The extreme experiment of current US monetary policy has evolved (as we noted yesterday [15]), from explicit end-dates, to unlimited end-dates, to threshold-based end-dates. Of course, this 'threshold' was no problem for the liquidty whores when unemployment rates were extremely high themselves, but as the world awoke to what we have been pointing out - that it's all a mirage of collapsing participation rates - the FOMC (and sell-side strategists) realized that the endgame may be 'too close'. Cue Goldman's Jan Hatzius, who in today's note, citing two influential Fed staff economists, shifts the base case and forecasts that the Fed will lower its threshold for rate hikes to 6.0% (and perhaps as low as 5.5%) as early as December (as a dovish forward-guidance balance to an expected Taper announcement).
Via Goldman Sachs,
- The most senior Fed staff economists for monetary policy analysis and domestic macroeconomics, William English and David Wilcox, have published separate studies that imply a strong case for a reduction in the 6.5% unemployment threshold for the first funds rate hike. We have proposed such a move for some time, but have been unsure whether it would in fact happen. And while the uncertainty around near-term Fed policy remains very considerable, our baseline view is now that the FOMC will reduce its 6.5% threshold to 6% at the March 2014 FOMC meeting, alongside the first tapering of QE. A move as early as the December 2013 meeting is possible, and if so, this might also increase the probability of an earlier tapering of QE.
It is hard to overstate the importance of two new Fed staff studies that will be presented at the IMF's annual research conference on November 7-8. The lead author for the first study is William English, who is the director of the Monetary Affairs division and the Secretary and Economist of the FOMC. The lead author for the second study is David Wilcox, who is the director of the Research and Statistics division and the Economist of the FOMC. The fact that the two most senior Board staffers in the areas of monetary policy analysis and domestic macroeconomics have simultaneously published detailed research papers on central issues of the economic and monetary policy outlook is highly unusual and noteworthy in its own right. But the content and implications of these papers are even more striking.
It will take us some time to absorb the sizable amounts of new analysis in the two studies, and we are only able to comment on a few selected aspects at this point. But our initial assessment is that they considerably increase the probability that the FOMC will reduce its 6.5% unemployment threshold for the first hike in the federal funds rate, either coincident with the first tapering of its QE program or before.
The first study, written by William English, David Lopez-Salido, and Robert Tetlow and entitled "The Federal Reserve's Framework for Monetary Policy--Recent Changes and New Questions," [16] uses a smaller version of the staff's large-scale econometric model FRB/US to analyze the optimal path for the federal funds rate. Using "small FRB/US," a set of assumptions about Fed preferences, and a set of assumptions about the baseline performance of the economy, the authors find that the theoretically optimal policy involves a commitment to hold the federal funds rate near zero until 2017, followed by a series of hikes that push the rate well above neutral by the early 2020s. In this simulation, the unemployment rate falls below the structural rate for a time, and inflation rises modestly above the 2% target. (The optimal policy in the English et al. study is more aggressive than that shown in Vice Chair Yellen's earlier set of optimal control simulations, which points to the first hike in early 2016; the reasons seem to include a lower assumption for the structural unemployment rate and a later baseline for the first hike in the funds rate.)
However, the authors note that such an optimal policy is possibly infeasible because it is complex and model-dependent, and because it simply assumes that policymakers are able to overcome the credibility problems associated with a commitment to a particular policy path far in the future. Hence, they investigate several different ways in which Fed officials might be able to approximate the optimal policy: (1) different sets of unemployment and inflation thresholds for the first hike, (2) a higher inflation target, and (3) a switch to a nominal GDP level target. They see potentially sizable benefits from a higher inflation target or a nominal GDP level target but also very sizable risks, and conclude that "it is hard to be confident" that such a change would enhance performance.
Regarding the unemployment and inflation thresholds, they simulate the performance of unemployment thresholds for the first hike ranging from 5.0% to 7.0% and inflation thresholds ranging from 1.5% to 3.0%. The results with respect to varying the inflation threshold are not very surprising. The authors find that the current 2.5% threshold performs no worse than other choices, and the differences are relatively minor (at least in the baseline simulation). But the results with respect to varying the unemployment threshold are much more striking. The key conclusion is that "…reducing the unemployment threshold improves measured economic performance until the unemployment threshold reaches 5.5 percent; a further reduction in the threshold to 5.0 percent, however, reduces welfare, as the control of inflation becomes notably less precise." In other words, a 6% unemployment threshold outperforms a 6.5% threshold, and a 5.5% threshold outperforms a 6% threshold!
