The new introduced leverage ratio could have
prevented the crisis
Tijs Groot 6053025
Economics and Business, Finance Track July 2015
Tanju Yorulmazer
Abstract
This paper provides new insights regarding the leverage ratio which is introduced in the Basel III Accord. The level of the leverage ratio is measured with respect to the performance of banks. Furthermore, the influence of the leverage ratio on stock return
is analyzed. The methodologies used are the mean comparison t-test and the ordinary least squares regression. This paper shows that a 5% leverage ratio is optimal. Furthermore, this paper does not find a clear relation between stock return and the
leverage ratio. These findings are empirical evidence in favor of the leverage ratio that is introduced in Basel III, these findings also provide a critic note to Basel II.
2
Statement of Originality
This document is written by Student Tijs Groot who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is
original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the
3
1. Introduction
In 2005 the Basel II Accord was introduced which was the follow up to the Basel I Accord. The Basel Accord is put in place to regulate and supervise the banking system, with the goal to create a more stable financial system. Apparently the Basel Accord has not fully succeed. Because in the last financial crisis banks needed the support of the governments to survive or even worse, they went bankrupt. Though all these banks seemed healthy before the crisis, as the banks met the regulatory requirements of Basel II Accord but actually the banks avoided these requirements. They did this by holding excessive amounts of off-balance sheet items, as these items do not fall under the capital requirements of Basel II (Barell et al., 2010). The increase in off-balance sheet debt led to an increase in risk and was eventually an important factor of the failing of the market. The last financial crisis disclosed that the Basel II framework had shortcomings. Therefore, recently a new framework was introduced, Basel III. The most important changes in the framework compared to Basel II are the following:
I. Introduction of the leverage ratio;
II. Introduction of the liquidity coverage ratio (LCR); III. Introduction of the net stable funding ratio (NSFR); IV. Raising the minimum capital requirement;
V. Introduction of the counter-cyclical capital buffer.
In this paper the focus is on the leverage ratio, which is the capital exposure divided by the total exposure. The capital measure is the regulatory tier 1 capital. The exposure measure consist of total assets, derivatives, financial securities and other off-balance sheet items. For a more detailed explanation of the construction of the leverage ratio I would like to refer to the appendix A. Appendix A presents all fields that ABN AMRO Bank uses to calculate the leverage ratio. The leverage ratio is mathematically stated below.
Leverage ratio = Capital Measure Exposure Measure
The motivation behind the introduction of the leverage ratio is to restrict the build-up of excessive leverage in the banking sector because this avoids the destabilizing deleveraging processes that can damage the financial system (BIS, 2014). The leverage ratio is interesting to analyze, because high leverage was one of the main causes for the recent financial crisis.
4 Furthermore, the leverage ratio can be a simple solution for the shortcomings of the complicated capital ratio. This is explained in more depth in the literature review. At last, already before the crisis multiple papers suggested a leverage ratio to be introduced in the Basel II Accord. Therefore, the aim of this paper is to answer the following research question.
Research Question: Could the leverage ratio have prevented the crisis?
This paper takes a closer look into the period of the financial crisis and what if the leverage ratio was introduced before the crisis. It not possible to exactly predict what would have happened when a leverage ratio was introduced before the crisis. But the intention of this paper is to find out if the leverage ratio would have been a good indicator for the performance of banks. A positive relation between underperforming and the leverage ratio indicates that the financial crisis could have been prevented by an earlier introduction of the leverage ratio. But when there is no or a negative relation between the leverage ratio and the performance of banks, the introduction of the leverage ratio in Basel III is doubtful. To answer the research question the following hypotheses are tested.
H1: The leverage ratio is a better indicator for banking performance than the capital ratio.
H2: The minimum required leverage ratio (3%) in the Basel III Accord should increase to create more financial stability.
H3: The leverage ratio has a negative relation with the return of banks.
The first hypothesis tries to analyze if the leverage ratio is a better and more accurate indicator for the performance of banks than the capital ratio. The second hypothesis tries to analyze if the minimum required leverage ratio should be increased to create more financial stability. The third hypothesis analyzes the effect that the leverage ratio has on the return of banks.
Data is collected from the database Bureau van Dijk (BvD) for the period of 2005 to 2013. The period of 8 years is chosen so enough data is available before, during and after the crisis. The dataset of Bankscope provides financial information for on- and off-balance sheet financials, what is needed for the calculation of the leverage ratio. Banks from Europe and the
5 United States of America are chosen to capture the main countries that are regulated by the Basel Accord III regulation. Furthermore, the database Compustat is used for data on the return of banks and the status (active/inactive) indicator of the banks.
The rest of the paper is organized as follows. In Section 2 a recapitalization of existing findings of empirical papers is presented. Section 3 explains the methodology. This is followed by the collection of data and descriptive statics in section 4. In section 5 the results are disclosed. Section 6 describes the robustness check. In section 7 the conclusion and the discussion is described. At the end of the paper the reference list and the Appendix are presented.
2. Related literature
In this section the findings of previous empirical papers are presented. To give a clear outlay of the existing information this section is split up into three subsections. The first subsection describes the evolvement of the Basel requirements over time. The following subsection describes the intention behind the introduction of the leverage ratio and the advantages. The last section describes the disadvantages of the leverage ratio.
2.1 The evolvement of the Basel requirements
As a result of the collapse of two large international banks, Franklin National Bank in the US and the German Bankhaus Herstatt in 1974, the stability of the international banking system became an official concern to the public. The collapse showed the vulnerability of the interdependent international banking system to the failure of one of its members (Spero, 1980). This vulnerability led to the introduction of the Basel committee on Banking Supervision. The Basel committee was assigned to regulate the cooperation between its member countries on banking supervisory matters (Basel, 2013). Starting members of this committee (who first were called committee on banking regulations and supervisory practices) are the G10 countries1. In 1988 the first step towards a more stable financial system was set by the introduction of Basel I. The Basel I framework was rested on a common standard of risk assessment, and was primarily focused on credit risk.
1The United States and Canada from North-America, Japan from Asia and Belgium, Netherlands, France, Germany, Italy, Sweden, United Kingdom and in
6 The model was named the Basel capital ratio and is mathematically presented below.
𝐵𝑎𝑠𝑒𝑙 𝐼 𝑐𝑎𝑝𝑖𝑡𝑎𝑙 𝑟𝑎𝑡𝑖𝑜 = 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 𝑅𝑖𝑠𝑘 − 𝑤𝑒𝑖𝑔ℎ𝑡𝑒𝑑 𝑎𝑠𝑠𝑒𝑡𝑠
The capital of the Basel I capital ratio consist of tier 1 and tier 2 capital. The risk-weighted assets are the assets multiplied by their risk weight. Under the Basel I accord, banks were required to hold more than 8% capital and at least 4% of tier 1 capital. Tier 1 capital is the core capital, and tier 2 capital is the supplementary capital. In table 2.1 the exact definition of tier 1 and tier 2 capital are displayed (BIS, 2014).
