Quantitative easing and forward guidance: what impact on financial markets? (Research of the month)

Eric Swanson, “Measuring the effects of Federal Reserve forward guidance and asset purchases on financial markets, » Journal of Monetary Economics, 2021

Abstract:

– In this article, Eric Swanson attempts to distinguish the effects of conventional monetary policy, quantitative easing, and forward guidance on different financial assets.

– His results show that unconventional monetary policy tools (quantitative easing and forward guidance) are just as effective as conventional monetary policy in changing financial asset prices, and that the effects are just as durable.

Although theFederal Reserve (Fed) began reducing its asset purchase program in November 2021, the « unconventional » monetary policy (UMP) instruments used since the 2007-2008 financial crisis are now an integral part of the Fed’s toolkit. These tools mainly include large-scale asset purchases (particularly Treasury bills) and forward guidance .[1] Numerous studies have concluded that these tools, particularly large-scale asset purchase programs, are generally effective: they stimulate economic activity by lowering long-term interest rates. Why do they have this effect? Is it because they send an important signal, namely that central bank short-term rates will remain low for a very long time (as long as the asset purchase program is not completed)? Or is it because asset purchases have an effect in themselves[2], to which the markets react? Or is it because both aspects are at work simultaneously? If so, does one of these two aspects dominate, meaning that central bankers could modify their monetary policy strategy to make it more effective?

Swanson, in his article entitled « Measuring the effects of Federal Reserve forward guidance and asset purchases on financial markets » (2021), answers these questions by assessing the effectiveness of these two monetary policy instruments.

Technical difficulties in measuring monetary policy instruments

To assess the effects of monetary policy on various variables, researchers must first identify monetary policy « surprises, » i.e., unexpected reactions by financial markets to announcements by central bankers. Why is it important to capture the « surprise » component in monetary policy announcements? Simply because central bankers’ decisions are influenced by the economic environment: when we capture the part of the announcement that was not anticipated by financial markets, it reflects only the markets’ response to the new monetary policy measures.

To capture this « surprise » component, Swanson collects data over a very short time interval (30 minutes), which corresponds to the response interval of interest rates of different maturities around monetary policy announcements. The idea is that only the FOMC announcement can influence their price movements within such a narrow time interval.

A major difficulty then arises in distinguishing the effect of the two aspects mentioned above: when an asset purchase program is announced, it is automatically accompanied by forward guidance , either explicitly or, very often, implicitly for the markets (Table 1).

Table 1: Examples of FOMC announcements combining forward guidance and quantitative easing tools

How, then, can we separate two intrinsically linked elements in the data? How can we know whether the measured effect is due to this signaling effect or to the transmission channels specific to quantitative easing? [5]

To do this, the economist first uses factor analysis. The principle is to synthesize the information contained in a large database into a small number of new variables (called « factors »). For example, if the responses for long-term rates (10, 20, 30 years) are highly correlated with each other, a new variable will be created summarizing the information from these responses within a first factor.

This allows the author to obtain variables describing monetary policy instruments, as it is impossible to measure them directly. In this article, the effects of monetary policy announcements on the entire yield curve are well summarized by three factors. These are, naturally:

• Conventional monetary policy;
• Quantitative easing; and
• Forward guidance.

These are the three main tools used by the Fed.

There is a slight complication here: once the factors have been obtained, it is sometimes necessary to impose certain restrictions to ensure that they accurately describe each monetary policy instrument.

To do this, we use the differences in the effects of monetary policy tools on the yield curve, as each instrument has a more or less pronounced impact depending on the maturity of the rates. For example, Swanson imposes the following restriction: forward guidance has no effect on the effective Fed funds rate[6]. This ensures that the  » forward guidance  » factor is separated from the « conventional monetary policy » factor. The three variables created above then have the expected effects on the various interest rates, ensuring that monetary policy instruments are correctly identified.

However, it is important to verify this by also observing whether the evolution of these factors is consistent with the monetary policy announcements concerning them.

This appears to be the case in this study (see graph above, taken from the article). Indeed, if we take the  » quantitative easing  » factor factor (LSAP factor, in orange above), it evolves in line with the FOMC announcements concerning it: it decreases sharply when the^first quantitative easing ( accommodative monetary policy) is announced, and increases when the reduction of the asset purchase program ( » taper tantrum » ) is announced. This characterizes a restrictive policy.

The effects of these various monetary policy tools on financial markets

Once the various monetary policy tools have been identified, the researcher can then study the impact of each of them on financial assets. The author first focuses on the very short-term effects of quantitative easing and forward guidance, using a very simple econometric model consisting of regressing changes in 2-, 5, 10, and 30 years—within a 30-minute window around the monetary policy announcement—on forward guidance and quantitative easing measures. The interest variable is then replaced by the euro/dollar exchange rate, movements in the S&P 500, and corporate bond rates. The author concludes that each component of the FOMC announcement has an impact on rates. The effect of forward guidance is mainly concentrated on 2- to 5-year maturities. The effect of quantitative easing is concentrated on longer maturities, with a peak on 10-year maturities. Swanson demonstrates that these two tools are « almost as effective as conventional monetary policy. » In fact, he finds that a surprise announcement of a $100 million asset purchase program would lower the 10-year rate by about 3 basis points. Under normal circumstances, a surprise 25 basis point cut in the Fed Funds rate would, in itself (i.e., once the associated forward guidance effect is isolated), have the same average effect of 3 basis points on the 10-year rate.

Are these effects persistent?

This is a common question in academic circles. We know, for example, that quantitative easing lowers rates on the day it is announced, but the effect may dissipate in the following days, with serious theoretical arguments supporting this view. Determining whether the effect persists is not as simple as it might seem: it is not enough to simply look at whether rates have risen after a QE announcement, since by definition other news may occur that causes rates to rise. The question is whether what happened on day D, the « treatment injected on day D, » is still present in the days to come.

To do this, the author uses special econometric methods (called « local projections »), focusing on cumulative returns over several days. This amounts to looking for a link between, for example, the 10-day return and the monetary policy announcement on day D, the 11-day return and the announcement, the 12-day return and the announcement, etc. He chooses a 120-day horizon. Using a large number of observations and econometric methods reduces the influence of news affecting yields after the announcement, and therefore provides a more relevant tool than a simple graphical observation. The estimates seem to show that the effect of each of these factors is lasting. Indeed, 120 days after an announcement, the effect of quantitative easing, like the effect of forward guidance, can still be seen in 5- and 10-year rates.

Conclusion

Assessing the effects of monetary policy on economic and financial variables is a complex task, particularly since the proliferation of monetary policy tools used since the 2007/2008 financial crisis.

However, assessing the effectiveness of these measures and their potential costs is essential for central banks. In his article, Swanson distinguishes between the effects of conventional monetary policy, forward guidance, and quantitative easing. He shows that these three components have a lasting effect on different assets, and that quantitative easing and forward guidance are just as effective as conventional monetary policy.

Article co-authored with Julien Pinter