This dissertation attempts to apply the KMV CreditMonitor Model to estimate default probability, measures correlation structures and tail dependencies for non-normal probability distribution of financial asset, estimates risk measure by generating and comparing the Extreme Value-at-Risk at different confidence levels for a sample of 150 commercial banks in the U.S. with varying sizes of assets in a given time period for acceptable accuracy. The paper attempts to examine not only what dependencies should exist amongst the default probabilities, default correlation, equity correlation, and Value at Risk, but whether asset returns exhibit heavy tailed behavior with a strong support of investigating more closely the properties of normality model, nonlinearity, stationary test, and skewness and kurtosis in order to making sure that the paper will get more accurate results, implement algorithm efficiently, and establish the validity of the conclusion.;The paper estimates the default probability and evaluates the asset return correlation matrix and the default correlation matrix through Monte Carlo Simulation with 10,000 replications for acceptable accuracy with Normal Copula, and estimates equity return volatility with using Log stock returns approach. Several conclusions seem warranted since the results confirm results found in other studies. Some results towards the end of this dissertation should give a better idea of answers EVT and Copula provide.;In summary, my results represent some clear implications including (1) larger banks tend to have lower default probabilities since the increase in asset sizes significantly reduces default probabilities (clearly inversely correlated to the default points); (2) the observed increase in size of banks is also a result of the ongoing close to zero in kurtosis; (3) large changes of financial returns tend to follow large changes, whereas small changes tend to follow small changes; (4) the fluctuation of equity returns seems to be positively correlated with the changes of volatility of each group; (5) larger banks tend to co move between equity return and equity volatility more than smaller banks; (6) the asset returns are either positively (weakly) or negatively (weakly) correlated to each other of each group while the default probabilities are only positively (weakly) correlated to each other of each group with a following result that the default correlation coefficients are not always lower than those of the asset returns; (7) asset correlation statistically affects the default correlation while the observed increase in co movement between asset correlation and default correlation is not due to the increase in size of the banks; (8) larger the bank is, larger the VaR with confirmation that larger banks tend to have bigger swings (higher the risk of the portfolio).;However, the results revealed some limitations that lead to recommendations for the further test outside banking industry, account for the cross-sectional parameters, and increase the sample size.
展开▼