Skip to main content
Log in

Monetary Conditions and Banks’ Behaviour in the Czech Republic

  • Research Article
  • Published:
Open Economies Review Aims and scope Submit manuscript

Abstract

This paper examines the impact of monetary conditions on the risk-taking behaviour of banks in the Czech Republic by analysing the comprehensive credit register of the Czech National Bank. Our duration analysis indicates that expansionary monetary conditions promote risk-taking among banks. At the same time, a lower interest rate during the life of a loan reduces its riskiness. While seeking to assess the association between banks’ appetite for risk and the short-term interest rate we answer a set of questions related to the difference between higher liquidity versus credit risk and the effect of the policy rate conditioned on bank and borrower characteristics.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

Notes

  1. A situation of a binary choice – a borrower with or without a bad history – calls for a discrete choice model such as probit.

  2. Generally, when controlling for unobserved heterogeneity we follow the flexible approach of (Heckmann and Singer 1984).

  3. We consider solely loans and overdrafts granted by the bank, and exclude unauthorized debits and loans bought from other banks.

  4. NACE is the European industry standard classification system (Statistical Classification of Economic Activities in the European Community).

  5. The same classification system as in the case of loans (the European industry standard classification system), although this time the code applies to the company’s industry.

  6. Inflation is measured by monthly consumer price indices (CPI).

  7. Monthly averages.

  8. The phenomenon of relationship lending in the Czech Republic is explored in detail in Geršl and Jakubik (2011).

  9. Loan currency t = 1 if the loan is granted in euros, dollars or pounds.

  10. We also experiment with one year prior to new loan origination and obtain the same positive dependence.

  11. In Model I differences between banks are captured by the “frailty effect”. Given the standard error of θ and the likelihood-ratio test statistic (\(\bar {\chi }^{2}_{(01)}\) = 47.25), we find a significant frailty effect, meaning that the correlation across loans grouped by banks cannot be ignored.

  12. To be precise, the figure would be 1.8 percentage points given a 3 percentage point drop in interest rates and a conservative change in the hazard rate of 0.6 percentage points.

References

  • Altunbas Y, Gambacorta L, Marques-Ibanez D (2009) Bank risk and monetary policy. Working papers 1075, European Central Bank

  • Altunbas Y, Gambacorta L, Marques-Ibanez D (2010) Does monetary policy affect bank risk-taking? mimeo 298. BIS working papers

  • Babouček I, Jančar M (2005) A VAR analysis of the effects of macroeconomic shocks to the quality of the aggregate loan portfolio of the Czech banking sector. CNB working paper series 1, Czech National Bank, Research Department

  • Bernanke BS, Gertler M (1995) Inside the black box, the credit channel of monetary policy transmission. J Econ Perspect 9(4):27–48

    Article  Google Scholar 

  • Bernanke BS, Gertler M, Gilchrist S (1996) The financial accelerator and the flight to quality. Rev Econ Stat 78(1):1–15

    Article  Google Scholar 

  • Borio C, Zhu H (2007) Capital, regulation, risk-taking and monetary policy: a missing link in the transmission mechanism. Mimeo bank for international settlements. Basel

  • Boyd JH, Runkle DE (1993) Size and the performance of banking firms: testing the predictions of theory. J Monet Econ 31:46–47

    Article  Google Scholar 

  • Campbell J, Cochrane J (1999) By force of habit: a consumption-based explanation of aggregate stock market behaviour. J Polit Econ 107(107):205–251

    Article  Google Scholar 

  • Chava S, Jarrow R (2004) Bankruptcy prediction with industry effects. Rev Financ 8:537–69

    Article  Google Scholar 

  • CNB (2010/2011) Financial stability report. Technical report, Czech National Bank

  • Cox D (1972) Regression models and life-tables. J R Stat Soc 34(2):187–220

    Google Scholar 

  • Dell’Ariccia G, Marquez R (2006) Lending booms and lending standards. J Financ 61(5):2511–46

    Article  Google Scholar 

  • Diamond DW (1984) Financial intermediation and delegated monitoring. Rev Econ Stud 51:393–414

