The three essays in this dissertation examine questions in the U.S. Treasury bond market and an asset-pricing anomaly in the stock market. They are unified by the theme relating quantities of assets with asset prices. These essays challenge the notion that because financial markets are highly liquid, quantities of (demand or supply) do not matter for predicting asset prices. Chapter 1 examines the time-varying impact of the US Treasury debt supply on bond risk premiums. I find that the elasticity of bond risk premium with respect to supply depends on the correlation between stock and bond returns. An increase in the supply of Treasury bonds raises the required bond risk premiums, but the effect is stronger as stock and bond returns become more positively correlated. I interpret this evidence within the context of a preferred-habitat asset pricing model where the arbitrageurs are the marginal investor for all bond maturities. Arbitrageurs demand higher compensation for maturity risk when the stock-bond correlation is positive as bonds are poor hedges for stocks. On the other hand, when the correlation turns more negative, an increased bond supply induces low or even negative risk premiums. The findings have practical implications for understanding the impact of the impending Federal Reserve;;s unwinding of its $4.5 trillion bond portfolio.Chapter 2 documents new empirical stylized facts about the postwar U.S. Treasury debt management policy. In particular, I document the puzzling fact that the US Treasury has tended to historically issue more long-term debt when the term spread is greatest. I propose a simple model that captures the practical incentives and constraints faced by the Treasury debt manager.The debt manager seeks to minimize borrowing costs while managing rollover risks. I calibrate the model to generate a measure of time-varying roll-over risks faced by the U.S. Treasury.Chapter 3 investigates the earnings announcement premium puzzle in Finland. Between 1999-2002 and 2006-2009, I find that stocks with earnings announcement earn excess returns over non-announcement stock in the 2 week window before the announcements that quickly dissipates post-announcement. Moreover, I find that the premium is significantly higher and persistent through a 30 day window around the financial statement releases. I find no premium around the interim earnings report and in fact accumulative losses. I also assess the relationship between announcement premium and trading volume. Using an administrative transaction-level data set, I find some supportive evidence for the attention-grabbing hypothesis. I find a positive correlation between the announcement premium and the net-buying trading volume among individual investors, especially around the financial statements.
【 预 览 】
附件列表
Files
Size
Format
View
Essays on the U.S. Treasury Debt Market and Asset-Pricing