impact of financial crisis on firms’ capital structure in ......university of kent, uk this study...

1

Impact of Financial Crisis on Firms’ CapitalStructure in UK, France, and Germany

Abdullah IqbalUniversity of Kent, UK

Ortenca Kume*University of Kent, UK

This study examines the impact of the recent financial crisis on the capitalstructure decision of UK, French and German firms. The results show thatoverall leverage ratios increase from pre-crisis (2006 and 2007) to crisis (2008and 2009) years and then decrease in the post-crisis (2010 and 2011) years.Both equity and debt levels change during the crisis and post-crisis years. Thefindings further reveal that firms with lower than industry average capitalstructure ratios in the pre-crisis period experience a gradual increase in theirleverage during crisis and post-crisis periods. However, firms with higher thanindustry average capital structure ratios in the pre-crisis periods experience asignificant decrease in the leverage ratios particularly in the post-crisis periodmainly due to changes in their equity levels. (JEL: G14, G15, G32)

Keywords: financial crisis; capital structure; leverage; UK; France; Germany

I. Introduction

Most of the extant literature shows that changes in capital structureaffect firm value. Prior studies (Demirgüç-Kunt and Maksimovic, 1999;Booth et al., 2001; Graham, Leary and Roberts, 2014, etc.) find that thecapital structure decision is influenced not only by firm-specific factorsbut also by institutional settings and macroeconomic uncertainty. Thisstudy examines changes in capital structures of non-financial/non-utilityEuropean firms around the crisis period of 2007–08. It focuses mainly

* Dr Ortenca Kume, Lecturer in Finance, Kent Business School, University of Kent, Canterbury, CT2 7PE, UK. Email: [email protected]

(Multinational Finance Journal, 2014, vol. 18, no. 3/4, pp. 249–280)© Multinational Finance Society, a nonprofit corporation. All rights reserved. DOI: 10.17578/18-3/4-3

Multinational Finance Journal250

on the UK, France, and Germany because these countries represent themost developed countries in Europe. Further, these countries havedifferent financial and institutional characteristics with UK being amarket-based economy (similar to the US) and France and Germany onthe other side being typical bank-based economies (similar to Japan).There are differences even among firms operating in bank-basedeconomies. For example, Brun et al. (2013) argue that German firmsemploy higher levels of debt than their UK and French counterparts dueto their long-term relationship with “HausBanks”. Bancel and Mittoo(2011) note that French economy was hit harder than UK and Germaneconomies during the crisis. They also show that unlike their Germancounterparts, French firms relied heavily on trade financing. However,none of these papers examines the time-series variation in leveragelevels across firms in all three countries (the UK, France, and Germany)around the recent financial crisis.

The major contribution of this study is that it provides an insight onthe impact of 2007–08 financial crisis on capital structure decisions ofnon-financial/non-utility firms in market-based (UK) and bank-based(France and Germany) economies. Secondly, it identifies the financingalternatives (short-term debt, long-term debt or equity) that are used byfirms operating in these economies around the crisis period. Finally, itshows how leverage ratios change during and after the crisis period forfirms with conservative and aggressive pre-crisis leverage ratios.

The results indicate that, on average, leverage ratios for firms in theUK and Germany increase from pre-crisis (2006–2007) to crisis period(2008–2009) but decrease from crisis (2008–2009) to post-crisis(2010–2011) period. However, leverage ratios from pre-crisis level topost-crisis level are not significantly different from each othersuggesting that post-crisis leverage ratios of these firms are back tosimilar levels as they were in the pre-crisis period with an adjustmentin the interim (i.e. crisis period). This study finds no evidence of asignificant change in the leverage ratios from pre- to crisis and fromcrisis to post-crisis period for French firms. In addition, leverage ratiosof firms with pre-crisis conservative capital structure across the threecountries increase significantly during the crisis period due to higherdebt levels. Contrarily, leverage ratios of firms with pre-crisisaggressive capital structure decrease from pre- to post-crisis period.While the gradual decrease in leverage ratio from pre- to post-crisisperiod for the UK firms is due to significant and gradual changes inboth debt and equity, the significant change in the leverage ratio for

251Impact of Financial Crisis on Firms’ Capital Structure

German and French firms occurs in the post-crisis period and is mainlydue to the increase in the levels of equity. The panel data regressionresults also confirm that crisis period has had a significant impact on theleverage ratios of both aggressive and conservative subsamples. Overall,these observations are consistent with those of Graham and Harvey(2001), Leary and Roberts (2005), Fosberg (2012), and Graham, Learyand Roberts (2014). Finally, the results show that changes in capitalstructure can be attributed to an active use of both short-term andlong-term debt as well as equity, during and after the crisis period.

The rest of the study is organised as follows. Section II reviews theexisting literature on leverage changes in different countries duringdifferent crisis periods. Section III describes data selection and researchmethodology. Section IV presents empirical findings. Finally, SectionV summarises and concludes the paper.

II. Literature review

The financial crisis that started at the end of 2007 in the subprime creditmarket led to a liquidity crisis in the short-term money markets(Brunnermeier, 2008; Fosberg 2012, etc). The crisis had itsconsequences in many European countries where troubled mortgageproviders or banks were rescued (Hodson and Quaglia, 2009; Alter andSchόler, 2012). Consequently banks were asked to issue additionalequity to maintain the minimum required capital ratios. Lack of banks’confidence on each other’s financial securities led to an increase in theinterbank lending rates and consequently to a reduction in the supply ofloans to non-financial firms (Fosberg 2012).

Kahle and Stulz (2013) argue that the financial crisis of 2007–08created a supply shock even in the equity markets due to the flight toquality in bond markets, which made it costly for high-levered firms toraise additional equity. In addition, this crisis resulted in a lowerdemand for consumption and higher uncertainty about economicrecovery thus leading to a decline in demand for products and servicesand resultantly a fall in debt and equity issuance and an increase in cashholdings due to postponement of investments. Graham, Leary, andRoberts (2014) report that during periods of economic downturns (oruncertainty), investment opportunities are rare and hence the need forexternal capital is weak leading to a reduction in firms’ leverage ratios.Based on a survey of U.S. chief financial officers’ views about the


impact of 2007–08 crisis on firms’ financial performance, Campello,Graham, and Harvey (2010) find that due to the reduction in demand aswell as cash flows, additional funding was perceived costly and difficultto raise thus there was no additional demand for funding.

Overall, empirical evidence on the impact of various financial criseson firms’ capital structure is sparse. Lim (2003) finds that large Koreanfirms left financial intermediaries and turned to capital markets after theKorean crisis. Similarly, Voutsinas and Werner (2011) note thatextreme credit supply fluctuations in Japan had a significant negativeimpact on Japanese firms’ financial leverage levels. This impact wasmore pronounced for bank-dependent firms. Consistent with theprevious studies, Balsari and Kirkulak (2010) report a negative impactof 1994 crisis on Turkish firms’ leverage ratios, but note that the rise inshort-term debt and the fall in Turkish firms’ equity levels during the2001–02 financial crisis led to a positive impact of this crisis on firms’leverage ratios.

Empirical studies considering the impact of 2007–08 financial crisison capital markets (Fosberg, 2012 and Kahle and Stulz, 2013) revealthat, in general, firms relied heavily on the use of public debt marketsduring the crisis. Kahle and Stulz (2013) note that net debt issuancesincreased during the first year of the 2007–08 crisis for bothbank-dependent and non-bank-dependent firms, but fell after 2008.Fosberg (2012) also report significant increase in debt ratios of USfirms over the pre-crisis period of 2006–08 followed by a gradualdecline in debt levels by the end of 2010 (i.e. post-crisis period) to thepre-crisis level. Pattani, Vera, and Wackett (2011) observe similarpatterns in public debt issued by UK firms. They also report an increasein public equity issuance in 2008–09 and a decline in 2009–10. Theyfurther note that the increase in equity issuance (not a first time issue)was mainly used for bank loan repayments because managers perceivedtheir firms to have high pre-crisis leverage ratios. On the contrary, theproceeds from equity issuance in 2010–11 were used to finance newprojects. Akbar, Rehman, and Ormrod (2013) find that the crisis did nothave a significant impact on the long-term financing channels for UKprivate firms, but impaired the financing channels of short-term debtand trade-credit for these firms. They also suggest that in order to hedgeagainst the negative impact of credit contractions, UK private firms heldmore cash and issued more equity. Similarly, Brun et al. (2013) arguethat the increase in equity of French firms after the crisis resultedmainly from the increase in retained earnings particularly for SMEs and


an increase in the issue premiums received by large firms. Overall, prior literature offers mixed results on changes in capital

structure during various crises periods in different countries/institutionalsettings. This study aims to shed further light on changes in leverageratios of European firms during and after 2007–08 financial crisis.

