 <

10 <

5 <


1.50

1.40

1.35

1.30

1.25

1.20

1.15

1.10

1.05

1.00


CR 4

CR < 100 0

CR < 90 1

CR < 80 2

CR < 70 5

CR < 60 7

CR < 55 8

CR < 50 7

CR < 45 9

CR < 40 20

CR < 35 24

CR < 30 48

CR < 25 55

CR < 20 49

CR < 15 33

CR < 10 5

CR < 5 0

AVE. CR = 28.4


< CP < 2.00 4

< CP < 1.50 3

< CP < 1.40 2

< CP < 1.35 3

< CP < 1.30 15

< CP < 1.25 34

< CP < 1.20 72

< CP < 1.15 77

< CP < 1.10 37

< CP < 1.05 10

AVE. CP = 1.17





64

Table 4-7
Distribution of Two-day Excess Returns


Announ.
Date


1

2

2

9

14

11

24

35

38

46

36

11

2

1

-1.43


Issue
Date


2

1

4

11

29

32

36

41

55

52

27

6

0

2

-0.61


Excess Returns (%)


10.00

8.00

6.00

4.00

2.00

1.00

0.00

-1.00

-2.00

-4.00

-6.00

-8.00

-10.00


< 2AR

< 2AR

< 2AR

< 2AR

< 2AR

< 2AR

< 2AR

< 2AR

< 2AR

< 2AR

< 2AR

< 2AR

< 2AR

2AR


10.00

8.00

6.00

4.00

2.00

1.00

0.00

-1.00

-2.00

-4.00

-6.00

-8.00

-10.00

AVE. 2AR





CHAPTER FIVE
EMPIRICAL DESIGN


This is a study of the impact on firm equity value on

the date convertible bonds are announced and the date they

are actually sold. As such an event study methodology will

be employed extensively. Following a discussion of the

event study methodology used, the discussion of the empiri-

cal design is divided into two sections: (1) a discussion

of the announcement date tests, and (2) a discussion of the

issue date tests.



Event Study Methodology


Abno rmal1 returns are obtained from the Center for

Research in Security Prices (CRSP) Daily Excess Returns

Tape. To estimate excess returns, CRSP first estimates

each security's beta. Beta is computed for a given year

only if the stock traded on at least half of the trading

days in that year. The individual beta is calculated as

follows:



Bi = (Covil,)/(VARm)









66

COVi,m = 1 (reti,tmret3t) (1/n)( C mret3t)
t t


VAR t (1rtmrtt/n) ( C mrett)( mret3t)
t t
where

reti,t = log (1 + return for security i on day t)
mrett = log (1 + value weighted market return on day t)
mret3t = mrett-1 mrett rt+ amvn vr
age market window)Bt~ a aig vr
n = number of observations for the year


Next, the firms are ranked and divided into 10 portfol-

ios according to beta. Betas used in the ranking are

either from the preceding year or, if not available, from

the current year. If neither is available, the firm is

placed in a portfolio of firms for which no excess returns

are calculated. Average daily returns are then calculated

over the year for each portfolio.



ARi It =1/m I: (ARj It) j = 1 to m
where

ARi It = the average return for portfolio i on day t

ARjlt = the CRSP daily return for firm j on day t
m = the number of firms in portfolio i


Daily excess returns for firm j are then the difference
between the firm's actual return on day t and the average

day t return for the portfolio containing firm j.





ERj~ = Rjl ARilt
where

ERjlt = the excess return for firm j on day t

Rjlt = the actual return for firm j on day t

ARi it = the average return for portfolio i on day t

For each firm in a given sample, 121 daily excess

returns are gathered, 60 days before to 60 days following

the day of the proposed event. The event will be either

the announcement of a convertible bond issue or the date

the issue is sold. These daily excess returns are then com-

bined to form 60 two-day excess returns for each firm, the

thirtieth, t = 30, of which contains the event date and the

preceding day. The two-day excess returns are then aver-

aged at each time t, t = 1 to 60, across the firms in the

sample.