Table 2.1 Definition of capital
Tier 1 – Paid-up capital
– Disclosed reserves (retained profits, legal reserves, etc.) Tier 2 – Undisclosed reserves
– Hybrid instruments (must be unsecured, fully paid-up) – Subordinated debt (max. 50% tier 1, min. 5 years – discount factor for shorter maturities)
Deductions – Goodwill (from Tier 1)
– Investments in unconsolidated subsidiaries (from tier 1 and tier 2) Laurent Balthazar (2006)
The risk-weighted assets (RWA) are a bank’s assets weighted according to the risk they carry. The assets of banks were classified in five categories according to their credit risk. The five categories of Basel I are defined in table 2.2.
7
Table 2.2 Risk-weight of assets
% Item
0 –Cash
–Claims on OECD central governments
–Claims on other central governments if they are denominated and funded in the national currency (to avoid country transfer risk) 20 –Claims on OECD banks and multilateral development banks
–Claims on banks outside OECD with residual maturity <1 year –Claims on public sector entities (PSE) of OECD countries
50 –Mortgage loans
100 –All other claims: claims on corporate, claims on banks outside OECD with a maturity >1 year, fixed assets, all other assets. Laurent Balthazar (2006)
With the introduction of this capital ratio the Basel committee was convinced that a measure was in place that strengthen the stability of the international banking system by reducing and monitoring the risk-taking of banks. Other goals were to increase capital adequacy and to decrease the competitive inequalities between countries (BIS, 2014). Because the cost of equity is generally much greater than the cost of debt, due to tax considerations, asymmetric information, bank safety net, and agency costs (Jones, 2000). Banks may view regulatory capital standards as a form of regulatory taxation because it requires banks to maintain higher equity than banks prefer (Donahoo and Shaffer, 1991).
During the active period of Basel I, the capital ratio showed shortcomings. The main shortcoming was that there was too little differentiation in the risk categories. This resulted that banks chose a higher risk investment within the same percentage category, which led to an increase in the vulnerability of the bank. This was not in line with the main goal of the Basel I accord. Other shortcomings are the following: no adjustments related to the changing nature of default risk, no consideration related to the maturity of risk and the lack of recognition of portfolio diversification (Balthazar, 2006). Furthermore, in the late 90s there was a huge technical development. Due to this technical development banks expanded their business to a new level. Because of the shortcomings of Basel I and the technical development, the Basel requirements needed to develop. This led to the introduction of the Basel II accord in 2004. The main objective of the introduction of Basel II was to update the existing capital requirement framework and to make it more comprehensive and risk sensitive. While the focus of Basel I was on credit risk only, Basel II also took into account market risk and operational risk (Von
8 Thadden, 2004). The Basel II framework is comprised of three pillars. Pillar I is the Solvency ratio, this describes the minimum capital requirements that banks are required to meet. Pillar II is the supervisory review and internal assessment, this requires banks to have their own internal processes to assess their overall capital adequacy in relation to their own risk profile. Pillar III is the market discipline, this requires banks to disclose information on their risk exposure, capital ratios and risk management.
The Basel II capital ratio is mathematically presented below.
𝐵𝑎𝑠𝑒𝑙 𝐼𝐼 𝑐𝑎𝑝𝑖𝑡𝑎𝑙 𝑟𝑎𝑡𝑖𝑜 = 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 𝑀𝑒𝑎𝑠𝑢𝑟𝑒
𝑅𝑖𝑠𝑘 − 𝑤𝑒𝑖𝑔ℎ𝑡𝑒𝑑 𝐴𝑠𝑠𝑒𝑡𝑠 + 𝑀𝑎𝑟𝑘𝑒𝑡 𝑟𝑖𝑠𝑘 + 𝑂𝑝𝑒𝑟𝑎𝑡𝑖𝑜𝑛𝑎𝑙 𝑟𝑖𝑠𝑘
Capital= tier 1 and tier 2 capital
Risk-weighted assets = Assets multiplied by their risk weight Market risk = the risk of adverse price movements.
Operational risk = the risk of loss resulting from inadequate internal processes or external effects.
In contrast to the Basel I requirements, the capital ratio extended with more risk categories and the total assets are based upon quality rather than on type (Balthazar, 2006). Banks are allowed to assess risk using external credit assessments, but they can also use their own internal system for rating credit risk (Basel Committee, 2004). The most common used systems are the standardized approach (STD) and internal rating based approach (IRB). The standardized approach for calculating the risk weights is based on external credit rating assessment. The risk weights for individual credit-risk assets can be found in Table 2.3. The reason for the introduction of the IRB approach is that in general banks are better in forecasting their risk compared to a standardized universal rule. The IRB approach is based on a bank’s own internal estimates of the creditworthiness of an asset.
Despite of the new Basel II accord, the financial crisis occurred. Banks needed the support of the government to survive or even worse, they went bankrupt. An underlying cause of the financial crisis was the build-up of excessive on- and off-balance sheet leverage in the banking system. This happened because leverage is said to be pro-cyclical (amplification of the effects of the business cycle). The reason is that banks actively manage their leverage ratio during the business cycle. In periods of economic boom asset prices increase rapidly and the return on investments are high. In order to maintain the same leverage ratio a bank should increase their debt position. However, this increase in debt also leads to an increase in the demand for assets, which further increases the price of assets. In order to maintain their leverage
9 ratio the banks have to increase their debt even further. This phenomenon is caught by the leverage cycle of Adrian and Shin (2010). A numerical example is provided in Appendix B. The opposite is true in economic downturn, because banks need to sell their assets (at fire sale prices) to maintain a stable leverage ratio. The large price drops will lead to losses on their initial position and result in higher margins. This leads to funding problems, and investors have to reduce their positions (Brunnermeier, 2009). The high reduction in positions causes an even bigger price drop, which implies the loss spiral/margin spiral that is presented in Figure 2.1.
Figure 2.1 Loss/Margin spiral (Brunnermeier, 2009)
The main shortcoming of Basel II was that banks could increase their leverage while maintaining strong risk-based capital ratios. They did this by moving credit risk off their balance sheet, aiming to increase their volume of operations without the need to put aside the capital coefficients that were required by Basel II. This can be done by acquiring protection against credit risk on the derivatives markets, by issuing credit securities whose return is dependent on the amortization paid by borrowers and by creating special investment vehicles. However, this was only possible because others (shadow banks) were willing to take on the risk (Brunnermeier, 2009). A second shortcoming of the Basel II accord is the pro-cyclical behavior of the capital ratio. The capital of banks will increase in economic booms, while the 8% threshold does not increase. Therefore, banks can increase their risk-taking. This becomes problematic if the economy is moving from an economic boom into a recession.
