Guidelines for the Empirical Analysis Analysis in R or STATA
Dissertation Empirical Analysis Project
- Plot the cross-sectional average of deposits/assets and Non-Deposit Debt/assets across time.
You can calculate Non-Deposit Debt = Assets – Deposits – Equity. How have the averages
evolved? How would you interpret your results?
- Run OLS regressions of quarterly loan growth on non-deposit debt/assets (one quarter lagged value) controlling for bank size (i.e. one quarter lagged natural logarithm of total asset) and
profitability (i.e. one quarter lagged return on assets) for the sub-sample of your data during the
financial crisis (i.e. 2008Q1 – 2010Q1). You should have Bank and Time fixed effects in your
regression. What is the sign and magnitude of the co-efficient on non-deposit debt/assets? Is the
coefficient significant? How will you interpret the co-efficient? Justify your findings. - Compute two measures / (ex-post) proxies for bank risk
a. Risk weighted asset divided by total assets
b. Non-performing loans divided by total loans - Plot the cross-sectional average of the above two measures across time. How have the averages
evolved across years? How would you interpret your results? - Run OLS regressions of the two ex-post measures of bank risk on equity over assets (one
quarter lagged values). Control for bank size (i.e. one quarter lagged natural logarithm of total
asset) and profitability (i.e. one quarter lagged return on assets) on the entire sample. You
should have Bank and Time fixed effects in your regression. What is the sign and magnitude of
the co-efficient on equity/assets? Is the coefficient significant? How will you interpret the coefficient? Justify your findings.
Note: Make sure you winsorize all your variables (per quarter at the 1st and 99th percentile) to
remove outliers.