The impact of country effects on the capital structure of European
firms during the recent financial crisis
Author: Randy Keehnen1 Date: July 2014 Supervisor: Dr. H. Gonenc Master Thesis University of Groningen Faculty of Economics and Business MSc Business Administration - Finance
The impact of country effects on the capital structure of European
firms during the recent financial crisis
By: Randy Keehnen
Abstract
This study analyzes data from 4013 firms divided over 22 European countries over the period 2005-2011, to investigate the relationship between a crisis period and the explanatory power of country-level determinants of leverage. In line with Booth et al. (2001), the results of this study show that country dummies have a significant influence in explaining the capital structure of firms as a country-level determinant. GDP growth also showed a positive impact on the explanatory power of models, but it was smaller.
The analysis in this paper clearly shows increased explanatory power of country dummies during a crisis period; however there is no statistical test to proof its significance.
Key words: capital structure, leverage, firm-level determinants, country effects, financial crisis
TABLE OF CONTENTS
1. Introduction ... 4
2. Literature review ... 5
2.1. Theories of capital structure ... 5
2.1.1. Trade-off theory ... 6
2.1.1.1. Static trade-off theory ... 7
2.1.1.2. Dynamic trade-off theory ... 7
2.1.2. Pecking order theory ... 8
2.1.3. Agency theory ... 9
2.2. Country-level determinants of leverage ... 10
2.2.1. Country dummies ... 10
2.2.2. GDP growth ... 10
2.2.3. Other country-level determinants of leverage ... 11
2.3. Firm-level determinants of leverage ... 12
2.3.1. Tangible assets ... 12
2.3.2. Size ... 13
2.3.3. Profitability ... 13
2.3.4. Growth opportunities ... 13
2.3.5. Others firm-level determinants of leverage ... 14
2.4. The crises since 2007 ... 14
2.4.1. The effect of a crisis on leverage ... 16
2.4.2. The effect of a crisis on country-level determinants of leverage ... 16
2.4.3. The effect of a crisis on firm-level determinants of leverage ... 18
3. Data & Methodology ... 19
3.1. Data ... 19
3.2. Methodology ... 23
4. Results ... 25
5. Conclusion ... 30
1. Introduction
Many things have been written about the recent crisis, but not many papers have investigated the effects of a crisis on the capital structure of firms. With the supply side of debt being more constrained and the equity markets liquidity going down, I expect that the crisis should have an effect. Therefore, to fill the existing gap and explore this expectation, this study will investigate the influence of the financial crises since 2007 on the capital structure of European firms.
During the subprime mortgage crisis, banks were faced with big write-offs on their mortgage-backed securities and collateralized debt obligations portfolios. (Lemke et al., 2013) Lehman Brothers, one of the biggest investment banks, even went bankrupt in 2008. Due to their weakened financial positions, banks became more reluctant in providing new loans and more risk-averse in providing debt. Long-term interest rate spreads started diverging as rates went down for countries like Germany, considered relatively safe by the market, while southern European countries saw their rates soar. As debt access became more constraint, leverage of firms should be affected by the crisis, and given the divergence of countries’ long-term interest rates, these effects are expected to be different across countries as well.
Several studies have investigated the influence of country-level determinants on leverage, finding that these determinants explain 33% (Gungoraydinoglu and Öztekin, 2011) to 43% (Booth et al., 2001) of the variation in capital structure. De Jong et al. (2008) even concluded that country specific factors have both a direct impact on leverage and an indirect impact on the importance of other determinants. This paper will investigate the influence of country-level determinants on leverage, by comparing several models including and excluding control variables for country, namely GDP growth and country dummies. Based on the literature, it is expected that a model with a country control variable will have a higher explanatory power than a model without such variable, displayed by a higher adjusted R-squared.
In this paper, I analyse firms from 22 different European countries over the period 2005-2011, using models similar to those of Booth et al. (2001) and De Jong et al. (2008). The period is chosen to cover a pre-crisis period (2005-2006), the subprime crisis period (2007-2009) and the Eurozone crisis period (2009-2011). This way the results can be compared across these periods. Note that the Eurozone crisis is considered to still be ongoing; however the data will only cover this crisis until and including 2011.
The results of the analysis show that adding country dummies cause a sharp increase in the explanatory power of a model that estimates firm leverage, in line with findings of Booth et al. (2001) and Gungoraydinoglu and Öztekin (2011). The analysis also shows that the explanatory power of models with country dummies is higher in crisis years than in pre-crisis years, in line with expectations based on previous literature. Finally, this paper also shows that even though GDP growth increases the explanatory power of the models, the increase is relatively small.
This study contributes to the evidence on the impact of country-level determinants on leverage, as it shows that models with these determinants have a higher explanatory power. It also provides initial insights in the effect of a crisis on the importance of these determinants, namely that during a crisis period, the explanatory power of country-specific factors increases.
This paper is organized as follows: Section 2 discusses the relevant background and literature, where section 2.1 will focus on the different theories of capital structure, followed by section 2.2 and 2.3 which will discuss the country-level and firm-level determinants of leverage. Section 2.4 will discuss the crises since 2007 and review literature concerning the effects of crises on leverage and its determinants. Section 3 discusses data and methodology. Section 4 presents the results. Section 5 concludes this paper.
2. Literature review
2.1. Theories of capital structure
chooses a certain proportion of debt and equity to finance its assets, which divides up the cash flows among investors. Assuming investors and firms have equal access to financial markets; investors can use homemade leverage to make sure they have the leverage they want. This classic arbitrage-based irrelevance proposition reasons that arbitrage by investors cause the value of a firm to be independent of its leverage. Another kind of capital structure irrelevance was shown in a paper by Miller (1977), which considered that both personal and corporate taxes determine the economy wide leverage ratio; however there are multiple equilibriums in which debt is issued by different firms.
Research since 1958 has shown that the Modigliani-Miller theorem fails under a variety of circumstances, as it assumes a perfect market with no taxes, transactions costs, bankruptcy costs, agency conflicts, adverse selections, and many other things which are not true in the real world. For this reason, the discussions around capital structure of firms evolved around these different topics influencing capital structure decisions, how capital structure decisions influence stock prices and ultimately research investigating the influence of several determinants of leverage.
2.1.1. Trade-off theory
One of the most well known theories is the trade-off theory. In their paper, Modigliani and Miller (1963) highlight the theoretical tax advantages of debt, which added to their previous theory would imply 100% debt financing by firms to shield their earnings from taxes. However, their data indicated that even in high tax years, there is no substantial increase in the use of debt, except relative to preferred stock. “The tax advantages of debt financing must be considerably less than the conventional wisdom suggests,” as Miller (1977) would write later on.
