Debt-Equity Choice
In Europe
Philippe GAUD
HEC, Universtiy of Geneva
Martin HOESLI
40, Bd. du Pont d’Arve
PO Box, 1211 Geneva 4
Switzerland
Tel (++4122) 312 09 61
Fax (++4122) 312 10 26
http: //www.fame.ch
E-mail: admin@fame.ch
HEC, University of Geneva, FAME and University of Aberdeen
André BENDER
HEC, University of Geneva, and FAME
Research Paper N° 152
June 2005
FAME - International Center for Financial Asset Management and Engineering
THE GRADUATE INSTITUTE OF
INTERNATIONAL STUDIES
Debt-Equity Choice in Europe
Philippe Gaud 1 , Martin Hoesli2 and André Bender3
June 2005
Abstract
Using a sample of over 5,000 European firms, we document the driving factors of
capital structure policies in Europe. Controlling for dynamic patterns and national
environments, we show how these policies cannot be reduced to a simple trade-off or
pecking order model. Both corporate governance and market timing impact upon
capital structure. European firms limit themselves to an upper barrier to leverage, but
not to a lower one. Debt constrains managers to payout cash, and equity may become
cheap during windows of opportunity. Internal financing, when available, is preferred
over external financing, but companies limit future excess of slack as it constitutes a
potential source of conflict.
JEL classification: G32
Keywords: dynamic capital structure; debt-equity choice; trade-off; agency; pecking
order
1
University of Geneva (HEC), 40 boulevard du Pont-d’Arve, CH-1211 Geneva 4, Switzerland, email:
philippe.gaud@hec.unige.ch
2
University of Geneva (HEC and FAME), 40 boulevard du Pont-d’Arve, CH-1211 Geneva 4,
Switzerland and University of Aberdeen Business School, Edward Wright Building, Dunbar Street,
Aberdeen AB24 3QY, UK, email: martin.hoesli@hec.unige.ch
3
University of Geneva (HEC and FAME), 40 boulevard du Pont-d’Arve, CH-1211 Geneva 4,
Switzerland, email: andre.bender@hec.unige.ch
Address correspondence to: Philippe Gaud, University of Geneva, HEC, 40 boulevard du Pont-d’Arve,
CH-1211 Geneva 4, Switzerland, email: philippe.gaud@hec.unige.ch, Tel: (+41) 22 379 8008, Fax:
(+41) 22 379 8104
We thank an anonymous referee, Yves Flückiger, Nicolas Schmitt, and participants at the 20 th
international meeting of the French finance association (AFFI) and at the University of Geneva (HEC)
finance research seminars for useful suggestions. The usual disclaimer applies.
Executive summary
Since the capital structure irrelevance proposition of Modigliani and Miller, hundreds of
papers have focused on financial policy. Theoretical works have brought three main classes of
models to test: The trade-off, the agency, and the pecking order hypotheses. Under the trade-off
hypothesis, the optimal financing policy consists in making adjustments toward the target debt
level provided that deviation costs exceed adjustment costs. The target leverage ratio balances the
marginal tax benefit with the marginal financial distress cost of debt. Under the agency
hypothesis, firms also face financial distress costs, but the level of debt becomes a governance
device due to informational asymmetries and divergences in the utility functions of stakeholders.
Under the pecking order hypothesis, informational asymmetries bias investment policy which in
turn affects the firm value. The adverse selection discount leads to rejecting positive net present
value projects. The resulting optimal financial policy first exhausts the least sensitive financing
source, i.e. internal financing, then debt and, as a last resort, equity.
In this paper, we investigate the debt-equity choice using a sample of more than 5,000
European firms over the period 1988-2000. We test trade-off, agency and pecking order models
through a panel analysis of firm- specific determinants of these choices. By focusing on the
possible adjustment to a target debt ratio and on the role played by operating and market
performance, we provide evidence that neither the trade-off nor the pecking order model, in their
most commonly accepted forms, offer a suitable description of the capital structure policies in
Europe.
We also document to what extent the national environment affects capital structure. Although
variables traditionally used in international studies consistently impact upon observed debt levels
in country subsamples, the national environment does matter. First, share repurchases appear to
be determined by interactions between national environments and more stylized facts, suggesting
an avenue for future research. Second, country dummy variables are found to be significant in
2
regressions pertaining to other types of events. This dummy control procedure allows an
investigation of the common factors shaping debt-equity choice of European firms.
Contrary to predictions of dynamic trade-off models, we do not report evidence of a lower
barrier to debt levels. Whereas firms do not suffer much from being significantly below the target
leverage ratio, they do not cross an upper debt level (i.e. the maximum debt capacity). Both
operating and market performance significantly affect debt-equity choice. In particular, debt
constrains managers to payout cash, and equity may become cheap during windows of
opportunity. Internal financing, when available, is preferred over external financing, but firms
limit future excess of slack because it potentially constitutes a source of conflict. Also in line with
the agency hypothesis, profitable firms prefer increasing dividends rather than decreasing debt
levels. We highlight that conflicts between shareholders and debtholders limit external equity
finance. Results further show that managers are trying to time the market, but also that financing
and investment activities interact. Debt does not constitute a suitable form of financing for firms
with value-enhancing investment projects. Instead, such firms issue equity. In contrast, when
there is a lack of profitable projects, debt disciplines managers as firms prefer issuing debt and
increasing dividends.
3
1 Introduction
Since the capital structure irrelevance proposition of Modigliani and Miller (1958), hundreds
of papers have focused on financial policy. Theoretical works have brought three main classes of
models to test: The trade-off, the agency, and the pecking order hypotheses. Under the trade-off
hypothesis, the optimal financing policy consists in making adjustments toward the target debt
level provided that deviation costs exceed adjustment costs. The target leverage ratio balances the
marginal tax benefit with the marginal financial distress cost of debt. Under the agency
hypothesis, firms also face financial distress costs, but the level of debt becomes a governance
device due to informational asymmetries and divergences in the utility functions of stakeholders.
Under the pecking order hypothesis, informational asymmetries bias investment policy which in
turn affects the firm value. The adverse selection discount leads to rejecting positive net present
value projects. The resulting optimal financial policy first exhausts the least sensitive financing
source, i.e. internal financing, then debt and, as a last resort, equity. On the contrary, a low or
negative discount leads to favoring the most sensitive financing source, i.e. external equity.
Empirical works find that some variables systematically affect capital structure, but they fail
to discriminate between the three above hypotheses. Traditional cross-sectional regressions of
observed leverage ratios have some explanatory power within countries, but inferences are
limited. Few studies test to what extent the results are robust to international samples being used.
It is difficult to document specific effects relating to one of the three hypotheses because a
significant impact of a give n proxy is often consistent with several theoretical explanations.
While theoretical hypotheses stress the dynamic nature of capital structure policies, capital
structure regressions are static and fail therefore to capture an important dimension.
A more suitable empirical analysis is to study debt-equity choice. Motivations that govern
corporate financial policy are analyzed by focusing on significant external changes of capital
structure. Several studies on debt-equity choice have appeared recently (Jung, Kim, and Stulz,
4
1996; Hovakimian, Opler, and Titman, 2001; Hovakimian, 2004; Hovakimian, Hovakimian, and
Tehranian, 2004). These studies highlight the economic role played by three important pieces of
the puzzle: Adjustment to a target leverage ratio, operating performance, and market
performance. Tests of the adjustment to a target leverage ratio are crucial because adjustment is
the cornerstone of dynamic trade-off models. An examination of the impact of operating and
market performance should highlight whether these are significant determinants of the target
leverage ratio and/or of deviations from this target. This analysis may also lead to the conclusion
that the impact of performance on debt-equity choice stems from other factors than the
adjustment to the target leverage ratio, as implied by pecking order and agency models.
Few studies focus on international samples to test capital structure models. Rajan and
Zingales (1995) and Booth, Aivazian, Demirgüç-Kunt, and Maksimovic (2001) are two
noticeable exceptions. Rajan and Zingales (1995) find similar levels of leverage across the G7
countries, thus refuting the idea that firms in bank-oriented countries are more leveraged than
those in market-oriented ones. However, they recognize that this distinction is useful in analyzing
the various sources of financing. They also find that the determinants of capital structure that
have been reported for the U.S. (size, growth, profitability, and importance of tangible assets) are
important in other countries as well. They show that a good understanding of the relevant
institutional context (bankruptcy law, fiscal treatment, ownership concentration, and accounting
standards) is required when identifying the fundamental determinants of capital structure. Booth
et al. (2001) suggest that the same determinants of capital structure prevail in ten developing
countries, but national environment is again important. These studies, however, do not shed any
light on the adjustment process of capital structure and fail to discriminate between the main
theoretical hypotheses.
Recent research (e.g. Miguel and Pindado, 2001; Gaud, Jani, Hoesli, and Bender, 2005)
benefits from advances in econometrics and highlights some dynamic features of the capital
structure policy of European firms. However, these papers document a wide array of adjustment
5
speeds. They rely upon institutional specifics to justify their results, but inferences are limited
since estimations are based on unique country samples.
This paper relates to these previous strands of literature. Its aim is to answer the following
questions: To what extent does the national environment affect debt-equity choices of European
firms? Beyond national effects, are there homogeneous determinants of debt-equity choices
across Europe? Do theoretical models help explain these determinants? Much work remains to be
done to investigate the dynamics of the capital structure of European firms and the present paper
is a step in this direction. A proper examination of the capital structure issue in Europe requires a
control for both the dynamic and institutional dimensions. We therefore study the debt-equity
choice using a panel of 5,074 firms for the period 1988-2000.
As debt-equity choice analysis focuses on motives which drive firms to significantly alter
their capital structure, it is dynamic and flexible in nature. To allow a test of the adjustment
toward a target leverage ratio, debt-equity choice analysis requires an estimation of the target.
Our estimation procedure of the target leverage is close to that of Hovakimian et al. (2001). First,
we analyze the determinants of observed debt leverage in country subsamples using empirical
proxies from previous international studies (Rajan and Zingales, 1995; Booth et al., 2001). We
then assign each variable either to target leverage equations or to debt-equity choice equations
according to theoretical priors and signs of estimated coefficients. The analysis of the debt-equity
choice requires a two-step estimation procedure: If a given variable does not enter directly in a
debt-equity choice regression, then it enters through the target leverage which is the fitted value
from the target leverage regression.