AERt = 1/n I ERj It

j = 1 to n t = 1 to 60
where

AERt the average 2-day excess return for the
n firms in the sample at time t

ERj,t = the 2-day excess return for firm j at
time t

n = the number of firms in the sample





68

To test the significance of each average two-day excess

return a Student's T-test will be performed. The standard

deviation of the 2-day abnormal returns is estimated from

the 59 2-day average excess returns from trading days -60

to -2 and +1 to +60. Calculating the standard deviation in

this manner avoids any potential effects of abnormal vari-

ance on the event date.

In most of the empirical tests the sample will be strat-

ified in some fashion to identify any equity price impact

related to the variable by which the firms have been strat-

ified. To test for a significantly different event date

impact between two strata, a new variable will be created

which amount s to the difference between the 2-day average

excess returns for the two strata. For example,


whereAERd,t = AERa,t AERblt t = 1 to 60

AERd,t= the difference in the average 2-day ex-
cess returns for strata a and b at time


AERa,t= the average 2-day excess return at time
t for the firms in stratum a

AERb,t= the average 2-day excess return at time
t for the firms in stratum b



Once the new variable is calculated, it is a simple

procedure to perform a T-test to determine significance.

As before, the standard deviation of the 2-day excess

returns (this time using AERd~t) is calculated after









69

omitting AERd,30, the difference between event date aver-

age excess returns. If the T-test determines that

AE~,30is significantly different from zero, this would
indicate a difference in the event date average 2-day ex-

cess returns for the two strata. That is, the variable

along which the strata have been formed is a factor in

event date equity price impacts.





Announcement Date Tests of Hypothesis



Hol: The announcement of an issue of convertible
bonds has no impact on the announcing firm's equity value.

This first test looks at the overall effects of an-

nouncing a convertible bond issue. No stratification or

classification of any kind is utilized. For all 232 firms

in the announcement date sample, 60 2-day excess returns

are gathered centering on the announcement date. A T-test

is performed as described above to determine whether the

announcement date excess return is significantly different

from zero. If the return is significant and positive (neg-

ative), this suggests that the announcement contains good

(bad) information.



Ho2: The announcement of an issue of convertible
bonds causes no increase in the variance of daily stock

returns.





70

To test whether a convertible bond announcement consti-

tutes the release of firm specific information, the distri-

bution of the announcement date 2-day excess returns will

be analyzed. On any given (non-event) day a firm may ex-

perience a non-sero excess return. Over many firms on that

same day or for the same firm over many non-event days, how-

ever, the average excess return will have an expected value

of zero with some positive variance. On an event day the

mean excess return for a portfolio of firms might be non-

zero, and increased uncertainty about the valuation of the

firms could cause an increase in cross-sectional stock re-

turn variance. The following test was utilized to compare

the cross-sectional variance at the announcement date with

that of a typical non-event period.

The standard deviation of the twenty-nine 2-day excess

returns preceding the announcement date is calculated for

each firm as follows:



Sj,a = (1/29 I (ERj~t)2)1/2

t = 1 to 29
where
Si,a= the standard deviation of pre-
announcement date excess returns for
firm j

ERjlt= excess return for firm j at time t

An F-test is performed for each firm that compares

announcement date variance with pre-announcement date

variance:





Vj = (ERj,30/S ,)~2


where


ERj,30 = the mean-adjusted announcement date
excess return for firm j

Sj~a = the pre-announcement date standard dev-
iation of excess returns for firm j

Vj = the ratio of the variance of the an-
nouncement date excess return to that of
the pre-announcement date period and is
distributed as F (1,30)


The expected

Therefore, we let


value of Vj is n/(n -2) = 29/27.