2.1.1.1. Static trade-off theory
For these reasons, Frank and Goyal (2008) split up Myers’ (1984) definition into two parts. A firm is said to follow the static trade-off theory if the firm’s leverage is determined by a single period trade-off between the tax benefits of debt and the deadweight costs of bankruptcy. They show that an increase in the costs of financial distress, an increase in non-debt tax shields, or an increase in the marginal bondholder tax rate reduces the optimal non-debt level. An increase in the personal tax rate on equity increases the optimal debt level of a firm. The effect of risk on leverage is ambiguous from a theoretical point of view, but Bradley et al. (1984) show that the relation between debt ratio and volatility is negative. It is important to note that this model is static, while firms operate over many periods. This makes the role of retained earnings and the interpretation of mean reversion particularly important. Retained earnings are not part of the model, but play a role in that profitable firms automatically create retained earnings and if the firm does not take offsetting actions, its leverage will automatically go down.
2.1.1.2. Dynamic trade-off theory
Another issue with this static model is that there is no other target for the firm than the solution of the model. It does not allow room for the target to float around over time and mean reverting towards the solution or small target adjustments based on expectations for the next period.
introduced analysis which included not only taxes and bankruptcy costs, but also transaction costs. As a result of the latter, firms choose to let their leverage drift much of the time, and it only takes action when it gets too far away from target to stay within a determined boundary; profits are used to repay debt and when the leverage becomes too low the firm issues new debt to rebalance. When losses increase debt automatically, equity will have to be issued when leverage hits the upper limit. Not surprisingly it is observed that profits and leverage are negatively related. Furthermore they show that volatility is negatively related to average leverage, because greater volatility results in a firm allowing the leverage ratio to fluctuate more, and the adjustment made when a boundary is crossed is also smaller. The most remarkable conclusion is that firms which continue to perform well will eventually issue debt to rebalance their capital structure.
Lewellen and Lewellen (2004) point out that when shares are repurchased by the firm, the amount of tax that is paid by shareholders depends on the capital gain since the time of purchase of these shares. Therefore firms with many long-term shareholders can be more reluctant to repurchase their shares as their shareholders would be expected to incur higher tax costs. Hennessy and Whited (2005) is one of the few papers that complete their model with the option for firms to dynamically deal with excess funds. They find that optimal leverage is path dependent and that profitable firms tend to be less highly levered.
2.1.2. Pecking order theory
According to the pecking order theory of Myers (1984), firms prefer internal to external financing and debt to equity if external financing is used. As noted before, Stiglitz (1973) showed that tax considerations can lead to this pecking order behaviour. The adverse selection motivation is based on the idea that the decision maker of the firm has more information about the firm’s value and growth opportunities than do outside investors. As a result the manager must consider the information signals resulting from his actions. If he decides to sell equity to the market, it implies he thinks that the market value of the firm is too high and thus shares are overpriced. Alternatively if he buys back shares from the market, it reveals that he thinks the share price is too low based on his inside information. Ross (1977) showed that firm value would be expected to rise with leverage, because increasing leverage increases the market’s perception of value.
debt works as well as internal financing, while risky debt is still a better option than equity (Myers, 1984). This is the pecking order. However, papers with different assumptions have found different results. Eckbo et al. (1990) found that in mergers with two-sided information asymmetry, firms sometimes prefer stock transactions over cash transactions. Halov and Heider (2003) argue that when the asymmetric information is about risk instead of firm value, the adverse selection argument focuses on debt and as a result external equity would be preferred over debt.
Jensen and Meckling (1976) defined the concept of agency costs and how it affects capital structure decisions. When trying to optimise value for the firm, he shows that in the example of an entrepreneur, retained earnings are preferred and debt can be equally good, but equity is inefficient. This is another version of the pecking order. They also argue that it makes a difference whether a firm focuses on shareholders or debt holders. If it is operated with the focus on shareholders, cash flows in bankruptcy states do not matter anymore, in which case a firm is more likely to accept riskier projects with bigger payoffs, as the downside of shareholders is limited to their investment and they do not care about the results for debt holders in case of bankruptcy. This behaviour is more likely to occur in firms that are already in trouble and feel forced to take the extra risk to turn the tide.
2.1.3. Agency theory
Harris and Raviv (1991) give a good summary of the agency cost related theory as put forward by Jensen and Meckling in 1976. There are two types of agency conflicts that influence capital structure decisions.
Conflicts between shareholders and managers arise because while managers are held responsible in case of losses, they do not receive the full amount of profits when they are doing well. This gives managers an incentive to reduce profits of the firm by spending more resources on expensive dinners, corporate jets and trips, et cetera. From the managers perspective, not doing so costs him these ‘benefits’ to the full extend, while he only receives a portion of the ‘gain’ it leads to through reduced costs. From a shareholder point of view, increasing debt could reduce this problem as it reduces the amount of free cash available due to interest payments.
unbalanced. Because debt holders are aware of this issue, they take into account the expected behaviour of equity holders, therefore demanding more interest, and thus ultimately the cost of the incentive falls on the shoulders of the equity holders who issue the debt. This effect is generally referred to as the ‘asset substitution effect’.
Based on these conflicts and the relevant responses in practice, the theory predicts that when a firm has a highly regulated environment, is part of a mature industry, is bigger, and/or has few growth opportunities, it will be more highly levered as the opportunities for asset substitution are lower. Also firms with higher operational cash flows should have more debt to reduce the amount of cash available to managers.
2.2. Country-level determinants of leverage
Even though country-level determinants of leverage are a much less investigated by academic literature than firm-level determinants, there are studies that investigate a great number of different factors like De Jong et al (2008) and Gungoraydinoglu and Öztekin (2011). The general finding is that country-level determinants play an important role in explaining firm leverage.
2.2.1. Country dummies
Booth et al. (2001) say that “Knowing the country of origin is usually at least as important as knowing the size of the independent variables for both the total and long-term book-debt ratios. Only for the market-debt ratio this is not true.” Their study shows that adding country dummies to the regression reduces the significance of the tax rate, tangibility and size coefficients; however it hardly changes the coefficients themselves. Their results indicate that their country dummies explain 43% of total-debt variation. Gungoraydinoglu and Öztekin (2011) investigate a great number of different country-level determinants and conclude that these explain one-third of the variation in capital structure, emphasising the importance of a country proxy in models explaining leverage.
2.2.2. GDP growth
a given year for a given country. Therefore, not many papers really go in depth on analysing this variable. What the impact of GDP growth as a country-level determinant should be on leverage much follows the discussion about growth opportunities. Therefore some papers expect and find a negative impact on leverage based on the dynamic trade-off theory (Kayo and Kimura, 2011), or argue this result is based on the pecking order theory (using retained earnings over debt, Bastos et al., 2009).
Others find a positive impact based on the pecking order theory (using debt over equity, De Jong et al., 2008). Interestingly, Hanousek and Shamshur (2011) find that GDP growth (used as a proxy for growth opportunities of a firm) has a positive impact on leverage in their total dataset, but a negative impact when they are only looking at the profitable firms in their dataset. This implies that GDP growth must have a relatively big positive impact on leverage for unprofitable firms. Huang and Ritter (2005) find that the real GDP growth rate is positively related to the likelihood of debt issuance, but is not reliably related to the likelihood of equity issuance.