The prerequisite analysis of the observed leverage ratio highlights some institutional effects
on capital structure. However, the relatively small number of significant debt-equity choices
during the period under review leads us to forego a country by country analysis. In other words,
we have to hypothesize a common European model of capital structure. We confirm this
hypothesis except for share repurchases. National environments also appear to be important
6
determinants of other debt-equity choices, but we do not document changes in sign on firmspecific variables, whether country control variables are included or not in regressions.
Furthermore, goodness-of- fit measures do not change dramatically whether we include these
dummy variables or not. Although the European market is in a consolidation phase, there still
exists considerable institutional and cultural diversity. Thus, strong economic forces appear to be
at work since results are found to be significant in spite of these institutional differences.
Results indicate that neither a simple pecking order model nor a simple trade-off model is
sufficient in understanding financial policy; they also highlight that agency and timing issues
impact upon capital structure. We document that the financing process is complex and dynamic.
In terms of debt ratios, we find that firms constrain themselves to an upper barrier only. As
implied by the role of debt capacity in pecking order models, firms refuse to exceed a maximum
debt level and prefer to repay debt or issue equity. Evidence of a lower leverage boundary is
inconclusive, contrary to predictions of dynamic trade-off models (which are models with two
barriers).
Both operating and market performance affect debt-equity choice since debt constrains
managers to payout cash, and since equity may become cheap during windows of opportunity. In
contrast to Hovakimian et al. (2004), we do not find that unprofitable firms seeking outside
financing prefer to issue equity. But we find that profitable firms issue debt to limit expropriation.
Internal financing, when available, is preferred over external financing, but firms limit future
excess of slack because it is a source of conflict. Profitable firms also prefer to increase dividends
rather than decrease their debt level. As Hovakimian et al. (2001), we find that wealth transfers
from shareholders to debtholders limit equity issues. Two distinct effects are embedded in market
performance results. Empirical evidence shows that managers attempt to time the market, but also
that financing and investment activities interact. In particular, debt does not constitute a suitable
form of financing for firms with value-enhancing investment projects. Instead, such firms issue
7
equity. When there is a lack of profitable projects, debt disciplines managers and firms choose
debt issues and dividend increases.
The remainder of the article is organized as follows. In section 2, we present the observed
debt levels and study their determinants for 13 European countries. In section 3, we develop the
framework for the debt-equity choice analysis. Section 4 presents the results of this analysis.
Finally, section 5 contains some concluding remarks.
2 Analysis of debt ratios
2.1 Measures and levels
Rajan and Zingales (1995) discuss the pros and cons of various leverage measures. To reduce
accounting and sector biases in an international setting, they advocate for consolidated debt to
capital ratios. They also adopt three other adjustments to further increase the degree of
comparability across their samples, but we do not implement them. They first adjust leverage
ratios for cash and cash equivalents. Opler, Pinkowitz, Stulz, and Williamson (1999) show that
cash should not be considered as negative debt; thus we consider cash as an important
determinant of capital structure. Rajan and Zingales (1995) also write off intangible assets against
equity due to the high level of intangibles in the United States. As we consider European firms
exclusively, intangibles are kept in the assets. They finally adjust equity for deferred taxes,
though they recognize that the extent to which differed taxes are equity-like may vary across
countries. Since the Worldscope database used to construct our sample does not disclose the
exact amount of differed taxes, no adjustment is made. Booth et al. (2001) make extensive use of
the long-term leverage ratio, arguing that it is a better proxy for the financial debt ratio than the
total liabilities to total assets ratio. Since Worldscope reports short-term and long-term financial
debt, we use total consolidated financial debt to capital as a leverage measure.
The sample includes listed companies in 13 member countries of the European Union (EU)
and the European Free Trade Association (EFTA). Financial institutions and companies whose
8
total book value does not exceed €5 million are excluded. Our sample thus contains primarily
industrial, commercial and service companies for which managers have considerable leeway
concerning financial decisions. After trimming 1 , our sample comprises 34,313 firm- year
observations for 5,074 firms over the 1988-2000 period.
Using the Worldscope database gives rise to a bias toward large and listed companies,
whereas such companies may represent a fraction only of the entire population. La Porta, Lopezde-Silanez, Shleifer, and Vishny (1997) show that this fraction is larger for the United Kingdom
and Scandinavian countries than for German and French civil law countries because the former
countries have relatively larger stock markets. 2 If public listing increases information releases,
accounting convergence and access to bond markets, our sample may be biased toward firms with
a more homogeneous financing behavior.
Table 1 presents mean debt levels by country where DTCM is the financial debt to market
capital ratio. Debt levels around Europe are fairly homogeneous. To the extent that comparisons
across studies are meaningful, leverage levels in our sample are somewhat smaller than those
reported in Rajan and Zingales (1995), but the overall ranking remains. In particular, our results
confirm that firms in Germany and the U.K. have smaller mean leverage ratios than those in
France and Italy.
Obviously, our adjusted measure of debt levels cannot account for the impact of all
institutional differences across countries on capital structure. Since we sample firms of developed
and conterminous countries, heterogeneity within the sample might be of less concern than in
broader studies (e.g. Booth et al., 2001; La Porta et al., 1997). Tax regimes may affect capital
structure through the tax shield (Modigliani and Miller, 1963). However, Miller (1977) shows
1
To minimize the impact of outliers, we exclude observations inside the 0.5% and 99.5% tails for the variables
defined in Table 2.
2
Capital structure choice of SME and unlisted companies constitutes a research question that is beyond the scope of
this paper.
9
that once personal taxes are taken into account, debt adds little value, if any, to the firm. Rajan
and Zingales (1995) argue that differences in tax regimes across and within countries render the
appraisal of the marginal tax rate tedious. La Porta, Lopez-de-Silanes, Shleifer, and Vishny
(1998) highlight how differences in investor protections against expropriation, as reflected by
legal rules and the quality of their enforcement, systematically vary across countries depending
on legal origins. La Porta et al. (1997) find that a better protection of outside shareholders
increases external equity finance. Concerning creditors, their measure broadly supports that a
greater protectio n increases debt finance.
2.2 Multivariate analysis
As mentioned previously, an analysis of the determinants of observed leverage ratios is a
prerequisite to investigating debt-equity choice in Europe. We regress leverage ratios on a set of
explanatory variables. Our aims are related both to the works of Rajan and Zingales (1995) and
Hovakimian et al. (2001). First, it is important to show to what extent these variables
homogeneously affect leverage ratios across countries. Perhaps more important is to understand
how the national environment and capital structure theories help explain these results. Second, we
use regression results and theoretical priors to affect a given explanatory variable either to the
target leverage ratio equation or to the debt-equity choice equation. The target leverage ratio is
the debt ratio that firms would choose in the absence of informational asymmetries, transaction
costs, or other adjustment costs.
We use the natural logarithm of sales as a proxy for firm size (SIZE). 3 This is a common
measure for size (Titman and Wessels, 1988; Rajan and Zingales, 1995; Booth et al., 2001). The
ratio of tangible assets to total assets (TANG) is used to proxy for tangible assets (Rajan and
Zingales, 1995). Two variables account for operating performance. Profitability (ROA) is defined
as the ratio of earnings before interest, taxes, depreciation and amortization (EBITDA) to total
assets (Rajan and Zingales, 1995; Miguel and Pindado, 2001). ROA proxies for internal finance
3
Size requires transformation to make it comparable in time and across countries. We thus use constant 2000 Euros.
10
capacity. As we do not correct leverage measures for cash, we include CASH in our set of
explanatory variables. It is the ratio of cash and cash equivalents to total assets. This ratio proxies
for past accumulation of financial slack. We use two variables to proxy for market performance:
MTB and RET. The market-to-book ratio (MTB) is a common measure for growth opportunities
(Rajan and Zingales, 1995; Booth et al., 2001). Since variations in MTB may have nothing to do
with changes in the relative weight of the various types of assets 4 , we introduce the relative
change in market value of equity (RET) to control for price effects. Finally, we also include the
ratio of amortization and depreciation to total assets (ATA) as an explanatory variable to proxy
for non-debt tax shields.
Regressions of determinants of observed debt levels are run using a panel Tobit estimator
with double censoring. Consistency is needed between these regressions and estimations of target
leverage ratios that are performed in section 4. Since target debt ratios are bounded between 0
and 1, censoring at these levels is required. Differences among jurisdictions, varying degrees of
exports by companies, the nature of business, and the risk profile of shareholders and managers
suggest that financial policies differ even within countries. Furthermore, it is likely that
macroeconomic shocks and changes in the institutional context have occurred in recent years. For
these reasons, we prefer panel data analysis, as it allows us to include time effects as well as to
control for the heterogeneity of firms through firm-specific effects. Table 2 contains the results of
country-based regressions of observed debt levels. These results make it possible to allocate
explanatory variables either to target leverage equations or to debt-equity choice equations, and
also to discuss differences across countries.
In each country, SIZE enters regressions with a positive sign. This result is in line with the
hypothesis that larger firms have more stable cash flows and higher target debt levels. Stable cash
flows decrease the probability of bankruptcy and therefore the costs of financial distress. They
also increase the probability that the debt tax shield can be fully exhausted. Since the sign of
4
For example, a discrepancy between the observed share price and the stock value in a perfect world affects MTB.
11
SIZE is in line with the static trade-off model, this variable is included in target leverage
equations.
TANG is positively signed in regressions, which is consistent with the hypothesis that
tangible assets act as collateral. In case of default, tangible assets have higher residual values than
other assets. Since debtholders can request the selling of assets, they will therefore request
tangible assets as collateral. As a result, the greater the tangible assets, the higher the target
leverage ratio. Given that the signs for TANG are consistent with the static trade-off model, this
variable is included in target leverage equations.
ROA and CASH enter regressions with a negative sign. High and stable operating
profitability decreases the probability of bankruptcy and increases the probability of fully
exhausting tax shields. CASH is equivalent to negative debt under the static trade-off hypothesis.