Uj = (Vj 27/29), and


E(U ) = 1.0

VAR(U ) = 2(29 3)/(29 6)


and


This is the standard variance formula for an F with

vl = 1 and v2 = 29 degrees of freedom after the
adjustment factor, 27/29, has been applied. See Warner, et

al (1987) for more on this procedure.

Dealing with a large sample of independent observations

allows the use of the Central Limit Theorem in testing

whether the mean of the Uj is significantly different
from one. Specifically,


1 (26/23))1/2


Z = ( 1( )/2


j = 1 to n





72

As noted above, the announcement date excess return for

each firm is mean-adjusted before performing the F-test.

Subtracting the average announcement date effect from each

firm removes the mean effect from the individual excess re-

turns. This allows testing of the variance alone, and

shows whether, aside from some average impact affecting all

firms, each firm has a specific equity price impact. In

other words, there are two potential impacts at the an-

nouncement of economically important information: (1) a

mean effect representing a shift in excess return distri-

bution due to some average impact each firm will exper-

ience, and (2) an increase in variance of equity returns

signalling a differing impact across firms beyond the aver-

age impact. That is,



ERjl = AERt + bjlt
where

ERjlt = the excess return for firm j at time t

AERt =the average excess return at time t for
the n firms in the sample

bj,t = the firm specific equity impact, either
positive or negative.


A significant Z, then, indicates that the signal
contained in a convertible bond announcement is valued

according to the firm making the announcement. The typical

bjlt term is large enough to significantly affect the
variance of the announcement date excess returns.





73

Ho3: Variation in the relative values of the debt
and equity components of the convertible bond has no affect

on any impact on the announcing firm's equity value.

Before this hypothesis can be tested, a proxy for the

debt and equity components has to be obtained. It should

be noted that since virtually all convertible bond issues

are sold at par, if a proxy for either the debt or equity

component can be obtained, the other value is automatic.

As a proxy for the equity component inherent in each

convertible bond, a ratio of the conversion price to cur-

rent price per share is used and is referred to as the CP

ratio. This ratio measures how far the conversion option

is "out of the money" or how likely the issue is to be con-

verted either voluntarily or through a forced conversion.

A large CP ratio indicates that the option is far out of

the money. The issue is less likely to be converted and

should impact more as an issue of straight debt. A small

CP ratio would indicate that the issue has a relatively

higher probability of being converted and should impact

more like common equity. In other words, those issues with

small CP ratios are considered to have a relatively larger

equity component than an issue with a larger CP ratio.

The conversion price for each issue is obtained from

Moody's Weekly News Reports during the week following the

issue date. Obtaining the conversion price very early in

the bond's life is important as it is released immediately





prior to the sale of the bonds and is subject to adjustment

for stock dividends or splits during the life of the bond.

The conversion price is then divided by the closing stock

price two days prior to the announcement date. This stock

price was utilized to help avoid any effects of the an-

nouncement on equity value.

The minimum and maximum CP ratios for the announcement

date sample were 0.75 and 2.03, respectively, with a mean

of 1.15. In option pricing jargon the 0.75 ratio is "in

the money", and the 2.03 ratio could be considered "far out

of the money." "In the money" would indicate that the firm

could immediately call the issue and force its conversion

to equity. It is important to remember, however, that

these CP ratios are based on conversion terms not available

until the issue date. Therefore, these announcement date

figures can only be considered expectations.

Once (expected) CP ratios are calculated for each firm,

the firms are divided into three approximately equal strata

according to CP ratio. T-tests are performed to test for

significance within as well as between strata. Signifi-

cance between strata indicates that the announcement date

equity price impact is a function of the (expected) rel-

ative values of the debt and equity components.



Ho4: Issue size has no affect on any equity price
impact at the announcement date.





75

To test whether the dollar size of the convertible

issue is a factor in the announcement date impact, the

firms will be stratified by the relative size of the

issue. This is defined as the dollar value of the announc-

ed issue divided by the market value of the firm's equity.

