1
The effect of interest rates on bank risk-taking in the US
AbstractA recent line of research consider the low interest rate environment of the pre-financial crisis as a contributing factor for excessive bank risk-taking in the search for yield. A large panel dataset containing of yearly observations on 1213 US banks in the period 2003-2016 has been constructed to analyze this ‘’bank risk-taking channel of monetary policy’’. The overall results do not provide evidence that a low interest rate environment incentivize bank risk-taking. Whereas non-performing loans do indicate a negative and significant relation with bank risk-taking, risk assets as well as various abnormal loan growth categories indicate a positive and significant relation and therefore results in less bank risk-taking. Finally, no differences in various bank loan growth categories were detected.
Name: Frank Grandia Student number: S1893122 Study Program: MSc Finance
Field key words: Interest rates, bank risk-taking, US banks, panel data Supervisor: Dr. J.O. Mierau
2 1. Introduction
The global financial crisis had an enormous impact on the global economy, where economic activity declined significantly and unemployment rose dramatically (Allen and Carletti, 2010). While there were many factors contributing to the global financial crisis, it is widely argued that excessive bank risk-taking in combination with weak supervision standards for banks contributed to the worst financial crisis since the Great Depression of the 1930s (Dell’Ariccia, Laeven and Marquez, 2014; Allen and Carletti, 2010; Acharya and Richardson, 2009). Allen and Carletti (2010) argue that excessive bank risk-taking was the result of the loose monetary policy prior the financial crisis, particularly by the United States Federal Reserve.
According to Rajan (2006), subsequent periods of low interest rates lead to the fact that risk-free assets become less attractive and therefore may lead to a portfolio consisting of riskier assets in the so called search for yield. Maddaloni and Peydró (2011) state that these periods of low interest rates have softened the lending standards of banks, which resulted in a period of excessive loan growth. Keeley (1990) argues that banks tend to underestimate risk during this period of excessive loan growth in the search for yield.
Borio and Zhu (2012) identified the role of the monetary policy before the financial crisis and investigated the effect of low interest rates on bank risk-taking, which they call the “risk-taking channel of monetary policy”. Delis and Kouretas (2011), Angeloni, Faia and Duca (2015), Maddaloni and Peydró (2011), Dell’Ariccia, Leaven and Marquez (2014) and Wang, Chen, Wan, Jin, and Mazzanti (2015) among others, present empirical evidence of the existence of this risk-taking channel, indicating that low interest rates indeed increase bank risk-taking.
3 Figure 1: Development of central-bank and long-term interest rates
This figure presents the bank interest rate and the long-term interest rate in the period 2000-2016. The central-bank rate is the annual average of the federal funds rate and the long-term rate is the annual average of the 10-year US government bond yield.
In this paper I will examine and provide further insight into the mechanism between low interest rates and bank risk-taking. This shall be done by measuring the effects of low interest rates on three bank risk-taking measures. The first bank risk-taking measure used in this paper will be risk assets, which is defined as the ratio of risk assets to total assets. According to Rajan (2006), periods of low interest rates may lead to a more riskier portfolio which then should be expressed in a higher ratio of risk assets. The second bank risk-taking measure used will be abnormal loan growth, which is defined as the difference between an individual bank’s loan growth and the median loan growth of banks from the same year. I will also divide the bank risk-taking measure abnormal loan growth in the subcategories mortgage loans, consumer loans, corporate loans and other loans. Finally, the last bank risk-taking measure used will be performing loans, which is defined as the ratio of non-performing loans to total loans. According to Maddaloni and Peydró (2011), periods of low interest rates lead to an increase in abnormal loan growth and the additional consequence of a higher ratio of non-performing loans.
The results of this paper differ between the three bank risk-taking measures. The bank risk-taking measures risk assets and abnormal loan growth show that low interest rates lead to less bank risk-taking. The bank risk-taking measure non-performing loans show that
4
low interest rates lead to more bank risk-taking. All these results are robust to different interest rates and include the involvement of bank-level control and macroeconomic variables.
Borio and Zhu (2012) argue that insufficient attention has so far been paid to the link between monetary policy and the perception and pricing of risk by banks. In addition, the current period of highly unusual low interest rates has not yet been reviewed by previous studies. This paper contributes to the existing literature by combining the period of low interest rates prior the financial crisis and the current period of low interest rates. Furthermore, this paper has tried to identify which category bank loan is most sensible for changes in interest rates.