III. Data and methodology

The sample in the study consists of firms from three major Europeancountries that are UK, France, and Germany over 2006–11 period. Therelevant data are extracted from Datastream. Initial sample is selectedusing the following criteria:

1. Firms are listed on London Stock Exchange for UK, EuronextParis for France, and Frankfurt Stock Exchange for Germany.

2. Firms operate in non-financial and non-utility sectors.

These criteria produce initial samples of 1748 firms for UK, 1622for France, and 1345 for Germany. Outliers and firms with negativemarket values and negative capital structure ratios are also excluded.Both for France and Germany, some firms appear twice or more withsimilar figures in the shortlisted sample. Upon investigation, it is foundthat Datastream uses separate codes for the same firm if it has issuedfurther equity, thus such French and German firms are also removed.Firms that do not belong to any Datastream Level-3 Industry Group arealso excluded. Finally, firms for which Datastream does not provideenough data to estimate the leverage ratio ((short-term debt + long-termdebt)/total capital) during either the ‘pre-crisis and crisis’ years or‘crisis and post-crisis’ years are removed too.

These restrictions produce samples of 871 firms for UK, 564 forFrance, and 392 for Germany across 13 Level-3 Datastream industrysectors, as reported in table 1. The study uses firm-year observations foranalysis, which vary depending on the availability of relevant data.

Capital structure or leverage ratio (LEV) is estimated for each yearacross all firms for each industry group. Similar to Al-Najjar andHussainey (2011), capital structure ratio is measured as total debt tototal capital ratio, where total debt is estimated as the sum of total (long-and short-term) loans and preference capital, and total capital includes


both total debt and total shareholders’ equity.1 This measure of leverageratio is also consistent with previous studies undertaken in this area(Harris and Raviv, 1991; Rajan and Zingales, 1995; Chen, 2004; Learyand Roberts, 2005; Huang and Song, 2006).

Financial years 2006 and 2007 are defined as ‘pre-crisis’ period,2008 and 2009 as ‘crisis’ period, and 2010 and 2011 as ‘post-crisis’period. The study employs t-tests for difference in means (assumingunequal variances) to identify if equally-weighted mean leverage ratiosare significantly different from each other during the three periodsnamely ‘pre-crisis to crisis’, ‘crisis to post-crisis’, and ‘pre-crisis topost-crisis’. Prior literature suggests that firms have target leverageratios (Leary and Roberts, 2005) and they use debt and/or equity toadjust their interim capital structure. This study uses t-test to examinewhether there are any significant changes in the levels of debt (short-and/or long-term) or equity across the three periods.

This study further investigates the changes in leverage ratios ofsample firms by classifying them into two subsamples based on whethertheir pre-crisis leverage ratios are higher or lower than their

TABLE 1. Distribution of sample firms by industry (Datastream Level-3) for eachcountry

Sector UK France Germany

Automobiles & Parts — 15 17Basic Resources 57 14 9Chemicals 15 15 22Construction and Materials 35 22 15Food and Beverage 35 42 17Healthcare 72 45 28Ind. Goods & Services 268 116 99Media 76 48 23Oil & Gas 61 13 25Personal & Household Goods 43 59 39Retail 50 30 20Technology 94 119 67Travel and Leisure 65 26 11

Overall (Total) 871 564 392

Note: The table reports the distribution of sample firms by country and industry usingDatastream Level-3 industrial classification, for the three countries.

1. We use alternative measures of leverage ratio and find similar results.


Level-3Datastream industry mean leverage ratios.2 Datastream industrymean leverage ratios. Firms with higher than industry mean leverageratios are identified as aggressive and those with lower than industrymean values are named as conservative subsamples. Again, t-test is usedto examine differences in mean leverage ratios and debt and equitylevels between pre-crisis, crisis and post-crisis periods for each of thesetwo subsamples across the three countries.

This paper also examines the impact of the financial crisis on firms’leverage ratios in a more formal setting. Similar to Lemmon, Roberts,and Zender (2008), a fixed-effect panel data regression model is used,including crisis and post-crisis dummies, to capture the impact of thefinancial crisis. The model also controls for other firm-specific factorsthat may have an impact on leverage ratios, as suggested by Frank andGoyal (2003):

0 1 2 3 4it it it it itLEV Tangibility Size MVBV Growth

5 6 7it it itAssetUniqueness BusinessRisk PPETA

(1) 8 9 10it i itROA CD POCD u e

where: asset tangibility (Tangibility) ratio is calculated as fixed assetsdivided by total assets while fixed assets are estimated as the differencebetween total and short-term assets; firm’s size (Size) is defined as thenatural logarithm of firm’s total assets; the market value to book valueratio (MVBV) is employed to capture firm’s growth opportunities;growth (Growth) is a proxy for firm’s growth in total assets andmeasured as the change in total assets; firm’s uniqueness(AssetUniqueness) proxy is defined as research and developmentexpenses divided by total assets; business risk variable (BusinessRisk)is defined as the coefficient of variation in sales over five-years onrolling basis (standard deviation of sales/average of sales); PPETArepresents the ratio of property, plant and equipment divided by totalassets; return on assets (ROA) is estimated as net income beforepreferred dividends divided by total assets; crisis period (CD), the mainvariable of interest, in the regression model is represented by a dummyvariable that takes a value of 1 for years 2008 and 2009 and zero for2006–2007 (pre-crisis period) and 2010–2011 (post-crisis period);post-crisis period dummy (POCD) variable takes the value of 1 for years

2. Please refer to table 1, for Level-3 Datastream industry classifications.


2010 and 2011 and zero for the other years in the sample period.To account and control for time-varying changes in firms’

observable determinants of leverage, crisis and post-crisis dummies areinteracted with control variables in Equation 1, as follows:3

0 1 2 3 4it it it it itLEV Tangibility Size MVBV Growth

5 6 7it it itAssetUniqueness BusinessRisk PPETA

8 9 10 11it itROA CD POCD Tangibility CD

12 13 14it it itSize CD MVBV CD Growth CD

15 16it itAssetUniqueness CD BusinessRisk CD

17 18it itPPETA CD ROA CD

19 20it itTangibility POCD Size POCD

21 22it itMVBV POCD Growth POCD

23 itAssetUniqueness POCD

24 25it itBusinessRisk POCD PPETA POCD

(2) 26 it i itROA POCD u e

IV. Empirical findings

A. Univariate results

Table 2 reports descriptive statistics of all variables across all firm-yearobservations used in this study. Similar to Dang (2013), the figuresshow that the mean leverage ratio (26.40%) in the UK (a market-basedeconomy) is lower than the respective mean ratios (32.40% and33.10%) in France and Germany (bank-based economies). Medianvalues of leverage ratios also depict a similar picture. It is also clear thatthe average level of debt as a percentage of equity is the lowest (about48% of equity) in the UK and highest (about 111% of equity) inGermany with France in the middle (about 83% of equity).

Figure 1 graphically presents annual mean leverage ratios for the fullsample of firms for each country from 2006 to 2011. It shows that the

3. We thank the anonymous referee for this suggestion.


TA

BL

E 2

.D

escr

ipti

ve s

tati

stic

s fo

r al

l var

iabl

es b

ased

on

firm

yea

r ob

serv

atio

ns f

or U

K, F

ranc

e an

d G

erm

any

Cou

ntri

esU

KF

ranc

eG

erm

any

Var

iabl

esM

ean

Med

ian

Std

. Dev

.O

bs.

Mea

nM

edia

nS

td. D

ev.

Obs

.M

ean

Med

ian

Std

. Dev

.O

bs.

LE

V r

atio

0.26

40.

239

0.21

145

970.

324

0.31

20.

220

3210

0.33

10.

310

0.23

123

36L

T d

ebt

2536

1930

8612

1025

545

9860

1155

8254

2140

425

3211

6525

7312

110

3351

129

2287

SH

T d

ebt

7440

519

5155

2815

4584

2530

3170

6215

5084

032

1139

6741

8217

3155

313

2259

Tot

al d

ebt

3273

9383

5815

8068

945

9985

4187

1817

631

9459

432

1110

4445

725

738

6322

098

2287

Equ

ity

6803

2334

786

4277

033

4596

1021

081

4804

440

2098

732

1294

4860

6312

037

9862

622

88T

angi

bili

ty0.