    Article  Google Scholar 

  • Diamond DW, Rajan RG (2006) Money in a theory of banking. Am Econ Rev 96(1):30–53

    Article  Google Scholar 

  • Duffie D, Saita L, Wang K (2007) Multi-period corporate default prediction with stochastic covariates. J Financ Econ 83:635–665

    Article  Google Scholar 

  • Finlay KL, Magnusson M (2009) Implementing weak instrument robust tests for a general class of instrumental variables models. Working papers No. 0901 Tulane University, Department of Economics

  • Geršl A, Hlaváček M (2007) Foreign direct investment, corporate finance, and the life cycle of investment. Czech Journal of Economics and Finance 57(9–10):448–464

    Google Scholar 

  • Geršl A, Jakubik P (2011) Relationship lending in emerging markets, Evidence from the Czech Republic. Comp Econ Stud 53(4):575–596

    Article  Google Scholar 

  • Geršl A, Jakubik P, Konecny T, Seidler J (2013) Dynamic stress testing. The framework for assessing the resilience of the banking sector used by the Czech National Bank. Czech Journal of Economics and Finance 63(6):505–536

    Google Scholar 

  • Glen J, Mondragón-Vélez C (2011) Business cycle effects on commercial bank loan portfolio performance in developing economies. Technical report, International Finance Corporation. World Bank Group

  • Gutierrez RG (2002) Parametric frailty and shared frailty survival models. Stata J 2(1):22–44

    Google Scholar 

  • Heckmann JJ, Singer B (1984) Econometric duration analysis. J Econ 24:63–132

    Article  Google Scholar 

  • Hellman T, Murdock K, Stiglitz JE (2000) Liberalization, moral hazard in banking and prudential regulation: are capital controls enough? Am Econ Rev 90 (1):147–65

    Article  Google Scholar 

  • Ioannidou V, Ongena S, Peydró JL (2007) Monetary policy and subprime lending: A tall tale of low federal funds rates, hazardous loans, and reduced loan spreads. Mimeo CentER-Tilburg University/European Central Bank, Tilburg

  • Ioannidou V, Ongena S, Peydró JL (2009) Monetary policy, risk-taking and pricing: evidence from a quasi-natural experiment. Technical report. CentER-Tilburg University, Tilburg

  • Ioannidou V, Ongena S, Peydró JL (2015) Monetary policy, risk-taking and pricing: Evidence from a quasi-natural experiment. Rev Financ

  • Jakubik P, Schmieder C (2008) Risk, stress testing credit. Is the Czech Republic different from Germany? CNB working paper series 9. Czech National Bank

  • Jiménez G, Ongena S, Peydró JL, Saurina J (2007) Hazardous times for monetary policy: What do twenty-three million bank loans say about the effects of monetary policy on credit risk? CEPR discussion paper 6514. Banco de España

  • Jiménez G, Ongena S, Peydró JL, Saurina J (2008) Hazardous times for monetary policy: What do twenty-three million bank loans say about the effects of monetary policy on credit risk-taking? Working papers 833. Banco de España

  • Jiménez G, Ongena S, Peydró JL, Saurina J (2014) Hazardous times for monetary policy. What do twenty-three million bank loans say about the effects of monetary policy on credit risk? Econometrica 82(2):463–505

    Article  Google Scholar 

  • Kadlcakova N, Keplinger J (2004) Credit risk and bank lending in the czech republic. Technical report, CNB Working Paper No. 6/2004

  • Kane EJ (1989) The S&L insurance mess: How did it happen? MIT Press, Cambridge, MA

    Google Scholar 

  • Keeley MC (1990) Deposit insurance, risk, and market power in banking. Am Econ Rev 80(5):1183–1200

    Google Scholar 

  • Kiefer NM (1988) Economic duration data and hazard functions. J Econ Lit 26:646–679

    Google Scholar 

  • Lown C, Morgan DP (2006) The credit cycle and the business cycle, new findings using the loan officer opinion survey. J Money, Credit, Bank 38(6):1575–1597

    Article  Google Scholar 

  • Maddaloni A, Peydró JL, Scope S (2009) Does monetary policy affect bank credit standards? Mimeo, European Central Bank