In a reaction to the financial crisis, Basel III was introduced in 2009. Basel III continues to build on the “three Pillars” of Basel II, but adjustments are made to improve the resilience of the banking sector. Pillar I is changed due to the introduction of a Net Stable Funding Requirements (NSFR), liquidity coverage ratio (LCR) and the leverage ratio. These changes
10 strengthen the regulation, supervision and risk management of the financial system. Other changes are the increase of the minimum capital requirement and the introduction of a counter-cyclical capital buffer. Pillar II is enhanced with some supplemental requirements for risk management and supervision. At last, Pillar III has been improved with a revised set of disclosure requirements for the market discipline (BIS, 2014).
2.2 The intention and the advantages of the introduction of the leverage ratio
The leverage ratio is an independent and simple to use requirement that supplements the risk-based capital requirements. The leverage ratio is the capital exposure divided by the total exposure. The capital measure only consists of regulatory tier 1 capital. For the exposure measure Basel looks at on-balance sheet exposures, exposure of derivatives, securities financing transaction exposures, and other off-balance exposure. The minimum leverage ratio is currently set at a 3% level, but in the Basel III accord is stated that this will increase to 4% in 2018 (BIS, 2014). The leverage ratio is mathematically stated below.
Leverage ratio = 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 𝑀𝑒𝑎𝑠𝑢𝑟𝑒 𝐸𝑥𝑝𝑜𝑠𝑢𝑟𝑒 𝑀𝑒𝑎𝑠𝑢𝑟𝑒
The advantages of the leverage ratio include the following. First of all, the leverage ratio is simple to apply and monitor. Therefore, it can be added quickly without high costs or requirements for expertise. Secondly, it is intended to constrain excess leverage build-up in the banking sector. This is to avoid the loss/margin spiral in periods of economic downturn (BIS, 2014). By setting a floor on the leverage ratio, it will limit the excessive build-up of leverage and reduce pro-cyclicality. Another advantage is that the leverage ratio is taking off-balance sheet items into account. Therefore, banks movement of on-balance sheet items to off-balance sheet items is limited by the leverage ratio. And last, according to Laeven and Valencia (2010) the leverage ratio is a better predictor of the failing of banks than the capital ratio. This is presented in figure 2.2 and 2.3.
11
Fig. 2.2 Risk-based capital ratios of major Fig. 2.3 Leverage ratios of major global banks, global banks, end-2006 end-2006
Although there is not a perfect relationship between the leverage ratio and failed banks, it appears that the failed banks are more centered to the lower bound of the graph. Therefore, the following hypothesis is conducted.
H1: The leverage ratio is a better indicator for banking performance than the capital ratio.
2.3 The disadvantages of the introduction of the leverage ratio
In this section the disadvantages of the leverage ratio are discussed. The first disadvantage is that the minimum required level of the leverage ratio (3%) is considered too low. Admati et al. (2011) argue for a higher leverage ratio, up to 20%. This reduces risk-taking by banks, because banks have a larger part of their own interest involved. Also the Federal Reserve (Fed)2 thinks 3% leverage ratio is too low (Fed, 2014). The Fed introduced a leverage ratio of 6% for the SIFIs (Systematically Important Financial Institutions) and a 5% leverage ratio for other bank holding companies. Notice that there are some small differences between the calculation of the leverage ratio for the U.S. compared to the leverage ratio in the Basel regulation. Even though in absolute value these differences don’t explain the difference within the level of the two ratios (Fed, 2014). Also within the Basel III accord the leverage ratio will be increased to 4% in 2018. Haldane (2012) of the Bank of England also argues in favor of a higher leverage ratio. In his research he discloses that during the crisis the leverage ratio should have been above 7% for
12 the largest banks to guard them against failing. Furthermore, Haldane argues that the leverage ratio in general needs to be around 4% to make a distinction between the performances of banks. The difference between the threshold set for the leverage ratio by the Basel committee and the threshold suggested by the Fed/Haldane results in the following hypothesis.
H2: The minimum required leverage ratio (3%) in the Basel III Accord should increase to create more financial stability
Blum (2007) explains a third downside of the leverage ratio. He argues that also safe banks (banks without any risky investments) are restricted by the leverage ratio. If the leverage ratio is binding for those safe banks, there are extra costs for those safe banks to increase their equity. In addition Groen (2014) argues that in Europe 34 banks have to increase their core capital (tier 1) with an amount of 29 billion. Compared to the amount of assets hold by those banks (2.430 billion), it seems modest. But considering that the aggregated net income of 2014 was only 18 billion it has great impact. The high increase in the amount of Tier 1 capital is also the main argument of the major banks against the leverage ratio. Furthermore, they argue that the high leverage ratio is negatively related to their profit. This argument is in line with the framework of Adrian and Shin (2010) because it is profitable to have high leverage in times of economic growth. This is the reason why this paper conducted the following hypothesis.
H3: The leverage ratio has a negative relation with the return of a bank.
The final disadvantage of the leverage ratio is discussed in World Bank (2012). They argue that there was already a leverage ratio in place in the U.S during the crisis. Therefore, the introduction of a leverage ratio would not have any effect. However, the leverage ratio that was in place was only based on on-balance sheet items. The off-balance sheet items were not taken into account. Because off-balance sheet debt was a main cause of the financial crisis, this leverage ratio was not a good predictor.
13
3. Methodology
To provide an answer to the research question “Could the leverage ratio have prevented the crisis? three hypotheses are formed based on the related literature. The hypotheses are formulated below.
H1: The leverage ratio is a better indicator for banking performance than the capital ratio
H2: The minimum required leverage ratio (3%) in the Basel III Accord should increase to create more financial stability.
H3: Leverage ratio has a negative relation with the return of banks.
To answer the first hypothesis the following variables are of interest: status and leverage ratio. The variable status is a dummy variable that indicates the performance of banks. The dummy variable is equal to 0 when the bank failed and equal to 1 when the bank survived. The leverage ratio is calculated by dividing the regulatory tier 1 capital by the total assets and the off-balance sheet items.
First a comparison is made between the leverage ratio and the capital ratio with respect to the performance of banks. This is in line with the research of Leaven and Valencia (2010). Furthermore, the means of the leverage ratio and capital ratio of failed and survived banks are compared. At last, a distinction is made between American and European banks because both nations proposed a different leverage ratio. The method that is used for the analysis is a mean comparison t-test. This is mathematically formulated on the next page.
14 Formula (1): 𝑡 = 𝑥̅1− 𝑥̅2 √𝑠12 𝑛1+ 𝑠22 𝑛2 Where:
𝑥̅1= The mean of the failed banks
𝑥̅2= The mean of the surviving banks 𝑠12
𝑛1 = Standard deviations of the failed banks means divided by the number of failed banks 𝑠22
𝑛2 = Standard deviations of the surviving banks means divided by the number of surviving banks
t = Value indicating if the difference is significant. Absolute t-value higher than 1.645 indicates a 10% significance. Absolute t-value higher than 1.96 indicates a 5% significance. Absolute t-value higher than 2.326 indicates a 1% significance.