2.2.3. Other country-level determinants of leverage
According to Taggart (1985), when inflation is expected to be high the real value on tax deductions on debt will be higher. Frank and Goyal (2009) show that expected inflation has a positive effect on firm leverage, in line with the trade-off theory. Bastos et al. (2009) finds no significant result. Gajurel (2006) finds that inflation is negatively related to the long-term debt ratio in his study focusing on Nepal. Fernandes (2011) did not find a significant difference in debt levels and their determinants between high- and low-inflation countries.
Stock market development of a country is expected to reduce firm leverage, as it allows companies better access to an alternative source of capital to finance investments and growth. Kayo and Kimura (2011) find this relationship with their analysis covering 40 different countries. However, they also find a negative coefficient for bond market development which is much less expected. De Jong et al. (2008) argue that with more bond market development, firms have more options for borrowing and creditors are more willing to provide debt. In their analysis they do find this positive relationship.
creditor right protection of a country has a significant negative impact on leverage. First expecting that this measure would be a sign of bond market development and thus have a positive impact, they argue that their results imply that higher creditor right protection implies that debt is more risky for firms increasing the risk of bankruptcy in times of financial distress. This would make firms more reluctant to borrow.
Based on the literature described above I will investigate whether adding country determinants to a model will explain leverage better, testing the following hypothesis;
Hypothesis I: Models containing country-level determinants will explain leverage better than
models not containing country-level determinants.
2.3. Firm-level determinants of leverage
Studies on firm-level determinants of leverage are plentiful. In this section I will discuss the most important firm-level determinants as well as the expected impact they have on the capital structure of firms. These will be used as control variables in the analysis.
2.3.1. Tangible assets
Tangible assets are probably the most investigated firm-level determinant of leverage in literature. If a company has more tangible assets, the liquidation value of the company is higher. As a result, cost of debt will be lower, because lenders suffer less risk and will therefore demand a lower premium. Therefore, the relative amount of tangible assets to total assets should have a positive relationship with the leverage ratio (Harris and Raviv, 1990). They also reason that managers use debt as a strategy to counter the high liquidation value from the tangible assets, by increasing the amount of debt and making liquidation of their company a less desirable strategy for shareholders.
countries. Firms with high levels of tangible assets are better able to raise finance in countries that provide strong creditor protection.
2.3.2. Size
Since larger firms are usually more diversified and have more stable cash flows, they have a lower risk of bankruptcy and lower bankruptcy costs according to the trade-off theory. Therefore they are expected to have higher levels of leverage. This result has been found by Ross (1977), Fama and French (2002), De Jong et al. (2008), Brav (2009), Frank and Goyal (2009), Kayo and Kimura (2011), Fernandes (2011) and many other papers.
On the other hand, however, the cost of issuing debt and equity is also related to firm size, and to issue new equity is much more costly for smaller firms, more so than long-term debt. Therefore, and because it allows firms flexibility to respond quickly to new opportunities (Shuetrim et al., 1993), smaller firms are expected to have relatively high leverage ratios, as found by Titman and Wessels (1988) based on the log of sales.
Rajan and Zingales conclude that they “do not really understand why size is correlated with leverage” (page 1457). The literature review work of Harris and Raviv (1991) concludes that even though different empirical results emerge from different papers, overall leverage increases with firm size.
2.3.3. Profitability
Frank and Goyal (2009), Fernandes (2011), and Kayo and Kimura (2011) find that more profitable firms have lower leverage when measured by total debt, but Fernandes also finds that these have more long-term debt. This is consistent with the pecking order theory and the dynamic trade-off theory, according to which more profitable firms should retain their earnings, because they are expected to have better investment opportunities (Frank and Goyal, 2008). Booth et al. (2001) finds the same result.
2.3.4. Growth opportunities
opposite results have been found by Titman and Wessels (1988), Rajan and Zingales (1995), and Allayannis et al. (2003).
2.3.5. Others firm-level determinants of leverage
Taxes have been a well debated characteristic, stemming from the early discussions about tax shields. Fernandes (2011) finds a significant negative relationship just like Booth et al. (2001), even though the static trade-off theory expects a positive relationship. They explain it by stating that the tax rate seems to be a proxy for profitability rather than tax shield effects, and therefore has the same negative sign. Fernandes also investigates the effect of business risk on leverage, but he finds no significant positive relationship.
2.4. The crises since 2007
The subprime mortgage crisis (2007-2009) was caused by falling housing prices and the effects this had on the value of mortgage-backed securities and collateralized debt obligations owned by banks and other institutions. (Lemke et al., 2013) After peaking in mid-2006, U.S. housing prices started declining steeply making it more difficult for borrowers to refinance their loans. With the increasing interest rates mortgage delinquencies rose to record highs and financial firms saw the value of securities backed with mortgages plummet. (Simkovic, 2011) During 2008 one of the biggest investment banks, Lehman Brothers, went bankrupt and other banks had to be saved to prevent a total collapse of the financial system. In Europe, big banks like HSBC and BNP Paribas were forced to do massive write offs on their investment funds and many banks were severely weakened in their financial positions. As a result, banks became more reluctant in providing new bank loans and, combined with concerns in the market about the financial system and the soundness of U.S. credit, this led to the recession of 2008 that was felt all over the world.
In response to the recession, the European Central Bank started lowering its minimum bid rate in October 2008, from 4.25% to 1.00% in May 2009.2 European banks rely heavily on the ECB liquidity since the crisis, due to the severe drop in interbank lending. As a result they took on this liquidity, but in turn they remained reluctant to invest this in newly issued government debt from European countries. As a result it became difficult or even impossible
for some countries to repay or refinance their government debt without the assistance of third parties like the ECB or IMF. (Shambaugh et al., 2012)
As government bonds got downgraded by rating agencies, a lot of European countries saw their bond yields rise to record highs, with long-term interest rates going as high as 36% for Greece and over 10% for countries like Cyprus, Ireland, and Portugal. Big economies like Spain and Italy saw their interest rates rise to over 6% causing growing concern about the financial stability of the euro area, while Germany saw its rate drop to 1.3% as it was considered relatively safe. As the European nations took measures like the European Financial Stability Facility, the European Stability Mechanism, and to deal with their budget deficits, interest rates started converging again in the second half of 2012, as shown in Fig. 1. (Source: Eurostat)
Figure 1 - Divergence and convergence of
long-term interest rates
0.00 5.00 10.00 15.00 2005 2006 2007 2008 2009 2010 2011 2012 2013 Year 1 0 -y e a r bo nd y ie ld (% )
Germany Netherlands Portugal Spain
In Spain, the economy went into a severe recession since the end of 2007. The unemployment rate reached 25% mid 2012 (Forelle and Steinhauser, 2012), the interest rate gap with Germany widened, and the Spanish economy had built up 2 million of empty dwellings, underlining the big real estate problems. During a long period of very low interest rates, households were encouraged to go on debt and had borrowed massively on variable interest rate loans; 98% of the loans are with variable interest rate and the private Spanish debt represented 64.6% of the GDP. (Gentier, 2012) While the Spanish banking system used to be considered one of the strongest of all Western economies, several banks got downgraded to “junk” status in May 2012 and the Bankia bank, Spain’s biggest mortgage lender, was nationalised requiring a bailout of €23.5 billion to cover losses from failed mortgages. (Bjork et al., 2012) In July 2012, Eurozone finance ministers agreed to provide Spanish banks with €100 billion of rescue loans.3
As many big banks across Europe had considerable exposure in Spain, this was considered the best way to prevent a domino effect that could potentially knock over the entire European banking system.