553

0.57

10.

239

4492

0.42

10.

405

0.21

531

970.

469

0.45

20.

210

2288

Siz

e11

.512

11.2

202.

210

4492

12.0

4211

.678

2.37

832

1012

.196

11.9

822.

187

2288

MV

BV

2.22

81.

520

3.17

645

022.

352

1.50

05.

532

3017

2.13

61.

510

2.58

021

36G

row

th0.

567

0.05

715

.363

4664

0.17

00.

048

1.37

831

4612

.465

0.04

957

7.58

722

90A

sset

0.01

70.

000

0.06

244

920.

058

0.02

50.

097

1124

0.02

00.

000

0.04

422

86B

usin

ess

0.37

50.

247

0.37

349

560.

255

0.17

20.

262

2988

0.49

20.

187

1.11

723

52P

PE

TA

0.24

10.

151

0.24

444

920.

008

0.05

60.

056

3060

0.22

80.

189

0.19

122

86R

OA

–0.0

220.

036

0.58

644

920.

040

0.05

20.

116

3115

0.03

40.

040

0.15

422

86

Not

e: T

his

tabl

e re

port

s so

me

desc

ript

ive

stat

isti

cs (

mea

n, m

edia

n, s

tand

ard

devi

atio

n, a

nd f

irm

yea

r ob

serv

atio

ns)

of a

ll th

e va

riab

les

used

in th

is s

tudy

for

the

UK

, Fra

nce

and

Ger

man

y fr

om 2

006

to 2

011.

The

var

iabl

es a

re a

s fo

llow

s: L

EV

rat

io is

mea

sure

d as

tota

l deb

t to

tota

l cap

ital

rati

o, w

here

tota

l deb

t is

mea

sure

d as

the

sum

of

tota

l (lo

ng-

– L

T d

ebt,

and

shor

t-te

rm –

SH

T d

ebt i

n £0

00)

loan

s an

d pr

efer

ence

cap

ital

, whe

reas

tota

l cap

ital

incl

udes

bot

h to

tal d

ebt a

nd to

tal s

hare

hold

er’s

equ

ity

(Equ

ity

in £

000)

; ass

et ta

ngib

ilit

y (T

angi

bili

ty) r

atio

is c

alcu

late

d as

fixe

d as

sets

divi

ded

by to

tal a

sset

s w

here

fixe

d as

sets

are

est

imat

ed a

s th

e di

ffer

ence

bet

wee

n to

tal a

sset

s an

d sh

ort-

term

ass

ets;

fir

m’s

siz

e (S

ize)

is d

efin

ed a

sth

e na

tura

l log

arit

hm o

f fir

m’s

tota

l ass

ets;

the

mar

ket v

alue

to b

ook

valu

e ra

tio

(MV

BV

) is

empl

oyed

to c

aptu

re fi

rm’s

gro

wth

opp

ortu

niti

es; g

row

th(G

row

th)

is a

pro

xy f

or f

irm

’s g

row

th i

n to

tal

asse

ts a

nd m

easu

red

as t

he c

hang

e in

tot

al a

sset

s; f

irm

’s u

niqu

enes

s (A

sset

uni

quen

ess)

pro

xy i

sde

fine

d as

rese

arch

and

dev

elop

men

t exp

ense

s di

vide

d by

tota

l ass

ets;

bus

ines

s ri

sk v

aria

ble

(Bus

ines

s R

isk)

is d

efin

ed th

e co

effi

cien

t of v

aria

tion

in s

ales

ove

r a

five

-yea

r ro

llin

g ba

sis

(sta

ndar

d de

viat

ion

of s

ales

div

ided

by

aver

age

of s

ales

); P

PE

TA

rep

rese

nts

the

rati

o of

pro

pert

y, p

lant

and

equi

pmen

t div

ided

by

tota

l ass

ets;

ret

urn

on a

sset

s (R

OA

) is

est

imat

ed a

s ne

t inc

ome

befo

re p

refe

rred

div

iden

ds d

ivid

ed b

y to

tal a

sset

s.


FIGURE 1.— Graph of annual mean leverage ratios for the fullsample for UK, France and GermanyNote: This graph shows annual mean leverage ratios for the UK, French and German samplefirms. Leverage ratio is determined as total debt to total capital ratio, where total debt ismeasured as the sum of total (long- and short-term debt) loans and preference capital, whiletotal capital includes both total debt and total shareholder’s equity of sample firms over theperiod 2006–2011.

mean leverage ratio is higher for German firms than for French and UKfirms across all years except in 2006. It also shows that leverage ratiosincrease for German and UK firms from pre-crisis (2006 and 2007) tocrisis years (2008 and 2009) and decline in the post-crisis years (2010and 2011). The trend in leverage ratios for both UK and German firmslooks similar during the sample period despite differences in theirinstitutional settings. On the other hand, mean leverage ratio for Frenchfirms does not show such a pattern. It decreases steadily from 2008onwards. Overall, by 2011, the mean leverage ratios for UK andGerman firms are back to their 2006 levels, whereas for French firmsthe mean leverage ratio in 2011 is slightly lower than the level in 2006.

Table 3 reports mean and t-test results for leverage ratios and itscomponents (short-term debt, long-term debt, and equity) across threeperiods: pre-crisis (2006–2007), crisis (2008–2009) and post-crisis(2010–2011) for the full sample (Panel A), conservative sub-sample(Panel B) and aggressive sub-sample (Panel C). For UK, Panel A figuresindicate that mean leverage ratio of 27.70% for the crisis period is

0.2

0.22

0.24

0.26

0.28

0.3

0.32

0.34

0.36

0.38

0.4

2006 2007 2008 2009 2010 2011

Germany

France

UK


higher than the level in pre-crisis (25.80%) and post-crisis (25.70%)periods. This trend shows an increase in leverage ratios during the crisisperiod and a decrease after the crisis with both of these changes beingstatistically significant at the 1% level. Additionally, t-test results showthat there is no significant difference in leverage ratios between the pre-and post-crisis periods.

The trend in leverage ratios for German firms is also similar to thatof UK with an initial increase during the crisis period and then areversal after the crisis years. These results are in line with thosereported by Leary and Roberts (2005) for the US market. Finally,leverage ratios in French firms show a steady downward trend in crisisand post-crisis years with none of the changes being statisticallysignificant. Overall, univariate results indicate that leverage ratios ofsample firms in all three countries change during the crisis and/or in thepost-crisis periods.

Panel A also shows whether it is equity or debt that causes thechange in capital structures of sample firms during and/or after thecrisis. The changes in equity are significant in UK firms during thecrisis years however this is not the case for French or German firms.4

The t-test results show a significant increase in equity levels for samplefirms in UK and France in the post-crisis years when compared to theirpre-crisis or the crisis years’ levels. For German firms, the change inequity is only significant from pre- to post-crisis period. Panel A showsthat mean levels for total debt also increase significantly from pre-crisisto both crisis and post-crisis periods for UK firms, but this increase issignificant only from pre-crisis to the post-crisis period for Frenchfirms. However, German firms do not experience any significant changein their total debt levels either in the crisis or post-crisis years. Thisshows that in Germany (a bank-based economy), debt levels did notchange however equity levels increase overall from pre- to post-crisisyears.

Akbar, Rehman, and Ormrod (2013) find that during the financialcrisis of 2007–08, changes in short-term debt led to changes in capitalstructures of UK private firms. This paper also examines if it is thelong-term or the short-term debt that leads to a change in the capitalstructure in the short-run. The results are mixed. For example, Panel A

4. This is consistent with the statistics reported by London Stock Exchange for furtherequity issues in the UK.


TA

BL

E 3

.M

ean

and

t-te

st f

or d

iffe

renc

e in

mea

n va

lues

bet

wee

n pr

e-cr

isis

, cri

sis,

and

pos

t-cr

isis

per

iods

A. M

ean

and

t-st

ats

for

the

full

sam

ple

of f

irm

s fo

r U

K, F

ranc

e an

d G

erm

any

UK

Fra

nce

Pre-

Cri

sis

Cri

sis-

Post

Pre

-Pos

tP

re-C

risi

sC

risi

s-Po

stPr

e-Po

st

Lev

. Rat

io

Mea

n0.