With hypothesis 2 I analyze what the optimal leverage ratio needs to be to create more financial stability. The method used to answer hypothesis 2 is a mean comparison t-test. The t-test compares the performance of banks with respect to the leverage ratio. The banks are grouped by leverage ratio as follows:
Group 1: Banks with a leverage ratio smaller than 3%
Group 2: Banks with a leverage ratio in between 3% and 4%
Group 3: Banks with a leverage ratio in between 4% and 5%
Group 4: Banks with a leverage ratio in between 5% and 6%
Group 5: Banks with a leverage ratio in between 6% and 7%
Group 6: Banks with a leverage ratio higher than 7%
The main focus is on the leverage ratios of 3%, 4%, 5% and 6%. These percentages are chosen in line with the level of the leverage ratio that will be introduced in Europe (2018) and the U.S. Therefore the following null hypothesis are tested:
1.
H0: Performance group 1 = Performance group2
H1: Performance group 1 ≠ Performance group2 2.
H0: Performance group 2 = Performance group3
15 3.
H0: Performance group 3 = Performance group 4
H1: Performance group 3 ≠ Performance group 4
By comparing the different groups it is possible to determine to what extend an increase in the leverage ratio causes an increase in the performance of banks. For example, if the H0 of section 2 is rejected, an increase of the leverage ratio from 4% to 5% results in an increase in the performance of banks. This shows that an increase of the leverage ratio from 4 to 5% is beneficial.
The method chosen to answer hypothesis 3 is the Ordinary Least Squares (OLS) regression model. This is in line with the method that is used in the paper of Kolk (2012). The Ordinary Least Squares (OLS) regression model is presented below. If the 𝛽1 coefficient is significantly higher than zero this will indicate that the leverage ratio has a positive impact on the performance of banks.
Regression (1):
𝑅𝐼 = 𝛼 + 𝛽1𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒 𝑟𝑎𝑡𝑖𝑜 + 𝛽2(𝑀𝐾𝑇 − 𝑅𝑓) + 𝛽3𝑆𝑀𝐵 + 𝛽4𝐻𝑀𝐿 + 𝛽5𝑀𝑂𝑀 + 𝛽6𝑃𝑟𝑒𝐶𝑟𝑖𝑠𝑖𝑠 + 𝛽7𝑃𝑜𝑠𝑡𝐶𝑟𝑖𝑠𝑖𝑠 + 𝛽8𝑃𝑟𝑒𝐶𝑟𝑖𝑠𝑖𝑠 ∗ 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒 𝑟𝑎𝑡𝑖𝑜
+ 𝛽9𝑃𝑜𝑠𝑡𝐶𝑟𝑖𝑠𝑖𝑠 ∗ 𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒 𝑟𝑎𝑡𝑖𝑜 + 𝜀
Where:
RI = Excessive return. The cumulative return index minus the risk free rate.
Leverage Ratio = Regulatory tier 1 capital divided by total assets and off-balance sheet items (%). MKT-RF= Fama French risk measure.
SMB = Fama French small market cap premium. HML = Fama French high book-to-market premium. MOM = Carhart momentum factor.
Pre-crisis = Dummy variable indicating 1 if the period is 2005-2007 and 0 otherwise. Post-Crisis = Dummy variable indicating 1 if the period is 2007-2013 and 0 otherwise.
Pre-crisis*leverage ratio = Interaction dummy variable between leverage ratio the period is 2005-2007. Post-crisis*leverage ratio = Interaction dummy variable between leverage ratio the period is 2009-2013.
In addition to the regression the option robust is added to correct for heteroskedasticity. As shown in the regression the factors of the Fama and French model are added. The Fama French model is an extension of the CAPM model. In the CAPM model only risk is an explaining variable of return. Fama and French (1993, 1995) extended the CAPM model with their
16 explanatory variables SMB (Small Minus Big), HML (High Minus Low) and Carhart’s (1997) MOM (momentum) variable. The risk is measured by the variable (MKT-RF). This variable measures the volatility of a stock in line with movements in the market. The SMB variable measures the level of higher excessive return that small market capitalized firms have over big market capitalized firms. The HML variable measures the level of higher excessive returns that high book-to-market companies have over low book-to-market companies. The MOM factor measures the advantage that stocks with a good past performance have over stocks with a bad past performance. These four variables have explanatory power regarding the return of a stock. For this reason the Fama French model is chosen combined with Carhart’s Momentum factor. By adding these variables to the regression a more realistic image is given of the relationship between return and our independent variable (leverage ratio).
The period of interest is split up in pre-crisis (2005-2007), crisis (2007-2009), post-crisis (2009-2013) and the total period (2005-2013). The periods pre-crisis and post-crisis are added as dummy variables to the regression (1). The periods are added because the related literature suggest that leverage is pro-cyclical, by performing regressions in different economic periods the influence of economic state can be determined. The period crisis is omitted from the regression to prevent omitted variable bias. This paper realizes that the crisis period for Europe took longer than only 2007-2009 but for banks the years 2007-2009 where most severe (Antoniades, 2015). Also added to the regression are the interaction variables of the specific periods with the leverage ratio. These interaction variables show if the influence of the leverage ratio is different between economic time periods. Furthermore, a distinction is made between U.S. and European banks, due to the different implementation in the level of the leverage ratio it is interesting to show if there are any differences between Europe and the U.S.
4. Data and descriptive statistics
This part presents the collection and description of the data used in this research. First the data collection is described, this is followed by the descriptive statistics.
17
4.1 Data collection
This research started with a dataset of Bankscope (Bureau van Dijk) of 32.405 banks all over the world. As mentioned earlier this research focuses only on Europe3 and the U.S. For the calculation of the leverage ratio the following variables need to be in the dataset: total assets, off-balance sheet items and total regulatory tier1 capital. First, all variables containing values in foreign currencies are converted into euro`s. The conversion is done manually using the sources OANDA or X-rates. The exchange rates that are used are the averages for the year. Furthermore, if a bank is missing data regarding one of the variables (total assets, off-balance sheet items and total regulatory tier1 capital) or the total assets are less than 10 billion, the bank is excluded from the dataset. Banks with total assets less than 10 billion are excluded because their influence is limited. After these corrections there are 2768 banks left, this includes banks from outside Europe or the U.S. but are active within these regions. Note that this process is separately done for every year, because only then the survivorship bias can be ruled out. The period of interest is from 2005-2013.
