1
Resilience and Profitability: the Effect of Regulatory Capital
Requirements on European Banks
University of Amsterdam
Faculty of Economics and Business BSc Business Administration Specialization Finance
Student: Jonah van Ravenzwaaij Studentnumber: 11709995 Supervised by Mr. O.C. Soons
2 Statement of originality
This document is written by Jonah van Ravenzwaaij, who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are 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 supervision of completion of the work, not for the contents.
3 Abstract
This thesis analyses the effect of regulatory capital on profitability of European banks, if the effect is influenced by the G-SIB label, and if the effect is different for banks that are located in Eurozone-countries compared to banks from non-Eurozone-countries. The research is conducted using a sample of 233 banks from 25 different European countries, over the time period of 2011 to 2018. The methodological structure of this research is based on conceptual frameworks and empirical results that are achieved through a fixed effects panel data analysis. The results do not show statistical evidence that allows us to conclude that regulatory capital affects bank profits. Furthermore, empirical evidence on an
interaction effect of regulatory capital with the G-SIB label and the Euro is absent. This research does find evidence that banks that possess the G-SIB label show higher average-assets, and that banks that are located in the Eurozone report higher return-on-average-assets and higher return-on-average-equity than banks that are located in non-Eurozone-countries.
4 Contents
1. Introduction 5
2. Literature review 7
2.1 The Basel Committee on Banking Supervision 7
2.2 Regulatory Capital 8
2.3 The G-SIB label 8
2.4 The Eurozone 9
2.5 Bank Profitability 10
2.5.1 Internal Determinants 10
2.5.2 External Determinants 11
3. Data and methodology 11
3.1 Research question 11 3.2 Data 11 3.3 Variables 13 3.3.1 Internal variables 14 3.3.2 External variables 15 3.4 Methodology 16 4. Empirical results 17 4.1 Descriptive statistics 17
4.2 Regression results & discussion 19
5. Limitations 22
6. Conclusion 22
7. Appendix 24
5 1. Introduction
Since the 2008 Financial Crisis, bank safety has become an increasingly important concern. Excessive risk-taking by financial institutions amplified the effect of the crisis on a global scale. Financial institutions had to be bailed out all over the world, and monetary and fiscal policies were employed in order to prevent the world financial system from crumbling. In 2010, as a response to the flaws in financial regulation revealed by the financial crisis of 2008, the Basel Committee on Banking Supervision developed the Basel III regulatory framework for more resilient banks and banking systems. The aim of Basel III is to
strengthen the regulatory capital framework in order to raise the resilience of the banking sector (BCBS, 2010). In addition to the measures regarding regulatory capital, Basel III increased the requirements for liquidity ratios and leverage ratios. The Financial Stability Board (2019) has reported that all 24 FSB jurisdictions have the risk-based capital and liquidity coverage ratio rules in place.
This research focuses mainly on the effect of regulatory capital on bank profitability. In light of the recent Basel III framework, the effect of regulatory capital is a topical issue. As Tung-Hao & Shu-Hwa (2013) point out, new regulations come at the cost of banks’ efficiency and profitability, through increasing solvency and liquidity. However, previous research is ambiguous regarding the effect of higher capital ratios on profitability in banking.
The banking sector is a key component of the overall financial system, and it is vital that the banking sector is able to endure negative shocks. Therefore, it is important to know which factors drive performance and profitability. Bank profitability and its determinants have been extensively examined in the literature. Plenty of previous research has shown that bank profitability is expressed as a function of bank-specific variables and
environmental variables. The goal of this research is to find the effect that regulatory capital measures have on the profitability of European banks, and whether this effect is moderated by the G-SIB label and/or whether or not the bank is located in the Eurozone. Consequently, the research question this thesis seeks to answer is: “Does regulatory capital have an effect on bank profitability, is this effect influenced by the G-SIB label and is the effect different for banks in Europe that are or are not located in countries with the Euro as its prime
currency?”.
In light of the financial crisis that had just taken place in 2008, the Financial Stability Board published a list of global systematically important financial institutions in November 2011 (FSB, 2011). Global Systematically Important Banks (G-SIBs) are banks whose failure might trigger a financial crisis. The Basel Committee on Banking Supervision assesses whether a bank qualifies as a G-SIB based on an indicator-based measurement approach.
FSB member authorities apply four requirements to Global Systematically Important Banks. First of all, G-SIBs are required to hold higher capital buffers by national authorities in accordance with international standards. In addition, they are required to meet the Total Loss-Absorbing Capacity standards, alongside the capital requirements from Basel III. Third,
6 they are obliged to meet resolvability assessments. Finally, G-SIBs must fulfil higher
supervisory expectations regarding risk management functions, risk data aggregation capabilities, risk governance, and internal controls.
As Kapopoulos & Siokis (2005) point out, the financial environment in the Eurozone is affected by the existence of a common currency and a common monetary policy. Banks are altering their strategies in order to do business and survive in this atmosphere. It is argued that common currency and monetary policy causes disintermediation, which forms a threat to the profitability of banks in the Eurozone. In addition, existing literature on the effect of regulatory measures on bank profitability shows diverging findings for banks from different countries. Therefore, this paper investigates whether being in the Eurozone affects
profitability for European banks, and whether the effect of regulatory measures is moderated by being in a country that utilises the Euro as its prime currency.
In order to answer the research question, an econometric model is adopted where bank profit is the outcome variable, which is measured in both return-on-average-assets and return-on-average-equity. The main independent variable of interest is the tier one capital ratio. In order to discover the isolated effect of regulatory capital, the model contains multiple control variables that through previous research have been shown to affect bank profitability. Internal variables that are included are size, credit risk, liquidity, and efficiency. The environmental variables that are contained in the model include GDP growth, inflation, and market concentration. The method of analysis will be a panel data analysis controlling for time fixed effects and country fixed effects.
The results show a negative but insignificant effect of tier one capital ratio on both return-on-average-assets and return-on-average-equity, therefore the null hypothesis (H0:
β1=0) is not rejected. Based on these findings, this paper can not conclude that measures
regarding regulatory capital have any effect on the profitability of European banks. Due to the insignificant coefficients on both interaction variables, it can also not be concluded that the effect of regulatory capital on bank profitability in Europe is moderated by either the G-SIB label or the Euro. However, the results of this research do show evidence that liquidity, efficiency, GDP growth, market concentration, and Euro all positively influence both ROAA and ROAE. In addition, the results show that on average, G-SIBs report higher ROAA than non-G-SIBs. The findings also show that both credit risk and size negatively affect profitability for European banks. The R-squared values of the regressions ranged between .326 and .364, which shows that there are other determinants of bank profitability that were not included in the model. Many factors that influence the profitability of banks are not straightforward and are difficult to quantify into variables. The main implications in this research are the issues of endogeneity and reverse causality which arise in the profitability equation.
