Stuart Briers, 40040306, Size: The most important leverage determinant? 1 Queen’s University Management School, Queen’s University Belfast Size: The most important leverage determinant? An empirical analysis of the U.S. Financial Crisis 2007-2009 Stuart Briers 40040306 BSc Finance John Turner Wed, 14 th May 2014 Abstract The Financial Crisis of 2007-2009 was the worst since the 1929 Great Depression and forced firms to reassess their capital structure and exposure to the market. Through this study of the top 1,200 U.S. firms during the period, firm size is assessed as the most important component in dictating leverage. The Pecking Order Theory is tested and analysed as the underlying reason why firm size affects leverage when considering retained earnings and debt. Existing evidence is compared on firm size and the Pecking Order Theory, with conclusions and future areas of research given based on the results of the sample. Previous studies in capital structure have been “hampered by a lack of consistent accounting and market information outside the United States” according to Rajan and Zingales (1995). Therefore, this paper will review capital structure changes in the U.S.A. in order to gain a broad understanding of how firms are structured; allowing inferences to be drawn on the effect firm size has on leverage. The U.S.A. is an ideal country to analyse due to its access to capital markets and as Myers (2001) points out; firms have the “broadest menu of financing choices”.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Stuart Briers, 40040306, Size: The most important leverage determinant?
1
Queen’s University Management School,
Queen’s University Belfast
Size: The most important leverage determinant?
An empirical analysis of the
U.S. Financial Crisis 2007-2009
Stuart Briers
40040306
BSc Finance
John Turner
Wed, 14th May 2014
Abstract
The Financial Crisis of 2007-2009 was the worst since the 1929 Great Depression and
forced firms to reassess their capital structure and exposure to the market. Through
this study of the top 1,200 U.S. firms during the period, firm size is assessed as the most
important component in dictating leverage. The Pecking Order Theory is tested and
analysed as the underlying reason why firm size affects leverage when considering
retained earnings and debt. Existing evidence is compared on firm size and the Pecking
Order Theory, with conclusions and future areas of research given based on the results
of the sample.
Previous studies in capital structure have been “hampered by a lack of consistent accounting
and market information outside the United States” according to Rajan and Zingales (1995).
Therefore, this paper will review capital structure changes in the U.S.A. in order to gain a
broad understanding of how firms are structured; allowing inferences to be drawn on the
effect firm size has on leverage. The U.S.A. is an ideal country to analyse due to its access to
capital markets and as Myers (2001) points out; firms have the “broadest menu of financing
choices”.
Stuart Briers, 40040306, Size: The most important leverage determinant?
2
This paper investigates if firm size is the most important component in dictating
leverage and if it is positively correlated to it due to the Pecking Order Theory. This theory,
which was first introduced by Donaldson (1961) and modified by Myers and Majluf (1984),
states that firms use retained earnings until depleted, at which point they issue debt. Equity
will only be used when it is not feasible or sensible to issue more debt. The paper will analyse
if the Pecking Order Theory is relevant in the Financial Crisis or if other theories can be used,
such as the Trade-off Theory (Kraus and Litzenberger 1973), where the firm chooses its
mixture of debt and equity as a function of the present value of both tax shields and
bankruptcy costs. Most researchers in capital structure have acknowledged firm size as a
significant factor affecting leverage, but few have called it the most important leverage
determinant. This paper will analyse evidence to attempt to prove this thesis.
Other factors will also be considered including tangibility of assets, market-to-book
value and profitability. The Financial Crisis of 2007-2009 has been described as the worst
since the Great Depression by the IMF1, but due to publishing lags, little is known about how
firm size affected capital structure during this time. Miglo (2013) points out its importance as
the crisis “forced financial economists to look critically at capital structure theory because the
problems faced by many companies stemmed from their financing policies”. This paper will
attempt to fill this gap in literature by showing that the size of a firm has the effect to
significantly reduce investment (i.e. leverage see fig. 7a) during a crisis when financial
institutions cut lending. This theory is very clear in the 2007-2009 Financial Crisis.
After the abstract and introduction, the following section discusses existing literature
on firm size and its importance in determining leverage. Reasons for its effect on leverage are
given by different researchers and also existing evidence on the strength of the Pecking Order
Theory. Other variables are also considered.