Since an increase in both operating performance variables should increase the target debt level,
the observed relationships are contrary to trade-off expectations. These variables are therefore
directly assigned to debt-equity choice equations. Note that agency considerations would also
lead to a positive impact of operating performance. If debt is a governance device on insiders and
if markets are sufficiently efficient in Europe, a raise in PROF will increase free cash flows,
which in turn will increase the leverage ratio. Managers have a greater preference for CASH than
shareholders because it reduces firm risk and increases their discretion. Excess CASH should thus
be counterbalanced by an increase in the debt ratio.
Empirically, the negative impact of operating performance variables on observed debt levels
does not come as a surprise as it is one of the most documented regularities in capital structure
studies. A common explanation stems from pecking order models where internal financing is
cheap because it avoids underinvestment costs (Myers, 1984; Myers and Majluf, 1984). All other
things equal –i.e. investments, dividends and the level of informationa l asymmetries – , for a
given and reasonable level of debt, a decrease in profitability increases financing needs. Once the
available slack is exhausted, debt finance is required, since debt value is less sensitive to
12
informational asymmetries than equity value. There are other explanations as argued by
Donaldson (1961). Lack of pressure on managers or transaction costs may also result in a
financing behavior where profitable (unprofitable) firms accumulate internal financing (debt)
without rebalancing.
MTB and RET have negative signs in regressions. Negative coefficients for market
performance are expected under several capital structure models. An increase in MTB decreases
the relative weight of assets in place against growth opportunities and therefore reduces the
relative residual value in case of liquidation. As a result, a lower target leverage ratio is expected
under the static trade-off model and MTB should be included in target leverage ratio equations.
However, the negative impact of market performa nce on the leverage ratio may not be due to
the target leverage and accordingly should be included in the debt-equity choice regressions.
Under agency models, a negative impact on the observed leverage is also expected. Agency costs
of debt (Jensen and Meckling, 1976; Myers, 1977; Zweibel, 1996) increase with the ratio of
investment projects to assets in place, while agency costs of managerial discretion (Jung et al.,
1996) diminish with profitable projects. When there are transaction costs, firms do not
systematically counterbalance changes in market value due to stock price effects with debt. If this
hypothesis holds, the effect of market performance on the market leverage ratio is due to a
simultaneity bias. Finally, price effects might also lead managers to try to time the market by
issuing (repurchasing) shares when prices are high (low) (Baker and Wurgler, 2002). The market
timing hypothesis is akin to a pecking order model where variable levels of asymmetric
information lead to negative underinvestment costs. If managers are active, price effects also
negatively impact upon the market leverage ratio. As a result, no straightforward conclusion can
be drawn concerning the impact of market performance on observed leverage ratios.
All else being equal, an increase in amortization reduces the tax burden, the marginal tax rate
and therefore the target debt level. Graham (2000) and Mackie-Mason (1990) find a significant
impact of taxes on the debt level. The ATA variable is not significant in most cases, and when
13
significant, the sign of the coefficient changes across leverage measures and countries. Titman
and Wessels (1988) argue that finding a suitable proxy for substitution effects between the tax
shields is difficult as it requires a measure of tax reductio n that excludes the effect of economic
depreciation expenses. As we do not have such a measure, we exclude ATA from the analyses
contained in subsequent sections.
2.3 Robustness and other further analyses
To evaluate the robustness of our country-based ana lysis of debt ratios, we rerun regressions
using the financial debt to book capital ratio. 5 Book and market based leverage ratios are common
in the literature. Book ratios may control for simultaneous effects between market leverage ratios
and explanatory variables based on stock prices. Managers might be concerned with book debt
levels when these ratios are determinants of bonus schemes, of bank credit policy or of agency
ratings.
Regarding the multivariate results in Table 2, the positive impact of SIZE remains when the
book leverage is used. This systematic sign might come as a surprise. One could expect a stronger
impact in countries where costs of financial distress are higher. Rajan and Zingales (1995) find a
negative impact of SIZE on the debt level of German firms. They fail to offer an explanation in
line with the costs of financial distress hypothesis, as they argue that creditor protection is strong
in Germany, an argument consistent with our results. Despite obvious caveats, a comparison of
estimated coefficients across countries is informative. We compute correlations between the
effect of a one standard deviation change of SIZE on the average debt level and measures of
creditor protection as defined by La Porta et al. (1998). Unreported results provide some support
to the costs of financial distress hypothesis. For example, the more legal rules restrict
reorganizations, the greater the impact of SIZE. When the probability of bankruptcy increases
due to higher creditor powers, stable cash flows of large firms may have a stronger impact on
leverage ratios. Results also show a negative correlation between legal reserve requirements and
5
To save space, results are not reported.
14
the impact of SIZE. The probability of bankruptcy may decrease in smaller firms with an increase
in the legal reserve requirement. When other measures of creditor protection are used, the sign of
correlations changes with measures of leverage ratios, and hence no inference can be drawn.
The positive impact of TANG remains with the book leverage, although it looses its
significance for the U.K. and Switzerland. Regarding cross-country analyses, one might expect a
more limited effect of TANG in countries where banks are important providers of finance (i.e. in
French and German civil law countries) because they have better information about investment
prospects. Again, with all the caveats in mind, we compute the effect of a one standard deviation
change of TANG on the average debt level. The strongest impact of TANG appears in
Scandinavian law countries, then in German civil law countries and next in French civil law
countries. However, the smallest impact is for the U.K., our sole common law country. Thus the
evidence on this issue is somewhat mixed.
There are no differences in estimated coefficients with respect to the two operating
performance variables, ROA and CASH. This is not the case, however, regarding market
performance variables. In most countries, when book leverage is used, one variable is no longer
significant, whereas the other still has a negative impact. The loss in significance of one market
performance variable, as well as the decrease of the impact on the average book leverage of a one
standard deviation change of the other variable, point to agency6 and/or price effects7 . In Austria
and the Netherlands, we document a positive impact of MTB on the book leverage. This effect
could be due to large assets in place which are either off balance sheet or fully depreciated. Such
impact stresses that a control for the national environment is required to assess the effect of
market performance on capital structure. In sum, we affect MTB to target leverage equations, and
MTB and RET to debt-equity choice equations in which we control for national environments.
6
Since the market leverage ratio includes market expectations, it is best suited to test agency models.
7
The book leverage is affected by price effects only if the market timing hypothesis holds.
15
Overall, results reported in Table 2 are consistent across countries and robust to the use of the
book leverage. Most variables enter regressions with the same sign in all European countries. In
particular, SIZE, TANG and MTB appear to be determinants of the target debt level, whereas
CASH and ROA do not. Despite institutional and cultural diversity in Europe, economic forces
appear to be at work in a consistent way on capital structure. However, much uncertainty remains
about the nature of these forces. Are profitable firms underlevered because managers, acting for
shareholders, minimize underinvestment costs or because internal financing gives managers
leeway and is less transparent? Are firms with high levels of market performance underlevered
because managers’ and shareholders’ interests are aligned or because ma nagers believe they can
time the market? Is debt a governance device on insiders? Are observed signs in Table 2 due to
omitted variables? We document these issues in the next sections.
3 The debt-equity choice method
The analysis of debt-equity choice provides for a better understanding of the issues
mentioned above and documents gaps in the empirical literature. Such analysis permits the testing
of the dynamic dimension of financing policy and an understanding of the dominant forces at
play. This is because choices encompass both financing and payout activities, and because
rational managers will modify capital structure only when benefits exceed costs. Dynamic panel
data estimators (Miguel and Pindado, 2003; Gaud et al., 2005) are another empirical design to test
dynamic capital structure models. They estimate endogeneous target leverage ratios and the
adjustment speed within the sample, but they are less flexible. Such estimators are constructed on
the a priori hypothesis that the target debt level exists; an assumption that is not relevant in a
pecking order framework. 8 They are also subject to survivor bias as they require unbalanced
panel data with long time series. The LOGIT estimator used to run debt-equity choice regressions
8
Shyam-Sunder and Myers (1999) point to the poor performance of a dynamic trade-off model compared to a simple
pecking order model to explain changes in debt ratios over time.
16
does not have these drawbacks. Explanatory variables are specific either to adjustments (if any)
to the target leverage ratio or to other theoretically identified forces.
However, the number of significant transactions over the sample period is too small to study
debt-equity choices within country subsamples. We have to assume a common European model
of capital structure and put it on test. Institutional and cultural specificities might affect financing
policy so dramatically that systematic differences across countries might exist. For example,
share repurchases have been legally authorized only since the end of the nineties in some
countries (e.g. Germany and Italy) and repurchase terms still differ across countries. Fortunately,
given the results of the previous section, one might expect that a high level of homogeneity
among capital structure policies in Europe is more than a working hypothesis. In order to control
for and to assess the impact of national environments, we report debt-equity choice regressions
under several specifications. If the national environment dramatically alters a specific choice,
high measures of goodness-of-fit are expected when the vector of explanatory variables is
restricted to time and country dummies. This is the null test of the ‘all national’ hypothesis. If the
national environment alters the capital structure choice, changes in sign on firm-specific variables
are also expected whether country dummies are included or not in the vector of explanatory
variables.
3.1 A two-step model
A shortcoming of traditional international studies of capital structure determinants – the
former section is no exception – is the difficulty to discriminate between the three main models of
capital structure. In particular, even if identical signs are reported for estimated coefficients
across samples, these signs might be consistent with several theoretical models. To avoid this
issue and to formulate discriminating tests of hypotheses, we implement a two-step estimation
design for debt-equity choices in an international setting:
L*it = δYit + ν i + λ t + ηit
(1)
17
~*
P( y it = 1) =
where
e α' L it +β' X it −1
1+ e
~
α' L*it + β ' X it −1
+ λ t + ρi + εit
(2)
L*it
:
target leverage ratio of firm i in year t
P(y it = 1)
:
probability of firm i operating externally in year t
rather than choosing an alternative transaction type
Yit , Χit
:
vectors of explanatory variables
νi
:
random effects for firm i
λt
:
vector of time dummy variables
ρi
:
vector of country dummy variables
ηit , ε it
:
stochastic error terms.