  • Markowitz HM (1952) Portfolio selection. J Financ 7(1):77–91

    Google Scholar 

  • Matsuyama K (2007) Credit traps and credit cycles. Am Econ Rev 97(1):503–16

    Article  Google Scholar 

  • Nkusu M (2011) Loans, nonperforming macrofinancial vulnerabilities in advanced economies, working paper 11/161. IMF

  • Rajan RG (2006) Has finance made the world riskier? European Financial Management 12(4):499–533

    Article  Google Scholar 

  • Royston P, Altman DG (1994) Regression using fractional polynomials of continuous covariates, parsimonious parametric modelling. J R Stat Soc 43(3):429–467

    Google Scholar 

  • Shumway T (2001) Forecasting bankruptcy more accurately: A simple hazard model. J Bus 74:101–24

    Article  Google Scholar 

  • Smith BD (2002) Monetary policy, banking crises, and the friedman rule. Am Econ Rev 92(2):128–34

    Article  Google Scholar 

  • Stiglitz JE, Greenwald B (2003) Towards a new paradigm in monetary economics. Cambridge University Press

  • Stomper A (2006) A theory of banks’ industry expertise, market power, and credit risk. Manag Sci 52(10):1618–1634

    Article  Google Scholar 

  • Čihák M, Heřmánek J, Hlaváček M (2007) New approaches to stress testing the Czech banking sector. Czech Journal of Economics and Finance 57(1–2):41–59

    Google Scholar 

Download references

Acknowledgments

This work was supported by the Czech National Bank (Research Project No. C4/2009) and the Grant Agency of the Czech Republic (projects GA CR No. 14-02108S and No. P402/12/G097). The authors thank Dana Hájková, Štěpán Jurajda, Xavier Freixas and two anonymous referees for useful comments, and Josef Brechler and Thomas Mitterling for excellent assistance. The opinions expressed in this article are only those of the authors and do not represent the official views of the institutions with which the authors are affiliated.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Adam Geršl.

Appendices

Appendix: A

Table 9 Definitions of variables
Table 10 Correlations between variables
Table 11 Weak instrument robust tests for IV probit
Table 12 Estimation results for probit model with clustered loans

Appendix: B

This section describes the steps involved in building the optimal survival and probit models developed as a robustness check for our probit and loan survival analysis. In the probit analysis we first evaluate the significance of each potential measure by considering its univariate probit fit. All covariates with p-values less than 25 % along with all those of known economic importance are initially included in the multivariable model. Following the fit of the initial model we verify the significance of each variable in the model to identify those which can be removed. In order to nominate covariates that might be deleted from the model we use the p-values from the Wald tests of the individual coefficients, and then examine the p-value of the partial likelihood ratio test to confirm that the deleted covariate is indeed not significant. Having eliminated all insignificant measures at this stage, we coarsely classify the discrete characteristics overly rich in their categories, such as the 72 firm regional affiliations. We fit a hazard model for each category and group the characteristics with similar parameter estimates and significance levels. Thereafter, we employ the method of fractional polynomials to suggest transformations of the continuous variables. To ensure the economic validity of the transformed continuous covariates, we limit our search for proper functional forms to the natural logarithm and powers of plus and minus one. Moreover, we use the fractional polynomials procedure as a tool for validating the variables’ significance once the optimal transformations have been incorporated. Finally, we determine whether our model necessitates interaction terms. We test the significance at the 5 % level of all economically plausible interaction terms formed from the main effects in our model. As previously, we examine the p-values from the Wald test and the partial likelihood ratio test.

