Interest rate and exchange rate exposures of banking institutions in pre-crisis Korea

Abstract

This study empirically investigates interest rate and exchange rate exposures of banking institutions in pre-crisis Korea. Using the sensitivity of stock returns as a measure of the exposure, it is shown that Korean commercial banks and merchant banking corporations had been significantly exposed to both interest rate and exchange rate risks, and that the subsequent profitability of commercial banks was significantly associated with the degree of pre-crisis exposure. The evidence suggests that, along with the negative exposure of banking institutions, the sharp depreciation of the Korean won and high interest rates at the end of 1997 further deteriorated the banking sector's capital adequacy worsening the financial crisis. The Korean case highlights the importance of upgrading financial supervision and risk management practices as a precondition for successful financial liberalization.

Acknowledgements

The author gratefully acknowledges a financial support from the Yonsei University and data support from the Korea Securities Research Institute.

Notes

For detailed descriptions of asymmetric information view of financial crises, see Mishkin (1996, 1997) among others.

Actual duration structures may vary across individual banking institutions depending upon the scope of businesses, regulation, and the risk attitude of bank managers, among others.

Deteriorations in bank asset quality ultimately reduce the accounting measure of bank equity capital as additional loan loss provisions are required. However, regardless of the timing of provisioning, the market value of bank capital would reflect actual bank asset quality in efficient financial markets.

Before April 1999, foreign exchange positions of commercial banks were regulated by the Bank of Korea, where the sum of the spot and future foreign exchange over-bought positions was limited to 15%, and over-sold positions to 10% of the previous month's bank capital.

In Korea, before the 1997 financial crisis, while the on-balance sheet foreign exchange exposure of commercial banks has been tightly regulated, the off-balance sheet exposure has not been closely monitored. Note that the amount of off-balance sheet transactions of Korean commercial banks had been rapidly increasing in the pre-crisis period from 77 trillion won in 1996 to 122 trillion won in 1997.

Hahm and Mishkin (2000) showed that at the onset of the 1997 crisis approximately 60% of total foreign currency denominated debts of non-financial corporate firms in Korea had been left unhedged.

In the case of exporting firms, the fact that non-financial corporations carry large amounts of net foreign currency debt may reflect their own hedging efforts against foreign currency cash flow risks. If this were the case, the total economic exposure of firms to the exchange rate risk and its effect on the borrower credit risk would be limited.

The argument does not necessarily imply that the interest rate and exchange rate exposure alone is the only cause of a full-blown financial crisis in emerging market countries. In Korea, the bank balance sheet had already been deteriorating substantially at the onset of the currency crisis due to the large terms of trade shocks in 1996 and resulting accumulation of non-performing loans. The argument implies that, in emerging market environment, the additional damage due to the bank exposure to interest rate and exchange rate risks could be critical in accelerating the financial crisis.

Strictly speaking, merchant banking corporations are not commercial banks but wholesale financial institutions engaging in underwriting commercial papers, leasing and short-term lending to corporate sector by funding themselves from issuing bonds and commercial papers and borrowing from inter-bank and foreign markets. The present study also investigates the financial exposure of merchant banking corporations because they not only issued quasi-deposits in the form of cash management accounts but also accounted for the largest share of short-term financial intermediation to the corporate sector by virtually monopolizing the commercial paper market. Note also that only commercial banks and merchant banking corporations were allowed for the full range of foreign exchange businesses in the pre-crisis period in Korea. In sum, the merchant banking industry was a quasi-banking industry.

The question on the degree of risk exposure and the question of whether these risks are properly priced in the capital market are separate issues. The present study focuses on the first question.

Note that the three-year corporate bond yield was used as a representative market interest rate in Korea as government bonds had not been actively traded in the secondary market due to the lack of trading volume. Note also that the won/dollar exchange rate was selected as a benchmark exchange rate based upon the observation that most of the foreign assets and liabilities of Korean financial institutions are denominated in US dollars. According to the Bank of Korea, the US dollar denomination accounted for 82.6% of total foreign currency assets and 88.7% of total foreign currency liabilities of Korean financial institutions at the end of 1998. Note also that the US dollar was used as a settlement currency for more than 85% of the export and import transactions in Korea in 1998.

In June 1993, the Korean government announced a comprehensive blueprint for Korea's financial deregulation and capital market opening over the following years. Starting from 1994, in accordance with the plan, Korea had implemented a series of deregulatory measures, and in December 1996, Korea attained membership in the Organization for Economic Cooperation and Development (OECD) satisfying most of the deregulation and market opening requirements.

To obtain the unanticipated component of interest rate and exchange rate changes, the third order univariate autoregressive model was estimated on the first difference in log interest rates and log exchange rates, respectively. The results were not sensitive to the variations in the order of autoregressive models. The above and subsequent regression models were also re-estimated using the whole values as independent variables. However, actual changes or innovations did not yield significantly different results both qualitatively and quantitatively. Hence, we report the results with innovations only.

Although not reported, in the case of two factor models where either interest rate or exchange rate alone is included in addition to the market factor, much stronger and significant estimates were obtained for interest rate and exchange rate exposures. In this sense, the estimates reported in Tables 7 and 8 must be interpreted as conservative measures for actual risk exposure.

In 1998, five commercial banks were merged with relatively sound banks by purchase and assumption (P&A) method and 17 merchant banking corporations were permanently closed down. Consequently, the 1998 return on asset data for those five commercial banks and 17 merchant banks are not available and those institutions were excluded from the sample for the present analysis.