There is a potential bias in Equation (2) as target leverage ratios in that equation are the fitted
values from Equation (1). Standard deviations of estimated coefficients are potentially downward
biased in two-step econometric models. This is because estimated target leverage ratios are
measured with errors. Murphy and Topel (1985) argue for a correction of the variance-covariance
matrix of estimated coefficients. We cannot correct the matrix since we use different samples to
estimate the target debt level. To control for the downward bias in standard deviations, we rerun
Equation (2) with target leverage ratios which are not based on regressions. In these regressions,
we use the mean country-year leverage ratio of the industry as a target.
Since we use annual consolidated accounts to identify debt-equity transactions, we lag all
explanatory variables by one year to limit simultaneity bias. Market timing, if it arises during
windows of opportunity, may be understated due to the lag in market performance variables.
3.2 Types of events
Studies of debt-equity choice focus on significant financing and payout transactions. We use
the traditional 5% cut-off criterion of book value of assets at the beginning of the year to identify
18
debt-equity choices. 9 In addition, a minimum of two years of data surrounding an event is
required. We use consolidated financial statements which make it possible to ana lyze private and
public external financing. Such data are particularly well suited for the European market where
bank finance and private equity are likely to play an important role. Hovakimian (2004) shows
that some transactions have specific determinants. Clear definitions of types of events that
exclude overlapping transactions are therefore necessary.
There are five basic transactions: Two pure financing transactions – equity issues and debt
issues – and three pure payout transactions – share repurchases, dividend increases and debt
reductions. 10 For dividends, we use a 5% cut-off because they are ‘sticky’ and because we want
to identify significant dividend increases. We include significant dividend increases in our event
set because they represent an important channel to pay out cash to shareholders in Europe. Other
transactions simultaneously affect the amount of debt and equity. Hovakimian et al. (2004)
control for market timing through regressions of the choice between equity issues versus
simultaneous debt and equity issues. In both cases, managers can try to time the market. As a
result, a negative effect of market performance can be unequivocally attributed to the investment
prospect (agency) hypothesis. Therefore, we introduce simultaneous debt and equity issues in our
set of events. Due to the small number of significant share repurchases in the sample, we exclude
simultaneous debt and equity repurchases.
9
Changes in equity capital, based on cash flow statements, are defined in Worldscope items as
[tf.SaleOfComAndPfdStkCFStmt - tf.PurchOfComAndPfdStkCFStmt - tf.CashDividendsCFStmt]. Changes in debt,
based on balance sheets, are defined as the difference between the value of [tf.STDebtAndCurPortLTDebt +
tf.TotalLTDebt] and its lagged value.
10
‘Issues (reductions) of debt’ comprise issues (repayments) of bonds and bank loans (repayments of loans).
19
All of the above events are either pure financing or pure payout transactions. 11 We also
consider the absence of an external transaction as a choice and report a ‘No transactions’ event
for firms that have not been active in capital markets for three consecutive years.
4 Empirical results
4.1 Variables
Since the target leverage ratio may explain part of the debt-equity choice, it has to be proxied
for in a first step. Table 3 presents regressions of the debt level on the three firm-specific
variables that have been identified in section 2 as determinants of the target debt ratio (SIZE,
TANG and MTB). Consistent with the static trade-off hypothesis, estimated coefficients have
identical signs across all countries. A low probability of bankruptcy implies higher target
leverage ratios for larger firms. Companies with high levels of tangible assets have higher target
debt levels since these assets can be pledged. High MTB implies higher growth opportunities and
thus higher costs of financial distress, which in turn reduce target debt levels. In Tables 4-6,
target debt ratios are the fitted values from country-based regressions.
Table 4 contains descriptive statistics. Debt levels of issuers are close to those of firms with
‘No transactions’. Debt issuers are more levered than equity issuers (mean DTCM of 26.8%
versus 23.9%), but they also have higher target debt ratios (mean TDTM of 30.0% versus 23.2%).
The spread between target and observed leverage ratios is in line with the adjustment hypothesis,
but its amount is not sufficient to cover the issue size (absOpsize 12 ). Mean absOpsize is
particularly large for equity issuers (24.0% versus 14.9% for debt issuers). The transaction size
matches the spread between target and observed debt ratios more closely for payout events. The
11
We do not deal with mixed debt-equity choices such as equity issues and debt reductions or debt issues and share
repurchases. We only seek to observe specific effects on equity and debt.
12
absOpsize is the ratio of the absolute value of the net amount of the transaction to total assets at the beginning of
the year.
20
average firm reducing debt is highly levered (41.9%) and its observed debt ratio is in excess of its
target (30.0%). Firms paying out cash to shareholders, either through share repurchases or
dividend increases, have low mean levels of debt (18.1% for repurchases and 6.4% for
dividends). These levels are much lower than their target ratio s (26.4% and 17.5%, respectively).
We also note significant negative spreads between observed and target leverage ratios for firms
with ‘No transactions’ (23.2% versus 30.6% on average). Overall, these large spreads suggest a
slow adjustment process, if any.
CASH and ROA levels of issuers are close to firms with ‘No transactions’. Debt issuers are
the most profitable firms (mean ROA of 13.8% versus 11.9% for equity issuers and 13.0% for
firms with ‘No transactions’), but they have the lowest accumulated slack (mean CASH of 9.0%
versus 10.6% and 12.5%, respectively). As for debt ratios, financing and payout events exhibit
contrasting results regarding operating performance. Firms paying out to their shareholders have
high levels of operating performance (mean ROA of 15.6% and 23.7% and mean CASH of
19.4% and 19.6%, for share repurchases and dividend increases, respectively). On the contrary,
firms paying out to their debtholders have modest levels of operating performance (mean ROA of
10.8% and mean CASH of 8.4%). The high proportion (8.2%) of firms with negative ROA
(d2=1) among firms reducing debt also underlines these financial troubles. There is also a high
proportion (10.5%) of firms with negative ROA among equity issuers, but few of them have a
MTB less than 1 (d1=1).
To better proxy for abnormal market performance, we use an adjusted measure of RET.
AdjRET is the RET minus the mean country-year RET. Equity issuers have a very high mean
AdjRET (24.0%), in contrast with other firms. Several types of firms have a negative mean
AdjRET (-7.1% for debt reduction, -6.5% for firms with ‘No transactions’, -4.9% for share
repurchases, and -0.1% for dividend increases). Marked differences exist as pertains to MTB
between these various types of firms. Firms that increase dividends have a high mean MTB
(2.376) and few of them have a MTB that is less than one. Share repurchasers also have a high
21
mean MTB (1.733), but two types of firms coexist since 30.1% of them have a MTB that is less
than one. Firms that reduce debt cumulate the worst negative mean AdjRET, the lowest mean
MTB (1.228) and the highest proportion of firms with a MTB that is less than one (33.6%). Firms
that issue debt are less in trouble since their mean AdjRET is positive (4.8%) and their mean
MTB is 1.458 (1.867 for equity issuers).
Finally, we introduce three other control variables that appear in debt-equity choice studies.
Following Hovakimian et al. (2004), we control for the size of transactions (absOpsize) in
regressions of external debt-equity cho ices. Since the issue size may be jointly determined with
the type of issue, there is a potential simultaneity bias. We exclude absOpsize from regressions
pertaining to the external versus internal choices as the bias is obvious in such regressions. In
external choice regressions, in particular when simultaneous issues are introduced, the size of the
transaction so dramatically changes the capital structure that absOpsize is required as a control
variable. 13 Like Hovakimian et al. (2001) and Hovakimian et al. (2004), we introduce a dummy
variable to control for earnings per share dilution (dEPdil 14 ) because dilution fallacies might
prevent equity issues. Since concerns of wealth transfers from shareholders to creditors might be
stronger with long-term debt, we also introduce the ratio of short-term financial debt to total
financial debt (DST).
4.2 Regression results
Regression results are shown in Tables 5 and 6. Table 5 displays financing choice regression
results, while Table 6 reports payout choice results. When explanatory variables include firmspecific variables, levels of pseudo-R2 are similar to those found in U.S. samples that are
presumably more homogeneous (Hovakimian et al., 2001; Hovakimian, 2004; Hovakimian et al.,
13
The omitting of absOpsize in regressions of pure external debt-equity choices does not affect results.
14
dEPdil is equal to 1 when the after tax cost of debt exceeds the ratio of net profits to market value of equity capital.
The estimated cost of debt is the ratio of interest expense on debt to total debt and the estimated tax rate is the ratio of
income taxes to taxable income.
22
2004). In line with these studies, we document a better explanatory power for payout choice
regressions. More importantly, when regressions are run with time and country dummy variables
exclusively, the explanatory power diminishes significantly, suggesting that the null hypothesis
of ‘all national’ behavior does not hold in our sample. There is an interesting exception however.
The pseudo-R2 does not change dramatically whether firm-specific variables are included or not
in regressions pertaining to share repurchases. Forces driving share repurchases are likely to
include interactions between specific regulations and effects identified in stylized capital
structure models. A simple control with time and country dummies is therefore not sufficient. An
in-depth cross-country analysis, however, is beyond the scope of this paper, and consequently, we
exclude share repurchases from our analysis.
Since we do not document changes in sign whether country dummies are included or not, and
since the low pseudo-R2 s reject the ‘all national’ hypothesis, we conclude that debt-equity
choices of European firms have common driving forces. This is not to say that national
environments do not matter. In fact, results reported in Table 2 and the highly significant Wald
tests on the country dummies in Tables 5 and 6 show that they do matter, but homogeneity in the
whole sample is sufficient for a dummy variable control.
The hypothesis of adjustment to a target debt level is corroborated for firms reducing the
level of leverage, but is rejected for firms increasing leverage. The choice to reduce debt is
positively (negatively) affected by DTCM (TDTCM) in ‘Debt reductions versus No transactions’
regressions. The same impacts are found for equity issues in ‘Equity issues versus No
transactions’ regressions. Although the size of the transaction is closer to the spread between
observed and target leverage ratios for payout transactions, correction of the leverage overload
also drives external equity financing. On the contrary, adjustment to an upper target debt level is
not a primary objective. For example, DTCM enters with a positive sign in ‘Debt issues versus
No transactions’ regressions and estimated coefficients for TDTCM are negative in ‘Dividend
increases versus No transactions’ regressions. These signs result from ‘No transactions’ firms
23
accumulating large downward deviations. As a result, regarding debt levels, we conclude that
firms have an upper barrier only.
