Essays on finance : asset pricing and bank risk

Abstract

Finance has an important role in working of economy. This two essays each deals with the different fields of finance. The first essay deals with bank risks and analyzes it with annual bank balance sheet data. There have been public policy proposals regarding bank supervision based on the hypothesis that subordinated note and debenture (SND) holders prevent banks to take high risk. This is the market discipline. As modern financial system and financial constructs become more and more complicated, regulatory discipline becomes ineffective. SND could be a good support if it acts as discipline as hypothesis suggests. Using Japanese commercial banks`` annual balance sheet data and monthly market capitalization data, this paper tested this hypothesis whether SND really makes banks to reduce risks. Dynamic panel data model was estimated using IV-GLS, IV-GMM, Random and fixed effect panel regression model for each future and current bank risk. The result supported market discipline hypothesis that high SND reduced future bank risks. In addition, the risk reduction is not followed by deteriorated bank return. Balance sheets of banks become superior after SND issuance. The second essay deals with pricing theory, particulaly, Capital Asset Pricing Model (CAPM). Since its development, the traditional Sharpe-Lintner CAPM has been widely used to explain stock returns variation. However, this model has several predictable errors, including that small stocks ("size effect") and high book-to-market-value stocks ("value effect") are likely to have higher expected returns than predicted. This systematic anomalies have been studied by abundant academics. This paper suggests that the errors are present as a result of mis-specified assumptions regarding investors`` preferences; that is, traditional CAPM holds when investors care about only the mean and variance of their portfolio. However, this paper argues that under other circumstances, for example, when the return has an asymme...