258

0.27

70.

277

0.25

70.

258

0.25

70.

322

0.33

10.

331

0.31

70.

322

0.31

7t-

stat

–2.5

27**

*2.

60**

*0.

205

–0.9

361.

501

0.54

8

Ln

(Equ

ity)

Mea

n10

.371

10.7

4110

.741

10.9

7110

.371

10.9

7110

.985

10.9

8110

.981

11.2

6710

.985

11.2

67t-

stat

–5.1

06**

*–2

.938

***

–7.6

90**

*0.

037

–2.8

37**

*–2

.738

***

Ln

(Tot

. Deb

t)

Mea

n9.

050

9.41

79.

417

9.62

09.

050

9.62

09.

893

9.97

29.

972

10.1

239.

893

10.1

23t-

stat

–3.4

99**

*–1

.794

**–5

.080

***

–0.6

12–1

.153

–1.7

15**

Ln

(LT

Deb

t)

Mea

n8.

667

9.05

39.

053

9.36

98.

667

9.36

99.

425

9.52

49.

524

9.62

99.

425

9.62

9t-

stat

–3.1

92**

*–2

.394

***

–5.4

32**

*–0

.728

–0.7

60–1

.437

Ln

(SH

T D

ebt)

Mea

n7.

681

7.99

17.

991

8.18

97.

681

8.18

98.

877

8.93

38.

933

9.10

58.

877

9.10

5t-

stat

–3.0

87**

*–1

.832

**–4

.807

***

–0.4

37–1

.367

*–1

.759

*

( C

onti

nued

)


TA

BL

E 3

.(C

onti

nued

)

A. M

ean

and

t-st

ats

for

the

full

sam

ple

of f

irm

s fo

r U

K, F

ranc

e an

d G

erm

any G

erm

any

Pre-

Cri

sis

Cri

sis-

Post

Pre

-Pos

t

Lev

. Rat

io

Mea

n0.

316

0.35

80.

358

0.31

90.

316

0.31

9t-

stat

–3.5

7***

3.34

5***

–0.2

58

Ln

(Equ

ity)

Mea

n11

.119

11.2

2511

.225

11.3

8711

.119

11.3

87t-

stat

–0.9

80–1

.510

–2.4

38**

Ln

(Tot

. Deb

t)

Mea

n 10

.219

10.3

9310

.393

10.4

2510

.219

10.4

25t-

stat

–1

.285

–0.2

37–1

.467

Ln

(LT

Deb

t)

Mea

n 9.

769

9.90

59.

905

10.0

559.

769

10.0

55t-

stat

–0

.927

–1.0

2–1

.918

*

Ln

(SH

T D

ebt)

Mea

n 9.

004

9.23

09.

230

9.18

69.

004

9.18

6t-

stat

–1

.630

0.32

3–1

.284

( C

onti

nued

)


TA

BL

E 3

.(C

onti

nued

)

B. M

ean

and

t-st

ats

for

the

cons

erva

tive

sub

-sam

ple

of f

irm

s fo

r U

K, F

ranc

e an

d G

erm

any

UK

Fra

nce

Pre-

Cri

sis

Cri

sis-

Post

Pre

-Pos

tP

re-

Cri

sis

Cri

sis-

Post

Pre

-Pos

t

Lev

. Rat

io

Mea

n0.

113

0.16

90.

169

0.17

50.

113

0.17

50.

168

0.20

30.

203

0.20

90.

168

0.20

9t-

stat

–8.6

94**

*–0

.774

–8.4

65**

*–3

.794

***

–0.6

16–4

.163

***

Ln

(Equ

ity)

Mea

n10

.042

10.4

0010

.400

10.5

7610

.042

10.5

7610

.689

10.6

7310

.673

10.9

2710

.689

10.9

27t-

stat

–3.8

95**

*–1

.771

**–5

.357

***

0.13

4–1

.993

**–1

.822

**

Ln

(Tot

. Deb

t)

Mea

n7.

657

8.31

18.

311

8.58

67.

657

8.58

68.

606

8.86

78.

867

9.07

28.

606

9.07

2t-

stat

–4.7

79**

*–1

.855

**–6

.260

***

–1.5

74*

–1.2

35–2

.714

***

Ln

(LT

Deb

t)

Mea

n7.

166

7.82

57.

825

8.26

87.

166

8.26

88.

134

8.50

78.

507

8.55

58.

134

8.55

5t-

stat

–4.1

08**

*–2

.541

***

–6.3

82**

*–2

.192

**–0

.279

–2.3

61**

*

Ln

(SH

T D

ebt)

Mea

n6.

576

7.07

47.

074

7.35

76.

576

7.35

77.

730

7.90

27.

902

8.21

77.

730

8.21

7t-

stat

–3.6

39**

*–1

.910

**–5

.362

***

–1.0

03–1

.887

–2.7

99**

*

( C

onti

nued

)


TA

BL

E 3

.(C

onti

nued

)

B. M

ean

and

t-st

ats

for

the

cons

erva

tive

sub

-sam

ple

of f

irm

s fo

r U

K, F

ranc

e an

d G

erm

any

Ger

man

y

Pre-

Cri

sis

Cri

sis-

Post

Pre

-Pos

t

Lev

. Rat

io

Mea

n0.

151

0.23

50.

235

0.21

90.

151

0.21

9t-

stat

–7.6

59**

*1.

216

–5.9

97**

*

Ln

(Equ

ity)

Mea

n 10

.921

10.9

7710

.977

11.1

6410

.921

11.1

64t-

stat

–0.3

78–1

.279

–1.6

30

Ln

(Tot

. Deb

t)

Mea

n9.

074

9.40

19.

401

9.52

99.

074

9.52

9t-

stat

–1.8

202

–0.7

09–2

.413

**

Ln

(LT

Deb

t)

Mea

n8.

588

8.85

58.

855

9.15

68.

588

9.15

6t-

stat

–1.3

57–1

.478

–2.7

89**

*

Ln

(SH

T D

ebt)

Mea

n8.

073

8.38

58.

385

8.37

58.

073

8.37

5t-

stat

–1.6

83*

0.05

4–1

.577

( C

onti

nued

)


TA

BL

E 3

.(C

onti

nued

)

C. M

ean

and

t-st

ats

for

the

aggr

essi

ve s

ub-s

ampl

e of

fir

ms

for

UK

, Fra

nce

and

Ger

man

y

UK

Fra

nce

Pre-

Cri

sis

Cri

sis-

Post

Pre

-Pos

tP

re-C

risi

sC

risi

s-Po

stPr

e-Po

st

Lev

. Rat

io

Mea

n0.

430

0.41

10.

411

0.35

70.

430

0.35

70.

4800

0.46

590.

4659

0.43

040.

4800

0.43

04t-

stat

2.03

4**

4.78

4***

7.01

8***

1.26

62.

953*

**4.

241*

**

Ln

(Equ

ity)

Mea

n10

.761

11.1

6211

.162

11.4

5910

.761

11.4

5911

.287

11.3

0611

.306

11.6

2411

.287

11.6

24t-

stat

–3.6

00**

*–2

.474

***

–5.8

80**

*–0

.124

–2.0

67**

–2.1

41**

Ln

(Tot

. Deb

t)

Mea

n10

.476

10.6

8910

.689

10.7

8010

.476

10.7

8011

.195

11.1

2111

.121

11.2

1111

.195

11.2

11t-

stat

–1.6

75**

–0.6

17–2

.193

**0.

439

–0.5

06–0

.094

Ln

(LT

Deb

t)

Mea

n10

.022

10.3

6710

.367

10.5

0210

.022

10.5

0210

.665

10.5

3610

.536

10.6

9710

.665

10.6

97t-

stat

–2.3

18**

*–0

.783

–2.9

42**

*0.

685

–0.8

23–0

.166

Ln

(SH

T D

ebt)

Mea

n8.

790

9.00

59.

005

9.10

98.

790

9.10

910

.002

9.96

79.

967

9.99

510

.002

9.99

5t-

stat

–1.7

14–0

.754

–2.4

21**

*0.

216

–0.1

640.

040

( C

onti

nued

)


TA

BL

E 3

.(C

onti

nued

)

C. M

ean

and

t-st

ats

for

the

aggr

essi

ve s

ub-s

ampl

e of

fir

ms

for

UK

, Fra

nce

and

Ger

man

y

Ger

man

y

Pre-

Cri

sis

Cri

sis-

Post

Pre

-Pos

t

Lev

. Rat

io

Mea

n0.