Results confirm, however, that firms do try to minimize target deviations once they have
decided to externally change their capital structure. DTCM (TDTCM) negatively (positively)
affects the decision to issue debt rather than equity, while signs are reverse in regressions
pertaining to debt reductions versus dividend increases. In these regressions, firms have already
chosen either to payout or to finance part of their assets. The decision to carry out a transaction
may be mainly motivated by other factors, such as financing needs or a market opportunity, even
though the observed signs on adjustment variables are as expected under the trade-off hypothesis.
This result on external choices is in line with Hovakimian (2004) for a sample of U.S. firms. 15
In summary, our results show that firms suffer little from being durably away from the target
ratio, except when they have to reduce an excess debt level due to the existence of a constraining
upper barrier to leverage. Thus, financial distress costs play an important role. Such behavior
does not reject a pecking order hypothesis. Though the target debt ratio is of secondary
importance in pecking order models, effective leverage affects capital structure choice due to debt
capacity. Highly levered firms can either choose to issue equity or to repay debt since operating
at levels close to debt capacity is expensive because of high bankruptcy costs (Myers and Majluf,
1984). Financial distress is also costly in agency models.
Debt-equity choice regressions help explain the negative impact of profitability on observed
debt levels. We argue that this empirical regularity occurs because two forces, the preference for
internal financing and the debt overhang issue, dominate a third one, i.e. the disciplinary capacity
of debt. With regard to financing transactions, estimated coefficients for the existing slack proxy
(CASH) highlight the preference for internal over external financing. CASH has a negative
impact on the probability of issuing both debt and equity versus ‘No transactions’. The more a
15
He refers to this weak hypothesis concerning the role of target ratios as the ‘debt-equity choice hypothesis ’.
24
firm stockpiles slack, the less it seeks external financing. In line with plain pecking order models,
external financing is not necessary if slack is sufficient to cover financing needs.
The two proxies for operating performance complement one another. Since CASH controls
for preference for internal financing, ROA allows to test whether debt is a governance device.
ROA enters with a positive sign in ‘Debt issues versus No transactions’ and ‘Debt issues versus
Equity issues’ regressions. This impact is specific to debt issues since it remains positive when
simultaneous issues are introduced. This is not consistent with a plain pecking order model where
a rise in ROA increases the need for debt financing. We conclude that for the more profitable
firms, pressures on European managers are strong enough for them to issue debt as a disciplinary
device. These results contrast with those of Hovakimian et al. (2004). In their regressions, which
do not include CASH, they find a specific negative effect of ROA on equity issues. They argue
that, on the one hand, unprofitable firms are likely to positively deviate from their target and will
therefore issue equity rather than debt and, on the other hand, profitable firms do not offset their
negative deviation as they prefer internal financing which is available. We argue that internal
financing, when available, is preferred over external financing, but that firms limit future slack by
issuing debt because it is a source of conflict between managers and external shareholders.
With regard to payout transactions, results for CASH and ROA are in line with agency
models. The observed signs could also be consistent with an excess of slack that has become too
costly. Estimated coefficients for ROA and CASH are positive in ‘Dividend increases versus No
transactions’ regressions and negative in ‘Debt reductions versus Dividend increases’ regressions.
In line with agency models, profitable firms try to raise debt levels and payout slack to their
shareholders, but the pecking order hypothesis cannot be rejected on the grounds that profitable
firms have to preserve their debt capacity. Since firms that choose to reduce their equity have low
leverage ratios and high levels of slack, they may not have debt capacity problems, but rather be
concerned with slack that is too costly. In ‘Debt reductions versus No transactions’ regressions,
estimated coefficients on ROA are positive while the ones on CASH are negative. When debt
25
levels are high, firms appear to prefer debt reductions as they preserve debt capacity and decrease
the costs of financial distress. The negative impact of CASH can be explained by the burden of
existing debt which limits the accumulation of slack.
Investment prospects and misvaluation of assets are two distinct driving forces shaping
capital structure policy that can be proxied by market performance. In Table 5, we observe a
dependence between investment and financing activities, which is consistent with the agency
hypothesis. The disciplinary role of debt should be of greater value to firms with mediocre
investment opportunities, while convergence of interests between managers and shareholders, as
well as agency costs of debt, should enhance the value of equity issues for firms with profitable
projects. We find that estimated coefficients on AdjRET and MTB are negative in ‘Debt issues
versus Equity issues’ regressions. To test whether MTB impacts differently firms with distinct
patterns of investment projects, we introduce d1(MTB) which interacts d1 and MTB. The sum of
the estimated coefficients for MTB and d1(MTB) points to a change in sign 16 from negative to
positive when firms turn from having a MTB higher than one to having a MTB less than one. The
agency hypothesis is also confirmed in regressions pertaining to simultaneous issues of debt and
equity, since we observe a specific negative impact of MTB on equity (debt) issues for low (high)
MTB firms.
The negative sign for estimated coefficients on MTB in ‘Debt issues and Equity issues versus
Equity issues’ regressions cannot stem from market timing strategies. Firms that have to choose
between the above two alternatives have already decided to issue equity, and thus timing is
irrelevant. Firms with mediocre investment projects avoid issuing equity due to lack of
convergence between managers’ and shareholders’ interests, whereas companies with sound
investment opportunities avoid raising the debt level to limit agency costs of debt. We
16
For each change in sign between an estimated coefficient and the sum of the estimated coefficient and that of the
interacted variable, we carry out a linear restriction test to reject the null hypothesis. It is rejected at the 1% level in
all cases.
26
additionally find that AdjRET has a positive impact on equity issues in ‘Debt issues versus Debt
issues and Equity issues’ regressions. Such a specific positive effect of AdjRET on equity issues
is in line with the hypothesis that managers implement market timing strategies. Various degrees
of information asymmetry across firms or time periods, slow assimilation of information or
segmentation of markets are possible explanations for this behavior.
As share repurchases have been excluded from our analysis, market performance cannot
stand for market timing in the payout choice regressions. AdjRET has a negative impact on
dividend increases, since coefficients are positive in ‘Debt reductions versus Dividend increases’
regressions and negative in ‘Dividend increases versus No transactions’ regressions. Since
dividend payers are healthy, they may react to a decrease in their value-enhancing growth
perspectives. This in turn increases their future slack which may lead them to distribute cash
steadily to shareholders, a behavior in line with the agency hypothesis. Market performance
variables appear to somewhat overlap since only d1(MTB) is negative in ‘Dividend increases
versus No transactions’ regressions. Firms in financial trouble cannot afford a payout. The
agency argument also helps explain debt reductions since AdjRET is positive in both ‘Debt
reductions versus Dividend increases’ and ‘Debt reductions versus No transactions’ regressions.
Better prospects increase agency costs of debt and alignment of interests between shareholders
and managers. Again, only healthy firms can afford payout as d1(MTB) enters negatively in
‘Debt reductions versus No transactions’ regressions. The more firms go into financial distress,
the less they are able to reduce debt since the burden becomes too heavy.
Wealth transfers from shareholders to long-term debtholders limit external equity financing.
The sum of the estimated coefficients for DST and d2(DST) 17 points to a change in sign in ‘Debt
issues versus Equity issues’ regressions. The changes in sign show that financially distressed
firms refrain from issuing equity to limit wealth transfers from shareholders to long-term
debtholders. This result is consistent with Hovakimian et al. (2001). In line with their findings
17
The variable d2(DST) interacts d2 and DST.
27
also, the dilution variable dEPdil enters with a negative sign in ‘Equity issues versus No
transactions’ regressions and with a positive sign in ‘Debt issues versus Equity issues’ and ‘Debt
issues versus Debt issues and Equity issues’ regressions, suggesting that the EPS dilution
diminishes the number of share issues.
4.3 Robustness checks
Our results and conclusions regarding (1) adjustment to the target debt ratio, (2) operating
performance and (3) the role played by market performance in financing choices are not affected
when the book leverage is used. Concerning payout events and market performance, our main
conclusions hold. MTB and d1(MTB) gain in significance with the book leverage. The agency
argument is confirmed for firms increasing dividends as MTB is negative (positive) and
d1(MTB) is positive (negative) in ‘Debt reductions versus Dividend increases’ (‘Dividend
increases versus No transactions’) regressions. Regarding debt reductions, the d1(MTB) variable
turns positive in the ‘Debt reductions versus No transactions’ regressions. A possible explanation
to this change in sign is that when MTB is less than one, the book leverage is lower than the
market leverage. As a second robustness check, we use the mean country-year leverage ratio of
the industry, where industry is identified using the three-digit SIC codes. This check is done for
both market and book measures. We do not find evidence that the potential downward bias in
standard deviations due to the use of fitted values hurts estimations.
5 Conclusion
In this paper, we investigate the debt-equity choice using a sample of more than 5,000
European firms over the period 1988-2000. We test trade-off, agency and pecking order models
through a panel analysis of firm- specific determinants of these choices. By focusing on the
possible adjustment to a target debt ratio and on the role played by operating and market
performance, we provide evidence that neither the trade-off nor the pecking order model, in their
28
most commonly accepted forms, offer a suitable description of the capital structure policies in
Europe.
We also document to what extent the national environment affects capital structure. Although
variables traditionally used in international studies consistently impact upon observed debt levels
in country subsamples, the national environment does matter. First, share repurchases appear to
be determined by interactions between national environments and more stylized facts, suggesting
an avenue for future research. Second, country dummy variables are found to be significant in
regressions pertaining to other types of events. This dummy control procedure allows an
investigation of the common factors shaping debt-equity choice of European firms.
