Asia-Pacific banks risk exposures: pre and post the Asian financial crisis

Abstract

In this article, we provide an insight into Asia-Pacific banks’ market, interest rate and exchange rate exposures using a market-based model, pre and post the Asian financial crisis. Our study provides a unique comparative analysis across 10 countries, for both short-horizon and long-horizon risk exposures. Overall, our findings reveal that bank portfolios in countries that are harder hit by the Asian crisis have higher market and short-term interest rate exposures post-crisis. With long-horizon returns, there are a larger number of significant interest rate (IR) and exchange rate (ER) exposures, which are consistent with the prior literature that long-horizon return measures economic exposures that are difficult to hedge. When the long-horizon regressions with an error correction model are carried out, the results obtained support the short-horizon results. Among the country groups, the newly industrialized economies display the greatest sensitivity to IR and ER changes during the post-Asian crisis period. Investigating bank regulation effects, we find evidence that bank portfolios that experience lower restrictions on their activities and ownership, and greater private monitoring have lower market risk.

Acknowledgement

The first author gratefully acknowledges support received from Monash University Postgraduate Publication Award.

Notes

¹ According to Mishkin (1999), there are two main reasons why excessive risk-taking occurred after financial liberalization in East Asia. First, bank managers often lacked the expertise to manage risk appropriately when new lending opportunities opened up after financial liberalization; second, depositors and foreign lenders to the banks have little incentive to monitor banks because they expect government to bailout failed banks. This led to imprudent lending and investment by banks.

² A common lender is the main source of funds for several countries. For example, Japanese banks were among the largest international lenders to other countries in the Asia-Pacific region, and as a result of their losses suffered in 1997, they were forced to reduce external lending to other countries to comply with capital adequacy requirements (Brownbridge and Kirkpatrick, 1999).

³ Chow et al . (1997b) highlighted that transactional (short-horizon) exposure can be managed by using financial derivative contracts while economic (long-horizon) exposure can only be managed through operational hedges (e.g. constructing overseas facilities) so that foreign currency inflows and outflows can be matched to reduce the amount of exposed cash flows.

⁴ In Japan, the main banks are the chief lenders to its client and recipient of fees for services. They are also characterized as the residual risk bearer due to their implicit obligations of financial support for their client in the case of financial distress (Saporoschenko, 2002).

⁵ Another branch of long run exposure studies regress short horizon stock returns on contemporaneous and lagged exchange rate factors; see for example, Bartov and Bodnar (1994). Chow et al . (1997b) argue that regressing long horizon stock returns on corresponding long horizon exchange rate changes are better suited for detecting long-term exchange rate effects on firm value than are short horizon stock returns.

⁶ Transaction exposure is the effect of exchange rate changes on cash flows between the time a transaction is ‘booked’ and when it is ‘settled’. It has a shorter time dimension (Chow et al ., 1997b).

⁷ Economic exposure is typically indirect and has a longer term dimension, encompasses the competitive and indirect effects of exchange rate risk. Economic exposure may arise from changes in the sales prices, sales volumes, and the cost of inputs of the firm and its competitors as a result of exchange rate changes (Martin and Mauer, 2003b).

⁸ Examining the currency exposure of developed (OECD) countries, Patro et al . (2002) found that Australia and Japan are significantly exposed to currency risk. Whether developing countries face greater currency risk relative to developed countries is an unexplored empirical question.

⁹ Blum (1999) argues that capital adequacy regulation potentially reduce banks’ profits. When future profits are lower, banks have a lower incentive to avoid default.

¹⁰ The addresses of the central bank or regulatory authority websites are as follows: Australian Prudential Regulation Authority http://www.apra.gov.au/ADI/ADIList.cfm#AOBC; Hong Kong Monetary Authority http://www.info.gov.hk/hkma/eng/site/index.htm; Bank of Japan http://www.boj.or.jp/en/; Bank Negara Malaysia http://www.bnm.gov.my/index.php?ch=13&cat=banking; Reserve Bank of New Zealand http://www.rbnz.govt.nz/banking/nzbanks/index.html. Central Bank of the Philippines http://www.bsp.gov.ph/resources/bankdir/bankdir2.asp; Monetary Authority of Singapore http://www.mas.gov.sg/frames/directory/index.htm; Financial Supervisory Commission of Korea http://www.fsc.go.kr/eng/financial/financial_list.asp; Central Bank of China (Taiwan) http://www.cbc.gov.tw/EngHome/Ebankexam/List.htm; Bank of Thailand http://www.bot.or.th/bothomepage/databank/Institutions/Institutions_e.htm.

¹¹ Stock price data are not available for these banks because they are either not listed on an organized stock exchange or have merged with other local banks.

¹² Since the complete bond return index data is unavailable for the Asian countries sample, the bond return index of Japan, adjusted for domestic currency, is used as a proxy of LTIR for all Asian countries, assuming that long-term interest rates are integrated for this group of Asian countries. The bond return index of Australia is also used as a proxy for bond return index for New Zealand.

¹³ The short-term 3-month interest rate used for each country is: the Treasury bill rate for Hong Kong, Singapore, the Philippines and New Zealand, the interbank rate for Malaysia and Thailand, the Gensaki 3-month rate for Japan, the commercial paper rate for Korea, the money market rate for Taiwan and the Australian bank accepted bill rate for Australia.

¹⁴ Valkanov (2003) proposed that the t-statistics obtained from OLS long-horizon regression should be rescaled by dividing the t-statistics obtained by the square root of the sample size. We do not employ this procedure as the critical value simulated by Valkanov (2003) is not applicable to our multivariate long-horizon regression.

¹⁵ The Asian crisis literature has identified a number of Asian countries as harder hit by the 1997 Asian crisis. These countries include South Korea, Malaysia, Singapore, the Philippines and Thailand (see for example Gochoco-Bautista et al ., 2000).

¹⁶ The sample period of Madura and Zarruk (1995) is 1988 to 1993 while the sample period of Saporoschenko's (2002) study is 1986 to 1992.

¹⁷ During the post-Asian crisis period, the Malaysian Central Bank kept the interest rates low to counter economic contraction (Hooy et al ., 2004).

¹⁸ New Zealand operates its exchange rate under a clean float, with the Reserve Bank rarely intervening in the foreign exchange market (Chen et al ., 2004).

¹⁹ http://www.worldbank.org/research/projects/bank_regulation.htm

²⁰ Unlike Barth et al . (2004), loan classification stringency and provisioning stringency is excluded from this category due to there being insufficient data in these areas for our sample of Asia-Pacific banks.

²¹ To conserve space, the analysis of the effect of bank regulation in the long-horizon return context is suppressed since the results are similar to those reported for the short-horizon analysis. The unreported results are available from the authors upon request.