Browsing by Author "Zhao, Binyu"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Essays in Financial Economics and Asset Pricing
    (2024-08-06) Zhao, Binyu; Back, Kerry; Tang, Xun
    The dissertation consists of three essays that investigate the information contents and the implications of the formulation process of financial asset prices. Chapter 1, The Information Content of Municipal Bond Auctions, explores the information contents of municipal bond auctions. Based on the auction data of municipalities in the U.S. from the year 2018 to 2019, this paper tests whether the dispersion of bidders’ values leads to overpricing of municipal bonds, and whether the price informativeness of municipal bonds decreases with the dispersion. An affiliated private value auction model is applied to estimate the variances of bidders’ values, and then regression analysis is performed. The main finding is that bonds issued by municipalities with larger dispersion tend to be overpriced. One standard deviation increase in the dispersion causes around a 0.0531 standard deviation decrease in spreads. Another finding is that the correlation between bond prices and local economy is significant only for counties with value variances in the second quartile, which suggests that the dispersion decreases the price informativeness of municipal bonds. Chapter 2, Information Precision, Liquidity, and Risk Premiums: Evidence from Stock Call Auctions in the U.S., investigates the effects of information precision on financial market outcomes. Using the data from NYSE Arca call auctions, this paper tests the empirical implications of the demand schedule submission model in Kyle (1989). After showing the solid predictive power of order books and estimating the predictive precision, this paper demonstrates that higher precision increases market liquidity and that the precision levels of different types of traders do not interact with each other to affect liquidity measures such as daily turnover. Another finding is that precision is priced in excess returns, but the relation between precision and risk premiums is not linear. The main channel for precision to take effects is marginal utility. The sensitivity of a trader's marginal utility to her trading quantity is determined by her risk aversion and her own information precision. Therefore, precision affects the elasticity of demand curves and thus affects the final market outcomes. Chapter 3, A Discussion on Call Auction Data: Explore the Predictive Power of Order Books, explores the predictive power of order book data during NYSE Arca closing call auctions by applying various machine learning algorithms, including Gradient Boosting, Random Forests, Lasso, and Ridge. For each algorithm, predictions are evaluated over different tuning parameter choices. Then grid search is applied to determine the optimal parameters for each algorithms, and the performances of different algorithms are compared. The findings include that the overall prediction performances could not increase monotonically with the model complexity, because the increase in the model complexity also leads to over-fitting and finally impairs prediction accuracy. For the order book data analyzed, Random Forests achieves the best prediction performance. In general, tree-based models perform better than linear models. It suggests that the relation between order book statistics and future equity returns could have a complex non-linear form.