The dollar size of the issue is obtained from Moody's

Industrial Manual, and the market value of the firm's e-

quity is calculated as the market price per share times the

number of shares outstanding. Both the price per share and

the number of shares are obtained from the CRSP Daily Stock

Master tape, and, to avoid any possible effects of the an-

nouncement on the share price, the closing price two days

before the announcement is used.

Once relative size is calculated for each firm for

which the data are available, the firms are divided into

three approximately equal strata. A significant difference

between strata indicates a difference in announcement date

equity price impacts based on relative size.



Ho5: Firm size has no affect on any equity price im-

pact at the announcement date.

To test whether the size of the issuing firm is a fac-

tor in the announcement date impact, the firms are strat-

ified by total market value of equity. As in the last

test, the market value of the firm is defined as the clos-

ing stock price two days before the announcement times the





76

number of shares on that date. Significance between strata

indicates that the equity price reaction to the announce-

ment is a function of the size of the issuing firm.



Issue Date Tests of Hypothesis



Hol: Sale of an issue of convertible bonds has no
impact on the issuing firm's equity value.

This is a general test of the stock price affects of

issuing convertible bonds on the date the bonds are actu-

ally sold. For the 298 securities in the issue date sam-

ple, the sale date is identified in Moody's Industrial

Manual. Significant issue date excess returns indicate

that a firm issuing convertible bonds will experience an e-

quity price impact at not only the announcement of the is-

sue but also at the sale date.



Ho2: The sale of an issue of convertible bonds
causes no increase in the variance of daily stock returns.

As with the announcement date tests, the cross-section-

al variance of excess returns on the issue date will be com-

pared to that of a typical non-event period. For this

test, however, the standard deviation of post issue date ex-

cess returns will be compared to issue date excess return

variance. The remainder of the test is exactly as its an-

nouncement date counterpart. A significant Z value will





77

indicate that firm specific information is released when

convertible bonds are actually sold.



Ho3: Variation in the relative values of the debt
and equity components of the convertible bond has no affect

on any issue date impact on the issuing firm's equity val-

ue.

Again the CP ratio is used as a proxy for the equity

component of the individual convertible bond. To calculate

the CP ratio the conversion price as stated in Moody's

Weekly News Reports the week after the issue sold is divid-

ed by the closing stock price for the issuing firm two days

before the issue date. The minimum and maximum CP ratios

at the issue date were 1.0 and 1.91, with an average of

1.17.

After the CP ratio is figured for the 257 firms for

which the data are available, the firms are divided into

three approximately equal strata. Significance between

strata indicates that the relative values of the debt and

equity components inherent in the convertibles are a factor

in the issue date equity impact.



Ho4: Issue size has no affect on any equity price im-

pacts at the issue date.

To test whether the value of the convertible relative

to the total market value of the issuing firm's equity is a





78

factor in any issue date equity price impact, the issuing

firms are divided into three approximately equal strata by

relative issue size. Relative size is defined as the dol-

lar size of the issue divided by the market value of the is-

suing firm's common equity. The dollar amount of convert-

ibles sold is obtained from Moody's Industrial Manual. The

value of the firm's equity is the closing price per share

two days preceding the issue date times the number of

shares on that date. Both of these figures are obtained

from the CRSP Daily Stock Master tape. Closing price is

obtained from two days before the event to avoid any ef-

fects of the release of the terms of the issue either on

the sale date or the day before.

A T-test is used to test for significance within each

stratum as well as between the strata. Significance be-

tween strata indicates that the issue date equity price im-

pact is a function of relative size.



Ho5: Firm size has no affect on any equity price im-
pact at the issue date.

The firms are divided into three approximately equal

strata according to market value of equity. Market value

of equity is defined as the closing price per share of com-

mon two days before the sale date times the number of

shares outstanding on that date. The source for these data

is the CRSP Daily Stock Master. Significance between the









79

strata indicates that firm size is a factor in the equity

price impact at the date convertible bonds are sold.