Contrary to predictions of dynamic trade-off models, we do not report evidence of a lower
barrier to debt levels. Whereas firms do not suffer much from being significantly below the target
leverage ratio, they do not cross an upper debt level (i.e. the maximum debt capacity). Both
operating and market performance significantly affect debt-equity choice. In particular, debt
constrains managers to payout cash, and equity may become cheap during windows of
opportunity. Internal financing, when available, is preferred over external financing, but firms
limit future excess of slack because it potentially constitutes a source of conflict. Also in line with
the agency hypothesis, profitable firms prefer increasing dividends rather than decreasing debt
levels. We highlight that conflicts between shareholders and debtholders limit external equity
finance. Results further show that managers are trying to time the market, but also that financing
and investment activities interact. Debt does not constitute a suitable form of financing for firms
with value-enhancing investment projects. Instead, such firms issue equity. In contrast, when
there is a lack of profitable projects, debt disciplines managers as firms prefer issuing debt and
increasing dividends.
29
6 References
Baker, M. and J. Wurgler, 2002, Market timing and capital structure, Journal of Finance 57, 1-32.
Booth, L., Aivazian, V., Demirgüç-Kunt, A., and V. Maksimovic, 2001, Capital structure in
developing countries, Journal of Finance 56, 87-130.
Donaldson, G., 1961, Corporate debt capacity: a study of corporate debt policy and the
determination of corporate debt capacity (Harvard Graduate School of Business
Administration, Boston)
Gaud, P., Jani, E., Hoesli, M., and A. Bender, 2005, The capital structure of Swiss firms: an
empirical analysis using dynamic panel data, European Financial Management 11, 51-69.
Graham, J. R., 2000, How big are the tax benefits of debt?, Journal of Finance 55, 1901-1940.
Hovakimian, A., 2004, The role of target leverage in security issues and repurchases, Journal of
Business 77, 1041-1071.
Hovakimian A., Hovakimian, G., and H. Tehranian, 2004, Determinants of target capital
structure: the case of dual debt and equity issues, Journal of Financial Economics 71, 517-540.
Hovakimian, A., Opler, T., and S. Titman, 2001, The debt-equity choice, Journal of Financial and
Quantitative Analysis 36, 1-24.
Jensen, M. and W. Meckling, 1976, Theory of the firm: managerial behavior, agency costs and
capital structure, Journal of Financial Economics 3, 305-360.
Jung K., Kim, Y., and R. Stulz, 1996, Timing, investment opportunities, managerial discretion,
and the security issue decision, Journal of Financial Economics 42, 159-185.
La Porta, R., Lopez-de-Silanes, F., Shleifer, A., and R.W. Vishny, 1997, Legal determinants of
external finance, Journal of Finance 52, 1131-1150.
La Porta, R., Lopez-de-Silanes, F., Shleifer, A., and R.W. Vishny, 1998, Law and finance,
Journal of Political Economy 106, 1113-1155.
MacKie-Mason, J. K., 1990, Do taxes effect corporate financing decisions?, Journal of Finance
45, 1471-1493.
30
Miguel, A. and J. Pindado, 2001, Determinants of capital structure: new evidence from Spanish
panel data, Journal of Corporate Finance 7, 77-99.
Miller M., 1977, Debt and taxes, Journal of Finance 32, 261-275.
Modigliani, F. and M.H. Miller, 1958, The cost of capital, corporate finance, and the theory of
investment, American Economic Review 48, 261-297.
Modigliani, F. and M.H. Miller, 1963, Corporate income taxes and the cost of capital: a
correction, American Economic Review 53, 433-492.
Murphy, K.M. and R.H. Topel, 1985, Estimation and inference in two-step econometric models,
Journal of Business and Economic Statistics 3, 370-379.
Myers, S.C., 1977, Determinants of corporate borrowing, Journal of Financial Economics 5, 147175.
Myers, S.C., 1984, The capital structure puzzle, Journal of Finance 34, 575-592.
Myers, S.C. and N.S. Majluf, 1984, Corporate financing and investment decisions when firms
have information that investors do not have, Journal of Financial Economics 13, 187-221.
Opler, T., Pinkowitz, L., Stulz, R., and Williamson R., 1999, The determinants and implications
of corporate cash holdings, Journal of Financial Economics 52, 3-46.
Rajan, R.G. and L. Zingales, 1995, What do we know about capital structure? Some evidence
from international data, Journal of Finance 50, 1421-1460.
Shyam-Sunder, L. and S.C. Myers, 1999, Testing static tradeoff against pecking order models of
capital structure, Journal of Financial Economics 51, 219-244.
Titman, S. and R. Wessels, 1988, The determinants of capital structure choice, Journal of Finance
43, 1-19.
Zweibel, J, 1996, Dynamic capital structure and managerial entrenchment, American Economic
Review 86, 1197-1215.
31
7 Tables
Table 1
Levels of leverage ratios in Europe
DTCM
United Kingdom
France
Germany
Sweden
Italy
Netherlands
Switzerland
Norway
Denmark
Spain
Belgium
Finland
Austria
Total
0.207
0.328
0.250
0.303
0.374
0.275
0.354
0.388
0.333
0.290
0.325
0.386
0.341
0.275
#
Observations
12,700
5,153
4,403
1,432
1,613
1,620
1,515
1,016
1,316
1,196
864
837
648
34,313
#
Firms
1,772
863
629
267
223
222
194
186
172
167
138
138
103
5,074
This table presents mean values of the leverage ratio by country
and overall. The data are from Worldscope® and the sample
contains firms of 13 European countries for the period 1988-2000.
DTCM is the ratio of total financial debt to total capital, where
capital is the sum of the book value of financial debt and the market
value of equity at the end of the year.
32
Table 2
Determinants of leverage ratios by country
United Kingdom
France
Germany
Sweden
Italy
Netherlands
Switzerland
Norway
Denmark
Spain
Belgium
Finland
Austria
SIZE
TANG
0.015
0.002***
0.034
0.002***
0.020
0.003***
0.027
0.003***
0.069
0.005***
0.036
0.003***
0.023
0.003***
0.023
0.003***
0.030
0.003***
0.034
0.006***
0.013
0.004***
0.028
0.003***
0.031
0.005***
0.066
0.012***
0.175
0.021***
0.253
0.022***
0.421
0.028***
0.182
0.039***
0.207
0.023***
0.189
0.025***
0.401
0.029***
0.203
0.030***
0.203
0.034***
0.311
0.038***
0.242
0.034***
0.276
0.036***
ROA
-0.347
0.012***
-0.560
0.027***
-0.363
0.028***
-0.406
0.046***
-0.685
0.065***
-0.740
0.051***
-0.592
0.061***
-0.163
0.050***
-0.753
0.058***
-1.034
0.068***
-0.324
0.066***
-0.452
0.069***
-0.407
0.088***
CASH
-0.287
0.014***
-0.317
0.023***
-0.374
0.027***
-0.245
0.042***
-0.180
0.055***
-0.482
0.036***
-0.201
0.032***
-0.138
0.038***
-0.304
0.036***
-0.233
0.058***
-0.332
0.050***
-0.223
0.050***
-0.270
0.064***
MTB
-0.040
0.002***
-0.061
0.003***
-0.060
0.003***
-0.030
0.004***
-0.060
0.007***
-0.021
0.005***
-0.046
0.005***
-0.056
0.006***
-0.049
0.003***
-0.051
0.009***
-0.073
0.009***
-0.042
0.006***
-0.067
0.015***
RET
-0.038
0.002***
-0.030
0.003***
-0.030
0.004***
-0.022
0.006***
-0.047
0.007***
-0.027
0.006***
-0.025
0.006***
-0.027
0.005***
-0.039
0.006***
-0.032
0.006***
-0.037
0.010***
-0.043
0.006***
-0.032
0.009***
ATA
-0.037
0.053
0.310
0.081***
-0.036
0.078
-0.542
0.136***
-0.409
0.214*
-0.186
0.107*
0.436
0.104***
-0.020
0.151
0.135
0.183
-0.019
0.253
-0.352
0.200*
-0.162
0.231
0.319
0.273
12,700
Log
likelihood
7,698
5,153
3,427
4,403
2,144
1,432
829
1,613
839
1,620
941
1,515
1,028
1,016
623
1,316
936
1,196
558
864
501
837
594
648
342
n
This table presents Tobit regressions of the DTCM ratio on firm-specific variables. The estimates are double censored at zero at
the lower end and one at the upper end. The data are from Worldscope® and the sample contains firms of 13 European countries
for the period 1988-2000. DTCM is the ratio of total financial debt to total capital, where capital is the sum of the book value of
financial debt and the market value of equity at the end of the year. SIZE is the natural logarithm of sales. TANG is the ratio of
tangible assets to total assets. ROA is the ratio of EBITDA to total assets. CASH is the ratio of cash and cash equivalents to total
assets. MTB is the ratio of the market value of assets (total assets plus market value of equity less book value of equity) to total
assets. RET is the ratio of the annual change in the market value of equity to the market value of equity at the beginning of the
year. ATA is the ratio of depreciation and amortization to total assets. All regressions include time dummies and random effects.
Standard deviations are reported in italics. *** indicates significance at the 1% level. ** indicates significance at the 5% level. *
indicates significance at the 10% level.
33
Table 3
Determinants of target leverage ratios by country
United Kingdom
France
Germany
Sweden
Italy
Netherlands
Switzerland
Norway
Denmark
Spain
Belgium
Finland
Austria
SIZE
TANG
0.014
0.002***
0.025
0.002***
0.020
0.003***
0.029
0.003***
0.055
0.003***
0.050
0.003***
0.024
0.003***
0.022
0.004***
0.038
0.003***
0.013
0.006***
0.022
0.004***
0.189
0.038***
0.132
0.036***
0.114
0.010***
0.206
0.022***
0.273
0.023***
0.484
0.027***
0.296
0.031***
0.315
0.023***
0.280
0.020***
0.474
0.030***
0.262
0.034***
0.306
0.035***
0.323
0.036***
0.030
0.004***
0.017
0.005***
MTB
-0.066
0.002***
-0.085
0.003***
-0.082
0.003***
-0.045
0.005***
-0.079
0.007***
-0.057
0.005***
-0.070
0.005***
-0.071
0.006***
-0.083
0.003***
-0.094
0.009***
-0.088
0.007***
-0.058
0.005***
-0.133
0.012***
12,700
Log
likelihood
6,745
5,153
3,005
4,403
1,978
1,432
763
1,613
764
1,620
784
1,515
944
1,016
594
1,316
766
1,196
438
864
453
837
542
648
317
n
This table presents Tobit regressions of the DTCM ratio on firm-specific
variables. The estimates are double censored at zero at the lower end and one
at the upper end. The data are from Worldscope® and the sample contains
firms of 13 European countries for the period 1988-2000. DTCM is the ratio of
total financial debt to total capital, where capital is the sum of the book value
of financial debt and the market value of equity at the end of the year. SIZE is
the natural logarithm of sales. TANG is the ratio of tangible assets to total
assets. MTB is the ratio of the market value of assets (total assets plus market
value of equity less book value of equity) to total assets. All regressions include
time dummies and random effects. Standard deviations are reported in italics.
