Income volatility of Indonesian banks after the Asian Financial Crisis

Abstract

This paper considers the factors that determine Indonesian bank risk both before and after the Asian Financial Crisis (AFC). In the pre-AFC period, bank capital holdings are positively associated with bank revenue risk, which is attributed to a combination of regulatory laxity as well as laxity of enforcement. In the post-AFC period, capital is found to reduce bank risk in a non-linear manner. Franchise value is associated with lower bank risk, but in a non-linear manner; low levels of franchise value are associated with increased bank risk, while higher levels of franchise value result in lower bank risk. It is also concluded the low-to-medium levels of bank loan growth are associated with lower bank risk; however, high levels of loan growth are risk increasing. These results point to the importance of enforcement of regulatory oversight in reducing bank risk.

Keywords:

• Indonesian banking
• Asian Financial Crisis
• revenue volatility

Acknowledgements

The author is grateful for the comments of Tom Smith, Gulasekaran Rajaguru, Ahmed Khalid and Noel Gaston as well as the participants at the First Annual joint workshop between NTU Singapore and the Globalisation and Development Centre, Bond University and two anonymous referees. All errors remain the responsibility of the author.

Notes

1. This delay in recognition of poor asset quality following a boom lending cycle is a global phenomenon, see Laeven and Majnoni (2003).

2. The experience of the Icelandic banking system during the 2008 financial crisis demonstrates the perils of less sophisticated bankers over-reaching themselves during asset booms involving sophisticated products (see http://en.wikipedia.org/wiki/2008%E2%80%932011_Icelandic_financial_crisis#Causes). The Icelandic experience provides a useful counterpoint to the Indonesian banking experience during the lead up to the 2008 crisis.

3. Examples of this stream of research include Sufian (2010), Omar, Majid, and Rulindo (2007), Thoraneenitiya and Avkiran (2009), Margono, Sharma, and Melvin (2010) and Hadad et al. (2011).

4. See, for example, Hutapea and Kasri (2010) as well as Shahimi, Ismail, and Ahmad (2006).

5. See also Berger, Herring, and Szego (1995), Gonzalez (2005) and Gennotte and Pyle (1991) for additional examples of this stream of research.

6. Empirical evidence examining the characteristics of banks with higher levels of non-interest income has found that they tend to be less well managed (Deyoung and Rice 2004b) and have higher risk loan portfolios (Lepetit et al. 2008).

7. As a bank holds less equity, it would be expected to be ceteris paribus, riskier and thus have more ex post volatility. However, in some short-run periods this choice to adopt a riskier portfolio of funding decisions can have the potential to yield higher ex post returns. This issue is particularly germane to the pre-AFC period in Indonesia.

8. As an alternative source of funds, many Indonesian banks were active in raising offshore wholesale funding (largely unhedged in foreign currencies). This exposure contributed to the banking crisis after the large rupiah devaluation during the AFC, thus franchise value also provides a control for foreign currency risk.

9. We are indebted to an anonymous referee for comments in this area.

10. Additional post-AFC sub-period analysis suffered from difficulties of reduced degrees of freedom.

11. Calculation of standardised coefficients for equity to total assets finds that a one-standard deviation increase in equity holding across the study period results in a 33% increase in ROA range volatility but also a 95% reduction in ROE volatility. However, given that the all years sample includes several sub-periods incorporating different regulatory and economic regimes the sub-period analysis may be more illuminating.

12. Consistent with Besanko and Kanatas (1996). BI choose not to enforce the limited regulations in place with respect to capital holdings.

13. Put another way, low levels of franchise value are associated with increased bank risk but as franchise value increases it acts to reduce risk in the manner posited by Marcus (1984) and Besanko and Thakor (1993).

14. Additional sub-period analysis broke the post-AFC period into two further sub-periods; 2000–2004 and 2005–2009. Both of these sets of additional analysis suffered from reduced degrees of freedom and as such are not reported in this paper. Tentative results indicated that post-AFC regulatory forbearance during the reconstruction period of 2000 to 2004 reduced the effectiveness of capital in mitigating bank risk. Bank non-interest income was found to be risk increasing in the sub-period 2005–2009. Further analysis to confirm these results would be a valuable follow-up to this study when more data are available.

15. This variable is used as it is reported in BankScope in the most consistent manner.

16. Derina (2011) also considered that the political implications of directed lending resulted in delays in recognition of impaired loans.

17. This paper will argue that those banks displaying higher variation in loan losses are inherently riskier.

18. It is possible that a firm with a positive ex post ROA has a low level of capital to assets, and so a positive z-score, reflecting higher returns to a risky leveraged position. This can create some noise when using z-scores as an ex post risk measure.

19. It should be noted that the data requirements of the robustness tests resulted in a considerable reduction in degrees of freedom (especially for the sub-period of 1991–1999).

20. Omitted in some regressions due to multicollinearity resulting from the small sample.