The rest of this section will discuss literature regarding the effects of crises on leverage and its determinants.
2.4.1. The effect of a crisis on leverage
While investigating the impact of the East Asian crisis (1997-1998) on the capital structure of emerging market firms, Fernandes (2011) finds that while total debt to equity ratios peaked during the crisis, they decrease substantially in the years after. This is expected from the market leverage ratio as the stock prices plummeted during the crisis to recover afterwards, however the same effects are found for book value leverage ratios as well.
Allayannis et al. (2003) show that, while investigating the East Asian crisis of 1997-1998, firms with more foreign EBIT, larger firms, and firms with foreign equity listing perform significantly better during the crisis. As the value of their equity remains relatively high compared to their peers, these companies would be expected to see their relative leverage go down.
2.4.2. The effect of a crisis on country-level determinants of leverage
Bris et al. (2004) investigate differences between crises on different continents. They show that the crises had a very significant positive effect on leverage in Asia and Latin
America, while the effect in Europe was also positive, but far less significant. This implies that country variables play an important part in capital structure decisions in times of crises. Greenlaw et al. (2008) states that contractions in financial institutions’ balance sheets cause a reduction in real GDP growth, showing an important link with crises.
In his study on the determinants of leverage, Fernandes (2011) uses a big dataset of 10,000 firms from 30 emerging markets. He finds that in the early 1990s, leverage was largely determined by country characteristics indicating that firms were dependant on their domestic environment to raise debt. However as emerging markets developed and got better integrated with the world economy, he shows that over time the importance of firm characteristics increases. At the end of his sample period, in 2007, capital structure determinants appear to be the same for emerging and developed markets, and country characteristics have become less important in explaining finance structures.
The opposite seems to happen in times of crisis, when financial market participants start acting suboptimal, resulting in sudden shocks in debt markets, equity markets, exchange rates and interest rates. As a result financial institutions get more stringent on providing loans as they lower their capacity and willingness to take on risk (Duchin et al., 2010). In the recent financial crisis, loan spreads skyrocketed while the higher post-crisis investment of cash-rich firms still seemed to be efficient, which means that most likely opportunities for growth were lost.
In their study, Allayannis et al. (2003) use the Asian financial crisis of 1997-1998 to investigate the role of debt type in firm performance. They show that firms with more foreign EBIT, larger firms, and firms with foreign equity listing perform significantly better during the crisis, indicating that firms with more diversification of their debt perform better as they are less influenced by the mentioned exchange rate and interest rate shocks. In long-term interest rates of European countries a similar result could be observed as investors shifted from Southern European countries to Northern European countries, widening spreads between countries making more diversified firms better protected. Fernandes (2011) shows that firms with ADRs have lower leverage ratios, because having a US cross-listing implies a reduction in the information asymmetry between managers and outside shareholders. Given how interest rate spreads between countries moved during the crisis, it seems obvious that the country of origin of a firm gained in importance when determining the access to debt.
the public debt market (indicated by having a credit rating) has a bigger impact on leverage than in other years. Banking crises hinder growth more in industries that are more dependent on external finance (Dell’Ariccia et al., 2008), also because firms forego profitable investment opportunities as a result of binding external financial constraints (Campello et al., 2010). Finally, during crises stock price declines are more severe for more financially constrained firms (Tong and Wei, 2008).
As the crisis leads markets to perform less optimal, I expect that the explanatory power of country coefficients will become bigger during a crisis period, which is what I will investigate in this paper. Based on this I formulate the following hypothesis;
Hypothesis II: Country-level determinants will have a bigger explanatory power on leverage
during a crisis period compared to before a crisis.
2.4.3. The effect of a crisis on firm-level determinants of leverage
There has also been some debate on the effects of crises on firm-level determinants of leverage. Fernandes (2011) shows that tangibility had almost no impact (coefficient close to zero) on leverage in the early years of his sample, but after the Asian crisis the effect became positive and very significant. Fernandes (2011) also finds that financial development of a country, measured by the ratio of stock market capitalization to GDP, has a negative impact on the importance of tangibility in explaining leverage ratios. One can argue that in a crisis, financial development becomes worse due to market inefficiency so that tangibility becomes more important to explain leverage.
Duchin et al. (2010) study the effect of the recent financial crisis on corporate investment. They find that corporate investment declined significantly (by 6.4%) and this decline is greatest for firms that have low cash reserves or high net short-term debt, are financially constrained, or operate in industries dependant on external finance (see also Rajan and Zingales, 1995). Duchin et al. show that, while net short-term debt has a negative effect on post-crisis changes in investment, long-term debt does not. This implies that this is a supply effect, caused by financial institutions becoming more risk averse after the initial substantial losses of the crisis. Their results also suggest that the higher post-crisis investment of cash-rich firms is efficient.
Dang et al. (2014) show that a firm tends to borrow less during an economic downturn as the value of its collateral declines and its debt capacity is reduced. Firms rebalance their leverage ratios more frequently in economic expansions than in economic recessions. Almeida et al. (2004) show that firms save cash out of their cash flows, (only) when they are financially constrained and run the risk of underinvesting in future states of the world.
3. Data & Methodology
3.1. Data
This paper will use a dataset containing 4013 firms from 22 European countries. Firm-specific data was collected from the Thompson Reuters Worldscope/Datastream database. GDP growth data was collected from Worldbank data of world indicators. The sample covers the years 2005-2011, to allow for the analysis of the pre-crisis period (2005-2006), the financial crisis period (2007-2009), and the Eurozone crisis period (2009-2011).
0.045. The last number implies that this sample has more profitable than non-profitable observations, even though the mean profitability is only 0.001. This low mean is mainly cause by some big firm-year losses (biggest loss is -4.407) which can be expected from a crisis period.
