In this commentary, we discuss the evolution of the term premium embedded in the ten-year Treasury yield after asset purchase programs were completed and prospects once the balance sheet begins normalizing. We distinguish between the initial impact of each balance sheet program on the term premium and the cumulative effect on the term premium of all the balance sheet programs, taking into account the diminishing effect on the term premium of each program over time and ultimately the end of the current reinvestment policy.1
The Initial Impact and Cumulative Effects of the Fed’s Balance Sheet Programs
Asset purchase and maturity extension programs have an initial impact on longer-term interest yields (for instance, the ten-year Treasury yield) principally by compressing term premia. The initial effect of each program begins to diminish immediately after it is completed. The cumulative effect of these programs is the remaining effect on the term premium at any point in time after the end of the program. We focus on the cumulative effect of all the programs, starting with the completion of QE3 in October 2014.
We use estimates of the cumulative effect from a Fed staff paper by Eric Engen, Thomas Laubach, and David Reifschneider.2 They use the same methodology as a Fed staff paper by Ihrig, Klee, Li, Schulte, and Wei, and extend the latter’s results from QE2 to QE3.3
The Cumulative Effect of QE Programs on the Term Premium Diminishes Over Time
The Fed’s asset purchases have expanded the size of the balance sheet and raised the average maturity of its portfolio, while its maturity extension programs raised the average maturity without increasing the size of the balance sheet.4 The remaining maturity of each security in the portfolio naturally declines with the passage of time, resulting in a decline in the average maturity of the portfolio when there are no active programs to acquire longer-dated securities.5 This is one source of the diminishing effect of the balance sheet programs
1 This commentary is the joint product of LH Meyer and Jonathan Wright, an LH Meyer senior adviser.
2 Engen, Eric M., Thomas Laubach, and David Reifschneider (2015). “The Macroeconomic Effects of the Federal Reserve’s Unconventional Monetary Policies,” Finance and Economics Discussion Series 2015-005. Washington: Board of Governors of the Federal Reserve System, January.
3Ihrig, Jane, Elizabeth Klee, Canlin Li, Brett Schulte, and Min Wei (2012). “Expectations about the Federal Reserve’s Balance Sheet and the Term Structure of Interest Rates,” Finance and Economics Discussion Series 2012-57. Washington: Board of Governors of the Federal Reserve System, July.
4 Both size and maturity distribution dimensions are captured by measuring changes in the balance sheet in terms of ten-year equivalents. In terms of ten-year equivalents, the Fed balance sheet went from roughly $300 billion in early 2009 to a peak of $3.1 trillion in March 2014 and was $2.7 trillion at the end of 2016.
5 Note that the average maturity of the portfolio diminishes even under the current reinvestment program because the Treasuries purchased by the Fed under that program have had (roughly) the same maturity distribution as total new Treasury issues, which have a lower weighted average maturity than the overall Fed balance sheet.
on the term premium. The cumulative effect also erodes as the expected end of the reinvestment program draws nearer.
Figure 1. Projection as of 2013:Q4 of the Cumulative Effect of
Balance Sheet Policies on the Ten-Year Treasury Term Premium
In Figure 1, we show the estimated path from Engen et al. of the projected cumulative effect of the Fed’s balance sheet programs on the term premium, beginning in the fourth quarter of 2013, when the cumulative effect peaked near 120 basis points. The projected cumulative effect is estimated to have declined to about 100 basis points in 2014:Q4, the quarter in which QE3 concluded. The paper estimated that the projected cumulative effect would only be 57 basis points at the end of 2016. When we turn in a forthcoming commentary to estimating the number of 25-basis-point hikes that balance sheet developments will substitute for in 2017 through 2019, we will use these estimates of the diminishing impact on the term premium in each of those years, shown in Table 1.
Table 1. Projection as of 2013:Q4 of the Reduction in the Cumulative Effect of
Balance Sheet Policies on the Ten-Year Treasury Term Premium (basis points)
|Point in Time||Cumulative Effect||Change from Year Ago|
Source: LH Meyer, Engen et al. (2015).
By the end of 2019, it is estimated that the projected cumulative effect of the balance sheet programs will have faded by 90 basis points relative to the peak effect, or by 80%, leaving very little erosion to take into account beyond 2019.
Assumptions, Assumptions, Assumptions
Projecting the impact of balance sheet policy on term premiums requires making assumptions about the timing of the end of the reinvestment program, how the FOMC will choose to shrink its balance sheet after the end of reinvestment, and what the size and composition of the normalized balance sheet will be. Figure 1 shows the estimated projected trajectory of term premium effects as of 2013:Q4, and it reflects market expectations at that time. As examples of how market expectations changed, the expected timing of the end of reinvestment was again and again moved later and the Fed’s guidance was interpreted as indicating that the normalized portfolio would feature a level of reserves that would have been viewed as normal in the pre-crisis operating regime. The fact that the eventual quantity of reserves will be sharply higher than assumed in Engen et al. means that the term premium when the balance sheet is normalized will end up lower than implied by their estimates in that paper.
The estimates also take into account the actual and projected issuance of Treasuries. The studies assume that the Treasury did not change its strategy for the maturity composition of new issues in response to the FOMC’s asset purchase programs. While it is true that the new issuance strategy had the opposite effect on the term premium as the balance sheet programs, this strategy was in place long before the balance sheet programs were initiated. The marginal effect of the balance sheet policies was still to reduce the term premium relative to what it otherwise would have been. Despite all these questions about the assumptions, we nevertheless view these as the best estimates available today. Indeed, these are the estimates that Fischer and Yellen have referenced when they have talked about the cumulative effect of balance sheet policies and how much the cumulative effect would fade going forward.6 This suggests that these may be the estimates that are likely, at least initially, to guide the FOMC’s expectations about the path of the term premium over the next few years and provide some guidance into how many rate hikes the diminishing effect will substitute for, a topic we will turn to in a forthcoming commentary.