This paper is structured as follows; section 2 provides a literature overview of the research in this field. Section 3 explains the data and methodology used for the empirical analysis. Section 4 presents the results and section 5 concludes and provides practical implications of this research.
2. Literature review
Maddaloni and Peydró (2011) investigated whether low levels of short and long-term interest rates soften bank lending standards in the Euro area and in the US for various periods. They find that low short-term interest rates do soften bank lending standards as opposed to low long-term interest rates. The softening in bank lending standards resulted in a period of excessive loan growth and is visible for corporate and household loans, but are most pronounced in the excessive loan growth of mortgage loans. Therefore, they state that the low interest rate environment has triggered excess bank risk-taking.
5
Gambacorta and Marques-Ibanez (2010) argue, securitization has led to the negative effect that is has increased the risk for the banking sector. In addition, the risk arising from securitization has become more difficult to measure and control because traditional risk measures might not be adequate to take these new developments into consideration.
Dell’Ariccia, leaven and Marquez (2014) argue that a low interest rate environment lead to an increase in bank risk-taking. They show that highly capitalized banks will decrease monitoring and soften their lending standards, whereas less capitalized banks will increase monitoring and tighten their lending standards. They argue that if monetary authorities had raised interest sooner and more aggressively, the consequences of the financial crisis would have been less severe.
More evidence on the risk-taking channel of monetary policy is provided by Angeloni, Faia and Duca (2015) and Ioannidou, Ongena and & Peydró (2015). Angeloni, Faia and Duca (2015) conclude that low interest rates, especially for extended periods, eventually increases bank risk-taking through an increase of the bank’s risk assets. Ioannidou, Ongena and & Peydró (2015) analyze the risk-taking channel of monetary policy in Bolivia in the period 1999 to 2003 through abnormal loan growth. They conclude that low interest rates lead to softening of the lending standards and consequentially leads to a substantial increase in riskier loans and provide evidence that these loans result in a higher ratio of non-performing loans.
Keeley (1990) argues that the search for yield arises from the fact that low interest rates decreases bank margins. He primarily focuses on the aspect of asymmetric information and he states that one of the key contributing factors for asymmetric information is the presence of the fixed-rate deposit insurance system. The explanation for excessive bank risk-taking in the presence of the deposit insurance system is that banks are able to borrow at the risk-free rate by issuing insured deposits and can then invest in risky assets with associated higher expected returns. The effect is an increase in risk assets and a decrease in capital which will increase the default probability.
6
deteriorated bank portfolios, more credit supply and overall lower returns which all in all increases the risk and financial instability in the banking sector.
Dell'Ariccia, Laeven and Suarez (2017) present more evidence of the risk-taking channel and find that short-term interest rates increase bank risk-taking through the mechanism of abnormal loan growth and non-performing loans. They examined bank risk-taking in the US during the period 1997-2011 and conclude that these results are less pronounced during periods of financial distress.
Delis and Kouretas (2011) build upon the theory of Keeley (1990) and the theoretical framework from Dell’ Ariccia and Marquez (2006) and examined bank risk-taking in the Euro area during the period 2001-2008. They argue that low interest rates lower bank margins and as a consequence, banks will engage in the search for yield by softening their lending standards. The softening in lending standards will result in a riskier portfolio and equivalently increase the probability of default. The results from Delis and Kouretas (2001) present a strong negative relationship between all sorts of different interest rates (short-term, long-term, bank-level and industry interest rates) and bank risk-taking and therefore conclude that the theoretical framework of Dell’ Ariccia and Marquez (2006) is confirmed. Furthermore, Delis and Kouretas (2011) find that the effect is smaller for specific bank individual bank characteristics such as highly capitalized banks and the effect is larger for banks engaging in non-traditional banking activities, which is visible by a higher volume of off balance sheet items.