494

0.49

00.

490

0.42

60.

494

0.42

6t-

stat

0.27

54.

310*

**4.

499*

**

Ln

(Equ

ity)

Mea

n 11

.324

11.4

9111

.491

11.6

2011

.324

11.6

20t-

stat

–1.0

62–0

.830

–1.8

38*

Ln

(Tot

. Deb

t)

Mea

n11

.338

11.4

5311

.453

11.3

4211

.338

11.3

42t-

stat

–0.6

750

0.63

6–0

.025

Ln

(LT

Deb

t)

Mea

n10

.841

10.9

2810

.928

10.9

2810

.841

10.9

28t-

stat

–0.4

740.

000

–0.4

61

Ln

(SH

T D

ebt)

Mea

n9.

887

10.1

2810

.128

10.0

049.

887

10.0

04t-

stat

–1.3

060.

693

–0.6

17

( C

onti

nued

)


TA

BL

E 3

.(C

onti

nued

)

Not

e: T

he ta

ble

repo

rts

mea

ns a

nd t-

test

res

ults

for

dif

fere

nce

in m

eans

for

dif

fere

nt v

aria

bles

for

the

full

(P

anel

A),

con

serv

ativ

e (P

anel

B)

and

aggr

essi

ve (P

anel

C) s

ampl

es o

f UK

, Fre

nch,

and

Ger

man

fir

ms

acro

ss th

ree

peri

ods.

Yea

rs 2

006

and

2007

are

def

ined

as

pre-

cris

is, 2

008

and

2009

as

cris

is, a

nd 2

010

and

2011

as

post

-cri

sis

year

s. F

irm

s wit

h le

ss (h

ighe

r) th

an in

dust

ry m

ean

of le

vera

ge ra

tio

are

clas

sifi

ed a

s th

e co

nser

vati

ve(a

ggre

ssiv

e) s

ubsa

mpl

e (p

leas

e re

fer

to t

able

1 f

or in

dust

ry g

roup

s).

For

eac

h co

untr

y, f

irst

col

umn

repo

rts

the

resu

lts

for

the

diff

eren

ce b

etw

een

pre-

and

cri

sis

year

s; s

econ

d co

lum

n fo

r cri

sis

and

post

-cri

sis

year

s; a

nd th

ird

colu

mn

for t

he p

re- a

nd p

ost-

cris

is y

ears

. The

var

iabl

es re

port

ed h

ere

are:

leve

rage

(Lev

.) ra

tio

mea

sure

d as

tota

l deb

t to

tota

l cap

ital

rati

o, w

here

tota

l deb

t is

the

sum

of t

otal

(lon

g- a

nd s

hort

-ter

m) l

oans

and

pre

fere

nce

capi

tal,

and

tota

l ca

pita

l in

clud

es b

oth

tota

l de

bt a

nd t

otal

sha

reho

lder

’s e

quit

y; n

atur

al l

og o

f eq

uity

; na

tura

l lo

g of

tot

al d

ebt;

nat

ural

log

of

long

-ter

m d

ebt;

and

nat

ural

log

of

shor

t-te

rm d

ebt,

resp

ecti

vely

. L

n re

fers

to

the

natu

ral

loga

rith

mic

val

ues.

Fin

ally

, **

*, *

*, a

nd *

rep

rese

ntst

atis

tica

l sig

nifi

canc

e at

1%

, 5%

, and

10%

leve

ls, r

espe

ctiv

ely.


shows that changes in both short-term and long-term debts aresignificant for UK firms during the crisis and post-crisis years butinsignificant for French and German firms.These results suggest thatchanges in both short- and long-term debts are mainly responsible forchanges in leverage ratios from pre- to crisis and post-crisis periodsin the UK, however this is not the case for France and Germany. Thisseems plausible in a market-based economy (such as UK) as comparedto a bank-based economy (such as France or Germany).

Panel B of table 3 shows that leverage ratios increase from pre- tocrisis and pre- to post-crisis periods for the conservative subsamplesacross all three major European countries. For UK, this increase is dueto increases in both equity and debt (short- and long-term) from pre- tocrisis and to post-crisis periods. For French subsample, this increase isdue to an increase in both short- and long-term debt and equity frompre- to post-crisis years, whereas for German subsample, it is due tochanges in long-term debt from pre- to post-crisis periods.

Panel C of table 3 shows that leverage ratios decrease steadily forUK, French and German aggressive subsamples during both crisis andpost-crisis periods. For UK subsample, a statistically significantdecrease from pre- to crisis period is observable but this is not the casefor French and German subsamples. However, the decrease becomesstatistically significant from pre- to post-crisis periods for all threesubsamples. This decrease is mainly due to an increase in equity, eitherin the crisis or post-crisis periods. For the UK, this study also finds asignificant increase in debt during the crisis period but the increase indebt is smaller than the increase in equity.

Figure 2 plots the graphs of the yearly averages (2006–2011) ofleverage ratios for conservative and aggressive subsamples in the threecountries. It shows a steady increase in leverage ratios for conservativesubsamples and a steady decrease for aggressive subsamples across thethree countries from pre-crisis to post-crisis years. It is also interestingthat the movement in leverage ratios for conservative firms in the pre-and crisis years and for aggressive firms in the post-crisis year arealmost similar for firms in the UK (a market-based economy) andGermany (a bank-based economy). The results for the conservative andaggressive subsamples are in line with those reported by Leary andRoberts (2005). They find that leverage ratios are more likely toincrease (decrease) if they are relatively low (high).

B. Regression results

Panel A of table 4 reports regression results for full samples in threecountries (UK, France, and Germany). The main variables of interest are


Conservative subsample:

Aggressive subsample:

FIGURE 2.— Graphs of annual mean leverage ratios forconservative and aggressive subsamples for UK, France and Germany.Note: The graphs show annual mean leverage ratios for UK, French and German firms fortheir conservative and aggressive subsamples, respectively. Firms with less than industrymean leverage ratio are identified as conservative and those with higher than industry meanvalue are named as aggressive subsamples (please refer to table 1 for industry groups).Leverage ratio is calculated as total debt to total capital ratio, where total debt is measuredas the sum of total (long- and short-term debt) loans and preference capital, while total capitalincludes both total debt and total shareholder’s equity of sample firms over the period2006–2011.

0.05

0.1

0.15

0.2

0.25

0.3

2006 2007 2008 2009 2010 2011

Germany

France

UK

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0.55

2006 2007 2008 2009 2010 2011

Germany

France

UK


the crisis (CD) and post-crisis dummies (POCD). Regression results(Equations 1 and 2) for full samples of UK and Germany show that the coefficients for CD for these two countries are positive and statisticallysignificant, but for French sample this coefficient is insignificant. These results suggest that leverage ratios are higher for UK and Germanfirms during the crisis period (2008–2009) than during pre-crisis periodeven after controlling for time-varying effects of determinants of capitalstructure, as shown in Equation 2. While this evidence is similar to thatof Fosberg (2012) for the US market, it is inconsistent with the resultsreported by Akbar, Rehman, and Ormrod (2013) for UK private firms.The coefficients for the post-crisis dummies across the three countriesare statistically insignificant for the full sample based on bothregression models. These findings support the univariate resultsreported earlier and are in line with Fosberg’s (2012) which suggeststhat leverage ratios for the US firms in post-crisis periods revert back totheir pre-crisis levels. Overall, these results are also consistent withthose of Kayhan and Titman (2007) and Leary and Roberts (2005) thatover time leverage ratios move towards target levels.

Regression results (Equations 1 and 2) in Panel B of table 4 forconservative subsamples show that coefficients for both the crisis (CD)and post-crisis (POCD) dummy variables are positive and statisticallysignificant for all three countries. The results highlight the fact thatleverage ratios for conservative firms increase during the crisis andpost-crisis periods across the three countries, despite controlling fortime-varying changes in factors affecting capital structure.

Panel C of table 4 reports regression results (Equations 1 and 2) foraggressive subsamples. The results indicate negative and statisticallysignificant coefficients for post-crisis dummy (POCD) variable acrossthe three countries for these subsamples. These findings are opposite tothose of conservative subsamples, as expected. Further, this study findsa marginally significant and negative coefficient for the crisis dummy(CD) for French subsample indicating that leverage ratio changes foraggressive subsample occurred in the post-crisis period. Overall, thesefindings are consistent with the results reported in table 3 Panels B andC and suggest that the changes in the leverage ratios for conservativeand aggressive subsamples are significant but in opposite directions.

