By Jonathan Wright1 and Larry Meyer
In a footnote to a recent speech, Yellen provided an empirical estimate, based on Federal Reserve staff research, of the effect of balance sheet normalization on long-term yields, and inferred from that how many rate moves it would substitute for.2 A very important and timely topic. We have questions about the link between the estimate she presents and the empirical research she cites and also about the size of the balance sheet developments she refers to.
The footnote follows this section from Yellen’s speech:
The downward pressure on longer-term interest rates that the Fed’s asset holdings exert is expected to diminish over time–a development that amounts to a “passive” removal of monetary policy accommodation. Other things being equal, this factor argues for a more gradual approach to raising short-term rates.
Here is footnote 17 (italics our emphasis):
Based on estimates generated using the term-structure model developed by Li and Wei (2013) and the procedure discussed in Ihrig and others (2012) and extended by Engen, Laubach, and Reifschneider (2015), the Federal Reserve’s holdings of Treasury securities and agency mortgage-backed securities continue to put considerable downward pressure on longer-term interest rates. However, this pressure is estimated to be gradually easing as the average maturity of the portfolio declines and the end-date for reinvestment draws closer. Over the course of 2017, this easing could increase the yield on the 10-year Treasury note by about 15 basis points, all else being equal. Based on the estimated co-movement of short-term and long-term interest rates, such a change in longer-term yields would be similar to that which, on average, has historically accompanied two 25 basis point hikes in the federal funds rate.
Yellen says that a 15-basis-point rise in the ten-year yield could result this year as the “average maturity of the portfolio declines and the end-date for reinvestment draws closer.” Two separate effects are apparently being referred to here. First, she is noting that, even with reinvestments, the SOMA average maturity is drifting down because the investments are at a shorter maturity than the existing stock. The qualitative story here is fine, but this dynamic has been going on since 2014, so there’s been an upward effect on longer-term yields for some time. It did not begin this year. It would be important to consider the cumulative effect. The second source of the impact on the ten-year yield is the anticipation of the end of the reinvestment program.
1 Jonathan Wright is a Professor of Economics at Johns Hopkins University and a Senior Adviser to LHMeyer. 2 Chair Janet L. Yellen at the Stanford Institute for Economic Policy Research, Stanford University, Stanford, California, January 19, 2017: The Economic Outlook and the Conduct of Monetary Policy.
As it gets closer, the anticipation effect rises. That’s fine, too, qualitatively. And it likewise may have been going on for years.
Yellen cites the Li and Wei model as underlying her calculation of a 15-basis-point effect. That model would imply that a $300-billion reduction in ten-year Treasury equivalents in the SOMA portfolio would raise the ten-year yield by about 15 basis points. A $300-billion drop in ten-year equivalents in 2017 is a puzzlingly large decline, if reinvestments are being maintained throughout the year. It is easier to understand this number if the run-off is assumed to begin this year, but it does not appear that this is what Yellen was referring to.
Moreover, it seems likely that the impact of a passive decline in ten-year equivalents should be smaller than the impact of large-scale outright purchases, whereas the methodology suggested in the footnote treats them as likely to have effects of equal magnitude and opposite sign.
Most importantly, the footnote might be understood by some as saying that a 15-basis-point increase in the ten-year yield has the same effects as a 50-basis-point increase in the funds rate. If so, that’s way too big. The literal statement is that a 50-basis-point increase in the fed funds rate has historically raised the ten-year yield by 15 basis points. But if we are talking about an increase in the ten-year yield without any corresponding increase in the fed funds rate, that has a much smaller impact on aggregate demand. Short rates matter, too. The idea that balance sheet developments this year, in the absence of an end to the reinvestment program, could substitute for two 25-basis-point hikes seems implausible to us. Did participants know this when they made their rate projections in December? If they did not know this when they made their projections in December, but know it now, then the dots would move down to just one hike in March! We doubt that very much.
We always assumed that the ten-year term premium would rise as we got closer to the expected end of the reinvestment program. While not the apparent purpose of Yellen’s footnote, it nevertheless highlights the importance of explicitly taking into account the effect of the runoff of the Fed’s portfolio on longer-term rates and hence on the number of funds rate hikes necessary to remove the monetary accommodation that prevails today. In LHMeyer’s most recent forecast, this consideration led us to cut the number of hikes in 2018 from four to three.