1 The Guardian: see http://www.theguardian.com/business/2008/apr/10/useconomy.subprimecrisis
In the above regressions leverage is once again regressed with log sales. Only half of
the industries are reported purely to show industry differences. The number of observations
differs because the sample selects the largest firms each year, so some firms may fall out or
fall in to the sample. Firm size plays an extremely insignificant role in the Energy sector
possibly because they are typically large multinationals proving hypothesis 3 that large firms’
leverage are not influenced by firm size. This is compared to the Healthcare and IT sectors
where it is highly significant. Consumer Staples is a potential area for more investigation as
firm size is insignificant outside the crisis, but played a significant role in leverage during it.
These goods are generally necessities (e.g.
food) so could identify that these firms
should reduce price because of less
disposable income, using increased leverage
to finance this. At this time deflation was a
major concern for the U.S. government (see fig.
1). Bradley et al. (1984) cited in Chaplinsky and Niehaus indicates industry factors are
important in capital structure as firms choose it “on the basis of underlying costs and benefits
that are similar within each industry”. As demonstrated here, there are differences in how
Fig. 1: U.S. Inflation Rate 2008-2012
Stuart Briers, 40040306, Size: The most important leverage determinant?
13
firm size affects capital structure depending on industry.
F. Analysing Leverage Ratios
Considering all firms again, figure 2 shows a
polynomial trend for each period (excluding outliers
to analyse the main sample). Leverage ratios appear
to follow a wave-like cycle with firms in the 200-400
category using the highest leverage. It is interesting
to note that ratios have risen for the top 600 firms meaning they are taking on more debt or
retiring equity through share repurchases which is becoming increasingly common since
2009 (figure 33).
Looking historically, Myers (2001) has similar
findings in that “more shares are extinguished in
acquisitions and share repurchase programs than
are created by new stock issues”. For the
remaining firms leverage fell sharply in 2009 before returning to pre-crisis levels in 2012.
These firms reduced debt or found it was unavailable to them due to lending reductions by
financial institutions (see fig. 6).
G. Does the Pecking Order Theory exist in the sample?
According to figure 44, the difference in
retained earnings is largest in the top 200
firms, where it increased during the periods
due to capital adequacy increases by firms
who can afford it. A wave-like pattern exists
3 Share Repurchases by firms listed on major exchanges with S&P 500 Index: see http://opesforge.com/?p=340 4 Line of best fit omits extreme observations for visual reference
Stuart Briers, 40040306, Size: The most important leverage determinant?
14
similar to leverage. The general trend for the Pecking Order Theory appears to exist as
leverage and retained earnings are inversely related across the sample. However in taking a
snapshot for example in looking at the 200th firm, leverage and retained earnings both grow
during each period. Looking at smaller firms, retained earnings for the firms in the 400-600
category fell sharply (making losses) coupled with the largest sample increase in leverage,
supporting the theory. In 2009, leverage ratios fell sharply for firms outside the top 400
although retained earnings remained steady. This could be seen to represent a large fall in
capital expenditure during the crisis.
H. Considering Other Variables Affecting Leverage
Table VII: Robust Regressions using major determinants of Leverage
2006 2009 2012
Leverage Leverage Leverage
Beta 8.50e-05 0.015 0.020**
(0.008) (0.010) (0.010)
Tangibility of Assets 0.237*** 0.333*** 0.331***
(0.019) (0.019) (0.020)
Market-to-Book Value -0.033*** -0.027*** -0.028***
(0.004) (0.005) (0.004)
Log Sales 0.004** 0.010*** 0.008***
(0.002) (0.002) (0.003)
Profitability ROA -0.001 0.001* 0.002**
(0.001) (0.000) (0.001)
Constant 0.073*** -0.067*** -0.055*
(0.025) (0.021) (0.030)
Observations 1,200 1,200 1,200
R-squared 0.319 0.396 0.342
Rajan and Zinagles’ variables (and also market beta) will now be considered to
determine their effect on leverage. This regression is used to show that firm size is not the
only important variable. On first glance, the R2’s seem consistent with previous studies where
Friend & Hasbrouck found the “overall explanatory power of the cross-sectional models is
quite low”. For example Carlton and Silberman (1977) report an unadjusted R2 of 0.3 and
Marsh’s is 0.37. Comparisons of R2 can be used the studies use broadly similar variables. The
model may explain greater variation in 2009 possibly because investors placed more
emphasis on firm size (highest of the three periods) shown by log sales (positive and
Stuart Briers, 40040306, Size: The most important leverage determinant?
15
significant throughout), because they wanted to retain earnings due to market conditions (as a
capital buffer) and therefore issue more debt. This potential explanation contradicts the
Pecking Order Theory. Debt issuers would be prudent and favour larger firms due to
increased credit risk from more bankruptcies in 20095.