*** indicates significance at the 1% level. ** indicates significance at the 5%
level. * indicates significance at the 10% level.
34
Table 4
Descriptive statistics of debt-equity choices in Europe
DTCM
TDTCM
ROA
CASH
AdjRET
MTB
DST
dEPdil
absOpsize
% d1=1
% d2=1
n
Debt issues
Equity issues
Share
repurchases
Dividend
increases
Debt
reductions
Debt and
Equity issues
No
transactions
0.268
0.300
0.138
0.090
0.048
1.458
0.464
0.526
0.149
21.6%
2.6%
5,739
0.239
0.232
0.119
0.106
0.240
1.867
0.446
0.303
0.240
12.4%
10.5%
667
0.181
0.264
0.156
0.194
-0.049
1.733
0.405
0.603
0.127
30.1%
2.7%
73
0.064
0.175
0.237
0.196
-0.010
2.376
0.537
0.439
0.084
3.7%
0.9%
923
0.419
0.300
0.108
0.084
-0.071
1.228
0.504
0.443
0.119
33.6%
8.2%
4,157
0.213
0.246
0.143
0.101
0.366
1.953
0.450
0.463
0.689
10.7%
5.1%
842
0.232
0.306
0.130
0.125
-0.065
1.304
0.427
0.443
0.025
31.0%
2.4%
5,917
This table presents the mean values of the variables used in the debt-equity choice analysis. The data are from Worldscope®
and the sample contains firms of 13 European countries for the period 1988-2000. All data except TDTCM and absOpsize
are lagged by one period. DTCM is the ratio of total financial debt to total capital, where capital is the sum of the book
value of financial debt and the market value of equity at the end of the year. TDTCM is the fitted value of regressions in
Table 3. ROA is the ratio of EBITDA to total assets. CASH is the ratio of cash and cash equivalents to total assets. AdjRET
is the difference between firm RET and the country-year RET, where RET is the ratio of the annual change in the market
value of equity to the market value of equity at the beginning of the year. MTB is the ratio of the market value of assets (total
assets plus market value of equity less book value of equity) to total assets. DST is the ratio of financial debt due in one year
to total financial debt. dEPdil is a dummy variable equal to 1 when the after-tax cost of debt exceeds the ratio of net profits
to market value of equity, and to zero otherwise. absOpsize is the ratio of the amount of the transaction to total book assets.
d1 is a dummy variable equal to 1 for firms with a MTB less than one, and to zero otherwise. d2 is a dummy variable equal
to 1 for firms with a ROA less than zero, and to zero otherwise.
35
Table 5
Determinants of financing choices in Europe
ROA
CASH
AdjRET
3.213
0.588***
4.300
0.530***
1.540
0.673**
1.443
0.660**
-0.636
0.424
-0.484
0.416
-0.207
0.057***
-0.202
0.057***
-0.168
0.066**
-0.115
0.063*
0.803
0.165***
0.606
0.159***
0.638
0.160***
0.711
0.156***
-0.889
0.373**
-0.886
0.358**
Debt issues versus
Debt issues and
Equity issues
-1.114
0.344***
-0.646
0.333**
1.706
0.626***
3.259
0.562***
0.225
0.744
-0.228
0.737
-0.196
0.524
0.380
0.535
-0.259
0.064***
-0.253
0.062***
0.003
0.081
0.068
0.080
0.747
0.181***
0.571
0.177***
0.457
0.167***
0.610
0.167***
Debt issues and
Equity issues versus
Equity issues
-0.144
0.418
-0.264
0.407
2.990
0.843***
2.506
0.666***
1.744
0.778**
1.827
0.782**
0.217
0.619
-0.021
0.595
0.038
0.086
0.039
0.084
-0.209
0.097**
-0.220
0.096**
0.103
0.221
0.124
0.218
2.361
0.142***
2.277
0.139***
-0.155
0.272
-0.467
0.230**
1.493
0.347***
1.622
0.344***
-3.139
0.232***
-3.054
0.223***
0.453
0.051***
0.424
0.048***
0.368
0.058***
0.364
0.056***
3.516
0.329***
3.292
0.304***
-2.871
0.708***
-5.220
0.677***
-1.737
0.696**
-1.152
0.682*
-2.465
0.544***
-2.659
0.526***
0.791
0.098***
0.749
0.099***
0.416
0.093***
0.313
0.101***
Debt issues versus
Equity issues
Debt issues versus
No transactions
Equity issues versus
No transactions
DTCM
TDTCM
-1.613
0.315***
-1.208
0.299***
MTB
d1(MTB)
DST
d2(DST)
dEPdil
absOpsize
Wald t
Wald c
0.766
0.098***
0.637
0.096***
-1.353
0.174***
-1.338
0.165***
94.5***
89.4***
-0.508
0.393
-0.681
0.391*
0.406
0.101***
0.260
0.099***
-5.005
0.341***
-4.968
0.326***
0.243
0.222
0.205
0.216
-0.317
0.528
-0.253
0.531
0.548
0.124***
0.559
0.121***
4.102
0.703***
4.102
0.699***
-0.665
0.066***
-0.595
0.065***
0.709
0.071***
0.586
0.069***
0.818
0.229***
0.840
0.230***
0.267
0.043***
0.273
0.041***
-1.606
0.178***
-1.363
0.167***
0.285
0.152*
0.168
0.151
1.368
0.385***
1.536
0.355***
-0.524
0.103***
-0.331
0.099***
113.3***
83.0***
34.8***
-1,870
0.126
-2,017
-1,674
0.058
0.335
-1,750
0.305
-2,372
-802
0.058
0.226
-810
0.218
-1,003
-7,256
0.031
0.102
-7,426
0.081
-7,806
-1,689
0.034
0.218
-1,802
0.165
-1,993
0.077
1,509
24.2**
344.3***
11,656
277.2***
174.0***
6,584
61.7***
36.9***
6,406
375.6***
38.4***
171.6***
253.0***
50.6***
Pseudo
2
R
0.150
6,581
26.2***
41.0***
171.4***
Log
likelihood
-1,820
277.2***
144.2***
43.2***
29.5***
26.8***
n
349.0***
This table contains the results using the LOGIT estimator for financing choice regressions. The data are from Worldscope® and the sample contains firms of 13 European countries for the period 19882000. All data except TDTCM and absOpsize are lagged by one period. DTCM is the ratio of total financial debt to total capital, where capital is the sum of the book value of financial debt and the market
value of equity at the end of the year. TDTCM is the fitted value of regressions in Table 3. ROA is the ratio of EBITDA to total assets. CASH is the ratio of cash and cash equivalents to total assets. AdjRET
is the difference between firm RET and the country-year RET, where RET is the ratio of the annual change in the market value of equity to the market value of equity at the beginning of the year. MTB is
the ratio of the market value of assets (total assets plus market value of equity less book value of equity) to total assets. d1(MTB) interacts d1 and MTB, where d1 is a dummy variable equal to 1 for firms
with a MTB less than one, and to zero otherwise. DST is the ratio of financial debt due in one year to total financial debt. d2(DST) interacts d2 and DST, where d2 is a dummy variable equal to 1 for firms
with a ROA less than zero, and to zero otherwise. dEPdil is a dummy variable equal to 1 when the after-tax cost of debt exceeds the ratio of net profits to market value of equity, and to zero otherwise.
absOpsize is the ratio of the amount of the transaction to total book assets. Estimated coefficients for the country and year dummy variables are not reported. Robust standard deviations are reported in
italics. *** indicates significance at the 1% level. ** indicates significance at the 5% level. * indicates significance at the 10% level. Wald t is a test of the joint significance of time dummy variables. Wald
c is a test of the joint significance of the country dummy variables. Wald t and c are asymptotically distributed as ?2 under the null hypothesis of no relationship.
36
Table 6
Determinants of payout choices in Europe
DTCM
TDTCM
ROA
CASH
AdjRET
MTB
d1(MTB)
Debt reductions
versus Share
repurchases
11.702
1.726***
12.007
1.723***
-6.695
2.824**
-5.283
2.093**
-0.913
2.064
-1.208
2.213
-6.827
1.132***
-5.644
0.962***
0.383
0.194**
0.473
0.226**
-0.217
0.168
-0.165
0.171
-2.025
0.384***
-1.984
0.381***
Debt reductions
versus Dividend
increases
18.421
1.442***
18.563
1.359***
-1.460
0.820*
-0.929
0.722
-7.806
1.014***
-8.216
0.993***
-3.201
0.656***
-2.761
0.579***
0.632
0.152***
0.605
0.153***
-0.018
0.097
0.068
0.081
Share repurchases
versus No
transactions
0.082
0.800
-0.256
0.775
2.460
2.154
0.174
1.588
3.317
4.408
3.058
3.951
3.862
0.933***
3.202
0.850***
-0.004
0.216
-0.106
0.208
Dividends increases
versus No
transactions
-3.184
0.519***
-3.189
0.499***
-3.895
0.702***
-4.610
0.673***
11.789
0.729***
11.857
0.726***
0.654
0.391*
0.520
0.358
Debt reductions
versus No
transactions
7.058
0.179***
6.913
0.174***
-4.362
0.352***
-4.840
0.305***
2.361
0.411***
2.387
0.402***
-1.665
0.256***
-1.558
0.241***
DST
d2(DST)
dEPdil
absOpsize
Wald t
Wald c
1.688
0.424***
1.612
0.405***
0.832
1.734
1.375
1.798
-0.577
0.278**
-0.589
0.273**
-5.048
2.114**
-6.113
2.103***
29.6***
43.3***
-0.379
0.292
-0.428
0.292
0.833
0.211***
0.770
0.204***
-0.950
0.650
-0.921
0.684
-0.477
0.134***
-0.491
0.132***
9.268
5.567*
8.982
5.294*
0.372
0.167**
0.273
0.149*
0.486
0.340
0.524
0.327
-0.288
0.392
-0.449
0.384
0.834
1.633
0.676
1.409
0.461
0.255*
0.577
0.278**
-0.334
0.155**
-0.369
0.163**
0.076
0.075
0.037
0.078
-1.086
0.225***
-0.917
0.224***
0.649
0.141***
0.649
0.136***
2.561
0.550***
2.379
0.544***
0.286
0.099***
0.370
0.094***
0.328
0.052***
0.278
0.049***
0.110
0.064*
0.100
0.064
-1.067
0.077***
-0.968
0.076***
1.455
0.088***
1.226
0.085***
1.733
0.242***
1.778
0.240***
0.000
0.053
0.031
0.051
26.3***
36.4***
40.1***
-240
0.349
-321
-843
0.129
0.650
-860
0.643
-2,273
-322
0.056
0.182
-355
0.100
-339
-1,746
0.140
0.355
-1,801
0.335
-2,514
-5,109
0.071
0.252
-5,295
0.225
-6,625
0.030
5,990
67.7***
220.2***
6,840
405.6***
457.8***
10,074
42.1***
62.4***
4,230
204.5***
71.2***
28.2***
50.3***
37.3***
Pseudo
2
R
0.400
5,080
26.4***
33.2***
22.6**
Log
likelihood
-221
51.5***
31.9***
39.1***
79.3***
33.2***
n
376.6***
This table contains the results using the LOGIT estimator for payout choice regressions. The data are from Worldscope® and the sample contains firms of 13 European countries for the period 1988-2000.