However, controlling for this is beyond the scope of this thesis. Future research should focus on adding explanatory variables to the profitability equation to improve estimations of profitability, and should deal with the statistical implications.
7 The literature review in the next chapter will elaborate on the theory that backs up this research. First, it will provide some insights on the financial crisis triggering the creation of the Basel III framework. After which, more details will be provided regarding theories and views on regulatory capital, the G-SIB label, and the Eurozone. Then the theoretical models and empirical findings on internal and external factors that influence bank profits will be addressed. Next, the data & methodology section will go deeper into variable selection, the hypotheses, and the adopted econometric model. The results section will provide a
discussion on the results and comparisons to previous empirical evidence. A separate chapter is dedicated to the limitations of this research paper. Finally, the last chapter will provide a conclusion on the subject.
2. Literature review
2.1 The Basel Committee on Banking Supervision
The Basel Committee on Banking Supervision (BCBS) is a policy research and
development entity focused on the global banking environment. The financial crisis of 2008 led to the creation of the Basel III regulatory framework for more resilient banks and banking systems in 2010. Basel III capital regulation is based primarily on risk-weighted asset ratios. Common Equity Tier 1 must make up 4.5% of risk-weighted assets, Tier 1 Capital must account for 6%, and banks must hold a total capital to risk-weighted assets ratio of at least 8% (BCBS, 2010). Next to
regulatory capital standards, Basel III imposed new requirements regarding liquidity and leverage ratios. The Financial Stability Board (2019) has reported that all 24 FSB jurisdictions have the risk-based capital and liquidity coverage ratio rules in place. Figure 1 shows a comparison of the Basel II and Basel III capital requirements.
The financial crisis in 2008 started with the housing bubble in the United States. The housing bubble was the result of low interest rates and the surplus of liquidity. While the lack of regulation played part in the crisis of 2008, strong governmental regulation also does not work wonders. Basel I and Basel II caused banks to participate in regulation arbitrage. Banks were engaging in the transferring of assets in order to circumvent regulation, creating the start of a future crisis (de Grauwe, 2011). The question arises whether regulatory
interventions solve the issues in the banking industries in the long run.
Figure 1 Comparison of capital requirements under Basel III and Basel II
Adapted from Zeman, Zoltan & Kalmar, Peter & Lentner, Csaba. (2018). Evolution of post-crisis bank regulations and controlling tools: A systematic review from a historical aspect. Banks and Bank Systems. 13. 130-140. 10.21511/bbs.13(2).2018.11.
8 2.2 Regulatory capital
In the literature, theories and empirical findings on the effect of regulatory capital on bank profitability are divergent. Goddart et al. (2004) argue that an increase in the level of capital can lead to an increase in profits caused by less insurance expenses on uninsured debt. On the other hand, they hypothesize that high capital ratios would suggest over-caution and ignorance of profitable opportunities. Berger (1995) argues that every bank has an
equilibrium capital ratio which determines the relationship between capital and profitability. He hypothesises that for riskier banks with relatively low capital, an increase in the level of capital is more likely to have a positive effect on the profitability of the bank.
It is generally understood that profitability is related to cost of capital. According to Molyneux (1993), higher levels of equity would reduce the cost of capital of banks and can therefore increase the profitability. On the other hand, lower regulatory capital ratios are a sign of riskiness of a bank, which would lead to an increase in financial distress costs, which in turn would reduce profitability of banks (Berger, 1995).
Through a Generalized Method of Moments estimation on a panel of Greek banks from 1985 to 2001, Athanasoglou et al. (2006) find a positive and significant relationship between the equity-to-assets ratio and bank profitability. They argue that banks with higher capital ratios are better able to pursue business opportunities and are more capable to deal with unexpected losses. In the same line of thought, Bourke (1989) speculates that well-capitalised banks enjoy cheaper funding. In a paper on the effect of capital regulation on banks in Africa, Ozili (2017) hypothesizes that capital ratios and bank profitability are
positively related. This paper argues that banks that take more risk in their pursuit of profits, will maintain higher levels of capital. In addition, banks with sufficient regulatory capital ratios performed better during the credit crisis, because they were better able to absorb adverse shocks and unexpected losses (Beltratti and Stulz, 2009).
On the contrary, Goddart et al. (2010) find negative relations between capital ratios and profitability. These findings are in line with the notion of a standard risk-return payoff. If capital ratios are a reflection of a bank’s risk, then more capitalised banks carry less risk, and therefore should generate lower returns. In their paper, Baker & Wurgler (2015) use the CAPM and Fama-French three factor model to assess the relationship between capital requirements and the cost of capital for banks. In general, higher capital requirements should lead to lower profitability through the increase in the cost of equity (Baker & Wurgler, 2015). However, through their research they find a “risk anomaly” where banks with less risk show higher profitability. Tran et al. (2016) show that regulatory capital negatively affects ROE and ROA, which strokes with the notion that higher capital requirements increase the cost of capital and lower the tax shield savings.
2.3 G-SIB label
In November 2011, the Financial Stability Board published a list of Global
9 Banks are banks whose failure might trigger a financial crisis. The Basel Committee on
Banking Supervision assesses whether a bank qualifies as a G-SIB based on an indicator-based measurement approach. Factors of importance are cross-jurisdictional activity, size, interconnectedness, substitutability of the financial institution’s infrastructure, and
complexity. As of November 2019 there are 30 banks in the world that possess the label of Global Systematically Important.
The G-SIB label causes the debt of some banks to be different in relation to the debt of other banks. For Global Systematically Important Banks, the public authorities would not allow failure of debt repayment. Therefore, the repayment of debt is implicitly guaranteed. For this reason, G-SIBs face lower cost of capital. This implicit guarantee creates competitive distortions, increases moral hazard and can cause excessive risk-taking. For these reasons, G-SIB banks are required to meet higher and more demanding standards than other banks. In addition, the supervision’s scope on G-SIBs is expanded (Schich and Toader, 2017).