The Trade-off Theory is
supported as leverage could be
affected by the present value of
bankruptcy costs. These results
differ from Rajan and Zingales as
they find with greater bankruptcy costs comes greater equity issuance. The regressions’
findings disagree with Harris and Raviv’s idea that greater bankruptcy will mean decreased
leverage. Looking at all firms in the sample, the mean leverage in 2006 and 2009 is 0.17
compared to 2012 at 0.19 as seen in table II. It is no surprise that firm size is positively
correlated considering Friend and Hasbrouck’s findings that larger firms have better access to
credit markets. This is because the U.S.A. has potentially the easiest access worldwide to
these markets for example the largest stock exchange in the world; NYSE6.
Market Beta values are only significant in 2012. As beta values increase by a unit,
the increase in leverage grows each period. Firms with higher betas (more risky) will take on
greater leverage. The significance should be treated cautiously due to only having 2006 data.
Tangibility of assets is highly significant, possibly acting as a proxy for increased
collateral required during and post crisis in order to fulfil stricter lending requirements due to
lower bank lending7 (fig. 6). This may also be seen as an opportunity to diversify, for
example, by purchasing land due to increased mortgage defaults.
5 Bankruptcy Statistics 2006-2012: see http://www.tradingeconomics.com/united-states/bankruptcies 6 New York Stock Exchange:
see http://www.investopedia.com/financial-edge/1212/stock-exchanges-around-the-world.aspx 7 Commercial & Industrial Loans 2006-2012: see https://research.stlouisfed.org/fred2/series/BUSLOANS/
Leverage is discussed in the previous section but importantly it hits a trough
during the crisis for smaller firms (deciles 6-10) because they have greater systematic risk
(due to the small firm effect) as seen with higher betas in fig. 7b. Due to data limitations, only
the 2006 beta is shown and as expected for the top decile (making up the greatest percentage
of the market) their beta is roughly 1.0.
The tangibility of assets increased during the crisis due to stricter lending
requirements on collateral, but for the top decile it was effectively unchanged. Apart from
these firms, tangibility of assets is greatest after the crisis, as a measure to prevent another
securitisation crisis as these are intangible products which can be price sensitive.
Market-to-book value levels are extremely high pre-crisis; due to the cheap supply of
credit available for investment or perhaps due to over-valuations, for example in property.
They fall during the crisis as expected because the outlook is more pessimistic and due to
lower lending figures there is less chance of obtaining this credit to invest. Firms in the 4th,
5th and 10th decile categories have negative retained earnings so the only method of
investment for these firms is through equity.
Log sales (firm size) stayed the same for the top 120 firms showing price inelastic
firms. Smaller firms (10th decile) see sales drop during the crisis but have recovered by 2012.
Profitability fell sharply for all firms in 2009 and has not recovered since. Again, the
smallest firms in the sample suffer the greatest decline in return on assets because of risk-
9 Return on Equity
0
2
4
6
8
10
12
1 2 3 4 5 6 7 8 9 100
2
4
6
8
10
12
1 2 3 4 5 6 7 8 9 10
Stuart Briers, 40040306, Size: The most important leverage determinant?
18
averse investors who favour safer (larger) firms; although due to financial amnesia10 the
situation will probably soon return to pre-crisis levels. Zwiebel finds firms with better
investment opportunities (high market-to-book values) and high profitability to have less
leverage due to requiring less debt to avert a takeover. This generally occurs in this sample
particularly towards the top firms.
IV) Conclusion
The results show that firm size is the most important leverage determinant however it
depends on market capitalisation and industry ranking. Firm size was found to be positively
correlated with leverage and the weight of debt, in each case being highly significant. Log
sales were relatively sticky during the three periods for the top decile showing price inelastic
firms. Smaller firms (10th decile) seen log sales drop during the crisis but recovering by 2012.
The firm needs to be considered within its relative placing in the economy as the very top
firms find firm size is insignificant in affecting leverage, whereas the bottom firms of the
sample find firm size plays a highly significant role in affecting leverage.
The type of industry a firm is placed in will also matter greatly as energy firms find
firm size to be insignificant, whereas in healthcare and IT
it is highly significant. In Consumer Staples, firm
size was only significant during the crisis
potentially due to changing consumer trends for
example a shift to discount dollar stores from
traditional stores (fig. 811). Firm size plays a more
important role in a crisis as documented by R2 in all regressions.
10 Financial Amnesia: see CFA July-Aug 2012 Publication
http://www.cfapubs.org/doi/pdf/10.2469/cfm.v23.n4.7 11 Wal-mart v Dollar Stores (Note that Wal-Mart’s sales numbers were divided by a factor of 10 to allow for a