Banking crises and nonlinear linkages between credit and output

Abstract

In this article, we analyse the asymmetric causality linkages between credit growth and output growth during banking crises. We employ a recently developed procedure, based on a bivariate Markov switching model, to test the hypotheses of independence, causality and asymmetric causality between credit and output. Using a sample of 103 banking crises around the world, we find that neither credit nor output takes precedence as a variable in calm and crisis periods, although there is evidence of instantaneous interdependence between the banking and real sector during crises. The results suggest that shocks propagate mostly within a year between the banking sector and the real economy. The linear link between credit growth and output growth is also regime dependent.

Acknowledgements

This article has benefited from very helpful comments by an anonymous referee, Martin Bohl, Bartosz Gębka, Nick Pilcher and participants of the seminars at the National Bank of Poland, Warsaw School of Economics and Bank of England.

Notes

¹ We use the expressions ‘states’ and ‘regimes’ interchangeably to discriminate between periods of calm and crisis.

² All the effects of shocks to the credit market explained in ‘Theoretical background’ may result in increased volatility of credit growth, especially given the intense character of these shocks during the banking crises.

³ We do not use quarterly data, because we expect lagged dependencies of an order higher than one when using such data. The Markov switching model and our tests are designed to test for lagged dependencies of order one. In addition, the quarterly seasonality of output growth and credit growth complicates analyses of causality between credit and output, because periods of prosperity and stagnation, and seasonal patterns of output and credit growth may be difficult to differentiate in our four-regime setting.

⁴ The changes in log values of our variables may significantly deviate from the percentage changes of these variables during financial crises. We use changes in logs to better fit (the left tails of) the assumed normal density functions of credit and output growth. Earlier studies related to this article also use logarithmic transformations (e.g. Psaradakis et al., 2005; Gambacorta and Rossi, 2010).

⁵ Considering different means or variances of the analysed variables for each country or crisis would significantly increase the number of estimated parameters and it would reduce the number of degrees of freedom. Therefore, we prefer to use macroeconomic control variables as factors influencing credit and output in each country.

⁶ The POLITY IV database is maintained through a partnership between the University of Maryland's Center for International Development and Conflict Management and the George Mason University Center for Global Policy.

⁷ In order to examine how our model fits the data, we use tests proposed by Breunig et al. (2003) and confirm that the parameters of sample means and variances simulated from our model are consistent with the original data. Detailed results are available upon request.

⁸ These are only rough approximations, because changes in log values are downward biased approximations of large percentage changes of credit and GDP during crises.