Moenninghof et al. (2015) conducted a study on the effect of regulatory measure announcements for G-SIBs. Their research showed that upon announcement of new
regulatory measures, G-SIBs’ stocks suffered negative abnormal returns compared to other banks. This confirms their hypothesis that the G-SIB regulations involve costs for the affected banks. However, they also find that being designated as a G-SIB leads to positive aggregate abnormal returns on the stock, which suggests that officially being labelled as a G-SIB creates value.
In their study on investors’ responses of the designation as a G-SIB, Dewenter & Hess (2013) found that the shareholders of these banks were affected negatively. Market
capitalization of the affected banks fell dramatically after they were designated as G-SIBs.
2.4 Eurozone
Next to adding the G-SIB factor to the profitability equation, this paper examines if the sought effects are different for banks from countries in Europe that do or do not have the Euro as their prime currency. As Kapopoulos & Siokis (2005) point out, the financial environment in the Eurozone is affected by the existence of a common currency and a common monetary policy. Banks are altering their strategies in order to do business and survive in this atmosphere. It is argued that common currency and monetary policy causes disintermediation, which forms a threat to the profitability of banks in the Eurozone.
Through a fixed effects analysis, Saeed (2014) found that capital ratio has a positive effect on bank profitability measures for 73 banks in the United Kingdom. In their study on 20 commercial banks in Sweden, Öhman & Yazdanfar (2018) find that capital adequacy, measured as a ratio of equity to total assets, is positively related to profitability. A research conducted on 28 commercial banks in the Republic of Croatia by Kundid (2012) shows that growth of regulatory capital and the capital adequacy ratio are negatively related to profitability of the Croatian banks. However, this paper does find a positive relationship between the equity to assets ratio and Return on Assets. Because of the diverging literature
10 regarding the effect of regulatory capital in non-Eurozone-countries, it is worth investigating if the effect of regulatory capital on bank profitability is different for banks in Europe that aren’t part of the Eurozone.
2.5 Bank profitability
In most theoretical models bank profitability is said to be determined by internal and external variables. Variables usually include size, liquidity, risk, efficiency, GDP growth, and inflation.
2.5.1 Internal determinants
Regehr and Sengupta (2016) argue that bank size and profitability are positively related due to economies of scale. They hypothesize that economies of scale can eliminate inefficiencies and reduce risks, and therefore might lead to a healthier banking system.
Athanasoglou et al. (2006), Ozili (2017), Dietrich & Wanzenried (2010), and Miller & Noulas (1997) find that credit risk, proxied by loan-loss-provisions ratio is negatively related to profitability for Greek, Swiss, African, and USA banks. This suggests an inverse relationship between risk and bank profitability. Abreu and Mendes (2001) show a positive relationship between the loans-to-asset ratio, which they used as a proxy for risk, and profitability. However, they explain these findings by arguing that the interest margins and profits are higher due to more deposits being transformed into loans.
In their research on the determinants of European bank profitability, Molyneux and Thornton (1992) find an inverse relationship between liquidity ratios and profitability, which is explained through the notion that liquid holdings represent a cost to the bank. However, in his study, Bourke (1989) found a positive relationship between liquidity ratios and profitability, which he deemed unexpected as common sense told him that liquid holdings should be considered as a cost to the bank.
Bank expenses and efficiency have been brought forward in the literature as profitability determinants as well. Bourke (1989) and Molyneux & Thornton (1992) found positive relationships between efficiency, measured by cost-to-income ratio, and
profitability.
Berger et al. (2000) show that bank profits have a tendency to persist of time. In addition, Goddard et al. (2011) argue that abnormal profits are not competed away
immediately, and that banks are capable of maintaining some of their abnormal profits each year. Through dynamic panel estimates on banks from 65 different countries, they find that profit persistence is negatively related to GDP growth. GDP growth lowers entry barriers for the banking sector, therefore increasing competition and lowering profit persistence. In addition, they show that market concentration is positively related to the persistence of bank profitability, which is in line with the Market Power Hypothesis, which is elaborated on below.
11 2.5.2 External determinants
The Structure-Conduct-Performance paradigm, first put forward by Chamberlin & Robinson (2003), states that the structure of a market influences the conduct of the firms within that market, which in turn determines the performance of the market. In line with this paradigm, is the Market Power Hypothesis, which states that increased market power can yield monopoly profits. From this theory comes the Relative Market Power Hypothesis, which suggests that only firms that have significant shares of the market and successfully apply product differentiation are able to earn abnormal returns and exert market power (Athanasoglou et al., 2006).
It is argued by Revell (1979) that the association between inflation and bank profitability might not be straightforward. In his book, he argues that the relationship
between inflation and bank profitability is moderated by the rate of growth of overhead and operating expenses. In addition, in their studies, Molyneux & Thornton (1992) and Bourke (1989) show the existence of positive relationships between inflation and profitability. Through their analysis on bank level data from 80 different countries, Demirguc-Kunt and Huizinga (1999) associate inflation with higher realized interest margins and higher profitability.
Tan & Floros (2012) find that GDP growth is negatively related to bank profitability in China over the period of 2003 to 2009. They argue that high economic growth reduces entry barriers for banks, therefore increasing competition and lowering profitability. However, it can be argued that economic growth increases bank activity. Economic growth leads to increases in customer deposits, loans granted, and interest margins, which all affect bank profitability in a positive way. Petria et al. (2015) find that GDP growth has a positive impact on bank profitability for commercial banks in the European Union.
3. Data and methodology 3.1 Research question
This thesis seeks to answer the following research question: “Does regulatory capital have an effect on bank profitability, is this effect influenced by the G-SIB label and is the effect different for banks in Europe that are or are not located in countries with the Euro as its prime currency?”.
3.2 Data
The original sample included data on 471 banks from 25 different European countries over the years 2011 to 2018 and was obtained through the thesis supervisor. Due to the Italian banking system being very different than other European systems regarding the number of banks, some banks have been dropped from the sample. In the original sample 255 of the 471 banks were located in Italy. In order for the results not to be driven by the sheer number of small Italian banks in the sample, all banks in the dataset that in one year showed total assets below $2,186,205,000 were deleted from the sample. This resulted in a
12 total of 238 banks being deleted from the sample, of which 203 were Italian banks. This leaves us with a strongly balanced panel of 233 banks from 25 different European countries over a time period of 8 years. Table 1 shows the number of observations per country in the final sample. The sample contains 11 banks that are labeled G-SIB (FSB, 2019). These banks are listed in table 2.