The data of the return of the banks is collected from Datastream. The return index (RI) shows the growth in value of a share that is held for a specific period of time. This under the assumption that dividends are reinvested to purchase additional units of the mutual fund at the closing bid price. The return index is mathematically formulated as follows:
Formula (2): 𝑅𝐼𝑡 = 𝑅𝐼𝑡−1∗ 𝑃𝑡+ 𝐷𝑡 𝑃𝑡−1 𝑃𝑡= 𝑐𝑙𝑜𝑠𝑖𝑛𝑔 𝑝𝑟𝑖𝑐𝑒 𝑃𝑡−1= 𝑐𝑙𝑜𝑠𝑖𝑛𝑔 𝑝𝑟𝑖𝑐𝑒 1 𝑚𝑜𝑛𝑡ℎ 𝑝𝑟𝑖𝑜𝑟 𝑡𝑜 𝑚𝑜𝑛𝑡ℎ 𝑡 𝐷𝑡= 𝐷𝑖𝑣𝑖𝑑𝑖𝑑𝑒𝑛𝑑𝑠 𝑟𝑒𝑖𝑛𝑣𝑒𝑠𝑡𝑒𝑑 𝑖𝑛 𝑚𝑜𝑛𝑡ℎ 𝑡 𝑅𝐼𝑡= 𝑅𝑒𝑡𝑢𝑟𝑛 𝑖𝑛 𝑚𝑜𝑛𝑡ℎ 𝑡 𝑅𝐼𝑡−1= 𝑅𝑒𝑡𝑢𝑟𝑛 1 𝑚𝑜𝑛𝑡ℎ 𝑝𝑟𝑖𝑜𝑟 𝑡𝑜 𝑚𝑜𝑛𝑡ℎ 𝑡
3 The countries included are: Belgium Poland Italy Sweden Austria Luxembourg Bulgaria Finland United Kingdom France Belarus Romania Russian Federation Ireland Netherlands Norway Bosnia And Herzegovina Liechtenstein Monaco Denmark Turkey Malta Spain Germany Switzerland Hungary Iceland Republic Of Moldova Slovakia Croatia Czech Republic Ukraine Serbia Latvia Portugal Estonia Slovenia Cyprus Greece Kosovo Montenegro Albania Lithuania Macedonia (Fyrom)
18 After collecting the values of the return index for the banks, the cumulative return index is calculated in the following way:
Formula (3):
𝑅𝐼𝑖,𝑡 = 𝑙𝑛 ( 𝑅𝐼𝑖,𝑡
𝑅𝐼𝑖,𝑡−1) + 1
The date for the return is the first day of the month, this is in line with the date of the Fama and French monthly data set. All banks that are listed in the U.S or Europe are collected. The data collection for Europe is done for every country individual. For this reason only data is collected for banks who are listed to the country-specific index. All banks that do not have data available for the variable return are deleted from the data set. The second variable of interest that is collected from Datastream is the variable Status. This variable contains information if a bank is active or inactive. The status inactive can be caused by various reasons: it can be due to bankruptcy, a merger, an acquisition or a change in name. For this reason a comparison is made with the variable Bank History that is collected from Bankscope. This variable gives a description of the history of the bank. If this variable contains information about a change in name during the period in interest, the bank is removed from the inactive list. The banks that are listed inactive due to a merger or acquisition are analyzed to determine if the merger or acquisition originates from a positive or negative event. An event is called negative when a bank was making a loss or was negative in news during the merger or acquisition, vice versa for a positive event. If the merger or acquisition comes from a positive event the bank is listed active and from a negative event the bank is listed inactive. Furthermore, a comparison is made with the list of failed banks provided by FDIC. The banks who are listed in more than 1 dataset are given the status inactive. After these corrections all banks that are listed as inactive are failed banks, all banks that are listed as active are surviving banks.
By the variables ISIN (identification code) and Name, which are in both datasets (Datastream, Bankscope), the two data sets are merged. After the merging 337 banks remain, which all have sufficient data. The decline in data is mainly caused by two actions. The first action is the limit of €10 bln of asset, the second is that the banks needs to be listed on a U.S or European index.
Furthermore, the Fama French factor variables are added into this data set. These variables are collected from the website of Fama and French. On their website also the
19 momentum factor of Carhart is collected. Furthermore, a split is made regarding these variables between European and U.S. data.
4.2 Descriptive statistics
In table 4.1 a summary of the statistics of the most important variables is presented. The table shows that the leverage ratio is slightly lower before the crisis than during and after the crisis. The mean of the leverage ratio of large banks is also slightly lower. The return is as expected negative during the crisis and positive before and after the crisis. The amount of banks that became inactive on the stock exchange is the highest during the crisis. Furthermore, the percentage of inactive banks is higher in Europe compared to the U.S. These statistics are primarily interesting for testing the hypotheses 1 and 2.
Table 4.1 Summary statistics
This table contains the mean and standard deviation of the leverage ratio (consist of regulatory tier1 capital), return (cumulative return index) and status (dummy variable indicating 1 when active). The data regarding these three variables is split in three period; pre-crisis (2005-2007), crisis (2007-2009) and post-crisis (2009-2013). Furthermore, a split is made if the banks are U.S. banks or European banks. Large banks indicate banks with a total asset higher than 100bln.
Leverage Ratio Return Status
Mean Std. Dev. Mean Std. Dev. Mean Std. Dev.
Pre-crisis 5.555 2.523 1.500 7.337 0.760 0.437 Crisis 6.257 5.456 -2.735 12.436 0.693 0.462 Post-crisis 6.357 3.380 1.032 13.646 0.833 0.373 U.S. 6.553 2.431 0.412 12.574 0.915 0.279 Europe 6.020 4.503 0.303 12.776 0.730 0.444 Large banks 5.623 4.529 1.117 13.996 0.843 0.364
Table 4.2 is the correlation matrix of the variables of interest for hypothesis 3. The Correlation between the excessive return and the Fama French factors are higher than the correlation of excessive return with the leverage ratio. Furthermore, the correlations between the leverage ratio and the Fama French Factors are low. The risk variable (MKT-RF) has a high correlation with the momentum factor and the book-to-market variable. But the correlations are still below 80% and therefore there is no multicollinearity.
20
4.2 Correlation Matrix
The excessive return is the monthly return minus the risk free rate. The leverage ratio is the capital divided by total exposure as discussed in this paper. RM-RF, SMB and HML are the Fama French factors. RM-RF is the market return minus the risk free rate. SMB is size premium, HML is the value premium. MOM is the momentum factor (Carhart).
Excessive Return Leverage ratio RM-RF SMB HML MOM Excessive Return 1 Leverage ratio 0.0309 1 RM-RF 0.3207 0.0068 1 SMB 0.3168 -0.0121 0.0997 1 HML 0.3581 0.0208 0.6472 0.2573 1 MOM -0.4486 -0.0188 -0.7159 -0.3919 -0.7202 1 5. Results
This part gives an overview of the most important results. First a comparison is made between the leverage ratio and the capital ratio as a performance indicator. Second, the results are presented with respect to the level of the leverage ratio. At last the relation between the leverage ratio and return is analyzed.