7 3. Methodology and data
The mechanism between low interest rates and bank risk-taking will be estimated by using a panel regression model. This paper builds upon the empirical model of Delis and Kouretas (2011) which is as follows:
(1) This model estimates the risk variable, , which depends on the interest-rate variable, , specific bank-level control variables, , and common macroeconomic control variables, , all in time of bank . The various bank risk-taking measures, interest rates and control variables will be explained in detail in section 3.1 to 3.3. The data in this paper will exist of an unbalanced panel set with annual data. All specific bank-level data are collected from Orbis Bank Focus database. The macroeconomic control variables GDP growth and inflation are retrieved from databank.worldbank.org. The short-term, long-term and central-bank interest rate are retrieved from Thomson Reuters Datastream.
Gambacorta (2005) and Delis and Kouretas (2011) both conducted their research using annual and quarterly data and find that the results were robust. Therefore, they conclude that the use of annual data is sufficient in order to examine the mechanism between low interest rates and bank risk-taking. For this reason, I will use annual data only. The panel dataset consist of 1213 commercial, savings and cooperative banks located in the United States during the period 2003-2016 with a total of 10053 unique yearly bank observations. I include these specialization banks for they are the only banks that take deposits which is needed in order to measure bank risk-taking. Furthermore, I have chosen to only include large banks with a minimum amount of assets of one billion USD because smaller banks are not representative for bank risk-taking for the entire banking sector.
8 Table 1: Descriptive statistics
Variable Mean Standard
deviation
Minimum Maximum
Risk assets 0.862 0.098 0.222 1.000
Abnormal loan growth 0.061 0.359 -1.013 10.853
Abnormal mortgage loan growth 0.076 0.465 -1.123 14.278
Abnormal consumer loan growth 0.092 0.608 -1.108 16.893
Abnormal corporate loan growth 0.094 0.600 -1.130 11.928
Abnormal other loan growth 0.122 0.713 -1.817 14.367
Non-performing loans 0.017 0.028 0.000 0.536
Capitalization 0.105 0.031 -0.036 0.497
Profitability 0.011 0.014 -0.169 0.195
Size 14.459 1.351 10.252 21.457
Off-balance sheet items 0.341 1.286 0.000 66.936
Economic growth 1.878 1.600 -2.776 3.786
Inflation 2.137 1.176 -0.356 3.839
Short-term rate 1.683 1.823 0.124 5.268
Long-term rate 3.328 0.999 1.803 4.792
Central-bank rate 1.443 1.760 0.090 5.020
Bank-level lending rate 0.071 0.034 0.000 1.027
9 Table 2: Correlation matrix
Capitalization Lagged Profitability Size Off-balance sheet items Economic growth Inflation Short-term rate Long-term rate Central-bank rate Bank-level lending rate Capitalization 1.000 Lagged Profitability 0.037 1.000 Size 0.104 0.041 1.000 Off-balance sheet 0.061 0.103 0.115 1.000 Economic growth 0.009 0.195 -0.038 0.031 1.000 Inflation -0.064 0.177 -0.113 0.002 0.384 1.000 Short-term rate -0.079 0.222 -0.108 0.002 0.124 0.630 1.000 Long-term rate -0.138 0.183 -0.184 0.007 0.176 0.576 0.815 1.000 Central-bank rate -0.073 0.231 -0.108 0.005 0.205 0.602 0.991 0.816 1.000 Bank-level lending rate -0.009 0.117 -0.089 0.004 0.011 0.217 0.303 0.333 0.300 1.000
10
3.1 Bank risk-taking
In this paper I will proxy the bank risk-taking using the following three measures. The first bank risk-taking measure is risk assets. Risk assets is the ratio of risk assets to total assets where risk assets are defined as all assets except risk-free assets. Risk-free assets are government securities, cash and balances due from other banks. This means that risk assets are all bank assets which are subject to changes in market conditions or credit quality. Therefore, an increase in risk assets implies a more risky bank portfolio which should be the result of more bank risk-taking. In my sample, risk assets has a mean value of 0.862, with a minimum value of 0.222 and a maximum value of 1.
The second bank risk-taking measure is abnormal loan growth. Abnormal loan growth is defined as the difference between an individual bank’s loan growth and the median loan growth of banks from the same year. Maddaloni and Peydró (2011) divide loans in the subcategories total loans, mortgage loans, consumer loans and corporate loans. In line with their paper, I will also investigate if softening of the lending standards is present and for which categories this softening of the lending standards is most visible. In addition, I will include the loan category other loan to the previous four subcategories.
