Investigating risk shifting in Islamic banks in the dual banking systems of OIC member countries: An application of two-step dynamic GMM

Abstract

In the last five decades, advances in information technology and in financial innovations have made possible the emergence of an immense capacity for banks to switch regimes from risk transfer to risk shifting. The devastating power of this capacity was amply pronounced in the financial crisis of 2007/2008. The fallout of which has intensified calls for a re-examination of current banking model and its risk management (or rather mismanagement). Risk shifting is, axiomatically, absent in an ideal Islamic financial system. The Islamic banking model, thus, provides unique paradigm with risk sharing at its core, potentially fostering financial inclusion and reducing the incidence of bank failures and the size of losses incurred by depositors and tax payers. However, the present formation of Islamic banking has grown out of conventional banking and reverse-engineers many of its techniques and instruments. The main objective of this paper is to empirically investigate risk management in Islamic banks in dual banking systems in member states of the Organization of Islamic Countries. The two-step dynamic difference GMM is applied to cater for the nature of Islamic banking data, which is characterized by a larger dynamic panel and a smaller timeframe. Findings tend to indicate that Islamic banking has a limiting effect on risk shifting. The effect however is not sufficient to fully nullify the overall risk shifting incentives. The evidence supports strengthening risk sharing and reforming Islamic banking configuration as the way forward.

Appendices

Appendices

Appendix 1

Table A1 Banks’ distribution by country

Appendix 2: IPP estimation

Merton (1977) characterizes deposit insurance as a put option written by the deposit insurer on bank’s assets and derives an implicit stock market-based price, as follows:

$$IPP \equiv N\left( {y + \sigma_{\text{v}} \sqrt T } \right) - (1 - \delta )^{n} \left( {\frac{V}{D}} \right)N\left( y \right),$$

where

$$y \equiv \frac{{\ln \left( {\frac{D}{{V(1 - \delta )^{n} }}} \right) - \sigma_{{{\text{v}} }}^{2} T/2 }}{{\sigma_{\text{v}} \sqrt T }}.$$

IPP is the actuarial value of insurance premium per dollar of insured deposits, V is the market value of bank assets, D is the face value of deposits, σ _v is asset risk, N is the cumulative standard normal distribution of a standard normal random variable, δ is the dividend per dollar of asset value, n is the number of times the dividend is paid per period, T is the unit of time until the expiry of the deposit insurance contract, it is assumed to be 1.

In this characterization, the face value of deposits (D) corresponds to the exercise price and the value of bank assets (V) corresponds to the market price.