Table 1 Observations per country
Country ISO code Freq. Percent Cum.
AT* 40 2.15 2.15 BE* 56 3.00 5.15 BG 32 1.72 6.87 CY* 8 0.43 7.30 CZ 16 0.86 8.15 DE* 320 17.17 25.32 DK 128 6.87 32.19 EE* 8 0.43 32.62 ES* 48 2.58 35.19 FI* 56 3.00 38.20 FR* 120 6.44 44.64 GB 168 9.01 53.65 GR* 32 1.72 55.36 HR 8 0.43 55.79 HU 16 0.86 56.65 IE* 40 2.15 58.80 IT* 416 22.32 81.12 LU* 8 0.43 81.55 MT* 8 0.43 81.97 NL* 128 6.87 88.84 PL 24 1.29 90.13 PT* 40 2.15 92.27 RO 8 0.43 92.70 SE 104 5.58 98.28 SI* 32 1.72 100.00
13 Table 2 G-SIBs
Country Bankname
Belgium Belfius
Germany Deutsche Bank
France Société Générale
BNP Paribas Groupe Crédit Agricole Groupe BPCE Great-Britain HSBC Standard Chartered Barclays Italy UniCredit
The Netherlands ING Bank
3.3 Variables
Table 3 lists the variables used in this study. Profitability can be measured in several ways, through return-on-average-assets, return-on-average-equity, or net interest margin. This paper focuses on return-on-average-assets (ROAA) and return-on-average-equity (ROAE). Both profitability measures have different benefits. While ROAE shows the net return on shareholder capital, ROAA shows the profit generated by the total assets of the bank and is considered a measure of management efficiency (Petria et al., 2015). Banks with lower leverage will generally show lower ROAE, and higher ROAA.
Since ROAE disregards the risk of high leverage, and leverage is often determined by regulation, Athanasoglou et al. (2006) argue that ROAA is the key variable to measure bank profitability. However, ROAA does not take into account off-balance-sheet activities.
Therefore, this paper follows the reasoning of Petria et al. (2015) and uses both measures of
14 profitability for analysis. Graph 1 shows that average ROAA and average ROAE show an upward trend in the data between 2011 and 2018.
3.3.1 Internal variables
Regulatory capital: The main independent variable of interest in this paper is regulatory capital, which is measured as a ratio of tier 1 capital to risk-weighted assets. Following Goddart et al. (2004), Molyneux (1993), Athanasoglou et al. (2006), and Beltratti & Stulz (2009) a positive relationship between regulatory capital and profitability is expected. Graph 2 visually shows that the average capital ratio of the banks in the data set has
increased between 2011 and 2018.
Size: The natural logarithm of total assets is used as a measure for size. The logarithm is used due extreme skewness of the distribution. As Regehr and Sengupta (2016) suggest, the expected relationship is positive.
Credit risk: The ratio of loan loss reserves to total loans is used as a measure for credit risk. This paper follows the reasoning of Athanasoglou et al. (2006), Ozili (2017), Dietrich & Wanzenried (2010), and Miller & Noulas (1997) and expects credit risk to negatively influence bank profitability.
Liquidity: The ratio of liquid assets over total assets measures the liquidity variable. As Bourke (1989) argues, it is conventional wisdom that liquid holdings act as a cost for banks. Therefore, the expected relationship between liquidity and profitability is negative.
Efficiency: The efficiency variable is measured through the cost-to-income ratio. Because a higher ratio mean that a bank is less efficient, the relationship between this variable and bank profitability is expected to be negative.
G-SIB label: The G-SIB label is measured through a dummy variable, equal to 1 if the bank is labeled as a G-SIB, and 0 otherwise. Due to ambiguity in previous research, there is no clear expectation on the relationship between the G-SIB label and bank profitability.
15 3.3.2 External variables
Inflation: The inflation variable is measured as an annual percentage change. In line with the findings of Demirguc-Kunt and Huizinga (1999), the expected relationship between inflation and bank profitability is positive.
GDP growth: The variable for GDP growth is measured as an annual percentage change. Petria et al. (2015) state that economic growth increases banking activity. Therefore, the relationship is expected to be positive.
Euro: This variable measures if the effects are different in countries that have adopted the Euro as their prime currency. It is measured as a dummy variable, equal to 1 if the bank’s origin lies in a country that has the Euro as its prime currency, and 0 otherwise. There is no clear expectation on the effect of this variable, due to ambiguous previous research.
Market concentration: The Herfindahl-Hirschman-index is used to measure market concentration. The structure-conduct-performance paradigm and the market-power hypothesis state that increased concentration leads to higher profits.
Table 3 Variables
Variable Measure Notation Hypothesized effect
Dependent
Profitability Profit after tax/total
assets
ROAA
Profitability Profit after tax/total
equity
ROAE
Independent internal
Regulatory capital Tier 1
capital/risk-weighted assets
Capital Positive
Size Natural logarithm of
total assets
Size Positive
Credit risk Loan loss
reserves/gross loans
Credit risk Negative
Liquidity Liquid assets/total
assets
Liquidity Negative
Efficiency Cost-to-income ratio Efficiency Negative
G-SIB label Dummy equal to one
for banks with the G-SIB label, and zero otherwise
G-SIB label ?
External
16 Market
Concentration
HHIndex (sum of the three largest banks as fraction of total bank assets)
Concentration Positive
GDP-growth Real GDP growth GDP growth Positive
Euro Dummy variable
equal to one for banks in the
Eurozone, and zero otherwise
Euro ?