CHAPTER SIX
RESULTS


This chapter is divided into four sections: (1) a dis-

cussion of the results of the announcement date tests, (2)

a discussion of the results of the issue date tests, (3) a

summary of the results of the announcement and issue date

tests, and (4) a discussion of additional tests performed.



Announcement Date Test Results



Hol: The announcement of an issue of convertible
bonds has no affect on the announcing firm's equity.

Table 6-la shows the average two-day excess returns for

the 121 day period surrounding the announcement of 232 is-

sues of convertible bonds. Observation 30 contains the an-

nouncement date, as identified in the Wall Street Journal,

and the preceding day. The average announcement date eq-

uity impact on the 232 firms is -1.43% and is significant

at the 1% level. Of the 232 two-day excess returns, 72%

(168) were negative.

The results indicate that the announcement of an up-

coming issue of convertibles is received as unfavorable





information by the market and has an economically signif-

icant impact on the market value of the average announcing

firm's equity. The null hypothesis that the announcement

of an issue of convertible bonds will have no effect on the

announcing firm's equity is rejected at the 1% level.

Table 6-1b shows average daily excess returns and aver-

age cumulative excess returns for 121 days surrounding the

announcement date. The average equity impacts for the an-

nouncement day and the preceding day are -0.727% and

-0.697%, respectively, and both are significant at the 1%

level. The cumulative excess returns are positive and sta-

tistically significant for most of the sixty days preceding

the announcement. The average cumulative excess return

reaches 4.25% on day -5 and is significant at the 1%

level. The negative returns for the five days preceding

and including the announcement date bring down the cum-

ulative excess return to about 2.4% which is only signif-

icant at the 5% level. By the first day following the

announcement, the average cumulative return is approaching

2.0% and only significant at the 10% level. This cumula-

tive return and significance is maintained for the ten days

following the announcement date, but by day +11 any signif-

icance is lost. This seems to indicate some validity in

the theory that management tries to time security





82

announcements. According to Myers and Majluf (1984) man-

agement will issue a security it deems to be overpriced.

The accumulation of over 4.0% in excess returns for the six-

ty days preceding the announcement and the subsequent loss

of this excess return would support Myers and Majluf. The

announcement of the impending convertible bond issue sig-

nals that the firm's equity is not worth as much as its cur-

rent stock price would indicate.

This test has determined that the announcement of an is-

sue of convertible bonds is a statistically as well as an

economically significant event. It will be the goal of the

following tests to help determine those characteristics of

the issue and/or the firm that affect the market's reaction

to the announcement of a convertible bond issue.



Ho2: The announcement of an issue of convertible
bonds causes no increase in the variance of daily stock re-

turns.

If the signal contained in an announcement of convert-

ible bonds is valued according to the announcing firm, the

equity impact could vary greatly from firm to firm. The re-

sulting movement of stock prices in differing amounts, both

positive and negative, will cause an increase in the cross-

sectional variance of stock returns. If the information

available at the announcement date is not complete enough

to form estimates of equity values, however, the market





83

might simply assess some more or less constant amount from

each announcing firm's equity. If this is the case, the

distribution of excess returns at the announcement date

would simply shift to the left with no increase in vari-

ance.

The Z-score from testing all 232 firms in the

announcement date sample was only 0.624, and the null

hypothesis cannot be rejected. Since there is no

significant increase in the cross-sectional variance of

equity values at the announcement date, it would appear

that firms are simply charged a flat amount for announcing

convertible bonds. This could have two explanations: (1)

there is no attempt by the market to estimate the terms and

other information not contained in the announcement, or (2)

the information omitted from the announcement is of no

economic importance. It only matters that the firm intends

to issue convertible bonds.