Table 1 Summary statistics of variables used
Mean Median Std.Dev. Min Max Observations Leverage 0.308 0.279 0.261 0.000 1.000 26316 Tangibility 0.238 0.176 0.220 0.000 0.919 26316 Size 12.168 12.041 2.168 4.564 16.893 26316 Profitability 0.001 0.045 0.279 -4.407 0.344 26316 Tobin's Q 1.647 1.260 1.168 0.621 6.371 26316 GDP growth 1.407 1.936 3.159 -8.539 9.157 26316
Table 2 presents the mean and median values of leverage and the firm-level determinants for each country. It shows that the highest leverage is found in Portugal, Italy, Greece and Spain, the group referred to often as the PIGS countries as they were hit hardest by the crisis and they required a lot of financial aid from other countries. Ireland was added to this group, but their crisis was mainly caused by their struggling housing market and their leverage does not stand out. Slovenia is also high on leverage but with only 9 Slovenian firms in the sample is not as notable.
Table 2 Cross-country summary statistics of leverage and firm-level determinants
Leverage Tangibility Size Profitability Tobin's Q GDP growth Austria Mean 0.375 0.294 13.032 0.028 1.458 1.667
Median 0.368 0.310 12.958 0.044 1.218 2.401 Belgium Mean 0.350 0.291 12.985 0.036 1.520 1.426
Median 0.353 0.234 12.665 0.047 1.152 1.769 Czech Republic Mean 0.157 0.419 12.952 0.045 1.295 3.425
Median 0.079 0.553 13.582 0.037 1.285 3.099 Denmark Mean 0.334 0.315 12.366 0.007 1.678 0.497 Median 0.310 0.264 12.256 0.050 1.212 1.577 Finland Mean 0.341 0.232 12.570 0.036 1.589 1.460 Median 0.351 0.182 12.287 0.058 1.300 2.916 France Mean 0.334 0.173 12.558 0.027 1.503 1.052 Median 0.318 0.110 12.240 0.050 1.214 1.826 Germany Mean 0.300 0.227 12.140 0.000 1.548 1.615 Median 0.265 0.178 11.947 0.038 1.224 3.269 Greece Mean 0.442 0.384 12.056 0.005 1.140 -0.353 Median 0.460 0.368 11.979 0.017 0.962 -0.214 Hungary Mean 0.200 0.353 11.802 -0.003 1.352 0.612 Mean 0.164 0.365 11.401 0.025 1.110 1.258 Ireland Median 0.323 0.252 12.751 0.039 1.677 1.345 Mean 0.337 0.195 12.943 0.054 1.298 1.431 Italy Median 0.444 0.224 13.201 0.014 1.311 0.024 Mean 0.451 0.169 12.997 0.027 1.159 0.931 Luxembourg Median 0.270 0.348 13.938 0.138 2.905 2.240 Mean 0.200 0.292 14.057 0.122 2.228 3.101 Netherlands Median 0.332 0.222 13.219 0.022 1.735 1.460 Mean 0.336 0.173 13.507 0.072 1.367 1.804 Norway Median 0.397 0.335 12.804 0.006 1.579 1.034 Mean 0.416 0.231 12.866 0.035 1.244 1.218 Poland Median 0.262 0.331 11.520 0.027 1.474 4.543 Mean 0.237 0.329 11.450 0.043 1.191 4.453 Portugal Median 0.609 0.324 13.369 0.006 1.258 0.389 Mean 0.627 0.341 13.359 0.017 1.119 0.775 Slovenia Median 0.426 0.515 14.009 0.038 1.249 2.188 Mean 0.487 0.504 13.982 0.026 1.051 3.589 Spain Mean 0.442 0.312 13.828 0.040 1.592 1.118 Median 0.441 0.279 13.725 0.039 1.251 0.892 Sweden Mean 0.255 0.155 11.473 -0.035 1.909 2.041 Median 0.203 0.071 11.347 0.049 1.433 3.161 Switzerland Mean 0.253 0.257 13.143 0.034 1.729 2.174 Median 0.216 0.206 12.905 0.065 1.353 2.695 Turkey Mean 0.277 0.359 12.234 0.045 1.450 4.843 Median 0.220 0.355 12.106 0.044 1.155 6.893 United Kingdom Mean 0.243 0.195 11.431 -0.038 1.922 0.869
Table 3 Sample description by country Country Number
of firms
Mean leverage (%) per year All years 2005 2006 2007 2008 2009 2010 2011 Austria 49 35.2 35.6 37.9 38.2 39.0 39.1 37.3 37.5 Belgium 79 37.9 36.6 34.8 36.2 34.1 33.3 32.2 35.0 Czech Republic 5 18.7 17.2 21.7 16.6 2.5* 15.6 15.4 15.7 Denmark 91 32.7 32.1 31.9 34.5 35.1 34.2 33.1 33.4 Finland 109 31.1 30.8 32.4 38.2 38.3 33.8 34.0 34.1 France 480 35.2 33.4 32.9 35.7 33.5 31.3 31.5 33.4 Germany 501 30.2 28.7 29.1 31.8 31.1 29.8 29.2 30.0 Greece 204 39.4 40.1 41.5 44.3 45.4 47.2 52.4 44.2 Hungary 21 20.7 17.9 16.3 20.0 20.6 21.0 23.5 20.0 Ireland 44 30.6 29.7 28.9 34.1 38.5 31.2 33.5 32.3 Italy 183 40.4 40.4 41.4 45.8 47.5 46.3 48.6 44.4 Luxembourg 12 23.6 32.6 31.5 28.5 27.8 24.2 19.2 27.0 Netherlands 100 30.1 31.0 32.1 38.4 34.7 32.4 33.9 33.2 Norway 117 34.1 38.6 39.2 43.5 40.8 40.5 40.0 39.7 Poland 177 24.8 23.2 24.1 28.3 27.7 26.5 27.7 26.2 Portugal 40 57.0 57.5 58.3 64.5 64.6 61.6 62.9 60.9 Slovenia 9 34.9 32.7 39.2 46.7 45.1 49.2 50.0 42.6 Spain 89 38.1 41.0 42.3 45.4 46.1 47.2 48.5 44.2 Sweden 266 24.6 23.3 24.7 27.6 25.3 26.3 26.7 25.5 Switzerland 148 26.6 24.9 24.6 26.2 26.1 24.1 24.7 25.3 Turkey 171 25.5 24.9 24.3 31.8 28.4 28.9 29.8 27.7 United Kingdom 1118 23.9 23.3 24.2 25.8 25.5 24.2 23.1 24.3 All countries 4013 30.0 29.3 29.8 32.7 31.7 31.0 31.2 30.8 Firm-year observations 3463 3750 3962 3980 3675 3830 3656 36316 *This shock in the 2009 mean leverage is caused by a missing firmyear for a high leverage company.
Figure 2 - Mean leverage before and during the crisis
0.250 0.300 0.350 0.400 0.450 0.500 0.550 2005 2006 2007 2008 2009 2010 2011 Year M e a n le v e ra ge
I will use GDP growth as a country-level determinant. Table 4 shows the cross-country GDP growth per year. As the sample has only European countries, the table shows that the economies slow down to no growth in 2008, followed by significant negative growth in 2009. In the years thereafter most countries recover from the crisis showing some modest growth. The table shows significant differences in GDP growth between countries each year, making it viable to have a real impact as a country determinant.