5.1 Leverage ratio versus capital ratio
This part describes the comparison between the leverage ratio and the capital ratio as a performance indicator. The figures and the tables are shown in Appendix C. Figure 5.1 shows the concentration of failed and surviving banks related to their leverage ratio. The graph shows that the concentration of failed banks is higher at the lower bound of the graph. This suggest that failed banks have on average a lower leverage ratio than surviving banks. Figure 5.2 shows the concentration of failed and surviving banks related to the capital ratio. In this graph it is harder to see if either the failed or surviving banks are concentrated at the lower bound of the graph. For this reason, tables 5.1 and 5.2 are presented. Table 5.1 presents the mean of the leverage ratio for the failed and the surviving banks. Table 5.2 presents the mean of the capital ratio for the failed and the surviving banks. In both tables the mean of the failed and surviving banks are compared. The differences in leverage ratio between failed and surviving banks is -1.189 and significant at a one percent level. This result shows that on average banks that failed had a 1.189% lower leverage ratio than the surviving banks. The difference in the capital ratio
21 is small and not significant. This indicates that there is no clear separation between failed and surviving banks with respect to the capital ratio.
In tables 5.3 to 5.6 the same analysis is used as in the previous tables but a separation is made between European and U.S. banks. For European banks that failed, the leverage ratio is significantly lower than the leverage ratio for surviving banks. Within the capital ratio no distinction can be made. For U.S. banks the same outcome is found in respect to the leverage ratio. But for U.S. banks the capital ratio is also significantly lower for failed banks compared to surviving banks. Note that the differences regarding the U.S. data are more sensitive as the number of banks is significantly lower than the number of banks in Europe.
5.2 Level of the leverage ratio
This part describes the results with respect to the level of the leverage ratio. The figures and the tables are shown in Appendix D. Figure 5.3 shows that 57.4% of the banks of group 1 survived. The percentage of surviving banks increases when you move to higher groups. This increase in the percentage of surviving banks indicates that banks with a higher leverage ratio survive more often. This is in line with results presented in table 5.1 of Appendix C.
The tables 5.7, 5.8 and 5.9 compare the means of the different groups. Table 5.7 shows that the mean of surviving for group 1 is 57.4% and the mean of surviving for group 2 is 72.5%. The difference between the means is 15.5% and significant. This shows that the percentage of surviving banks with a leverage ratio between 3% and 4% (group 2) is higher compared to banks with a leverage ratio below 3% (group 1). Table 5.8 shows that also between group 2 and 3 there is significant increase in surviving. This increase is not notable in the comparison between group 3 and 4 (table 5.9). This result is in line with figure 5.3, as the line of the graph flattens out when moving to higher groups. The results show that the percentage of surviving banks significantly increases if you move up to a leverage ratio of 5%. But an increase in leverage ratio above 5% does not show an increase in the percentage of surviving banks.
5.3 Relation of the leverage ratio with excessive return
In appendix E the regressions show the influence of the leverage ratio on excessive stock return. The first regression shows the raw (without any control variables) influence of the leverage ratio on excessive stock return. The coefficient value of the leverage ratio is 0.027. This indicates that a one percent point increase in the leverage ratio will increase the return by
22 0.027%. But the coefficient is not significantly different from zero. In the second regression the Fama French factors and the momentum factor are added. With the exemption of the leverage ratio all variables are significant. The MKT-RF, SMB, HML all have positive influence on excessive stock return, the momentum factor has a negative influence on the excessive return. In the third regression the periods of interest are added. The variables pre-crisis and post-crisis are both positive and significant. The coefficient 3.4 for pre-crisis and the coefficient of 2.5 for post-crisis show that the average excessive return is 3.4% higher before the crisis and 2.5% higher after crisis compared to the crisis. Furthermore, the leverage ratio has no significant influence during, before and after the crisis on excessive return because the variable leverage ratio and both interaction variables are not significant. In regression 4, 5 and 6 the focus is on: European banks (4), banks with a total asset larger €100 bln (5) and banks with a leverage ratio higher than 5% (6). In regression 5 the coefficient of the interaction variable of post-crisis with the leverage ratio is 0.166 and significant. This shows that a one percent increase in leverage ratio after the crisis increases the excessive return with 0.166%. In regression 7 the focus is on European banks with total assets larger than 100bln. The coefficient of the interaction variables of leverage ratio with pre-crisis and post-crisis are significant. The leverage ratio reduced the excessive return with -0.097% with a one percent increase before the crisis. After the crisis, a one percent point increase in the leverage ratio increases the excessive return with 0.162%. In the 8th regression, where the data set is reduced by all limits (Europe, Total assets> €100 bln,
Leverage ratio > 4 percent), the leverage ratio has no longer a positive effect on excessive return. The negative influence on excessive return before the crisis remains.
6. Robustness Check
Currently the capital measure of the leverage ratio consists of regulatory tier 1 capital. But the Basel III Accord states that during the period of 2013-2018 the Basel committee will collect data to track the impact of using either common equity or regulatory tier 1 capital. For this reason this section repeats hypotheses 1 and 2 but with a capital measure consisting of common equity. The outcomes are compared to the outcomes of the leverage ratio with regulatory tier 1 capital.
In table 8 the correlation is stated between the leverage ratio consisting of regulatory tier 1 capital and common equity. The ratio of 0.890 indicates that the ratios are highly correlated. Therefore the expectation is that the results are similar for both ratios.
23
Table 6.1 Correlation Matrix
leverage ratio (REG) leverage ratio (CET) leverage ratio (REG) 1 leverage ratio (CET) 0.890 1
Table 6.2 in appendix F presents the mean of failed and surviving banks. The mean of the leverage ratio of failed banks is 6.03%.This leverage ratio is on average -0.98% lower than the 7.01% of surviving banks. Table 6.3 and 6.4 show the means of the leverage ratio of only European and U.S. banks. In both situations the mean of failed banks is significantly lower than the mean of the surviving banks. These results are in line with the results when the capital measure of the leverage ratio consist of regulatory tier 1 capital.
Figure 6.2 gives an indication for the determination of the level of the leverage ratio. This figure presents the percentage of surviving banks per group. The largest increase is from group 2 to group 3. This shows that the percentage of surviving banks increases if the leverage ratio increases from 4% to 5%. After group 3 the line flattens out, which is in line with results of the leverage ratio consisting of regulatory tier 1 capital. The following tables: 6.3, 6.4 and 6.5 compare the means of the groups. The increase from group 2 to group 3 is the only increase that is significant. The value 0.198 shows that on average the percentage of banks that survived is 19.8% higher for banks from group 3.
7. Discussion/conclusion
This paper analyzes the explanatory power of the leverage ratio on the performance of banks. Also it analyzes what the optimal level of the leverage ratio needs to be to increase financial stability. At last this paper analyzes the relation between excessive stock return and the leverage ratio. The research question of the paper is: Could the leverage ratio have prevented the crisis?