Empirical evidence of the mechanism between abnormal loan growth and bank risk-taking is provided by Foos, Norden and Weber (2010) and Amador, Gómez-González and Pabón (2013). Foos, Norden and Weber (2010) study the effect of loan growth on bank risk. They elaborate on the mechanism between the importance of loan growth on the one hand, and the possible negative effects of a too rapid increase in loan growth on the other hand. The results indicate that loan growth represents an important driver of bank risk, especially loan growth in subprime mortgage lending and conclude that abnormal loan growth is an appropriate bank risk-taking measure. Amador, Gómez-González and Pabón (2013) provide additional evidence in line with Foos, Norden and Weber (2010) that abnormal loan growth leads to an increase in bank risk. In addition, they indicate that high levels of abnormal loan growth is most often the result of softening the lending standards. In my sample, abnormal loan growth has a mean value of 0.061, with a minimum value of -1.013 and a maximum value of 10.853.
11
quality of the assets of a bank and is a proxy for credit risk. Higher levels of non-performing loans result in more losses for the bank and is therefore an appropriate proxy for bank risk-taking. Non-performing loans has a mean of 0.017, with a minimum value of 0 and a maximum value of 0.536.
3.2 Interest rates
This paper examines the mechanism between low interest rates and bank risk-taking. Therefore, I will include various interest rates in order to analyze the potential differences between these interest rates. Angeloni, Faia and Duca (2015) argue that since banks rely mostly on short-term funding, low short-term interest rates may increase bank risk-taking more than low term interest rates. In this paper I will include a short-term rate, long-term rate, central-bank rate and a bank-level lending rate. The short-long-term interest rate is the annual average of the 3-month interbank rate, the long-term interest rate is the annual average of the 10-year US government bond yield and the central-bank rate is the annual average of the federal funds rate. It should be noted that the short-term interest rate and the central-bank interest rate has a correlation coefficient of 0.991 and therefore little difference in the results between these rates can be expected. The trend for the various interest rates is that the rates have been increasing up to 2007 and strongly decreasing afterwards.
12
low interest rates result in a higher ratio of risk assets and abnormal loan growth, in contrast to non-performing loans, where low interest rates result in lower non-performing loans.
Figure 2: Bank-level lending rate and risk assets
This figure presents the non-parametric regression between risk assets and the bank-level lending rate. Risk assets is measured by the ratio of risk assets to total assets and bank-level lending rate is measured by the ratio of interest income to total customer loans. This figure reports all values of the bank-level lending rate up to 0.25 and the regression line shows a negative relationship between risk assets and the bank-level lending rate.
Figure 3: Bank-level lending rate and abnormal loan growth 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 .00 .05 .10 .15 .20 .25
Bank-level lending rate
R is k a s s e ts -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 .00 .05 .10 .15 .20 .25
Bank-level lending rate
13 This figure presents the non-parametric regression between abnormal loan growth and the bank-level lending rate. Abnormal loan growth is measured by the difference between an individual bank’s loan growth and the median loan growth from the same year and bank-level lending rate is measured by the ratio of interest income to total customer loans. This figure reports all values of the bank-level lending rate up to 0.25 and the regression line shows a negative relationship between abnormal loan growth and the bank-level lending rate.
Figure 4: Bank-level lending rate and non-performing loans
This figure presents the non-parametric regression between non-performing loans and the bank-level lending rate. Non-performing loans is measured by the ratio of non-Non-performing loans to total loans and bank-level lending rate is measured by the ratio of interest income to total customer loans. This figure reports all values of the bank-level lending rate up to 0.25 and the regression line shows a positive relationship between non-performing loans and the bank-level lending rate.
3.3 Control variables
I will include several control variables in the estimated equation to avoid the omitted variables bias. These variables can be labeled in bank-level control and macroeconomic variables. The bank-level control variables will consist of capitalization, profitability, size and off-balance sheet items and the macroeconomic variables will consist of economic growth and inflation. The bank-level control variables are included because these bank characteristics could affect bank risk-taking. The macroeconomic variables are included because bank risk-taking could be influenced by the macroeconomic environment.