Ho3: Variation in the relative values of the debt
and equity components of the convertible bond has no effect

on any impact on the announcing firm's equity value.

According to Myers and Majluf (1984) managers prefer to

issue debt, and issuing equity or convertible bonds is con-

sidered a negative signal. It follows from their reasoning

that the impact from announcing convertible bonds might be

due to the equity component inherent in the issue. This





test divides the issues into three strata according to eq-

uity value. The equity component of each bond in the sam-

ple is measured by its CP ratio, the ratio of conversion

price to stock price at issue date. Table 6-2 shows the re-

sults of this stratification.

The third of the firms with the highest CP ratios exper-

ienced an announcement date 2-day excess return of -1.04%

which is significant at the 1% level. The middle and low-

est strata experienced 2-day excess returns of -1.63% and

-1.84%, respectively. Both are significant at the 1% lev-

el. Sixty-nine percent (49 of 71) of the high strata 2-day

excess returns, seventy-three percent (52 of 71) of the mid-

dle strata 2-day excess returns, and seventy-eight percent

(55 of 70) of the low strata 2-day excess returns were neg-

ative. T-tests revealed no significant difference between

announcement date reactions in any of the strata. No sta-

tistical relationship is established between the announce-

ment date impact and the equity component of the convert-

ible bond. The T-test comparing the announcement date eq-

uity impacts for the high and low strata produced a t-value

of 1.43, and, therefore, the null hypothesis cannot be re-

jected at the 10% level. However, a pattern in an-

nouncement date impacts does seem to be present. The ab-

solute magnitude of the announcement appears to be a neg-

ative function of CP ratio. That is, as the CP ratio in-

creases, the equity impact decreases. This would seem to





indicate that even though the conversion terms are not

released until the issue date, the market is capable of

predicting where management will set the terms or at least

makes an attempt to do so.



Ho4: Issue size has no affect on any equity price im-

pact at the announcement date.

Miller and Rock (1985) claim that any unexpected an-

nouncement of external financing signals a short-fall in op-

erating funds. A necessary implication of their theory is

that the impact on the announcing firm's equity should be a

positive function of the size of the issue of securities

sold. If this is the case there should be a larger impact

on the equity of the firm selling a relatively larger issue

of convertible bonds. Table 6-3 presents the results when

the firms are stratified according to the relative size of

the convertible bond issue defined as the dollar value of

the issue divided by the market value of the issuing firm's

equity.

The firms were divided approximately into thirds. The

70 firms with the relatively largest issues experienced an

impact of -1.75% which is significant at the 1% level.

Seventy six percent (53 of 70) had negative 2-day excess

returns. The next stratum experienced an average 2-day ex-

cess return of -1.21%, again significant at the 1% level.

Sixty-eight percent (48 of 71) of the 2-day excess returns









86

returns were negative for this group. The 71 firms in the

stratum with the relatively smallest issues experienced an

average 2-day excess return of -1.12%. This was signif-

icant at the 1% level and 73% (52 of 71) of the excess re-

turns were negative.

To test for a significantly different impact across the

strata, a T-test was performed. The null hypothesis that

the impacts were equal across the three strata could not be

rejected at the 10% level. This would indicate that the

relative size of the issue of convertible bonds is not a

factor in the announcement date impact on the issuing

firm's equity, and it does not support the theory of Miller

and Rock (1985).



Ho5: Firm size has no affect on any equity price im-

pact at the announcement date.

According to Myers and Majluf (1984) management will is-

sue securities when they are overpriced. Further, manage-

ment would prefer to issue debt. The announcement of the

sale of convertible bonds would be a negative signal to the

market and would be met with a negative impact on stock

prices. Miller and Rock (1985) maintain that any unex-

pected announcement of external funding is a negative sig-

nal and would be met with a negative reaction in stock

prices. An implication of both is that the predictability

of any announcement will affect the reaction by the market.