3.4 Methodology
To determine the effect of regulatory capital on bank profitability, the following linear model is specified:
𝑃𝑟𝑜𝑓𝑖𝑡𝑖,𝑡 = 𝛼𝑖,𝑡 + 𝛽1𝐶𝐴𝑃𝐼𝑇𝐴𝐿𝑖,𝑡 + 𝛽2𝑆𝐼𝑍𝐸𝑖,𝑡+ 𝛽3𝐶𝑅𝐸𝐷𝐼𝑇𝑅𝐼𝑆𝐾𝑖,𝑡+ 𝛽4𝐿𝐼𝑄𝑈𝐼𝐷𝐼𝑇𝑌𝑖,𝑡
+ 𝛽5𝐸𝐹𝐹𝐼𝐶𝐼𝐸𝑁𝐶𝑌𝑖,𝑡+ 𝛽6𝐺𝑆𝐼𝐵𝐿𝐴𝐵𝐸𝐿𝑖,𝑡+ 𝛽7𝐼𝑁𝐹𝐿𝐴𝑇𝐼𝑂𝑁𝑗,𝑡
+ 𝛽8𝐶𝑂𝑁𝐶𝐸𝑁𝑇𝑅𝐴𝑇𝐼𝑂𝑁𝑗,𝑡+ 𝛽9𝐺𝐷𝑃𝐺𝑅𝑂𝑊𝑇𝐻𝑗,𝑡+ 𝛽10𝐸𝑈𝑅𝑂𝑗,𝑡+ 𝜀𝑖,𝑡
In order to capture the possible interaction effect between CAPITAL and G-SIBLABEL an interaction term is added to the above specified model:
𝑃𝑟𝑜𝑓𝑖𝑡𝑖,𝑡 = 𝛼𝑖,𝑡 + 𝛽1𝐶𝐴𝑃𝐼𝑇𝐴𝐿𝑖,𝑡 + 𝛽2𝑆𝐼𝑍𝐸𝑖,𝑡+ 𝛽3𝐶𝑅𝐸𝐷𝐼𝑇𝑅𝐼𝑆𝐾𝑖,𝑡+ 𝛽4𝐿𝐼𝑄𝑈𝐼𝐷𝐼𝑇𝑌𝑖,𝑡
+ 𝛽5𝐸𝐹𝐹𝐼𝐶𝐼𝐸𝑁𝐶𝑌𝑖,𝑡+ 𝛽6𝐺𝑆𝐼𝐵𝐿𝐴𝐵𝐸𝐿𝑖,𝑡+ 𝛽7𝐼𝑁𝐹𝐿𝐴𝑇𝐼𝑂𝑁𝑗,𝑡
+ 𝛽8𝐶𝑂𝑁𝐶𝐸𝑁𝑇𝑅𝐴𝑇𝐼𝑂𝑁𝑗,𝑡+ 𝛽9𝐺𝐷𝑃𝐺𝑅𝑂𝑊𝑇𝐻𝑗,𝑡+ 𝛽10𝐸𝑈𝑅𝑂𝑗,𝑡
+ 𝛽11(𝐶𝐴𝑃𝐼𝑇𝐴𝐿 ∗ 𝐺𝑆𝐼𝐵𝐿𝐴𝐵𝐸𝐿) + 𝜀𝑖,𝑡
The third equation contains an adjustment in order to capture the possible interaction effect between CAPITAL and EURO:
𝑃𝑟𝑜𝑓𝑖𝑡𝑖,𝑡 = 𝛼𝑖,𝑡 + 𝛽1𝐶𝐴𝑃𝐼𝑇𝐴𝐿𝑖,𝑡 + 𝛽2𝑆𝐼𝑍𝐸𝑖,𝑡+ 𝛽3𝐶𝑅𝐸𝐷𝐼𝑇𝑅𝐼𝑆𝐾𝑖,𝑡+ 𝛽4𝐿𝐼𝑄𝑈𝐼𝐷𝐼𝑇𝑌𝑖,𝑡
+ 𝛽5𝐸𝐹𝐹𝐼𝐶𝐼𝐸𝑁𝐶𝑌𝑖,𝑡+ 𝛽6𝐺𝑆𝐼𝐵𝐿𝐴𝐵𝐸𝐿𝑖,𝑡+ 𝛽7𝐼𝑁𝐹𝐿𝐴𝑇𝐼𝑂𝑁𝑗,𝑡
+ 𝛽8𝐶𝑂𝑁𝐶𝐸𝑁𝑇𝑅𝐴𝑇𝐼𝑂𝑁𝑗,𝑡+ 𝛽9𝐺𝐷𝑃𝐺𝑅𝑂𝑊𝑇𝐻𝑗,𝑡+ 𝛽10𝐸𝑈𝑅𝑂𝑗,𝑡
+ 𝛽11(𝐶𝐴𝑃𝐼𝑇𝐴𝐿 ∗ 𝐸𝑈𝑅𝑂) + 𝜀𝑖,𝑡
Where profit is measured by ROAA and ROAE, and the subscripts represent bank i, located in country j, in year t.
For this analysis, a fixed effects model is adopted in order to control for unobservable time effects and country effects. If we assume that banks in the same country have the same organizational structure and culture over time, we can control for the above mentioned unobservable characteristics by adopting a fixed effects model. The results for the Hausman test are included in the Appendix and show that the fixed effects model is the appropriate choice. Through a Generalized Method of Moments estimation, Goddart et al. (2011) show that bank profits persist over time. This implies that profits are partly determined by the profits in the preceding years. In addition, there are some other unobservable time effects
17 that influence bank profitability, changes in cultural attitudes towards banks for example. By applying a fixed effects model, it is possible to control for these unobservable effects. The
variables have been winsorized at the 1st and 99th percentile to reduce the effect of extreme
outliers.
4. Empirical results 4.1 Descriptive statistics
Summary statistics on the included variables are presented in table 4. Table 4 provides interesting information. On average, the banks in the sample had a assets of 0.28% over the period of 2011 to 2018. The mean return-on-average-equity of 3.77% is way larger and also shows a greater spread. On average, the capital ratio for the banks is 17.18%. However, the capitalization differs among banks, the least
capitalized bank had a capital ratio of 7.04%, whereas, for the best capitalized bank, tier one capital covered 80.64% of risk-weighted assets. The cost-to-income ratio has a mean of around 64%, ranging from 8.38% for the most efficient bank to 166.17% for the least
efficient bank. The loan-loss-reserves ratio, which is the measure for credit risk, ranges from .01% to 24.67%, with a mean of 4.12%. Liquidity ratios range from .38% to 88.52%, with a mean of 19.78%, for the different banks over the years from 2011 to 2018. The average HHIndex, which is a measure of market concentration, is 69.30. Over the period of 2011 to 2018, the average annual GDP growth and inflation were 1.39% and 1.31%, respectively.
