Essays in financial economics and econometrics
This dissertation studies topics in capital markets from the perspectives of financial econometrics and empirical finance. It consists of three essays, each corresponding to one chapter. In the first chapter, we analyze whether the systematic asymmetry factor, proxied by skewness of the market, represents risk source that commands premium. It is shown that the asymmetry factor exhibits clear cyclical behavior and is closely related to the macroeconomy. We then demonstrate that the factor influences the cross-section of expected stock returns, and it earns a significant annual premium of 8%, after controlling for market, size, value, momentum, and liquidity. Comprehensive robustness checks are performed, all of which reinforce the findings. The investment implication is then explored, and it is shown that the tradable asymmetry factor can expand the investment universe and can significantly improve the Sharpe ratio of the optimal portfolio. In the second chapter, we analyze a modern three factor model of yield curve. The three latent factors, level, slope, and curvature, are extracted from the yields of all Treasury bonds with different characteristics. We demonstrate that this model does not belong to the affine class, which justifies its good forecasting performance. Next, it is found that the three factors capture systematic risk in the bond market and are hence priced. Based on these findings, we generalize the traditional Macaulay duration into a vector, whose components hedge against level, slope, and curvature respectively. We prove empirically that the generalized duration offers superior hedging performance for bond portfolios over alternative duration measures. In the last chapter, we generalize the important realized volatility methodology to high dimensional cases of crucial importance. We show that combining the informational content in intra-day data and the flexible factor structure offered by shrinkage approach can significantly improve the performance of large covariance estimator. This is especially so in situations where we want the inverse of the covariance matrix, such as portfolio optimization. However, the estimator's performance is less impressive in the Value-at-Risk exercise. We apply our estimator to equity portfolio optimization, and find it beats other popular covariance estimators.
Year of publication: |
2005-01-01
|
---|---|
Authors: | Ji, Lei |
Publisher: |
ScholarlyCommons |
Saved in:
freely available
Saved in favorites
Similar items by person
-
Ji, Lei, (2020)
-
Diebold, Francis X., (2006)
-
Knowledge-intensive business services and economic growth with endogenous market structure
Huang, Chienyu, (2013)
- More ...