Search

The first two chapters of my dissertation study the US-listed reverse merger firms. In the past decade, a growing number of foreign firms were listed in the United States through reverse merger, a non-IPO listing technique that requires less information disclosure. Using a hand-collected data set of US-listed Chinese firms, this paper answers three questions. First, are the foreign firms still of quality as the bonding theory suggests? No, I find widespread delistings in these firms and lawsuits related to their misconduct. Second, why do the bad firms list? I find that the firms’ directors profit from fast stock sales after listing. Moreover, these firms tend to be US-incorporated reverse mergers that are headquartered in small cities, are audited by small firms, and that change their auditors frequently. Third, how did they learn this technique? Using a social network analysis, I find that the firms are assisted by professionals to help them circumvent the US regulations. Further, I find that the social network of the linked directors facilitates the spread of their misconduct. During the wrongdoers’ listings, the investors in these firms lost at least $811 million. However, the penalties imposed on the wrongdoers only accounted for 4.19% of this loss. In the third chapter, I apply a rule-based textual analysis in the Chinese market to study how the textual information in the news and analyst reports affect the stock return. According to the literature, the market does not underreact to the public news, but the textual information in the analyst reports is not incorporated into the market price immediately in the US, where the analyst reports are not freely public. It remains unknown whether the analyst reports’ text information affects stock return instantly when it becomes freely public. Using a unique Chinese dataset of free and public analyst reports and applying the rule-based textual parsing technique, this chapter investigates the relationship between the abnormal stock return and the corresponding financial news and analyst reports. I find strong evidence that the textual information in the news affects the adjusted overnight return quickly after the news release and there is no effect after. It takes longer for the textual information in the analyst reports to affect the abnormal return; the reports do have significant predictability on a four-week horizon, but not immediately after the report release. The simulation shows that utilizing the market underreaction is a profitable trading strategy. I therefore reject the hypothesis that the underreaction to the analyst reports is because they are not public.

In this paper, we build the rationale of the financial intermediate's decision of making loans to potential home buyers over an infinite time horizon. In the first period "good" borrowers with stable future income flows receive loans and buy homes. In later periods, the intermediate securitizes the loans to raise new capital and makes loans to some of the "bad" borrowers with uncertain future income flows. Currently, we simplify the securitization as a tool to raise capital without cost over time. This unrealistic simplification should be improved in later work. The financial intermediate calculates the expected payoffs in different scenarios under the realizations of uncertainty to decide whether to make loans to a new borrower and whether to liquidate a house if the owner is short of liquidity in the short run. After clarifying the sequence of moves of different agents within each period, we compute the financial intermediate's decision rule described by a Bellman equation. Then we simulate borrowers' income realization and produce a figure of house price as well as value function over time.