Table 4 Descriptive Statistics
Variable Obs Mean Std.Dev. Min Max
ROAA ROAE 1864 1863 .28 3.77 1.084 12.442 -13.52 -66.77 8.91 34.27 Capital 1864 17.181 10.417 7.04 80.64 Size 1864 17.402 1.801 14.606 21.753 Credit risk 1760 4.121 4.753 .01 24.67 Liquidity 1864 19.777 17.344 .381 88.515 Efficiency 1862 64.316 25.224 8.38 166.17 Inflation 1864 1.308 1.119 -1 4.6 GDP growth 1864 1.389 1.735 -2.9 8.1 Concentration 1864 69.304 12.917 38.811 95.397
Table 5 shows the correlation matrix for the independent variables. The paper of Grewal et al. (2004) points out that problems regarding multicollinearity arise when correlation among two predictor variables is severe. Correlation coefficients of more than 0.80 are considered to be too high. Table 4 shows that there is no reason to be concerned about multicollinearity in this analysis.
18 Table 5 Pairwise correlations
Var iab les (1 ) (2 ) (3 ) (4 ) (5 ) (6 ) (7 ) (8 ) (1 ) Ca p ital 1 .00 0 (2 ) Liq u id ity 0 .15 6 *** 1 .00 0 (3 ) Cr edi t risk -0 .15 4 *** -0 .05 1 ** 1 .00 0 (4 ) G D P g rowth 0.20 4 *** 0 .08 0 *** -0 .12 5 *** 1 .00 0 (5 ) In fla tion -0 .07 9 *** 0 .05 2 ** -0 .21 5 *** -0 .12 3 *** 1 .00 0 (6 ) Con cen -tr a tion 0 .18 0 *** -0 .05 4 ** -0 .17 4 *** 0 .07 3 *** -0 .09 8 *** 1 .00 0 (7 ) Effic ien cy -0 .18 4 *** 0 .07 7 *** -0 .01 8 0 .07 6 *** 0 .06 5 *** -0 .16 3 *** 1 .00 0 (8 ) Siz e -0 .10 6 *** 0 .19 8 *** -0 .12 4 *** 0 .02 3 0 .09 6 *** -0 .20 9 *** 0 .08 4 *** 1 .00 0 *** p<0 .0 1 , ** p<0 .0 5 , * p<0 .1
19 4.2 Regression results
Table 6 shows the results of the analysis. All the regressions include time fixed effects and country fixed effects, in order to control for unobservable factors. All of the observed effects are larger when ROAE is the outcome variable as opposed to when ROAA is the outcome variable.
All of the regressions show a negative but statistically insignificant effect of capital on profitability. The negative sign is in line with the findings of Goddart et al. (2010), Baker & Wurgler (2015), and Tran et al. (2016) and can be explained by several factors. The standard risk-return payoff model for example. As Goddart et al. (2010) argue, higher capital should imply less risk, and should therefore lower the returns. At first sight, this notion would be inconsistent with the negative effect of credit risk on both ROAA and ROAE that is found, which is significant at the 1% level in all of the regressions. However, credit risk measures the probability of losses due to debtors not fulfilling their obligations. Therefore, higher credit risk would lead to more losses on loans and lower overall profitability. Specifically, these findings report that a percentage point increase in loan-loss-reserves ratio leads to a .067 percentage point decrease in ROAA and a 1.17 percentage point decrease in ROAE. Therefore, banks should be cautious when it comes to loan-loss-reserves, as these are a threat to profitability.
Liquidity has an unexpected positive effect on Both ROAA and ROAE (p < 0,01). Specifically, one percentage point increase in liquidity ratio leads to increases in ROAA and ROAE of .007 and .093 percentage points, respectively. These findings are in line with those of Bourke (1989). However, a negative effect was expected as liquid holdings should
represent a cost to the bank (Molyneux and Thornton, 1992). The positive effect of liquidity can possibly be explained through the notion that more liquid holdings can lead to lower cost of capital and can allow banks to take advantage of profitable business opportunities faster. Hence, The liquidity requirements that Basel III imposes on banks might even be beneficial regarding the profitability of banks.
The negative effect of the variable efficiency on profitability (p < 0.01) is in line with expectations, as efficiency is measured as income ratio. An increase in the cost-to-income ratio of one percentage point leads to a decrease in ROAA of .014 and a decrease in ROAE of .172 percentage points. The higher the cost-to-income ratio is for a bank, the less efficient the bank is. Therefore, the reported effect fulfils expectations.
The effect of size is negative and significant (p < 0.01) when ROAA is the outcome variable. The size variable was measured as the natural logarithm of total assets. This means that for an increase in total assets of 1%, the return-on-average assets goes down by about .069 percentage points. When ROAE is used as profitability measure, the size variable is weakly significant (p < 0.1). A percentage increase in total assets leads to a decrease in ROAE of about .48 percentage points. This is contradictory with the argument of Regehr &
Sengupta (2016) that economies of scale should increase profitability. The negative coefficients can be explained by the fact that larger banks are constrained by more
20 regulations. Since bank size is inversely related to profitability, banks should be cautious when trying to grow their assets quickly, as this might come at the cost of profitability.
GDP growth has the expected positive effect on profitability (p < 0.01). This is in line with the findings of Petria et al. (2015) and shows that economic growth indeed leads to increased banking activity and higher profits for banks. The reported effect of GDP growth is especially high when ROAE is the profitability measure. For a unit increase in GDP growth, ROAE increases by about 1.88 percentage points. This paper does not find an effect of inflation on bank profitability.
The coefficient on market concentration is positive and significant (p < 0.01). Therefore, the findings of this paper are in line with the structure-conduct-performance paradigm and market-power-hypothesis put forward by Chamberlin & Robinson (2003), which state that more market power leads to higher profits.
Furthermore, this paper finds that being located in a country in Europe that lies within the Eurozone as opposed to being located in a European country that does not utilize the Euro as its prime currency leads to an increase in profitability (p < 0.01). More
specifically, banks in the Eurozone report ROAA that is 1.32 percentage points higher than banks that are not in the Eurozone. The effect of being in the Eurozone on ROAE is even bigger, namely 11 percentage points.
The positive coefficient on the G-SIB variable shows that banks that possess the G-SIB label generally report higher ROAA (p < 0.05). This shows that these banks benefit from an implicit debt guarantee provided by being Too-Big-To-Fail. The cost of capital of these banks is lower and therefore they are more profitable.