Capitalization is the ratio of equity capital to total assets. A tradeoff between capitalization and bank risk-taking can be expected because higher equity capital implies
.00 .05 .10 .15 .20 .25 .30 .00 .05 .10 .15 .20 .25
Bank-level lending rate
14
cautious bank behavior. Capitalization has a mean value of 0.105, with a minimum value of -0.036 and a maximum value of 0.497.
Profitability is the ratio of profits before tax to total assets. The expectation of the impact of profitability on bank risk-taking is uncertain. On the one hand, higher profitability could be the result of higher levels of risk assets. These higher levels of profits can then be used for supplying new loans in the following period. On the other hand, this increase in loans could lead to riskier loans with a higher probability of default, which then can result in lower profitability in the following period. For this reason, I will include the variable lagged profitability instead of profitability. Profitability has a mean value of 0.011, with a minimum value of -0.169 and a maximum value of 0.195.
Size is the natural logarithm of total assets. It should be noted that the variable size depicts the natural logarithm of total assets and therefore the descriptive statistics do not represent the true values of total assets. The expectation of the impact of size on bank risk-taking is that larger banks are more risk averse than smaller banks. Size has a mean value of 14.459, with a value of 10.252 for the smallest bank and a value of 21.457 for the largest bank.
balance sheet items is the ratio of off-balance sheet items to total assets. Off-balance sheet items can be seen as a non-traditional banking activity. Non-traditional banking activities have increased significantly over the past few years and are associated with high levels of risk. Off-balance sheet items has a mean value of 0.341, with minimum value of 0 and a maximum value of 66.936.
Economic growth is annual GDP growth. Positive and high levels of GDP growth indicate more favorable economic conditions. Favorable economic conditions are associated with an increase in bank risk-taking through an increase in bank loans. Therefore, a positive relation between GDP growth and risk assets and abnormal loan growth is to be expected. In addition, a negative relation between GDP growth and non-performing loans is to be expected because favorable economic conditions should decrease the probability of not being able to repay a loan and vice versa. Economic growth has a mean value of 1.878, with a minimum value of -2.776 and a maximum value of 3.786.
15 4. Results
I have performed the Hausman test for the panel regression model and concluded that the preferred model is the fixed effects regression model.
The results of the fixed effects regression model are presented in table 3 to 5. The first interesting fact is that the bank-level lending rate does not seem to be a representative proxy for the various interest rates, which is in contrast to the result of Delis and Kouretas (2011). Where the bank-level lending rate accurately represent the various interest rates in their study, it can be seen that the bank-level interest rate consistently switch signs in comparison with the various interest rates. Therefore, the bank-level lending rate, which represents the pass through interest rate of banks to customers, cannot be used to further analyze the mechanism between low interest rate and bank risk-taking.
In addition, the short-term and the central-bank interest rate are quite identical which is in accordance with my expectation. Overall, the long-term interest rate seems to have a larger impact on bank risk-taking where the coefficient of the long-term interest rate is larger compared to the short-term and central-bank interest rate. In contrast, the results regarding the various bank risk-taking proxy’s differ quite a lot. it should be noted that the coefficients of the various interest rates are, although significant, relatively small and therefore the results should be interpreted with caution because there is little economic significance.
With regards to risk assets as the proxy for bank risk-taking, the various interest rates are positive and significant. This indicates that a decrease in interest rates lead to a decrease in risk assets and hence less bank risk-taking. These results are in contrast to my expectation that low interest rates lead to more bank risk-taking.
With regards to abnormal loan growth as the proxy for bank risk-taking, the various interest rates are positive and significant. This indicates that a decrease in interest rates lead to a decrease in abnormal loan growth and hence less bank risk-taking. These results are in contrast to my expectation that low interest rates lead to more bank risk-taking.
16
and because all interest rates do not correspond with the main objective of this paper, I will not elaborate further on the differences between the various loan categories.
With regards to non-performing loans as the proxy for bank risk-taking, the various interest rates are negative and significant. This indicates that a decrease in interest rates lead to an increase in non-performing loans and hence more bank risk-taking. These results correspond to my expectation that low interest rates lead to more bank risk-taking.