To select the covariates for the survival analysis we employ essentially the same methods as those used in the probit regression. We begin with the bivariate analysis of the association between all plausible variables and the loan survival time. For all potential predictors we compute the first, fifth, tenth, fifteenth and twentieth percentiles of the survival times. No estimates of higher survival quantiles are needed, as the loan data are typically characterized by low default occurrence. In our dataset the default ratio does not exceed 20 % in specific sub-groups and is approximately 2 % on average. For descriptive purposes, we break continuous variables into ten and twenty quantiles and compare the survivorship experience across the groups so defined. We examine the equality of the survivor functions using a set of available non-parametric tests, but we mostly rely on the log-rank test. Additionally, we consider the partial likelihood ratio test obtained in the estimation of each covariate’s group-specific impact on the time to loan failure. Evidently, the same type of bivariate analysis is performed for categorical predictors. All variables with log-rank and partial likelihood ratio test p-values less than 20 % along with all those that are economically vital are initially included in the multivariable model. Thereafter, we repeat all the steps already described for the probit variable selection. We fit the initial model, remove insignificant covariates, coarsely classify the discrete characteristics and apply the method of fractional polynomials to the multivariable proportional hazards regression model. Next, we determine whether any economically plausible interaction terms need to be added. Finally, we check the model’s validity and its adherence to the proportionality assumption.

The methodology of fractional polynomials due to Royston and Altman (1994) offers an analytical way of determining the scale of the continuous predictors. Royston and Altman (1994) introduce a family of curves called fractional polynomials with power terms limited to a small predefined set of values and show how to find the best powers yielding the best-fitting and parsimonious model. In a single covariate case, a fractional polynomial of degree m is defined as:

$$\begin{array}{@{}rcl@{}} \phi_{m}(X; \xi, p)=\xi_{0}+\sum\limits^{m}_{j=1}\xi_{j}X^{p_{j}} \end{array} $$
(B.1)

where m is a positive integer, p = (p 1,...,p m ) is a vector of powers with p 1 <...<p m , ξ = ( ξ 0, ξ 1,..,ξ m ) are coefficients and \(\phantom {\dot {i}\!}X^{p_{j}}\) signifies:

$$\begin{array}{@{}rcl@{}} X^{p_{j}}=\left\{{~}_{\ln(X)\ if\ p_{j}=0}^{\,\,X^{p_{j}}\ if\ p_{j}\neq 0} \right. \end{array} $$
(B.2)

Expressions B.1 and B.2 combined and generalized can be rewritten into:

$$ \phi_{m}(X; \xi, p)=\sum\limits^{m}_{j=0}\xi_{j}H_{j}(X) $$
(B.3)
$$ H_{j}(X)=\left\{{~}_{H_{j-1}(X)\ln(X)\ if\ p_{j}=p_{j-1}}^{\qquad X^{p_{j}}\ if\ p_{j}\neq p_{j-1}}\right. $$
(B.4)

Royston and Altman (1994) advocate that p= { −2, −1, −0.5, 0, 0.5, 1, 2, 3} is a set of powers sufficiently rich to handle many practical cases. The best model is the one with the largest log likelihood. We use the fractional polynomials routine extended for multivariable specifications and implemented in STATA. An iterative search of scale within multivariable models involves checking for the scale of each covariate. To briefly illustrate the process, let’s consider m=2. For each variable the routine tests the best J=2 model versus the linear model, the best J=2 versus the best J=1 fractional polynomial model and the linear model versus the model excluding the tested covariate. Having checked each predictor, the procedure repeats for each variable using the outcome of the first cycle for all covariates other than the one currently being tested in the second cycle. The reiteration aims to ascertain whether changing the functional form of one covariate alters the transformation of the other covariates. The routine runs until no further transformation is suggested. Table 13 contains the definitions of the optimally chosen covariates, while tables 1416 present their descriptive statistics.

Table 13 Robust specification: Definitions of variables
Table 14 Robust probit model: Data descriptive statistics
Table 15 Robust survival model: Data descriptive statistics
Table 16 Robust models: Correlations between variables
Table 17 Estimation results for robust probit model with clustered loans
Table 18 Robust probit results for firm turnover controls
Table 19 Robust probit results for firm employment controls
Table 20 Robust probit estimation results for loan collateral types
Table 21 Robust probit estimation results for loan purpose

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Geršl, A., Jakubik, P., Kowalczyk, D. et al. Monetary Conditions and Banks’ Behaviour in the Czech Republic. Open Econ Rev 26, 407–445 (2015). https://doi.org/10.1007/s11079-015-9355-y

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11079-015-9355-y

Keywords

JEL Classification

Navigation