Next to investigating the stand-alone effect of regulatory capital on bank profitability, this research looked into an interaction effect between regulatory capital and the GSIB-label, and an interaction effect between regulatory capital and the Euro. However, no interaction effect was found. This implies that it can’t be concluded that the effect of regulatory capital on profitability is either influenced by the GSIB-label, or the Euro.
21 Table 6 : Regression results
(1) (2) (3) (4) (5) (6)
ROAA ROAA ROAA ROAE ROAE ROAE
CAPITAL -0.002 -0.002 -0.006 -0.003 -0.003 -0.034 (0.003) (0.003) (0.004) (0.040) (0.040) (0.055) LIQUIDITY 0.007*** 0.007*** 0.007*** 0.093*** 0.093*** 0.092*** (0.002) (0.002) (0.002) (0.033) (0.033) (0.034) CREDITRISK -0.067*** -0.067*** -0.067*** -1.173*** -1.172*** -1.171*** (0.019) (0.019) (0.019) (0.229) (0.229) (0.229) GDPGROWTH 0.179*** 0.179*** 0.177*** 1.887*** 1.899*** 1.876*** (0.044) (0.044) (0.043) (0.525) (0.526) (0.519) INFLATION 0.013 0.013 0.017 0.306 0.305 0.342 (0.038) (0.038) (0.038) (0.498) (0.498) (0.506) CONCENTRATION 0.028*** 0.028*** 0.029*** 0.302*** 0.303*** 0.305*** (0.009) (0.009) (0.009) (0.089) (0.089) (0.089) EFFICIENCY -0.014*** -0.014*** -0.014*** -0.172*** -0.172*** -0.173*** (0.002) (0.002) (0.002) (0.022) (0.022) (0.023) LOGSIZE -0.069*** -0.069*** -0.067*** -0.484* -0.480* -0.471* (0.020) (0.020) (0.020) (0.267) (0.268) (0.267) GSIB 0.159** 0.120 0.156** 1.482 -3.316 1.459 (0.065) (0.468) (0.065) (1.189) (7.917) (1.189) EURO 1.324*** 1.324*** 1.216*** 11.014*** 11.030*** 10.183*** (0.319) (0.319) (0.325) (3.192) (3.191) (3.228) CAPGSIB 0.003 0.357 (0.034) (0.561) CAPEURO 0.006 0.045 (0.006) (0.079) _cons -1.019 -1.020 -0.991 -9.537 -9.664 -9.327 (0.968) (0.967) (0.967) (10.715) (10.726) (10.671) Obs. 1758 1758 1758 1757 1757 1757 R-squared Time dummies Country dummies 0.326 Yes Yes 0.327 Yes Yes 0.327 Yes Yes 0.363 Yes Yes 0.364 Yes Yes 0.364 Yes Yes Standard errors are clustered by bank and presented in parentheses.
*** p<0.01, ** p<0.05, * p<0.1
As Garcia-Herrero et al. (2009) point out, in the analysis of bank profitability,
problems arise concerning endogeneity. It is argued that by retaining profits, banks are able to increase their equity. The retention of profits is easier for more profitable banks. In addition, more profitable banks will be better able to grow their size through increased spending on marketing and advertisements. Another problem that Garcia et al. (2009) bring up, is unobservable heterogeneity across banks. The data sample contains banks from 25 different European countries, all with different cultures regarding corporate governance. Arguably, these banks differ in a number of unobservable characteristics. The correct way to
22 deal with these problems is through adopting a Generalized Method of Moments estimator (Hansen 1982). That is, however, outside of the scope of this thesis.
5. Limitations
It should be noted that this research has limitations. First of all, the econometric method used was a fixed effects panel data analysis. However, recent research has pointed out that a General Method of Moments estimator might be the better way to go when investigating the profitability of banks. It is possible that the results of this paper are biased due to incorrect specification of the model, and should therefore be interpreted with care. Furthermore, the results of this research suffer from sample selection bias. In order for the results not to be driven solely by the relatively large number of banks that fill up the Italian financial landscape, small banks were excluded from the sample. Therefore, the sample is not representative for the entire European banking environment. Future research should focus on correctly specifying the econometric model and carefully selecting the method of analysis. In addition, future research should select the sample in such a manner that it is generalizable to European banks, without being biased by the sheer number of small banks in Italy.
6. Conclusion
The goal of this thesis was to answer the following research question: “Does
regulatory capital have an effect on bank profitability, is this effect influenced by the G-SIB label and is the effect different for banks in Europe that are or are not located in countries with the Euro as its prime currency?”. The answer to this question was investigated through a panel data fixed effects analysis on a sample of 233 banks from 25 different European countries over the time period from 2011 to 2018. This thesis complements existing
literature by investigating whether there exists an interaction effect on bank profits between regulatory capital and the G-SIB label, and between regulatory capital and the Euro.
In order to answer the research question, an extensive review on the existing literature was given. After which, an empirical analysis was performed on the data. The findings of this research don’t allow us to conclude that regulatory capital has an effect on bank profitability. In light of these findings, banks might not be influenced that much by the measures imposed by the Basel III framework regarding capital requirements. Next to the capital requirements, Basel III imposed new liquidity standards on banks. Since this research does find a positive relationship between liquidity ratios and bank profits, there is a
possibility that the Basel III framework affects banks in a positive manner. Future research should focus on investigating if it is possible to increase the standards imposed by Basel III to an even higher level, without reducing the attractiveness of the banking industry. There is a good chance that there are optimal levels of capital ratios, liquidity ratios, and leverage ratios that provide a perfect trade-off between profitability and safety of the financial
23 it is feasible to increase requirements imposed on banks even further without harming profits.
24 7. Appendix
8. Bibliography
Abreu, M., & Mendes, V. (2001). Commercial bank interest margins and profitability: evidence for some EU countries. In Pan-European Conference Jointly Organised by the IEFS-UK & University of Macedonia Economic & Social Sciences, Thessaloniki, Greece, May (pp. 17-20).
Ana, K. (2012). How Much is the Choice of Capital Structure Important for Bank Profitability in Croatia?. Zagreb International Review of Economics & Business, (1), 53-68.
Athanasoglou, P. P., Brissimis, S. N., & Delis, M. D. (2008). Bank-specific, industry-specific and macroeconomic determinants of bank profitability. Journal of International Financial Markets, Institutions and Money, 18(2), 121–136.
Baker, M., & Wurgler, J. (2015). Do Strict Capital Requirements Raise the Cost of Capital? Bank Regulation, Capital Structure, and the Low-Risk Anomaly. American Economic Review, 105(5), 315–320.