Bank capitalization is positively related with the bank risk-taking proxy’s risk assets and abnormal loan growth, in contrast to a negative relation with non-performing loans. Higher equity capital implies more cautious bank behavior and therefore a tradeoff was to be expected. Thus, this trade-off is present only for non-performing loans. Higher equity capital could be the consequence of for example stricter capital requirements.
Lagged profitability is positively related with risk assets and negatively related to abnormal loan growth and non-performing loans. This result shows that banks used the profits of the previous period not to increase risk assets but to increase their loan growth, which subsequently lead to more risky loans. The increase in loan growth, accompanied with an increase in non-performing loans can be explained by softening the lending standards.
Size is positively related to all bank risk-taking proxy’s, suggesting that larger banks exhibit more bank risk-taking, where the theory suggest that larger banks are more risk averse. A possible explanation for this surprising but clear result could be an indication that larger banks are financial healthy and therefore have higher capital reserves to capture the possible losses.
Off-balance sheet items is positive and significant for risk assets as a proxy, implying that banks with a higher amount of off-balance sheet items also exhibit more bank risk-taking in their traditional banking activities. In contrast, the effect of off-balance sheet items are small and insignificant for the other bank risk-taking proxy’s, indicating that is has little influence on abnormal loan growth and non-performing loans. This results seems intuitive because risk assets and balance sheet items can be seen as interrelated, whereas off-balance sheet items has no direct relation with bank loans.
17
conditions. An explanation for the negative relation between economic growth and risk assets could be that when the economic conditions in the US worsen, banks search for other ways to increase their returns through riskier assets. Inflation is positively and significant related to all bank risk-taking measures, suggesting that higher inflation leads to more bank risk-taking.
Table 3: Interest rates and risk assets: fixed effects regression
Risk assets Risk assets Risk assets Risk assets Capitalization 0.237*** (0.028) 0.259*** (0.029) 0.234*** (0.028) 0.162*** (0.028) Lagged profitability 0.339*** (0.055) 0.364*** (0.055) 0.354*** (0.055) 0.480*** (0.054) Size 0.019*** (0.001) 0.023*** (0.002) 0.018*** (0.001) 0.012*** (0.001) Off-balance sheet items 0.020*** (0.002) 0.020*** (0.002) 0.020*** (0.002) 0.024*** (0.002) Economic growth -0.005*** (0.000) -0.005*** (0.000) -0.005*** (0.000) -0.006*** (0.000) Inflation 0.002*** (0.001) 0.003*** (0.001) 0.003*** (0.001) 0.007*** (0.001) Short-term rate 0.003*** (0.000) Long-term rate 0.006*** (0.001) Central-bank rate 0.003*** (0.000) Bank level lending rate -0.481*** (0.030) Observations 10053 10053 10053 10053 R-squared 0.713 0.713 0.713 0.720
18 Table 4: Interest rates and abnormal loan growth: fixed effects regression
Abnormal loan growth Abnormal loan growth Abnormal loan growth Abnormal loan growth Capitalization 1.905*** (0.176) 2.113*** (0.177) 1.900*** (0.175) 1.612*** (0.176) Lagged profitability -1.978*** (0.340) -2.154*** (0.337) -1.948*** (0.340) -1.465*** (0.332) Size 0.006 (0.008) 0.046*** (0.010) 0.006 (0.008) -0.017** (0.008) Off-balance sheet items -0.004 (0.015) -0.007 (0.015) -0.004 (0.015) 0.009 (0.015) Economic growth 0.002 (0.002) 0.002 (0.002) 0.000 (0.002) -0.004* (0.002) Inflation 0.012*** (0.004) 0.009** (0.004) 0.014*** (0.004) 0.032*** (0.004) Short-term rate 0.012*** (0.003) Long-term rate 0.046*** (0.005) Central-bank rate 0.011*** (0.003) Bank level lending rate -1.920*** (0.188) Observations 10053 10053 10053 10053 R-squared 0.184 0.188 0.183 0.191
19 Table 5: Interest rates and non-performing loans: fixed effects regression
Non-performing loans Non-performing loans Non-performing loans Non-performing loans Capitalization -0.093*** (0.011) -0.100*** (0.011) -0.091*** (0.011) -0.077*** (0.011) Lagged profitability -0.696*** (0.021) -0.734*** (0.021) -0.701*** (0.021) -0.772*** (0.020) Size 0.006*** (0.001) 0.005*** (0.001) 0.006*** (0.001) 0.008*** (0.001) Off-balance sheet items 0.001 (0.001) 0.001 (0.001) 0.001 (0.001) 0.000 (0.001) Economic growth -0.004*** (0.000) -0.003*** (0.000) -0.003*** (0.000) -0.003*** (0.000) Inflation 0.003*** (0.000) 0.002*** (0.000) 0.003*** (0.000) 0.000** (0.000) Short-term rate -0.003*** (0.000) Long-term rate -0.003*** (0.000) Central-bank rate -0.002*** (0.000) Bank level lending rate 0.056*** (0.012) Observations 10053 10053 10053 10053 R-squared 0.507 0.496 0.506 0.493
