Book Description

This dissertation consists of three essays in the field of monetary and financial economics. Specifically, we use high-frequency financial data to study monetary policies with a focus on the information effect, namely, that some of the interest rate movements around central bank announcements are not policy-driven, but are results of the market becoming aware of the central bank's view about future economic prospects. Understanding the role played by the information effect will help us apprehend monetary policy implications in both normal times and extraordinary situations. Chapter 1 evaluates the impact of unconventional monetary policy in the newly developed instrumental variable structural Vector Autoregression (VAR) framework. In the current low interest rate environment, central banks must resort to using unconventional monetary policies, such as forward guidance and quantitative easing, to flight recessions. To empirically evaluate the effectiveness of these unconventional policies, we need to rely on the clean policy shock. A prominent concern is that the often used high-frequency interest rate surprises not only reflect unexpected policy changes, but also contain the information effect. We contribute to the literature by using a heteroskedasticity identification approach, taking advantage of changes in the relative dominance of economic shocks around different macroeconomic announcements. Analysis based on clean policy shocks suggests that the unconventional policies successfully aided the recovery in the U.S. More importantly, we show that the information effect, while it may introduce bias, is rather modest when it comes to estimating the real impact of unconventional monetary policies. Chapter 2 studies the stock return pattern after the U.S. Federal Open Market Committee (FOMC) announcement. This research is motivated by recent literature that documents stock returns drifts, both before and after FOMC announcements, according to policy rate surprises. Indeed, research has shown that the information contained in the central bank announcement is multifaceted: its current monetary policy stances (monetary policy news) and news about future economic prospects (non-monetary policy news). Our contribution is to combine these two strands of literature. To the best of our knowledge, no study has looked at stock market reactions to the non-monetary news stemming from policy announcements. We identify both good and bad news events using a combination of sign restriction with high-frequency financial prices. The novel finding is that following bad FOMC announcements, that is the market interpreted the Fed announcements as revealing negative information about the economy, we observe significant positive stock returns in a 20-day period. We call this the ``post-FOMC drift.'' Further analysis suggests that the drift is likely caused by relatively heightened risks associated with bad announcements, although the drift is consistent with market overreactions as well. Moreover, the post FOMC drift is a market-wide phenomenon and can be exploited in an easy-to-implement trading strategy with a historical record of earning 40\% of the annual equity premium. In Chapter 3, we explore the channels through which the FOMC announcements affect the financial market. While much of the existing literature measures the surprise components with only changes in policy rates (surrounding the announcement), we contribute to the existing literature by taking a broader view through examining unexpected changes in longer-term yields, corporate credit spreads, and inflation expectations (a proxy for growth prospects), using high-frequency financial data. Through a regression analysis, our findings show that these additional surprises provide orthogonal information and sharply increase the goodness of fit in explaining stock returns around FOMC announcements, with the inclusion of inflation expectations having the biggest contribution. The important role of inflation expectation suggests that the current literature, which uses stock prices together with nominal rates to disentangle the information contents of central bank announcements, may be too limited in the scope of information it uses.