Basel Committee. (2010). Basel III: A global regulatory framework for more resilient banks and banking systems. Basel Committee on Banking Supervision, Basel.
Beltratti, A., & Stulz, R. M. (2009). Why did some banks perform better during the credit crisis? A cross-country study of the impact of governance and regulation (No. w15180). National Bureau of Economic Research.
Berger, A. N. (1995). The Relationship between Capital and Earnings in Banking. Journal of Money, Credit and Banking, 27(2), 432.
25 Berger, A. N., Bonime, S. D., Covitz, D. M., & Hancock, D. (2000). Why are bank profits so persistent? The roles of product market competition, informational opacity, and
regional/macroeconomic shocks. Journal of Banking & Finance, 24(7), 1203–1235.
Bourke, P. (1989). Concentration and other determinants of bank profitability in Europe, North America and Australia. Journal of Banking & Finance, 13(1), 65-79.
Demirguc-Kunt, A., & Huizinga, H. (1999). Determinants of Commercial Bank Interest Margins and Profitability: Some International Evidence. The World Bank Economic
Review, 13(2), 379–408.
Dewenter, K. L., & Hess, A. C. (2013). The Financial Stability Board and Global Systemically Important Banks: Unintended Consequences?. Available at SSRN 2338665.
Dietrich, A., & Wanzenried, G. (2011). Determinants of bank profitability before and during the crisis: Evidence from Switzerland. Journal of International Financial Markets, Institutions and Money, 21(3), 307–327.
FSB. (2011, November). Policy Measures to Address Systemically Important Financial Institutions . Retrieved from
https://www.fsb.org/wp-content/uploads/r_111104bb.pdf?page_moved=1
FSB. (2019, November). 2019 list of global systemically important banks (G-SIBs).
Retrieved from https://www.fsb.org/wp-content/uploads/P221119-1.pdf
García-Herrero, A., Gavilá, S., & Santabárbara, D. (2009). What explains the low profitability of Chinese banks? Journal of Banking & Finance, 33(11), 2080–2092.
Global systemically important banks: revised assessment methodology and the higher loss absorbency requirement. (2018, July). Retrieved from
https://www.bis.org/bcbs/publ/d445.pdf
de Grauwe, P. (2011). The Banking Crisis: Causes, Consequences and Remedies. World Scientific Studies in International Economics, 23-46
Grewal, R., Cote, J. A., & Baumgartner, H. (2004). Multicollinearity and Measurement Error in Structural Equation Models: Implications for Theory Testing. Marketing Science, 23(4), 519-529.
Goddard, J., Liu, H., Molyneux, P., & Wilson, J. O. S. (2010). Do Bank Profits Converge? European Financial Management, 19(2), 345–365.
Goddard, J., Liu, H., Molyneux, P., & Wilson, J. O. S. (2011). The persistence of bank profit. Journal of Banking & Finance, 35(11), 2881–2890.
26 Goddard, J., Molyneux, P., & Wilson, J. O. S. (2004). The profitability of european banks: a cross-sectional and dynamic panel analysis. The Manchester School, 72(3), 363–381.
Iannotta, G., Nocera, G., & Sironi, A. (2007). Ownership structure, risk and
performance in the European banking industry. Journal of Banking & Finance, 31(7), 2127– 2149.
Implementation and Effects of the G20 Financial Regulatory Reforms . (2019,
October). (5). Retrieved from https://www.fsb.org/wp-content/uploads/P161019.pdf
Kapopoulos, P., & Siokis, F. (2005). Market Structure, Efficiency and Rising
Consolidation of the Banking Industry in the Euro Area. Bulletin of Economic Research, 57(1), 67–91.
de Larosière, J., et al., e. (2009). Report of the de Larosière group - European Commision. Brussels: European Commision
Lee, C.-C., & Hsieh, M.-F. (2013). The impact of bank capital on profitability and risk in Asian banking. Journal of International Money and Finance, 32, 251–281.
Lee, T.-H., & Chih, S.-H. (2013). Does financial regulation affect the profit efficiency and risk of banks? Evidence from China’s commercial banks. The North American Journal of Economics and Finance, 26, 705–724.
Miller, S. M., & Noulas, A. G. (1997). Portfolio mix and large-bank profitability in the USA. Applied Economics, 29(4), 505–512.
Moenninghoff, S. C., Ongena, S., & Wieandt, A. (2015). The perennial challenge to counter Too-Big-to-Fail in banking: Empirical evidence from the new international regulation dealing with Global Systemically Important Banks. Journal of Banking & Finance, 61, 221– 236.
Molyneux, P., & Forbes, W. (1995). Market structure and performance in European banking. Applied Economics, 27(2), 155–159.
Molyneux, P., & Thornton, J. (1992). Determinants of European bank profitability: A note. Journal of banking & Finance, 16(6), 1173-1178.
Öhman, P., & Yazdanfar, D. (2018). Organizational-level profitability determinants in commercial banks: Swedish evidence. Journal of Economic Studies.
Ozili, P. K. (2016). Bank Profitability and Capital Regulation: Evidence from Listed and non-Listed Banks in Africa. Journal of African Business, 18(2), 143–168.
27
evidence from EU 27 banking systems. Procedia economics and finance, 20(15), 518-524.
Regehr, K., & Sengupta, R. (2016). Has the relationship between bank size and profitability changed?. Economic Review (01612387), 101(2).
Revell, J. “Inflation and Financial Institutions,” Financial Times, London, 1979.
Saeed, M. S. (2014). Bank-related, industry-related and macroeconomic factors affecting bank profitability: A case of the United Kingdom. Research journal of finance and accounting, 5(2), 42-50.
Schich, S., & Toader, O. (2017). To Be or Not to Be a G-SIB: Does It Matter?. Journal of Financial Management, Markets and Institutions, (2), 169-192.
Smirlock, M. (1985). Evidence on the (Non) Relationship between Concentration and Profitability in Banking. Journal of Money, Credit and Banking, 17(1), 69.
Tan, Y., & Floros, C. (2012). Bank profitability and GDP growth in China: a note. Journal of Chinese Economic and Business Studies, 10(3), 267–273.
Tran, V. T., Lin, C.-T., & Nguyen, H. (2016). Liquidity creation, regulatory capital, and bank profitability. International Review of Financial Analysis, 48, 98–109.