The coefficients for the crisis dummy (CD) in equation 1 for the fullsamples in UK and Germany indicate that the increase in leverage ratios(between 2 and 3 percentage points) though statistically significant iseconomically small.


TA

BL

E 4

.P

anel

dat

a re

gres

sion

res

ults

of

leve

rage

rat

ios

A. R

egre

ssio

n re

sults

for

ful

l sam

ples

for

UK

, Fra

nce,

and

Ger

man

y

LE

V R

atio

(F

ull S

ampl

e)U

KF

ranc

eG

erm

any

Equ

atio

n 1

Equ

atio

n 2

Equ

atio

n 1

Equ

atio

n 2

Equ

atio

n 1

Equ

atio

n 2

Tan

gibi

lity

0.06

6***

0.06

9***

0.12

5*0.

129*

*0.

284*

**0.

282*

**Si

ze0.

010*

0.01

0*0.

059*

**0.

056*

**0.

082*

**0.

082*

**M

VB

V0.

003*

**0.

003*

**0.

005*

**0.

005*

**0.

018*

**0.

018*

**G

row

th0.

0002

0.00

010.

051*

**0.

050*

**–0

.000

1–0

.000

1A

sset

Uni

quen

ess

–0.0

76–0

.077

–0.0

52–0

.058

0.24

10.

240*

Bus

ines

sRis

k–0

.030

***

–0.0

31**

*–0

.039

–0.0

38–0

.003

–0.0

03P

PE

TA

0.03

90.

044.

025

4.31

40.

045

0.03

8R

OA

0.00

20.

002

–0.5

24**

*–0

.522

***

0.03

30.

033

Cri

sis

Dum

my

(CD

)0.

024*

**0.

046*

**0.

006

0.00

60.

029*

**0.

030*

**P

ost-

Cri

sis

Dum

my

(PO

CD

)0.

003

0.00

8–0

.004

–0.0

03–0

.003

–0.0

03T

angi

bili

ty*(

CD

)–

–0.0

05–

0.00

8–

0.00

6Si

ze*(

CD

)–

0.00

2–

0.00

03–

0.00

04M

VB

V*(

CD

)–

–0.0

04**

*–

0.00

02–

0G

row

th*(

CD

)–

–0.0

001

–0.

009

–0.

002

Ass

etU

niqu

enes

s*(C

D)

–0.

089

––0

.056

––0

.188

**B

usin

essR

isk*

(CD

)–

0.03

6**

––0

.010

–0.

001

PP

ET

A*(

CD

)–

0.01

3–

–0.0

23–

0.00

4R

OA

*(C

D)

–0.

016

––0

.037

––0

.011

Tan

gibi

lity

*(P

OC

D)

–0.

029

––0

.025

––0

.008

Size

*(P

OC

D)

––0

.002

–0.

001

––0

.001

( C

onti

nued

)


TA

BL

E 4

.(C

onti

ued)

A. R

egre

ssio

n re

sults

for

ful

l sam

ples

for

UK

, Fra

nce,

and

Ger

man

y

LE

V R

atio

(F

ull S

ampl

e)U

KF

ranc

eG

erm

any

Equ

atio

n 1

Equ

atio

n 2

Equ

atio

n 1

Equ

atio

n 2

Equ

atio

n 1

Equ

atio

n 2

MV

BV

*(P

OC

D)

––0

.002

*–

–0.0

003

–0.

001

Gro

wth

*(P

OC

D)

–0.

001

––0

.002

–0.

001

Ass

etU

niqu

enes

s*(P

OC

D)

––0

.077

–0.

247

––0

.096

Bus

ines

sRis

k*(P

OC

D)

–0.

049*

**–

0.02

7–

0.02

0**

PP

ET

A*(

PO

CD

)–

–0.0

05–

–0.0

17–

0.04

4**

RO

A*(

PO

CD

)–

–0.0

03–

–0.0

02–

–0.0

33C

onst

ant

0.10

20.

104

–0.4

92**

–0.4

61*

–0.8

71**

*–0

.870

***

No

of o

bser

vati

ons

4180

4180

860

860

2026

2026

No

of g

roup

s85

385

319

919

937

837

8R

-squ

are

over

all

0.70

30.

700

0.82

00.

818

0.79

70.

798

F-t

est

7.07

***

3.94

***

27.4

8***

10.6

9***

14.5

6***

17.2

9***

( C

onti

nued

)


TA

BL

E 4

.(C

onti

nued

)

B. R

egre

ssio

n re

sult

s fo

r co

nser

vati

ve s

ubsa

mpl

es f

or U

K, F

ranc

e, a

nd G

erm

any

LE

V R

atio

(Con

serv

ativ

e su

bsam

ple

)U

KF

ranc

eG

erm

any

Equ

atio

n 1

Equ

atio

n 2

Equ

atio

n 1

Equ

atio

n 2

Equ

atio

n 1

Equ

atio

n 2

Tan

gibi

lity

0.08

4***

0.08

4**

0.29

8***

0.30

1***

0.24

9***

0.25

0***

Size

0.00

60.

050.

050*

*0.

046*

*0.

078*

**0.

077*

**M

VB

V0.

006*

**0.

005*

**0.

024*

**0.

024*

**0.

019*

**0.

019*

**G

row

th0

–0.0

001

0.06

9***

0.07

0***

–0.0

03–0

.004

Ass

etU

niqu

enes

s–0

.187

***

–0.1

88**

*–0

.121

–0.1

150.

253

0.26

0B

usin

essR

isk

–0.0

09–0

.009

–0.0

51**

–0.0

53**

–0.0

07–0

.006

PP

ET

A0.

066*

0.06

6*–0

.258

–0.4

230.

101*

0.10

3**

RO

A–0

.019

–0.0

18–0

.398

***

–0.3

84**

*–0

.001

0.00

9C

risi

s D

umm

y (C

D)

0.05

4***

0.06

3***

0.04

4***

0.04

3***

0.06

0***

0.06

1***

Pos

t-C

risi

s D

umm

y (P

OC

D)

0.05

6***

0.05

8***

0.04

0***

0.03

9***

0.04

2***

0.04

4***

Tan

gibi

lity

*(C

D)

–0.

022

–0.

031

–0.

012

Size

*(C

D)

––0

.002

–0.

000

–0.

001

MV

BV

*(C

D)

–0.

002

–0.

001

––0

.002

Gro

wth

*(C

D)

–0.

002

–0.

005

–0.

003

Ass

etU

niqu

enes

s*(C

D)

–0.

027

–0.

091

––0

.223

*B

usin

essR

isk*

(CD

)–

0.00

6–

–0.0

24–

0.00

9P

PE

TA

*(C

D)

––0

.004

–2.

271

––0

.007

RO

A*(

CD

)–

–0.0

16–

–0.0

04–

–0.1

56**

( C

onti

nued

)


TA

BL

E 4

.(C

onti

nued

)

B. R

egre

ssio

n re

sult

s fo

r co

nser

vati

ve s

ubsa

mpl

es f

or U

K, F

ranc

e, a

nd G

erm

any

LE

V R

atio

(Con

serv

ativ

e su

bsam

ple

)U

KF

ranc

eG

erm

any

Equ

atio

n 1

Equ

atio

n 2

Equ

atio

n 1

Equ

atio

n 2

Equ

atio

n 1

Equ

atio

n 2

Tan

gibi

lity

*(P

OC

D)

–0.

058*

*–

–0.0

67–

0.01

0Si

ze*(

PO

CD

)–

–0.0

04**

–0.

003

–0.

000

MV

BV

*(P

OC

D)

–0.

001

–0.

000

–0.

001

Gro

wth

*(P

OC

D)

–0.

001

–0.

050

––0

.003

Ass

etU

niqu

enes

s*(P

OC

D)

––0

.178

––0

.014

––0

.020

Bus

ines

sRis

k*(P

OC

D)

–0.

047*

*–

–0.0

04–

0.00

4P

PE

TA

*(P

OC

D)

–0.

006

–1.

727

––0

.004

RO

A*(

PO

CD

)–

0.00

5–

–0.0

60–

–0.0

63**

*C

onst

ant

–0.0

090.

