Since early 1980s, the need to identify determinants of banking crisis becomes more urgent than ever before. The systemic banking sector problems emerged frequently around the world and its contagious effects always come up with budgetary burden, output loss as well as long-run depression. Various studies about banking crisis show that despite the crisis in banking sector varies across the countries and overtime, some factors play an important role in explaining the logic behind many crises. In this study, multivariate logit model is applied to identify these factors.
The author finds that banking crisis is more likely to erupt as macro-conditions are weak, especially when economy experience low growth GDP. Our regression results also indicate that high rate of inflation increases the risk of banking sector problems. As mentioned in theoretical section, inflation increases risk of crisis through nominal interest rate that makes bank system more difficult to serve its maturity transformation function. Therefore, restrictive monetary policies are quite necessary in the context of inflationary economy. One thing should pay more attention is that an inflation stability program is likely to cause volatile rise on real interest rate. Empirical evidences previous chapter tells us that increased risk of banking sector could be result of high real interest rate. Hence, inflation control program should be designed carefully, particular in balancing between effectiveness and a cost of possible banking crisis.
In our specifications, a sharp devaluation of exchange rate is also associated with banking volatility. The risk of crisis due to a sharp depreciation of domestic currency is particularly high when bank‘s liability side dominated with foreign deposits. In such case, exchange rate risks will double debt obligations’ banks with depositors and turn the banks to be insolvent. One new findings in this paper is that implementing absolute value to evaluate contribution of exchange rate to risk of banking crisis is seems unsuitable. Descriptive statistic evidences show that although
- 46 -
exchange rate risk is acknowledged as one of the key factors causing banking fragility in Thai banking system, almost influences of exchange rate on banking crisis are reflected less correctly in Thai case. In other words, the model tends to ignore effect of exchange rate when absolute value of exchange rate are low and vice versa.
Unlike Demirguc-Kunt and Detragrache (l998a), our findings suggest that there is no empirical evidence to confirm real likelihood of explicit insurance and banking fragility. As discussed above, scientific studies are inconsistent with each other to explain effect of deposit insurance scheme on probability of banking crisis.
Demirguc- Knut and Detragrache (l998a) find that crises are more likely in countries with a deposit insurance system take place while Hagen and Ho (2003) do not recognize any persuasive evidence in their research. This contradiction not only comes from implementing different approaches in their studies, but it also belongs quite much to what determinants they use. For instance, Demirguc-Kunt and Detragrache (l998a) use dummy variable and quality of law enforcement to test effect of deposit insurance scheme, but Hagen and Ho (2003) use dummy variable only. Moreover, experts also believe that deposit insurance variable may be insignificant as amount of bail-out package are quite low for each depositor. Thus, under so many unreliable conditions it seems impossible to address likelihood of deposit insurance and banking fragility. Further researches are required to have more confident results.
Another important determinant expected to have great contribution to volatility of banking sector is liberalization. Theory of banking crisis has shown that effect of liberalization program is likely to spread banking sector through at least three channels including financial system opening, interest rate deregulation, and bank loan deregulation. Its usual consequences are sharp rise of real interest rate and credit booming in countries where financial liberalization program introduced. In our study, both real interest rate and credit booming are used to verify how closely financial liberalization program connects with probability banking crisis. Even though there is a
- 47 -
strongly empirical evidence to confirm relationship between banking crisis and real interest rate, it can not conclude that financial liberalization program facilitate crisis in banking sectors as real interest rate could be a consequence of group of factors other than financial liberalization programs. However, our regression results indicate that (not very strong) liberalizing financial system could generate more risks for credit market. In that case, credit booming will deteriorate real sector, and afterwards, spread to banking sector.
Using absolute value of exchange rate found is unappropriate technique to access banking sector problem, particularly relating to prediction power of model.
The model tends to ignore crises that originating surge in exchange rate. In other words, the prediction probability is high in countries which domestic currency is low to US dollar and vice versa. Thus, implementing percent rate is obviously good alternative solution. Surely, this is not great finding, but it is worth for noting in further researches.
There are some shortcomings existing in our study. Firstly, our definition of banking crisis is probably not precise. Hagen and Ho (2003) show that such definition could define the crisis too late because the crisis may star before it plagues banking system. Secondly, our study also ignores some errors in timing crisis periods . Lastly, exploiting annual data to explain banking crisis is obviously a limitation when evidences in Thailand and Uruguay during crisis periods indicate that crisis could be accessed in month. This, thus, leaves a question about accurate time of the crisis.
Dataset of banking crisis episodes used in this study is mainly developed by Caprio and Klingebiel. Even though most of banking crisis periods in this dataset receive support of many experts, this dataset can not
- 48 -
avoid shortcomings as the most useful rule for banking crisis is liquidity in banking system, not experts’
opinion.
- 48 -
Demirguq-Kunt et a1 (2006b),
Garcia (1999), International association of deposit insurance (2008)
Dummy variable for presence of explicit insurance, it takes
1 when a full or partial insurance for depositors is introduced otherwise it equals 0.
Deposit Dummy Ambiguous
Credit
growth Ratio of credit to private
sector to GDP Ratio GDP- IFS line 99B.
M2 is money + quasi M2/Reser
ve M2 to foreign exchange
reserves of central bankmoney IFS-line 34+35, reserves come from IFS-line Id.d IFS-line 60.. or line 60L minus inflation rate collected from WDI 2007
Ratio
Nominal interest rate minus inflation rate.
Realinterest %
National currency over US
dollar
Ex Change of exchange rate IFS-line ae.
Rate change of consume price index
Change in terms of trade
Inflation IFS- line 64.x
TOT Ratio WDI 2007
APPENDIX:
- 49 -