This table reports coefficients and standard errors (in parentheses). Dependent variable is non-performing loans which is the ratio of non-performing loans to total loans. The explanatory variables are as follows: capitalization is the ratio of equity capital to total assets, lagged profitability is the ratio of profits before tax to total assets of the previous year, size is the natural logarithm of total assets, off-balance sheet items is the ratio of off-balance sheet items to total assets, economic growth is annual GDP growth, inflation is annual CPI inflation, short-term rate is the annual average of the 3-month interbank rate, long-term rate is the annual average of the 10-year US government bond yield, central-bank rate is the annual average of the federal funds rate, bank-level lending rate is the ratio of interest income to total customer loans. Observations is the number of observations and r-squared measures the goodness of fit of the model. * significant at the 10 percent level, ** significant at the 5 percent level, *** significant at the 1 percent level.
5. Conclusion
20
again fulfill an important role in understanding the effect of the current low interest rate environment. The purpose of this paper is to provide a comprehensive assessment of the effect of the past and current low interest rates on bank risk-taking.
The overall results show that low interest rates do not increase bank risk-taking in the period 2003-2016 in the US. The results do not correspond with the results of Delis and Kouretas (2011) for European banks. Of the three bank risk-taking measures, only non-performing loans comply with the results of Delis and Kouretas (2011) and indicate that lower interest rates lead to more bank risk-taking. Risk assets and abnormal loan growth as bank risk-taking measures present strong evidence that low interest rates lead to less bank risk-taking. All these results are robust to different interest rates and include the involvement of bank-level control and macroeconomic variables.
One of objectives of this paper was to provide further understanding in the differences of the various loan growth categories. The results show that there are no structural differences in mortgage, consumer, corporate and other loan categories. A possible explanation for this results is that these loan growth categories are interrelated and that these categories are all equally sensible to changes in interest rates and banks do not make a distinction between the various loan categories.
Another important implication of the results is that the bank-level lending rate is not a representative proxy for the short-term, long-term and central-bank interest rate. The results clearly show that the bank-level lending rate consistently switch signs in comparison with the three interest rates. For this reason, future research should be cautious about implementing a bank-level lending rate as a proxy for interest rates.
21
One of the limitations of this paper is that the empirical model used could suffer from the persistence of bank risk-taking. If this risk indeed persists then a static model could be biased, and a dynamic panel regression model would be better suited to analyze bank risk-taking. Therefore, a generalized method of moments model developed by Arellano and Bover (1995) and Blundell and Bond (1998) could have been used to allow for a dynamic and persistence nature of bank risk-taking behavior.
Another limitation of this paper is that it combined two periods of low interest rates, namely the period prior the financial crisis and the current period, where isolating one period of low interest rates could yield different results. In addition, Dell'Ariccia, Laeven and Suarez (2017) concluded that the results of bank risk-taking are less pronounced during periods of financial distress.
The current period of unusual low interest rates has not yet been reviewed by previous studies. More research is necessarily to provide further understanding whether the current period of low interest rates will result in the same excessive bank risk-taking behavior as the period prior the financial crisis or that banks have indeed learned from their previous mistakes and act more cautiously. This should also answer the question whether monetary authorities should raise the current interest rates more aggressively in order to soften the potential negative consequences and give more explanation about the current loose monetary policy by the United States Federal Reserve. To conclude, future research could provide further understanding in the visibility of softening the lending standards and whether the loan growth categories are truly interrelated and equally sensible to changes in interest rates.