001

–0.5

65**

–0.5

30**

–0.9

27**

*–0

.927

***

No

of o

bser

vati

ons

2297

2297

492

492

1033

1033

No

of g

roup

s46

246

211

611

619

419

4R

–squ

are

over

all

0.53

80.

537

0.79

80.

793

0.66

20.

661

F-t

est

18.9

7***

7.83

***

21.7

2***

8.46

***

34.1

8***

13.6

1***

( C

onti

nued

)


TA

BL

E 4

.(C

onti

nued

)

C. R

egre

ssio

n re

sult

s fo

r ag

gres

sive

sub

sam

ples

for

UK

, Fra

nce,

and

Ger

man

y

LE

V R

atio

(Agg

ress

ive

sub

sam

ple)

UK

Fra

nce

Ger

man

y

Equ

atio

n 1

Equ

atio

n 2

Equ

atio

n 1

Equ

atio

n 2

Equ

atio

n 1

Equ

atio

n 2

Tan

gibi

lity

–0.0

07–0

.001

–0.0

60–0

.031

0.28

3***

0.27

6***

Size

0.00

30.

002

0.06

3*0.

054

0.06

7***

0.06

6***

MV

BV

0.00

10.

001

0.00

4***

0.00

4***

0.01

6***

0.01

5***

Gro

wth

0.00

00.

000

0.03

20.

032

0.00

00.

000

Ass

etU

niqu

enes

s0.

566*

*0.

572*

*0.

353

0.28

60.

238

0.24

3B

usin

essR

isk

–0.0

58**

*–0

.063

***

0.03

90.

041

–0.0

03–0

.003

PP

ET

A–0

.082

–0.0

8617

.517

**18

.473

***

0.00

5–0

.007

RO

A0.

008*

0.00

8*–0

.653

***

–0.6

60**

*0.

073*

**0.

066*

**C

risi

s D

umm

y (C

D)

–0.0

110.

029

–0.0

23*

–0.0

23*

–0.0

04–0

.003

Pos

t-C

risi

s D

umm

y (P

OC

D)

–0.0

62**

*–0

.058

***

–0.0

45**

*–0

.044

***

–0.0

48**

*–0

.048

***

Tan

gibi

lity

*(C

D)

––0

.026

––0

.038

–0.

008

Size

*(C

D)

––0

.003

–0.

003

––0

.001

MV

BV

*(C

D)

––0

.006

***

–0.

000

–0.

001

Gro

wth

*(C

D)

–0.

000

–0.

014

–0.

004

Ass

etU

niqu

enes

s*(C

D)

–0.

150

––0

.093

––0

.127

Bus

ines

sRis

k*(C

D)

–0.

066*

*–

–0.0

32–

0.00

0P

PE

TA

*(C

D)

–0.

002

––0

.043

–0.

030

RO

A*(

CD

)–

0.04

4–

–0.0

74–

0.04

1

( C

onti

nued

)


TA

BL

E 4

.(C

onti

nued

)

C. R

egre

ssio

n re

sult

s fo

r ag

gres

sive

sub

sam

ples

for

UK

, Fra

nce,

and

Ger

man

y

L

EV

Rat

io

(Agg

ress

ive

sub

sam

ple)

UK

Fra

nce

Ger

man

y

Equ

atio

n 1

Equ

atio

n 2

Equ

atio

n 1

Equ

atio

n 2

Equ

atio

n 1

Equ

atio

n 2

Tan

gibi

lity

*(P

OC

D)

–0.

022

–0.

020

––0

.016

Size

*(P

OC

D)

–0.

000

––0

.001

––0

.001

MV

BV

*(P

OC

D)

––0

.002

––0

.005

–0.

000

Gro

wth

*(P

OC

D)

–0.

007

––0

.002

–0.

004

Ass

etU

niqu

enes

s*(P

OC

D)

–0.

094

–0.

483*

––0

.246

*B

usin

essR

isk*

(PO

CD

)–

0.04

2–

0.06

1–

0.02

2***

PP

ET

A*(

PO

CD

)–

–0.0

39**

–0.

024

–0.

075*

**R

OA

*(P

OC

D)

––0

.022

–0.

014

––0

.021

Con

stan

t0.

423*

**0.

448*

**–0

.415

–0.3

17–0

.553

***

–0.5

34**

*

No

of o

bser

vati

ons

1883

1883

368

368

993

993

No

of g

roup

s39

239

283

8318

418

4R

-squ

are

over

all

0.58

20.

586

0.67

90.

670

0.73

10.

734

F-t

est

8.89

***

4.57

***

15.0

1***

5.96

***

19.2

7***

8.32

***

( C

onti

nued

)


TA

BL

E 4

.(C

onti

nued

)

Not

e:

Thi

s ta

ble

repo

rts

fixe

d-ef

fect

pan

el d

ata

regr

essi

on r

esul

ts u

sing

Equ

atio

ns (

1) a

nd (

2) f

or t

he f

ull

sam

ples

(P

anel

A),

con

serv

ativ

esu

bsam

ples

(P

anel

B)

and

aggr

essi

ve s

ubsa

mpl

es (

Pan

el C

) fo

r U

K, F

ranc

e an

d G

erm

any.

The

var

iabl

es a

re a

s fo

llow

s: L

EV

rat

io (

the

depe

nden

tva

riab

le) i

s m

easu

red

as to

tal d

ebt t

o to

tal c

apit

al ra

tio,

whe

re to

tal d

ebt i

s m

easu

red

as th

e su

m o

f tot

al (l

ong-

– L

T d

ebt,

and

shor

t-te

rm –

SH

T d

ebt)

loan

s an

d pr

efer

ence

cap

ital

, and

tota

l cap

ital

incl

udes

bot

h to

tal d

ebt a

nd to

tal s

hare

hold

er’s

equ

ity

(Equ

ity)

; ass

et ta

ngib

ilit

y (T

angi

bili

ty)

rati

ois

cal

cula

ted

as f

ixed

ass

ets

divi

ded

by to

tal

asse

ts w

here

fix

ed a

sset

s ar

e es

tim

ated

as

the

diff

eren

ce b

etw

een

tota

l ass

ets

and

shor

t-te

rm a

sset

s;fi

rm’s

siz

e (S

ize)

is d

efin

ed a

s th

e na

tura

l log

arit

hm o

f fir

m’s

tota

l ass

ets;

the

mar

ket v

alue

to b

ook

valu

e ra

tio

(MV

BV

) is

empl

oyed

to c

aptu

re fi

rm’s

grow

th o

ppor

tuni

ties

; gro

wth

(G

row

th)

is a

pro

xy f

or f

irm

’s g

row

th in

tota

l ass

ets

and

mea

sure

d as

the

chan

ge in

tota

l ass

ets;

fir

m’s

uni

quen

ess

(Ass

et u

niqu

enes

s) p

roxy

is d

efin

ed a

s re

sear

ch a

nd d

evel

opm

ent e

xpen

ses

divi

ded

by to

tal a

sset

s; b

usin

ess

risk

var

iabl

e (B

usin

ess

Ris

k) is

def

ined

the

coef

fici

ent o

f va

riat

ion

in s

ales

ove

r a

five

-yea

r ro

llin

g ba

sis

(sta

ndar

d de

viat

ion

of s

ales

div

ided

by

aver

age

of s

ales

); P

PE

TA

rep

rese

nts

the

rati

o of

pro

pert

y, p

lant

and

equ

ipm

ent d

ivid

ed b

y to

tal a

sset

s; re

turn

on

asse

ts (R

OA

) is

esti

mat

ed a

s ne

t inc

ome

befo

re p

refe

rred

div

iden

ds d

ivid

edby

tota

l ass

ets.

Cri

sis

dum

my

(CD

) rep

rese

nts

cris

is p

erio

d th

at ta

kes

a va

lue

of 1

for y

ears

200

8 an

d 20

09 a

nd z

ero

for 2

006–

2007

(pre

-cri

sis)

and

2010

–201

1 (p

ost-

cris

is)

and

post

-cri

sis

dum

my

(PO

CD

) ta

kes

valu

es o

f 1

for

year

s 20

10 a

nd 2

011

and

0 fo

r th

e ot

her

year

s in

the

sam

ple

peri

od.

In E

quat

ion

(2),

eac

h of

the

inde

pend

ent v

aria

ble

is in

tera

cted

wit

h bo

th C

D a

nd P

OC

D. F

inal

ly, *

**, *

*, a

nd *

rep

rese

nt s

tati

stic

al s

igni

fica

nce

of th

e co

effi

cien

ts a

t 1%

, 5%

, and

10%

leve

ls, r

espe

ctiv

ely

(see

Equ

atio

ns 1

&2)

.