Table 4 GDP growth (in %) per country per year
2005 2006 2007 2008 2009 2010 2011 Austria 2.40 3.67 3.71 1.44 -3.82 1.77 2.83 Belgium 1.75 2.67 2.88 0.99 -2.80 2.32 1.77 Czech Republic 6.75 7.02 5.74 3.10 -4.51 2.47 1.82 Denmark 2.45 3.39 1.58 -0.78 -5.67 1.58 1.10 Finland 2.92 4.41 5.34 0.29 -8.54 3.36 2.73 France 1.83 2.47 2.29 -0.08 -3.15 1.72 2.03 Germany 0.68 3.70 3.27 1.08 -5.15 4.01 3.33 Greece 2.28 5.51 3.54 -0.21 -3.14 -4.94 -7.10 Hungary 3.96 3.90 0.11 0.89 -6.80 1.26 1.60 Ireland 5.88 5.40 5.45 -2.11 -5.46 -0.77 1.43 Italy 0.93 2.20 1.68 -1.16 -5.49 1.72 0.48 Luxembourg 5.25 4.94 6.59 -0.73 -5.56 3.10 1.90 Netherlands 2.05 3.39 3.92 1.80 -3.67 1.53 0.94 Norway 2.59 2.30 2.65 0.07 -1.63 0.48 1.22 Poland 3.62 6.23 6.79 5.13 1.60 4.07 4.45 Portugal 0.78 1.45 2.37 -0.01 -2.91 1.94 -1.25 Slovenia 4.01 5.85 6.87 3.59 -8.01 1.38 0.70 Spain 3.58 4.08 3.48 0.89 -3.83 -0.20 0.05 Sweden 3.16 4.30 3.31 -0.61 -5.03 6.56 2.93 Switzerland 2.69 3.75 3.85 2.16 -1.94 2.95 1.79 Turkey 8.40 6.89 4.67 0.66 -4.83 9.16 8.77 United Kingdom 3.23 2.76 3.43 -0.77 -5.17 1.66 1.12 All 2.73 3.55 3.41 0.22 -4.27 2.40 1.75 3.2. Methodology
out as Fernandes found no significant results for these, and tax rate was not included considering its expected correlation with profitability (Allayannis et al., 2003);
t i t i t i t i t i t i TobinQ y ofitabilit Size y Tangibilit Leverage , , 4 , 3 , 2 , 1 , P r . (1)
To the basic model I add all possible combinations of year, industry, and firm fixed effects controls, resulting in the models described by Eq. (2) to Eq. (6);
t i t t i t i t i t i t i D TobinQ y ofitabilit Size y Tangibilit Leverage , 5 , 4 , 3 , 2 , 1 , P r , (2) t i I t i t i t i t i t i D TobinQ y ofitabilit Size y Tangibilit Leverage , 5 , 4 , 3 , 2 , 1 , P r , (3) t i I t t i t i t i t i t i D D TobinQ y ofitabilit Size y Tangibilit Leverage , 6 5 , 4 , 3 , 2 , 1 , P r , (4) t i i t i t i t i t i t i D TobinQ y ofitabilit Size y Tangibilit Leverage , 5 , 4 , 3 , 2 , 1 , P r , (5) t i i t t i t i t i t i t i D D TobinQ y ofitabilit Size y Tangibilit Leverage , 6 5 , 4 , 3 , 2 , 1 , P r . (6)
Firm fixed effects and country dummies cannot be combined within one model, as it causes perfect multicollinearity. The same problem would occur when combining firm and industry fixed effects. For that reason you will not find results of such models in the tables.
Then I run these models with either GDP growth or country dummies added as country level determinant, as described by Eq. (7) and Eq. (8), to be able to test whether these increase the accuracy of the model in explaining firm leverage.
t i I t t i t i t i t i t i t i D D GDPGrowth TobinQ y ofitabilit Size y Tangibilit Leverage , 7 6 , 5 , 4 , 3 , 2 , 1 , P r . (7) t i I C t t i t i t i t i t i D D D TobinQ y ofitabilit Size y Tangibilit Leverage , 7 6 5 , 4 , 3 , 2 , 1 , P r . (8)
thus would make the results more difficult to interpret (De Jong et al., 2008). α is the constant. In case of the use of dummies in the model this constant also includes the effects for the dummy that was omitted from the regression. For my regressions I always omit the first dummy of the batch, which in the case of this study are the 1XX industry, the country Austria and the year 2005. Tangibility is calculated as total assets minus current assets divided by total assets, following Alayannis et al. (2003) and Booth et al. (2001). Size is calculated as the log of assets in USD. Profitability is calculated as return (EBITDA, Shuetrim et al. 1993) on assets. Growth opportunities (TobinQ) are calculated as the ratio of the market value of assets to book value of assets, also known as Tobin’s Q (Fernandes, 2011). The GDP Growth rate is taken from Worldbank data of world indicators. Firm (Di), industry (DI), year (Dt), and
country (DC) dummies are added to control for the respective fixed effects.
First, I will run OLS regressions for the full sample, including all years and firms, using the models described above. These regressions will show whether the firm determinants have the expected signs. Comparing the adjusted R-squared of the models with and without country-level determinants will show whether adding country proxies adds to the explanatory power of the models, answering hypothesis I.
Second, I will perform the OLS regressions separately per year, to be able to determine the explanatory power for each year and to check the differences between years. The differences in adjusted R-squared per year will show whether the explanatory power of each model increases in the crisis period or not.
Finally, I follow Fernandes (2011) in estimating basic regressions with firm-level determinants, country dummies, and industry dummies respectively as the only explanatory variables. The changes over time of the adjusted R-squared of these regressions will show which determinants have become more important over time, to answer hypothesis II.
4. Results
opportunities (Tobin’s Q) have lower leverage, consistent with the theory that excessive leverage will force firms to forego profitable investment opportunities. Especially in times of crisis this is to be expected.
Adding industry dummies (model 3 and 4) and controlling for firm fixed effects (model 5 and 6) increases the adjusted R-squared by a lot, up to 71.43%. The year dummies (model 2, 4 and 6) do however not add much to the adjusted R-squared, and are only significant in the firm fixed effects model (6).
Table 5 The determinants of leverage per model
Table 6 shows the model that includes controls for industry and year fixed effects (Eq. (4)), compared to the models extended with country-level proxies using GDP growth (Eq. (7)) and country dummies (Eq. (8)). GDP growth has a significant negative impact on leverage, in line with the expectations considering growth opportunities show the same sign, and that the majority of the sample firms are profitable. Hanousek and Shamshur (2011) showed that a dataset with profitable firms is more likely to show a negative coefficient for GDP growth.
