This thesis includes two essays on equity analysts’ agent role and one essay on investors’ inattention to good news. From a broader economic perspective, equity analysts are essentially agents acting on behalf of multiple principals including their employers, investors and issuers (Fisch & Sale, 2003). Classic agency theory predicts that analysts selectively provide coverage and report their expectations. In the first essay, I examine empirically if incremental investment value can be uncovered from analysts’ choices between silence and speech, measured as the level of analyst reporting not explained by size or turnover. I find that “silence” negatively, and “speech” positively predicts future stock returns. More importantly, as “speech is silver, silence is golden”, the observed price shift is mainly driven by silence, providing evidence that analysts’ inaction can impede price discovery process. This is consistent with the claims that analysts’ expectations are based on valid information, that analyst self-selection is pervasive due to the principal-agent conflicts, and that the loss of information with analyst silence has resulted in some mis-valuation which can be viewed as a form of classic agency cost. The second essay tests if analysts are systematically less forthcoming in reporting bad earnings news when the principal-agent conflicts are exacerbated. I find that analysts’ downward consensus earnings forecast revisions are less informative than their upward revisions; that less is more when analysts report bad news - extreme downward revisions contain little incremental information beyond momentum compared with moderate downward revisions; and that the differential richness of information in good and bad news revisions is more pronounced among bigger, more heavily covered stocks and stocks with higher institutional holdings, namely, stocks that are typically more prone to the analyst agency problem. Thus the loss of information in bad news revisions and extreme bad news revisions’ lagging behind price action can be viewed as another form of agency cost. In the third essay, I investigate how negativity bias in information processing affects the positive-negative-asymmetry in the stock price continuation phenomenon. Psychology literature document that negative stimuli elicit more attention and negative information is generally processed more thoroughly and is weighed more heavily in impression formation, memory, learning and decision making than positive information (Baumeister, Bratslavsky, Finkenauer, & Vohs, 2001; Rozin & Royzman, 2001). Insofar as people are cognitive misers, all else being equal, investors tend to pay less attention to good news than to bad news. Using earnings announcement as the information shock, I document evidences that investors incorporate bad earnings news to fuller extent than they do with good earnings news. Furthermore, given that psychological biases are typically increased when there is more uncertainty (Hirshleifer, 2001) and ambiguity or uncertainty is often associated with higher risk and the possibility of hostile manipulation, I also find more pronounced asymmetry in post announcement drift when information uncertainty is greater. / published_or_final_version / Business / Doctoral / Doctor of Philosophy

Financial contagion, in this study, refers to spreading of crisis across markets in different locations. The observable consequence is usually in the form of increase in co-movement of asset prices in two markets after a crisis event. The causes of financial contagion have been studied for over twenty years, however, up till now, results have been mixed. One unsettled issue is whether market fundamentals alone can explain financial contagion. Pure fundamental based explanation suggests that the financial, economic and trade linkages are solely responsible for the transmission of crisis across markets. On the other hand, the behavioral finance researchers propose that herding behavior also plays an important role in explaining financial contagion. This issue cannot be easily resolved since it is difficult to empirically distinguish linkage effect and herding behavior. This thesis contributes to this unresolved issue by examining financial contagion in the stock market and indirect real estate market. In the stock market, both fundamental linkages and herding are likely to exist. However some securities are less prone to herding than others. Herding across international markets is likely to be less serious when there is less information asymmetry between investors and management. In addition, compared with foreign investors, local investors are more confident in the link between market fundamentals and the corresponding securities. Real Estate Investment Trusts (REITs) are likely to suffer from less information asymmetry problem since the REITs market has more stringent regulatory requirements for information disclosure. Furthermore, the pricing of real estate asset, the main type of assets held by the REITs, often requires local knowledge. Local investors investing in REITs are less likely to mimic the investor behavior in another overseas REITs market. Listed property companies also share some similarities with REITs, although they are less immune to herding compared with REITs as information disclosure is less stringent for listed property companies. Since the asset prices of real estate are affected by the economic performance, fundamental linkages amongst all indirect real estate still likely to exist and are similar to other types of listed companies. If market fundamental is the only source of financial contagion (i.e. no herding), financial contagion in the global stock and indirect real estate markets should be similar. This thesis uses the 2008 global financial crisis (GFC) as the crisis event to examine financial contagion across the world’s major equity markets. Our empirical results show that financial contagion is stronger in the entire stock markets than in the indirect real estate markets and that financial contagion is the weakest in the REITs markets, which support the herding behavior hypothesis and reject the pure fundamental explanation. This reasoning does not require indirect real estate to be totally immune from herding. All that is needed is that indirect real estate is less prone to herding compared with the common stocks. Herding behavior can be rational or irrational. The latter refers to revision of asset prices by following the pricing behavior of other markets irrespective of market fundamentals. Our empirical evidence cannot reject irrational herding behavior in the indirect real estate market since contagion effect becomes stronger when windows of observations are lengthened. That is when more time was allowed for investors to react to the pricing behaviors in other markets, financial contagion became stronger. However, no similar results were found in the stock market. This impl\ies that compared with the indirect real estate market, herding is more serious in the stock market but such herding is also more rational than that in the indirect real estate market. / published_or_final_version / Real Estate and Construction / Master / Master of Philosophy