However, this paper finds that the corresponding increase forconservative subsamples during the crisis period is economically larger(between 4.4 and 6.0 percentage points) across the three countries. Theeconomic impact of the crisis is also captured by the post-crisis dummy(POCD) coefficients across the three countries showing an increase of4.0 and 5.6 percentage points for the conservative subsamples and adecrease of 4.5 and 6.2 percentage points for the aggressive subsamples,as compared to the pre-crisis period. Results from equation 2, across allsubsamples, show that the significant impact of crisis and post-crisisdummies persists even when the interaction dummies with the controlvariables are included.

V. Summary and Conclusion

This study contributes to the extant literature by examining the capitalstructure or leverage ratios of non-financial/non-utility firms in threemain European countries (UK, France and Germany) around 2007–08financial crisis. The sample period is segregated into three periods ofinterest: 2006–2007 as the pre-crisis period, 2008–2009 as the crisisperiod, and 2010–2011 as the post-crisis period. Univariate and paneldata regression results reveal that, on average, leverage ratiossignificantly increase from pre- to crisis periods and then revert topre-crisis levels in the UK and Germany. These changes areinsignificant across the three periods for French firms. This study findsthat during the crisis period, the increase in leverage ratios for UK firmsis mainly due to the use of both short- and long-term debt, whereas inthe post-crisis period the reversion back to the pre-crisis levels is mainlydue to the use of equity. There are no significant changes in either short-or long-term debt of full samples of French and German firms, either inthe crisis or post-crisis years.

Considering the availability of excessive leverage for firms beforethe financial crisis (for example, Fosberg, 2012), the sample firms ofeach country are segregated into two subsamples (conservative andaggressive) based on the levels of their pre-crisis leverage ratios ascompared to their industry average ratios during this period. Firms areconsidered to follow a conservative (aggressive) funding strategy in thepre-crisis period if they had lower (higher) than industry mean leverageratio during that period. The evidence on conservative subsamplesshows a significant increase in leverage ratios from pre- to crisis period.


Furthermore, t-test results reveal that the increase in leverage ratios tocrisis period for UK conservative firms is mainly due to the significantincrease in debt rather than in equity. This study does not find suchstrong evidence on the use of debt or equity for French and Germanconservative firms.

Findings from the aggressive firms indicate that overall the leverageratios in the post-crisis period are significantly lower than theirpre-crisis levels across the three countries. While this decrease is morepronounced from the crisis to post-crisis period for French and Germanaggressive sub-samples, the leverage ratios for UK subsample havesignificantly and gradually fallen from pre- to crisis and then inpost-crisis period. This gradual decline is mainly due to the significantreduction in equity levels across the three periods. Additionally, t-testresults indicate a significant increase in debt (though lower than theincrease in equity) for UK aggressive subsample from pre- to crisisperiod. Finally, this study does not find any significant evidence onincreases in debt or equity levels for either French or German aggressivesub-samples.

Overall, the results show that 2007–08 financial crisis had asignificant impact on leverage ratios of firms in both market-based (theUK) and bank-based (Germany and France) economies. Leverage ratiosof sample firms in the post-crisis period revert back to their pre-crisislevels with adjustments during the crisis years. Consistent with Learyand Roberts (2005), this paper finds that firms with lower thanindustry-average debt ratios experience an increase in debt ratios andthose with higher than industry-average debt ratios experience adecrease in these levels from pre- to the post-crisis periods.

Accepted by: Prof. P. Theodossiou, Editor-in-Chief, July 2014

References

Akbar, S.; Rehman, S.; and Ormrod, P. 2013. The impact of recent financialshocks on the financing and investment policies of UK private firms.International Review of Financial Analysis 26(4): 59–70.

Al-Najjar, B., and Hussainey, K. 2011. Revisiting the capital-structure puzzle:UK evidence. Journal of Risk Finance 12(4): 329–338.

Alter, A., and Schόler, Y. S. 2012. Credit spread interdependencies of Europeanstates and banks during the financial crisis. Journal of Banking & Finance36(12): 3444–3468.


Balsari, C. K., and Kirkulak, B. 2010. Effect of financial crises on the capitalstructure choice: Evidence from Istanbul Stock Exchange (ISE). Workingpaper. Izmir: Dokuz Eylul University.

Bancel, F., and Mittoo, U. 2011. Financial flexibility and the impact of theglobal financial crisis: Evidence from France. International Journal ofManagerial Finance 7(2): 179–216.

Booth, L.; Aivazian, V.; Demirguc-Kunt, A.; and Maksimovic, V. 2001. Capitalstructures in developing countries. The Journal of Finance 56: 87–130.

Brun, M.; Chai, F.; Elgg, D.; Esteban, A.; van Gastel, G.; Körting, T.; andVivet, D. 2013. Profitability, equity capitalization and net worth at risk.Working Paper, European Committee of Central Balance Sheet Data Office.http://www.banque-france.fr/fileadmin/user_upload/banque_de_france/Economie_et_Statistiques/WP_Study_Group_ECCBSO_2013_V1.0.pdf

Brunnermeier, M. K. 2008. Deciphering the liquidity and credit crunch2007–08. NBER working paper no.14612. Cambridge, Mass: NationalBureau of Economic Research.

Campello, M.; Graham, J. R.; and Harvey, C. R. 2010. The real effects offinancial constraints: Evidence from a financial crisis. Journal of FinancialEconomics 97(3): 470–487.

Chen, J. J. 2004. Determinants of capital structure of Chinese-listed companies.Journal of Business Research 57(12): 1341–1351.

Dang, V. A. 2013. Testing capital structure theories using error correctionmodels: evidence from the UK, France and Germany. Applied Economics45(2): 171–190.

Demirgüç-Kunt, A., and Maksimovic, V. 1999. Institutions, financial marketsand firm debt maturity. Journal of Financial Economics 54: 295–336.

Fosberg, R. H. 2012. Capital structure and financial crisis. Journal of Financeand Accountancy 11:46–52

Frank, M. Z., and Goyal, V. K. 2003. Testing the pecking order theory ofcapital structure. Journal of Financial Economics 67(2): 217–248.

Graham, J. R., and Harvey, C. R. 2001. The theory and practice of corporatefinance: Evidence from the field. Journal of Financial Economics 60:186–243.

Graham, J. R.; Leary, T. M.; and Roberts, M. R. 2014. A century of capitalstructure: The leveraging of corporate America. NBER working paper no.19910. Cambridge, Mass: National Bureau of Economic Research.

Harris, M., and Raviv, A. 1991. The theory of capital structure. The Journal ofFinance 46(1): 297–355.

Hodson, D., and Quaglia, L. 2009. European perspectives on the globalfinancial crisis: Introduction. Journal of Common Market Studies 47(5):939–953.

Huang, G., and Song, F. M., 2006, The determinants of capital structure:Evidence from China. China Economic Review, 17(1), 14–36.


Kahle, K. M., and Stulz, R. M. 2013. Access to capital, investment, and thefinancial crisis. Journal of Financial Economics 110(2): 280–299.

Kayhan, A., and Titman, Sh. 2007. Firms’ histories and their capital structures.Journal of Financial Economics 83: 1–32.

Leary, M. T., and Roberts, M. R. 2005. Do firms rebalance their capitalstructures? The Journal of Finance 60(6): 2575–2619.

Lemmon, M. L.; Roberts, M. R.; and Zender, J. F. 2008. Back to the beginning:persistence and the cross-section of corporate capital structure. The Journalof Finance 63(4): 1575–1608.

Lim, Y. 2003. Sources of corporate financing and economic crisis in Korea: Amicro-evidence. NBER working paper no. 9575. Cambridge, Mass:National Bureau of Economic Research.

Pattani, A.; Vera, G.; and Wackett, J. 2011. Going public: UK companies’ use of capital markets. Bank of England Quarterly Bulletin 51(4): 319–330.

Rajan, R., and Zingales, L. 1995. What do we know from capital structure?Some evidence from the international data. The Journal of Finance 50:1421–1460.

Voutsinas K., and Werner R.A. 2011. Credit supply and corporate capitalstructure: Evidence from Japan. International Review of Financial Analysis20(5): 320–334.

impact of financial crisis on firms’ capital structure in ......university of kent, uk this study...

Documents