Table 6 Firm and country determinants of leverage
Model (4) (7) (8) Constant -0.2020*** -0.1774*** -0.1293*** 5.46 4.81 3.34 Tangibility 0.2231*** 0.2275*** 0.2277*** 27.46 28.10 27.65 Size 0.0325*** 0.0321*** 0.0290*** 40.57 40.20 35.38 Profitability -0.1110*** -0.1069*** -0.1002*** 19.23 18.59 17.68 Tobin's Q -0.0198*** -0.0187*** -0.0163*** 14.36 13.65 11.94 GDP growth -0.0119*** 15.41 Year Dummies: 2006 -0.0039 0.006 -0.0036 0.70 1.07 0.65 2007 -0.0083 0.0003 -0.0059 1.51 0.05 1.10 2008 0.0084 -0.0202*** 0.0127** 1.52 3.46 2.33 2009 -0.0024 -0.0845*** 0.0033 0.42 10.94 0.60 2010 -0.0031 -0.0063 -0.0001 0.56 1.14 0.02 2011 -0.0047 -0.0153*** -0.0008 0.83 2.71 0.14 Dummies: Country x Industry x x x Observations 26316 26316 26316 Adjusted R-squared 0.1902 0.1975 0.2245
The results of the model are a bit better when GDP growth is added, but less than expected. This could be caused by the strong similarities with Tobin’s Q in terms of impact on leverage, and therefore it is possible that GDP growth acts as more of a replacement of Tobin’s Q explanatory power, causing the adjusted R-squared of the model not to go up as much.
The explanatory power of the model is much better when country dummies are used instead of GDP growth. The adjusted R-squared increases from 19.02% to 22.45% when adding the country dummies, which shows that hypothesis I is correct for these dummies; models containing country dummies do indeed explain leverage better than models not containing country-level determinants. Comparing the models without industry dummies (Eq. (2)), without year dummies (Eq. (3)), and without both (Eq. (1)) respectively, shows comparable results. As an extra robustness check, I also compare the three models over subsamples, running the regressions separately per year. The results are the same; the model with country dummies (Eq. (8)) shows the best explanatory power.
Adding GDP growth to the firm fixed effects model (Eq. (6)) increases the adjusted R-squared only slightly, from 71.58% to 71.65%. This confirms the earlier finding that the influence of GDP growth is limited in this sample. Firm fixed effects and country dummies cannot be combined in one model as it will cause multicollinearity.
Next I run the regressions separately for each year. Table 7 shows the results for the model including both industry and country dummies. Year dummies had to be dropped from model (8) as it obviously cannot be used in yearly regressions. Again all firm determinants are significant and with the expected signs. Only in 2008 is Tobin’s Q not significant, the year in which leverage was raised most in the sample. As companies get financially constrained they can no longer act on all investment opportunities and thus these have a smaller impact on the leverage of firms. The table also shows that from the start of the subprime mortgage crisis (2007), the explanatory power of the model is greater than before the crisis and remains higher for all years. This indicates that hypothesis II might be correct as well.
power on leverage during a crisis period. Therefore hypothesis II does seem to be correct; however there is not enough data to show proof that this increase is significant.
Table 7 Determinants of leverage per year
Year 2005 2006 2007 2008 2009 2010 2011 Constant -0.119 -0.157 -0.152 -0.207** -0.252** -0.040 -0.027 1.11 1.53 1.62 2.08 2.22 0.40 0.25 Tangibility 0.191*** 0.198*** 0.219*** 0.239*** 0.257*** 0.238*** 0.244*** 8.11 8.88 10.42 11.11 11.66 11.07 11.00 Size 0.028*** 0.029*** 0.033*** 0.033*** 0.032*** 0.024*** 0.026*** 11.92 13.58 15.77 15.50 14.33 11.05 11.69 Profitability -0.119*** -0.084*** -0.107*** -0.051*** -0.158*** -0.112*** -0.112*** 6.40 5.39 7.19 4.14 9.99 6.88 6.96 Tobin's Q -0.018*** -0.017*** -0.016*** -0.002 -0.017*** -0.025*** -0.016*** 4.90 5.54 5.17 0.44 4.21 6.80 3.91 Dummies: Country x x x x x x x Industry x x x x x x x Observations 3463 3750 3962 3980 3675 3830 3656 Adjusted R-squared 0.1918 0.2152 0.2262 0.2190 0.2348 0.2162 0.2273 Absolute values of the t-statistic are presented below the coefficients. ***, ** and * denote that a
coefficient is significant at the 1%, 5%, and 10% level respectively.
Figure 3 - Explanatory power of firm, country, and industry characteristics over time.
0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 2005 2006 2007 2008 2009 2010 2011 Year A dj R -s qu a re d
Firm Country Industry
5. Conclusion
This paper considers the impact of country determinants on leverage during a crisis period. GDP growth and country dummies are used as country proxies to determine their influence on the explanatory power of the models. To perform the analysis, a generally accepted model is used, which controls for tangibility, size, profitability and growth opportunities as firm level determinants. Fixed effects for firms, industries, and years are also controlled. The results show that adding country dummies to a model explaining leverage of firms significantly increases the adjusted R-squared of the model, in line with previous research. The results also show that the influence of country level determinants on leverage is higher during a crisis period.
Analysis in this paper shows that explanatory power of models including GDP growth as country-level determinant increases only slightly during a crisis period. Models including country dummies do show higher explanatory power during the crisis period. Finally, when using country dummies as the only explanatory variable of leverage, the adjusted R-squared becomes much higher during crisis years.
The result that GDP growth has hardly any impact could be explained by its correlation in explaining leverage with growth opportunities of a firm. Hanousek and Shamshur (2011) already used GDP growth as a proxy for a firm’s growth opportunities. The result that country dummies improve the explanatory power of the model was expected based on Booth et al. (2001) who indicated that country dummies explain 43% of total-debt variation. Gungoraydinoglu and Öztekin (2011) also found that country-level determinants explain one-third of the variation in capital structure.
Based on Fernandes (2011) I expected to find a higher explanatory power of country level determinants during a crisis period. Fernandes showed that for East Asian countries, the explanatory power of country factors decreased over time as their markets got better integrated with the world economy. I argued that during crisis periods markets work less optimally and thus the reverse effect could be expected. The analysis shows that the explanatory power of country-specific factors does increase during the crisis period; however there is no statistical test to proof its significance.
Investigating a sample containing more years could provide more accurate results. Furthermore, no correction was made for ADRs even though Fernandes (2011) found that ADRs have lower leverage ratios. It can be assumed that the control variable size corrects for this as firms with an ADR are usually bigger in size, but it might be insufficient. More importantly, the analysis did not provide robustness checks where specific countries were excluded, for example the PIGS countries, which could have an impact on the results as well.
For future research it might be interesting to look at a broader scope of country-level determinants for their effect on leverage and the effect of a crisis on these determinants. Inflation (Frank and Goyal, 2009; Gajurel, 2006), stock market development (Kayo and Kimura, 2011), bond market development and creditor right protection (De Jong et al., 2008) have been documented to have a significant impact on leverage and could be interesting to investigate in this regard.
6. Reference List
Allayannis, G., Brown, G., Klapper, L., 2003. Capital structure and financial risk: evidence from foreign debt use in East Asia. Journal of Finance 58, 2667-2709.