87

That is, if the market is able to predict the announcement

of the need for external funding and the type of security

to be issued, there should be no perceptible impact in con-

junction with the announcement of the security offering.

The availability of information is then critical in deter-

mining the impact on the announcing firm's equity.

If the availability of information varies from firm to

firm, market reactions could be firm specific. If, for in-

stance, information on large firms is more readily access-

ible than that for smaller firms, the market would be bet-

ter able to predict the actions of larger firms. For this

test the firms were divided into three strata according to

the market value of equity. A T-test was performed to de-

termine significance within as well as among the strata.

Table 6-4 shows the results of this stratification.

For the 71 firms in the stratum of the largest firms, the

average 2-day announcement date excess return was -1.48%

and was significant at the 1% level. Seventy-six percent

(54 of 71) of the 2-day excess returns were negative. The

middle strata experienced a significant (1%) reaction of

-1.06% while that for the smallest firms was -1.55%, also

significant at the 1% level. The percentages of negative

2-day excess returns were sixty-nine and seventy-three,

respectively, for the middle and smallest strata. As with

relative size, there was no statistical difference between

the reactions of any two strata. This indicates that the





size of the issuing firm is not a factor in determining the

announcement date impact, and the null hypothesis cannot be

rejected.



Table 6-5 presents a summary of the results for the an-

nouncement date tests. These tests have determined that an-

nouncing an issue of convertible bonds is an economically

important event for the announcing firm, and there is some

evidence that management attempts to time the announcement

to take advantage of abnormal pricing. The average firm ex-

periences an announcement date equity price impact of near-

ly -1.5%. Factors such as firm size and relative size of

the issue have been shown generally not to affect announce-

ment date equity impacts. Even though the size of the eq-

uity component inherent in convertible bonds was not demon-

strated to have a statistically significant affect on an-

nouncement date equity price impacts, the observed pattern

of excess returns may indicate that the market attempts to

estimate the terms of the issue before they are announced.





Issue Date Test Results



Hol: Sale of an issue of convertible bonds has no

impact on the issuing firm's equity value.





89

Table 6-6a presents 60 two-day residuals around the

date of sale for 298 issues of convertible bonds. Obser-

vation thirty contains the issue date, as identified in

Moody's Industrial Manual, and the preceding day. The

average issue date equity price impact is -0.61% and is

significant at the 1% level. Of the 298 firms issuing con-

vertibles, 72% (168) experienced a negative issue date

equity impact. This indicates that the date convertible

bonds are sold is an important event for the selling firm,

and the null hypothesis is rejected at the 1% level.

Table 6-6b shows average daily and cumulative average

abnormal returns for the same period. The average daily

abnormal returns for days -40, -14, and -10 show statis-

tical significance. The average abnormal return on the day

the bonds were sold is -0.48% and is significant at the 1%

level. Besides the issue date, however, there are no signi-

ficant daily excess returns between day -10 and day +60,

and none of the average cumulative abnormal returns from

day -12 to day +60 is significant.

Although the issue date itself has been shown to be a

statistically significant event for the average convertible

bond issuing firm, there are no patterns in the cumulative

returns before or after the event which may be of help in

explaining either the firm's motive for selling vs. with-

drawing the issue or even why the issue date is an event.

It is important to make note at this time that the





90

convertible bond issues used in this study were located in

Moody's Bond Record and, hence, were already outstanding.

This, by definition, excludes any issues that were announc-

ed and subsequently withdrawn by the announcing firm. In-

cluding canceled issues and treating cancellation dates as

issue dates might alter the results of the test. As the

results of Mikkelson and Partch (1986) imply, the can-

cellation of convertible bonds may be viewed as a positive

signal and cause positive average cancellation date re-

turns. In that case, including cancelled offerings could

push the average issue date abnormal return to zero. Also,

at the date the cancelled issues are announced, the market

may accurately predict that the issues will be withdrawn,

and the announcing firms will experience much smaller an-

nouncement date equity price impacts. This could greatly

affect the overall average announcement date equity price

impacts for firms announcing convertible bonds. This is

only conjecture at this point, however, but it warrants

future testing.

