Expectations of Large-Scale Asset Purchases

During and after the recent financial crisis, the Federal Reserve took several unprecedented actions in an attempt to bolster the economy. Before the crisis, monetary policy typically consisted of the Federal Open Market Committee (FOMC) setting a target for the federal funds rate, the overnight interest rate at which banks lend to one another. Once the federal funds rate reached its effective zero lower bound, however, the FOMC turned to a number of unconventional tools to stimulate the economy. One such tool, large-scale asset purchases (LSAPs)--often referred to as quantitative easing (QE)--consists of the Federal Reserve purchasing U.S. Treasury securities and agency mortgage-backed securities (MBS) with the aim of driving down longer-term interest rates, thereby stimulating economic activity.

Because LSAPs are an unconventional tool for the FOMC, their effectiveness remains uncertain due to several factors. First, because the use of LSAPs has been limited, any assessment of their impact relies on relatively few observations. Second, after the financial crisis, many factors had substantial impact on the U.S. economy--including the European debt crisis, changes in U.S. fiscal policy, and the use of other unconventional FOMC tools such as forward guidance for the federal funds rate. Thus, isolating the impact of LSAPs is difficult. Third, the extent to which markets expected the FOMC's announcement of LSAPs is uncertain, and these market expectations can affect estimates of LSAP efficacy.

Attempts to estimate the effectiveness of LSAPs often use an "event study" approach. This approach focuses on a short interval around key program announcements, which isolates the impact of the announcements and documents interest rate movements during that interval. However, event studies often ignore the possibility that financial market participants expected the announcements. Such expectations can affect interest rates well ahead of the interval considered.

This article presents evidence from surveys of market participants, news articles, and Internet searches that indicate financial markets expected LSAP announcements prior to the various Federal Reserve programs. The article concludes that, because event studies inadequately control for expectations, they likely understate the effects of LSAPs. The first section of the article discusses the event study approach and how it typically excludes the effects of expectations. Section II presents evidence on expectations from several sources: surveys, news stories, and Internet searches. Section III discusses how expectations affected interest rates before Federal Reserve announcements of recent asset purchase programs.

I. EVENT STUDIES AND EXPECTATIONS

In theory, asset purchases stimulate the economy by lowering interest rates at various time horizons. When the Federal Reserve purchases either Treasuries or agency MBS, which are securities that represent claims to the cash flow of sets of mortgages, this additional source of demand for these assets pushes up their price, hence lowering their yield. As a result, the interest rates associated with these assets fall. The decline in interest rates pushes investment from Treasuries or MBS into other sectors of the economy, which in turn stimulates economic activity and drives growth. Lower interest rates also make mortgages less expensive, fueling home purchases.

Since the start of the financial crisis, the Federal Reserve has implemented several rounds of LSAPs. Table 1 shows the four major programs along with key announcement dates. (l) The first round, labeled QE1, came in November and December 2008, followed by an extension in March 2009. The second round, QE2, officially was announced in November 2010. However, many people consider a speech three months earlier by Federal Reserve Chairman Ben S. Bernanke in Jackson Hole, Wyoming, as having already provided a strong signal that the FOMC would pursue additional LSAPs. The Jackson Hole speech, part of the Economic Policy Symposium conducted by the Federal Reserve Bank of Kansas City, had been delivered in August 2010. After QE2, a third round of purchases, called the Maturity Extension Program (MEP), used the proceeds from sales of short-term Treasuries to finance longer-term Treasury purchases. The MEP began in September 2011 and was extended in June 2012. The fourth and latest round, QE3, was announced in September 2012 and extended in December 2012.

Estimating the effects of asset purchases on interest rates often relies on an event study. Event studies typically focus on an LSAP announcement, tracing interest rate changes over a short time frame around the announcement. Assuming no other major economic news, announcements, or developments occur within the given time frame, only the LSAP announcement would be expected to produce interest rate movements, and any such movements are therefore attributed to the LSAP.

Much of the literature uses the Federal Reserves early LSAP programs to estimate the impact of their announcements on interest rates. Table 2 shows several estimates of how asset purchases affected 10-year Treasury interest rates. The estimates vary depending on the program and the exact methodology, but the negative values all suggest asset purchases lowered long-term interest rates, as theory predicts. For example, Gagnon and others estimate the effects of QE1 by measuring the cumulative change in interest rates from the end of the day prior to an LSAP announcement to the end of the day of the announcement. Because interest rates on 10-year Treasuries declined up to 91 basis points during the announcement days, Gagnon and others conclude the announcement caused this reduction.

Choosing an adequate time frame for event studies presents a trade-off. On one hand, too large a window allows contamination from other news or economic developments. On the other hand, too small a window potentially ignores some of the effects of purchases.

The time frames chosen in event studies may lead to a misestimation of the effects of LSAPs in two ways. First, they may capture only the immediate response by markets to an announcement rather than the long-run effects. The immediate market response may be an initial overreaction that is offset or extended some time later with a partial reversal, or it may be an initial underreaction that is extended some time later with further movement in the same direction. Either way, an event study focusing on a time frame that captures only the initial reaction and not the subsequent correction may misestimate the total impact of the announcement.

Second, event studies typically ignore how the markets' expectations of an announcement may have altered interest rates well before the time frame under consideration. Event studies that examine a window of time starting at the time of the announcement, or just briefly before it, may ignore the possibility that interest rates at the time already reflect markets' expectations that an announcement might occur. A hypothetical measure of the full impact of asset purchases on interest rates would need to include the pre-announcement effects stemming from advance expectations, as well as the subsequent response to an actual announcement. In an extreme case where all market participants fully expect an announcement, interest rates will begin to fall in advance of the announcement. The decline in rates will occur as traders, anticipating heightened demand for bonds, begin buying bonds and thus bidding up the bonds' prices, which move inversely with their yields. By the time the LSAPs are announced officially, their impact already will have been factored into bond prices and yields--and thus interest rates--and the actual announcement will have no further effect. …

To read the full text of this article and others like it, subscribe today!