Almeida, H., Campello, M., Weisenbach, M., 2004. The cash flow sensitivity of cash. Journal of Finance 59, 1777-1804.
Bastos, D., Nakamura, W., Basso, L., 2009. Determinants of capital structure of publicly-traded companies in Latin America: the role of institutional and macroeconomic factors. Journal of International Finance and Economics 9, 24-39.
Bjork, C., House, J., Schaefer Muñoz, S., 2012. Spain pours billions into bank. The Wall
Street Journal, 25 May 2012. Available at:
http://online.wsj.com/news/articles/SB10001424052702304840904577425730120083816
Bokpin, G., 2010. Financial market development and corporate financing: evidence from emerging market economies. Journal of Economic Studies 37, 96-116.
Bradley, M., Jarrell, G., Han Kim, E., 1984. On the existence of an optimal capital structure: theory and evidence. Journal of Finance 39 (3), 857-878.
Brav, O., 2009. Access to capital, capital structure, and the funding of the firm. Journal of Finance 64 (1), 263-308.
Bris, A., Koskinen, Y., Pons, V., 2004. Corporate financial policies and performance around currency crises. Journal of Business 77 (4), 749-795.
Campello, M., Graham, J., Harvey, C., 2010. The real effects of financial constraints: evidence from a financial crisis. Journal of Financial Economics 97, 470-487.
Dang, V., Kim, M., Shin, Y., 2014. Asymmetric adjustment toward optimal capital structure: evidence from a crisis. International Review of Financial Analysis 33, 226-242.
De Jong, A., Kabir, R., Nguyen, T., 2008. Capital structure around the world: the roles of firm- and country-specific determinants. Journal of Banking & Finance 32, 1954-1969.
Dell’Ariccia, G., Detragiache, E., Rajan, E., 2008. The real effect of banking crises. Journal of Financial Intermediation 17, 89-112.
Dewally, M., Shao, Y., 2014. Liquidity crisis, relationship lending and corporate finance. Journal of Banking & Finance 39, 223-239.
Duchin, R., Ozbas, O., Sensoy, B., 2010. Costly external finance, corporate investment, and the subprime mortgage credit crisis. Journal of Financial Economics 97, 418-435.
Eckbo, B., Giammarino, R., Heinkel, R., 1990. Asymmetric information and the medium of exchange in takeovers: theory and tests. Review of Financial Studies 3, 651-675.
Fischer, E., Heinkel, R., Zechner, J., 1989. Dynamic capital structure choice: theory and tests. Journal of Finance 44, 19-40.
Forelle, C., Steinhauser, G., 2012. Latest Europe rescue aims to prop up Spain. Wall Street
Journal, 11 June 2012. Available at:
http://online.wsj.com/news/articles/SB10001424052702303768104577458562351966868
Frank, M., Goyal, V., 2008. Trade-off and pecking order theories of debt. In: Eckbo, B. (Ed.), Handbook of Corporate Finance: Empirical Corporate Finance, Vol. 2. North Holland, Amsterdam, pp. 135-202.
Frank, M., Goyal, V., 2009. Capital structure decisions: which factors are reliably important? Financial Management 38, 1–37.
Gajurel, D., 2006. Macroeconomic influences on corporate capital structure. Unpublished working paper. University of Tasmania, Hobart, Australia.
Gentier, A., 2012. Spanish banks and the housing crisis: worse than the subprime crisis? International Journal of Business 17 (4), 342-351.
Greenlaw, D., Hatzius, J., Kashyap, A., Shin, H., 2008. Leveraged losses: lessons from the mortgage market meltdown. U.S. Monetary Policy Forum Report No. 2.
Gungoraydinoglu, A., Öztekin, Ö., 2011. Firm- and country-level determinants of corporate leverage: some new international evidence. Journal of Corporate Finance 17, 1457-1474.
Halov, N., Heider, F., 2011. Capital structure, risk and asymmetric information. Quarterly Journal of Finance 1 (4), 767-809.
Hanousek, J., Shamshur, A., 2011. A stubborn persistence: is the stability of leverage ratios determined by the stability of the economy? Journal of Corporate Finance 17, 1360-1376.
Harris, M., Raviv, A., 1991. The theory of capital structure. Journal of Finance 46 (1), 297-355.
Haugen, R., Senbet, L., 1978. The insignificance of bankruptcy costs to the theory of optimal capital structure. Journal of Finance 33 (2), 383-393.
Hennessy, C., Whited, T., 2005. Debt dynamics. Journal of Finance 60 (3), 1129-1165.
Huang, R., Ritter, J., 2009. Testing the market timing theory of capital structure. Unpublished working paper. University of Florida, Gainesville.
Jensen, M., Meckling, W., 1976. Theory of the firm: managerial behaviour, agency costs, and ownership structure. Journal of Financial Economics 3, 305-360.
Judge, A., Korzhenitskaya, A., 2012. Credit market conditions and the impact of access to the public debt market on corporate leverage. International Review of Financial Analysis 25, 28-63.
Kayo, E., Kimura, H., 2011. Hierarchical determinants of capital structure. Journal of Banking & Finance 35, 358-371.
Kraus, A., Litzenberger, R., 1973. A state-preference model of optimal financial leverage. Journal of Finance 33, 911-922.
Lemke, T., Lins, G., Picard, M., 2013. Mortgage-backed securities. Thomson West, Eagan, Minnesota, United States.
Lewellen, J., Lewellen, K., 2004. Taxes and financing decisions. AFA 2005 Philadelphia Meetings Paper. Dartmouth College, Hanover, United States.
Modigliani, F., Miller, M., 1958. The cost of capital, corporation finance, and the theory of investment. American Economic Review 48, 261-297.
Modigliani, F., Miller, M., 1963. Corporate income taxes and the cost of capital: a correction. American Economic Review 53, 433-443.
Myers, S., 1984. The capital structure puzzle. Journal of Finance 39, 575-592.
Rajan, R., Zingales, L., 1995. What do we know about capital structure? Some evidence from international data. Journal of Finance 50 (5), 1421-1460.
Ross, S., 1977. The determination of financial structure: the incentive-signalling approach. Bell Journal of Economics 8, 23-40.
Shambaugh, J., Reis, R., Rey, H., 2012. The euro’s three crises. Brooking Papers on Economic Activity Spring 2012, 157-231.
Shuetrim, G., Lowe, P., Morling, S., 1993. The determinants of corporate leverage: a panel data analysis. Research Discussion Paper 9313, Reserve Bank of Australia.
Simkovic, M., 2011. Competition and crisis in mortgage securitization. Indiana Law Journal 88, 213-271.
Stiglitz, J., 1973. Taxation, corporate financial policy and the cost of capital. Journal of Public Economics 2, 1-34.
Titman, S., Wessels, R., 1988. The determinants of capital structure choice. Journal of Finance 43 (1), 1-19.