References
Acharya, V., Richardson, M., 2009. Causes of the financial crisis. Critical Review 21, 195-210. Allen, F., Carletti, E., 2010. An overview of the crisis: causes, consequences, and solutions.
International Review of Finance 10, 1-26.
Altunbas, Y., Gambacorta, L., Marques-Ibanez, D., 2010. Bank risk and monetary policy. Journal of Financial Stability 6, 121-129.
22
Angeloni, I., Faia, E., Duca, M., 2015. Monetary policy and risk taking. Journal of Economic Dynamics and Control 52, 285-307.
Arellano, M., Bover, O., 1995. Another look at the instrumental variable estimation of error-components models. Journal of Econometrics 68, 29-51.
Blundell, R., Bond, S., 1998. Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics 87, 115-143.
Borio, C., Zhu, H., 2012. Capital regulation, risk-taking and monetary policy: a missing link in the transmission mechanism? Journal of Financial Stability 8, 236-251.
Data.worldbank.org., 2017. Inflation, consumer prices (annual %). Available at: http://data.worldbank.org/indicator/FP.CPI.TOTL.ZG?end=2016&locations=US&start =2003, Accessed 28-02-2017.
Data.worldbank.org., 2017. GDP growth (annual %). Available at:
http://data.worldbank.org/indicator/NY.GDP.MKTP.KD.ZG?end=2015&locations=US& start=2003, Accessed 28-02-2017.
Delis, M., Kouretas, G., 2011. Interest rates and bank risk-taking. Journal of Banking & Finance 35, 840–855.
Dell’Ariccia, G., Laeven, L., Marquez, R., 2014. Real interest rates, leverage and bank risk-taking. Journal of Economic Theory 149, 65-99.
Dell'Ariccia, G., Laeven, L., Suarez, G., 2017. Bank leverage and monetary policy's risk‐taking channel: evidence from the United States. Journal of Finance 72, 613-654.
Dell’Ariccia, G., Marquez, R., 2006. Lending booms and lending standards. Journal of Finance 61, 2511–2546.
Foos, D., Norden, L., Weber, M., 2010. Loan growth and riskiness of banks. Journal of Banking & Finance 34, 2929-2940.
Gambacorta, L., 2005. Inside the bank lending channel. European Economic Review 49, 1737-1759.
Ioannidou, V., Ongena, S., Peydró, J., 2015. Monetary policy, risk-taking, and pricing: evidence from a quasi-natural experiment. Review of Finance 19, 95-144.
23
Maddaloni, A., Peydró, J., 2011. Bank risk-taking, securitization, supervision, and low interest rates: evidence from the Euro-area and the U.S. lending standards. Review of Financial Studies 24, 2121-2165.
Rajan, R., 2006. Has finance made the world riskier? European Financial Management 12, 499-533.
Wang, Z., Chen, J., Wan, Y., Jin, Y., Mazzanti, J., 2015. Information disclosure and bank risk‐ taking under a partially implicit deposit insurance system: evidence from China. Australian Economic Review 48, 163-176.
Appendix
25 Abnormal other loan growth Abnormal other loan growth Abnormal other loan growth Abnormal other loan growth Capitalization 2.360*** (0.357) 2.665*** (0.361) 2.351*** (0.357) 1.867*** (0.359) Lagged profitability -1.429** (0.693) -1.596** (0.687) -1.389** (0.692) -0.519 (0.677) Size -0.040** (0.017) 0.019 (0.020) -0.040** (0.017) -0.079*** (0.017) Off-balance sheet items 0.005 (0.031) 0.002 (0.031) 0.006 (0.031) 0.029 (0.031) Economic growth 0.004 (0.005) 0.003 (0.005) 0.001 (0.005) -0.007 (0.005) Inflation 0.017** (0.008) 0.016** (0.008) 0.021*** (0.008) 0.052*** (0.007) Short-term rate 0.021*** (0.005) Long-term rate 0.070*** (0.011) Central-bank rate 0.020*** (0.005) Bank level lending rate -3.173*** (0.383) Observations 10053 10053 10053 10053 R-squared 0.142 0.145 0.142 0.148