The next test of hypothesis tests the nature of the

issue date equity price impact. Specifically, whether the

issue date impact is a result of the release of new infor-

mation or simply the result of resolution of the uncer-

tainty surrounding sale of the issue.



Ho2: The sale of an issue of convertible bonds
causes no increase in daily stock return variance.





91

As discussed in the announcement date tests, an

increase in the cross-sectional variance of issue date

excess returns will indicate the presence of a firm

specific equity impact in addition to the average impact

felt by all firms. Comparing the cross-sectional variance

at the issue date with the average cross-sectional variance

of the 30 post-issue date 2-day excess returns yielded a Z-

score of 1.10, and the null hypothesis cannot be rejected

at the 10% level. There is no evidence to support the al-

ternative hypothesis that each firm has a potentially dif-

ferent issue date equity price impact. The results seem to

indicate that firms that sell previously announced convert-

ible bonds are assessed a more or less constant amount.



Ho3: Variation in the relative values of the debt
and equity components of the convertible bond has no affect

on any issue date impact on the issuing firm's equity

value.

Table 6-7 shows the results from stratifying the firms

according to CP ratio, the ratio of conversion price to cur-

rent stock price. The stratum of firms with the large CP

ratios experienced an insignificant average issue date im-

pact of +0.19%. Approximately 44% of the firms in this

stratum, 37 of 84, had negative issue date excess returns.

A test of whether the number of negative returns was

significantly different from 50% yielded an insignificant

Z-





92

of 1.09. This indicates that the probability of a negative

issue date equity impact for the firms in this stratum is

only 50%.

The firms in the medium stratum experienced an average

equity impact of -0.59%, significant at the 5% level. Over

65% of these firms, 56 of 86, experienced negative

returns. The firms in the small stratum experienced an

average issue date equity price impact of -1.40% which is

significant at the 1% level. Almost 78% of these firms, 67

of 86, had negative issue date 2-day excess returns. The

hypothesis that the probability of a negative return is

greater than 50% is rejected for both strata at the 1%
level.

Testing between strata yielded significant results.

The hypothesis that the issue date 2-day excess return for

the stratum with the small CP ratios is equal to that of

the medium stratum is rejected at the 10% level. That the

small stratum equity price impact is equal to that of the

the large stratum was rejected at the 1% level. This is

strong statistical evidence of a relationship between CP

ratio and issue date equity price impact. Firms in the

stratum that set low CP ratios experience the greatest

equity impact followed by the firms in the stratum of

middle CP ratios. The firms that set their CP ratios

relatively high had no significant impact on their equity

value at the issue date.





93

This test would indicate that the equity impact for

firms selling convertible bonds is related to where man-

agement sets the CP ratio. There are two possible inter-

pretations of these results: (1) The equity valuation of

the convertibles directly affects the issue date equity

price impact, the CP ratio captures the equity valuation,

and investors are not able to predict the issue's terms at

the announcement date. (2) the equity valuation of the con-

vertible is unrelated to the issue date equity price im-

pact, and the CP ratio is actually masking the effects of

another variable related to the firm or the issue.

The goal of the remaining tests in this section is to

determine whether there are variables other than the CP

ratio which affect the issue date equity price impact. In

each test of hypothesis the firms will be stratified accord-

ing to the variable in question, and the average issue date

equity impacts will be compared across the strata. Rela-

tionships between these and other variables and the CP

ratio will be tested in the last section, "Additional

Testss.'



Ho4: Issue size has no affect on any equity price

impacts at the issue date.

Table 6-8 shows the results from stratifying the firms

according to relative size of the issue of convertible

bonds. The stratum of firms selling the relatively largest




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - Version 2.9.9 - mvs