This research question is analyzed by the following three hypotheses:
H1: The leverage ratio is a better indicator for banking performance than the capital ratio
24
H2: The minimum required leverage ratio (3%) in the Basel III Accord should increase to create more financial stability.
H3: Leverage ratio has a negative relation with the return of banks.
7.1 Conclusion
This paper shows that banks that failed during the period 2005-2013 had on average a lower leverage ratio compared to banks that survived. On average the leverage ratio was 1.2% lower for failed banks. The result shows that the leverage ratio is a better indicator of survival than the complicated capital ratio. This result is in line with the related literature. Many papers (Blum 2008, Hadan (2009), Blundell-Wignall, A. and Atkinson, P. (2010b) predicted that the leverage ratio is a better risk indicator than the capital ratio. The result is in line with hypothesis 1.
In addition the paper shows that moving to a higher leverage ratio increases the percentage of surviving banks. This increase in surviving banks is notable up to 5%. Moving to a higher leverage ratio than 5% does not show a significant increase in the percentage of surviving banks. This is in line with the proposed level of the leverage ratio of the FED (2010) but higher than the leverage ratio introduced by Basel (2010). This result also contradicts the arguments of Admati et. all (2011) who argue that the leverage ratio should exceed the 5%. The result of a leverage ratio of 5% is what is hypothesized in hypothesis 2.
At last, this paper shows the relation between the leverage ratio and excessive return. The theory of Adrian and Shin (2010) suggest that high leverage during economic growth is beneficial. Therefore banks argue that a high leverage ratio during economic growth is negatively related to return. This papers does not find a clear relationship between excessive return, the leverage ratio and economic state. But a negative relation between the leverage ratio and excessive return is found when the dataset is reduced to only European Banks with a total asset higher than 100 bln in the period before crisis. Furthermore, a positive relation between the leverage ratio and excessive return is found when the dataset is reduced to only Banks with a total asset higher than 100 bln in the period after crisis. These results are line with the theory of Adrian and Shin but as stated above this only occurs when the dataset is reduced. The results do not show a clear relationship between leverage ratio and return.
The Basel III Accord states that during the period of 2013-2018 the Basel committee will collect data to track the impact of using either common equity or regulatory tier 1 capital. This paper shows that in both situation the leverage ratio is lower for failed banks compared to
25 surviving banks. Furthermore, for both leverage ratios the percentage of surviving banks increases up to 5% and flattens out for a leverage ratio above 5%.
7.2 Discussion
The first point of discussion is that in most of the cases there is no relationship between the leverage ratio and the excessive return within a certain economic state. The first explanation can be that the periods of pre-crisis, crisis and post-crisis are not well defined. The crisis period in this paper is from 2007 to 2009. The duration of the crisis in most European countries was longer, this might have influenced the stock return of banks post-crisis. However, this does not explain why there was not in all regressions a negative relation between excessive return and the leverage ratio before the crisis. The result in line with the theory (Adrian and Shin, 2010) is the negative relationship between the leverage ratio and excessive return before the crisis (when the data is reduced to only European banks with a total assets higher than 100 bln). The negative relationship can be an argument for the large European banks against the leverage ratio. This shows that a higher leverage ratio reduces the return for a bank. Society might respond to the high cost that were involved in saving those banks during and after the crisis, and that the introduction of the leverage ratio (up to 5%) reduces the changes for a bank to default. This implies that the costs for banks do not outweigh the benefits of society. Although the higher cost for banks can be transferred back to society when banks increase the prices of their products.
The second point of discussion is the introduced level of the leverage ratio in the Basel III accord (2010). The current leverage ratio in the Basel III Accord is 3% and will increase in 2018 to 4%. The paper shows that the leverage ratio should increase to 5% because this increases the percentage of surviving banks. This assumes that the introduced leverage ratio in the Basel III Accord is too low. However, the paper only looks at the isolated influence of the leverage ratio and does not include the other new regulations of the Basel III Accord. This suggests that in combination with the other new requirements, a leverage ratio of 4% can be high enough.
The last point of discussion is the calculation of the exposure measure of the leverage ratio. This paper calculates the exposure measure of the leverage ratio based on total assets and off-balance sheet items. But as shown in Appendix A the exposure measure consist of more items. By taking into account these items the leverage ratio can change and result in different outcomes.
26
8. References
Admati, A.R. DeMarzo, P. M, Hellwig, M. F. & Pfleiderer, P. (2011). “Fallacies, Irrelevant Facts, and Myths in the Discussion of Capital Regulation: Why Bank Equity is Not Expensive”, Rock Center for Corporate Governance at Stanford University, Working Paper No. 86.
Adrian, T. and Shin, H, S. (2010). “Liquidity and Leverage”, Journal of Financial Intermediation, 19, pp. 418-437.
Antoniades, A. (2015). “Commercial bank failures during the great recession: the real (estate) story”, Working paper European Central Bank, pp. 2-55
Barrell, R. Davis, E.P. Karim, D. and Liadze, I. (2010). “Evaluating off-balance-sheet exposures in banking crisis determination models”, National Institute Economic Review, 213, pp. 39-64.
Basel Committee on Banking Supervision. (2014). “Basel III: Leverage ratio framework and disclosure requirements”, Bank for international settlements [online]. Available from: < http://www.bis.org/publ/bcbs270.pdf> [Accessed 6 February 2015].
Basel Committee on Banking Supervision. (2010). “Basel III: A global regulatory framework for more resilient banks and banking systems”, Bank for international settlements [online]. Available from: < http://www.bis.org/publ/bcbs189.htm> [Accessed 6 March 2015].
Basel Committee on Banking Supervision. (2006). “Basel III: International convergence of capital measurement and capital standards”, Bank for international settlements [online]. Available from: < http://www.bis.org/publ/bcbs128.htm> [Accessed 5 March 2015].
Basel Committee on Banking Supervision (2014). “History of the Basel Committee”, Bank for international settlements [online]. Available from: < http://www.bis.org/bcbs/history.htm> [Accessed 4 July 2015].
Blum, J.M. (2008). “Why Basel II’ may need a leverage ratio restriction”, Journal of Banking and Finance, 32, pp. 1699–1707.
27 Blundell-Wignall, A. and Atkinson, P. (2010b). “Thinking beyond Basel III: Necessary solutions for capital and liquidity”. OECD Journal: Financial Market Trends 2010 (1)
Brunnermeier, M. (2009). “Deciphering the 2007/8 Liquidity and Credit Crunch”, Journal of Economic Perspectives 23, pp. 77-100.
Carhart, M. (1997) “On Persistence in Mutual Fund Performance”, Journal of Finance, LII, pp. 57-82.