All data except TDTCM and absOpsize are lagged by one period. DTCM is the ratio of total financial debt to total capital, where capital is the sum of the book value of financial debt and the market value
of equity at the end of the year. TDTCM is the fitted value of regressions in Table 3. ROA is the ratio of EBITDA to total assets. CASH is the ratio of cash and cash equivalents to total assets. AdjRET is the
difference between firm RET and the country-year RET, where RET is the ratio of the annual change in the market value of equity to the market value of equity at the beginning of the year. MTB is the
ratio of the market value of assets (total assets plus market value of equity less book value of equity) to total assets. d1(MTB) interacts d1 and MTB, where d1 is a dummy variable equal to 1 for firms with
a MTB less than one, and to zero otherwise. DST is the ratio of financial debt due in one year to total financial debt. d2(DST) interacts d2 and DST, where d2 is a dummy variable equal to 1 for firms with
a ROA less than zero, and to zero otherwise. dEPdil is a dummy variable equal to 1 when the after-tax cost of debt exceeds the ratio of net profits to market value of equity, and to zero otherwise.
absOpsize is the ratio of the amount of the transaction to total book assets. Estimated coefficients for the country and year dummy variables are not reported. Robust standard deviations are reported in
italics. *** indicates significance at the 1% level. ** indicates significance at the 5% level. * indicates significance at the 10% level. Wald t is a test of the joint significance of time dummy variables. Wald
c is a test of the joint significance of the country dummy variables. Wald t and c are asymptotically distributed as ?2 under the null hypothesis of no relationship.
37
The FAME Research Paper Series
The International Center for Financial Asset Management and Engineering (FAME) is a private foundation created
in 1996 on the initiative of 21 leading partners of the finance and technology community, together with three
Universities of the Lake Geneva Region (Switzerland). FAME is about Research, Doctoral Training, and Executive
Education with “interfacing” activities such as the FAME lectures, the Research Day/Annual Meeting, and the
Research Paper Series.
The FAME Research Paper Series includes three types of contrib utions: First, it reports on the research carried out
at FAME by students and research fellows; second, it includes research work contributed by Swiss academics and
practitioners interested in a wider dissemination of their ideas, in practitioners' circles in particular; finally,
prominent international contributions of particular interest to our constituency are included on a regular basis.
Papers with strong practical implications are preceded by an Executive Summary, explaining in non-technical terms
the question asked, discussing its relevance and outlining the answer provided.
Martin Hoesli is acting Head of the Research Paper Series. Please email any comments or queries to the following
address: Martin.Hoesli@hec.unige.ch.
The following is a list of the 10 most recent FAME Research Papers. For a complete list, please visit our website at
www.fame.ch under the heading ‘Faculty and Research, Research Paper Series, Complete List’.
N°151
Spatial Dependence, Housing Submarkets, and Housing Prices
Steven C. BOURASSA, University of Louisville, USA; Eva CANTONI, Department of Econometrics, University of
Geneva; Martin HOESLI, University of Geneva, HEC, FAME and University of Aberdeen
N°150
Estimation of Jump-Diffusion Precesses via Emirical Characteristic Function
Maria SEMENOVA, HEC Lausanne and FAME; Michael ROCKINGER, HEC Lausanne and FAME
N°149
Suggested vs Actual Institutional Allocation to Real Estate in Europe: A Matter of Size?
Martin HOESLI, HEC, University of Geneva and FAME and University of Aberdeen; Jon LEKANDER, Aberdeen
Property Investors, Stockholm
N°148
Monte Carlo Simulations for Real Estate Valuation
Martin HOESLI, HEC Geneva and FAME and University of Aberdeen; Elion JANI, HEC Geneva; André BENDER,
HEC Geneva and FAME
N°147
Equity and Neutrality in Housing Taxation
Philippe THALMANN, Ecole Polytechnique Fédérale de Lausanne
N°146
Order Submission Strategies and Information: Empirical Evidence from NYSE
Alessandro BEBER, HEC Lausanne and FAME; Cecilia CAGLIO, U.S. Security and Exchange Commission
N°145
Kernel Based Goodness-of-Fit Tests for Copulas with Fixed Smoothing Parameters
Olivier SCAILLET, HEC Geneva and FAME
N°144
Multivariate wavelet-based shape preserving estimation for dependent observations
Antonio COSMA, Instituto di Finanza, University of Lugano, Lugano, Olivier SCAILLET, HEC Geneva and
FAME, Geneva, Rainier von SACHS, Institut de statistique, Université catholique de Louvain
N°143
A Kolmogorov-Smirnov type test for shortfall dominance against parametric alternatives
Michel DENUIT, Universitéde Louvain, Anne-Cécile GODERNIAUX, Haute Ecole Blaise Pascal Virton, Olivier
SCAILLET, HEC, University of Geneva and FAME
N°142
Times-to-Default: Life Cycle, Global and Industry Cycle Impacts
Fabien COUDREC, University of Geneva and FAME, Olivier RENAULT, FERC, Warwick Business School
International Center FAME - Partner Institutions
The University of Geneva
The University of Geneva, originally known as the Academy of Geneva, was founded in 1559 by Jean
Calvin and Theodore de Beze. In 1873, The Academy of Geneva became the University of Geneva with the
creation of a medical school. The Faculty of Economic and Social Sciences was created in 1915. The
university is now composed of seven faculties of science; medicine; arts; law; economic and social sciences;
psychology; education, and theology. It also includes a school of translation and interpretation; an institute
of architecture; seven interdisciplinary centers and six associated institutes.
More than 13’000 students, the majority being foreigners, are enrolled in the various programs from the
licence to high-level doctorates. A staff of more than 2’500 persons (professors, lecturers and assistants) is
dedicated to the transmission and advancement of scientific knowledge through teaching as well as
fundamental and applied research. The University of Geneva has been able to preserve the ancient European
tradition of an academic community located in the heart of the city. This favors not only interaction between
students, but also their integration in the population and in their participation of the particularly rich artistic
and cultural life. http://www.unige.ch
The University of Lausanne
Founded as an academy in 1537, the University of Lausanne (UNIL) is a modern institution of higher
education and advanced research. Together with the neighboring Federal Polytechnic Institute of Lausanne,
it comprises vast facilities and extends its influence beyond the city and the canton into regional, national,
and international spheres.
Lausanne is a comprehensive university composed of seven Schools and Faculties: religious studies; law;
arts; social and political sciences; business; science and medicine. With its 9’000 students, it is a mediumsized institution able to foster contact between students and professors as well as to encourage
interdisciplinary work. The five humanities faculties and the science faculty are situated on the shores of
Lake Leman in the Dorigny plains, a magnificent area of forest and fields that may have inspired the
landscape depicted in Brueghel the Elder's masterpiece, the Harvesters. The institutes and various centers of
the School of Medicine are grouped around the hospitals in the center of Lausanne. The Institute of
Biochemistry is located in Epalinges, in the northern hills overlooking the city. http://www.unil.ch
The Graduate Institute of International Studies
The Graduate Institute of International Studies is a teaching and research institution devoted to the study of
international relations at the graduate level. It was founded in 1927 by Professor William Rappard to
contribute through scholarships to the experience of international co-operation which the establishment of
the League of Nations in Geneva represented at that time. The Institute is a self-governing foundation
closely connected with, but independent of, the University of Geneva.
The Institute attempts to be both international and pluridisciplinary. The subjects in its curriculum, the
composition of its teaching staff and the diversity of origin of its student body, confer upon it its
international character. Professors teaching at the Institute come from all regions of the world, and the
approximately 650 students arrive from some 60 different countries. Its international character is further
emphasized by the use of both English and French as working languages. Its pluralistic approach - which
draws upon the methods of economics, history, law, and political science - reflects its aim to provide a
broad approach and in-depth understanding of international relations in general. http://heiwww.unige.ch
Prospect Theory
and Asset Prices
Nicholas BARBERIS
University of Chicago
Ming HUANG
40, Bd. du Pont d’Arve
PO Box, 1211 Geneva 4
Switzerland
Tel (++4122) 312 09 61
Fax (++4122) 312 10 26
http: //www.fame.ch
E-mail: admin@fame.ch
Stanford University
Tano SANTOS
University of Chicago
2000 FAME Research Prize
Research Paper N° 16
FAME - International Center for Financial Asset Management and Engineering
THE GRADUATE INSTITUTE OF
INTERNATIONAL STUDIES