Federal Reserve. (2014). “Agencies adopt enhanced supplementary leverage ratio finale rule and issue supplementary leverage ratio notice proposes rulemaking”. [Online] Available from: <http://www.federalreserve.gov/newsevents/press/bcreg/20140408a.htm [Accessed 6 March 2015]
Fama, E.F and French, K.R. (1993). “Common Risk Factors in the Returns on Stocks and Bonds", Journal of Financial Economics, 33, pp. 3–56
Fama, E.F and French, K.R. (1995). “Size and book-to-market factors in earnings and returns”, Journal of Finance, 50, pp. 131-155
Groen, W.P. (2014). “Was the ECB’s Comprehensive Assessment up to standard?”, CEPS policy brief Available from:
<http://aei.pitt.edu/57215/1/PB_No_325_WPdG_ECB_Comprehensive_Assessment.pdf> [Accessed 29 March 2015]
Gropp, R. and Heider, F. (2009). “The Determinants of Bank Capital Structure”, European Central Bank WP, No. 1096.
Haldane, G.A. (2012). “The dog and the Frisbee”, given at the Jackson Hole 36th economic policy symposium.
Kolk van der. K. (2012). “The effects of the important variables controlled in Basel 3 on banking performance”, University of Amsterdam, pp. 2-36
28 Koudstaal, M. and van Wijnbergen, S. (2012). “On risk, leverage and banks: do highly leveraged banks take on excessive risk?”, Duisenberg school of finance, pp. 2-25
Lane, P.R. (2012). “The European sovereign debt crisis”, Journal of Economic Perspectives – 26-3, pp. 49-68
Laeven, L. and Valencia, F. (2010). “Resolution of Banking Crises: The Good, the Bad and the Ugly”, IMF Working Paper, No. 146.
Spero, J. (1980). “The Failure of the Franklin National Bank: Challenge to the International Banking system”. New York: Columbia University Press.
Von Thadden, E.L. (2004). “Bank capital adequacy regulation under the new Basel Accord”, Journal of Financial Intermediation 13, pp. 90-95.
29
Appendix A: All fields that ABN AMRO uses to calculate the leverage ratio
Table 1.1
(in millions) 31-Dec-14 31-Dec-13
Tier 1 capital 15.435 14.087
Exposure measure
On-balance sheet exposures 386.867 372.022 Off-balance sheet items 36.018 33.543 On-balance sheet netting 37.709 54.959
Derivative exposure -11.783 -2.667
Securities financing exposures -13.217 -10.472
Other regulatory measures 744 644
Exposure measure 436.338 448.028
30
Appendix B: Adrian-Shin (2010) numerical example
Assume an investment of $1000. This investment is levered for 95%, so there is $950 debt and $50 equity. The result is a 5% leverage ratio.
Table 2.1 Adrian-Shin (2010) Table 2.2 Adrian-Shin (2010)
Assets Liabilities
Investment (1000) Debt (950)
Equity (50)
A one-to-one relation between equity and the investment exist, if the market value of the debt stays constant. For example, a 10$ increase in the investment will lead to a 10$ increase in equity. Due to this increase, the leverage ratio goes up to 5.94%. In the case of commercial banks, the balance sheet is actively managed. Therefore, it is possible to keep the leverage ratio stable, related to the investment. To maintain a constant leverage in terms of market value, commercial banks can increase their debt. Due to this increase in debt, a positive relationship is notable between economic upturn and debt. In order to hold the leverage ratio constant the debt has to increase to $1105. This increase in debt also results in an increase in demand of assets, which causes an increase in the prices of assets. As a result banks have to increase their debt even further to maintain a constant leverage ratio. This framework shows the pro-cyclical nature of the leverage cycle.
Assets Liabilities
Investment (1010) Debt (950)
31
Appendix C: Leverage ratio failed and surviving banks
Figure 5.1 Leverage ratio
This figure shows the leverage ratio of the failed and surviving banks. The data set consist
32
Figure 5.2 Total regulatory capital ratio
This figure shows the capital ratio of the failed and surviving banks. The data set consist of 337 banks over a period from 2005-2013.
34
36
Appendix E: Excessive return for banks
Table 5.10 Return
This table looks at the determinants of excessive return for a subset of banks. In column 1 the raw influence of the leverage ratio on excessive return. In column 2 the Fama French factors are included in the regression. In column 3 the dummy variables pre-crisis and post-crisis are added, besides the interaction variables of those variables with leverage ratio are added. The variable pre-crisis will take the value 1 for the period of 2005-2007, the variable post-crisis will take value 1 for the period 2009-20013. In regression 4-6 the data is limited by the following: Only European banks (4), only banks with total assets >100 bln (5) and only banks with a Leverage ratio>4%. The regressions is monthly data from 2005-2013. Standard errors are between the brackets. * p value < 0.1, ** p value < 0.05,*** p value < 0.01. In all regressions a correction has been made with respect to heteroskedasticity.
(1) (2) (3) (4) (5) (6) (7) (8) Leverage ratio 0.027 0.024 0.017 0.025 0.038 -0.016 0.039 0.002 (0.021) (0.021) (0.033) (0.036) (0.039) (0.041) (0.039) (0.044) MKT-RF 0.097*** 0.044** 0.164*** 0.124*** 0.023* 0.193*** 0.262*** (0.019) (0.019) (0.022) (0.031) (0.025) (0.034) (0.046) SMB 0.672*** 0.597*** 1.155*** 0.916*** 0.483*** 1.254*** 1.547*** (0.046) (0.047) (0.059) (0.082) (0.058) (0.090) (0.131) HML 0.261*** 0.317*** 0.020 0.219*** 0.335*** 0.034 -0.071 (0.050) (0.050) (0.064) (0.084) (0.062) (0.096) (0.124) MOM -0.267*** -0.248*** -0.396*** -0.472*** -0.187*** -0.539*** -0.613*** (0.039) (0.039) (0.056) (0.081) (0.043) (0.098) (0.141) Pre-crisis 3.427*** 3.059*** 3.522*** 2.979*** 3.125*** 3.440*** (0.381) (0.397) (0.492) (0.580) (0.528) (1.119) Post-crisis 2.465*** 1.226*** 1.891*** 2.253*** 1.60*** 0.260 (0.363) (0.398) (0.547) (0.536) (0.604) (1.087) Lev*Pre-crisis -0.075 -0.059 -0.096 -0.039 -0.097* -0.119* (0.054) (0.055) (0.059) (0.063) (0.058) (0.067) Lev*Post-crisis 0.061 0.051 0.166** 0.086 0.162** 0.128 (0.048) (0.049) (0.075) (0.056) (0.082) (0.110) Constant 0.063 0.071 -2.115*** -1.237*** -1.756*** -1.800*** -1.163*** .261 (0.158) (0.159) (0.289) (0.322) (0.408) (0.426) (0.452) (0.843)
Europe Yes yes yes
Total
Assets>100mld yes yes yes
Leverage ratio >4% yes yes
37
Appendix F: Robustness
Figure 6.1 Leverage ratio
This figure shows the leverage ratio of the failed and survived banks. The data set consist of 337 banks over a period from 2005-2013. The capital measure of the leverage ratio is common equity.