The second paper, written by David Reifschneider, Willam Wascher, and David Wilcox and entitled "Aggregate Supply in the United States: Recent Developments and Implications for the Conduct of Monetary Policy," [17] discusses the considerable evidence that the US economy's supply-side performance has deteriorated significantly since 2007, with very slow potential GDP growth and an increase in structural unemployment. They estimate that real potential GDP growth has only averaged 1.3% since 2007, the output gap is currently about 3% of GDP, and the structural unemployment rate had risen to 5.75% by 2012 (although it is now again on a slight downward trend). They then use a modified version of FRB/US with an added role for "hysteresis" in labor markets--that is, a gradual transformation of cyclical unemployment into structural unemployment and/or labor force withdrawal --to analyze the sources of this deterioration, using a simulation in which the model economy is hit by a major financial crisis that is calibrated to match the size of the 2007-2009 episode. In a nutshell, they find that the post-crisis period "features a noticeable deterioration in the economy's productive capacity" and that about 80% of the deterioration "…represents an endogenous response to the persistently weak state of aggregate demand."
The authors then go on to discuss the "optimal" policy response to the crisis, taking into account not just the cyclical position of the economy but also the implications of weakness in aggregate demand for the supply side and therefore the longer-term structural position of the economy. They are careful to note that optimal control solutions are sensitive to the specification of the model economy, assumptions about the baseline performance of the economy, and assumptions about the Fed's objectives. But even with these caveats, it is striking that the optimal control solutions imply a period of near-zero interest rates that lasts until the unemployment rate has fallen to a level that is somewhere around the structural rate. This finding reflects the usual optimal control logic, namely that the central bank's objectives over the simulation period as a whole are best served by allowing for a period of modest overheating and overshooting of inflation in the more distant future in order to frontload a greater amount of monetary stimulus early on in the simulation period. But this logic is now enhanced by the hysteresis effects from weakness in aggregate demand on the supply side of the economy, which further increase the longer-term costs of failing to return to full employment.
The implications of the results strengthen further under an alternative assumption about the Fed's objectives. This is that Fed officials interpret the employment side of their mandate in terms of what we have called the total employment gap. As the authors note, "…optimal policy should become even more accommodative if the central bank did not target the unemployment gap but instead aimed at keeping the employment-to-population ratio near the trend level that would prevail in the absence of hysteresis effects and exogenous (but ultimately transitory) shocks to the natural rate." In fact, their simulations show that under this alternative assumption the funds rate remains at its current near-zero level until the unemployment rate has fallen about 1 percentage point below its structural rate.
Taking the two studies together, our preliminary takeaways are as follows:
First, the studies suggest that some of the most senior Fed staffers see strong arguments for a significantly greater amount of monetary stimulus than implied by either a Taylor rule or the current 6.5%/2.5% threshold guidance. To be clear, both studies contain a significant amount of caveats and offsetting considerations, as well as a disclaimer that they only reflect the views of the authors. But given the structure of the Federal Reserve Board, we believe it is likely that the most senior officials--in particular, Ben Bernanke and Janet Yellen--agree with the basic thrust of the analysis.
Second, the studies provide two complementary reasons for why additional easing is warranted, which correspond closely to our own recent analysis: (a) optimal control considerations that argue for "credibly promising to be irresponsible" and (b) the possibility that the economy is underperforming its "deep" structural potential--that is, the level of potential output that would have obtained in the absence of the demand weakness of the past five years--by much more than suggested by the current unemployment gap alone. These two reasons are additive in the sense that each provides a separate rationale for further easing, and taken together they provide a very strong rationale for such a move.
Third, the studies suggest that the most likely form of this additional easing would be a reduction in the 6.5% unemployment threshold, i.e. a further ramping-up of the primary form of forward guidance that the committee has already chosen. The discussion of QE--the other key form of unconventional policy currently in place--is quite scant; however, the English et al. paper notes that "uncertainty about the level of costs and efficacy…would lead to a reduced level of purchases." This discussion admittedly does not give us much to go on, but we would view it as broadly consistent with the idea that a reduction in the unemployment threshold might be accompanied by a tapering of QE.
The upshot from our perspective is that the probability of an outright reduction in the unemployment threshold has increased by enough to make this our baseline expectation. Admittedly, the uncertainty around what the committee will do to strengthen the guidance remains considerable. On the one hand, the discussion in the minutes of the September 17-18 FOMC meeting seemed to reveal more sympathy for an inflation lower bound--an option that does not receive any attention in the staff studies--than for a reduction in the unemployment threshold. On the other hand, both the English et al. study and at least parts of the Wilcox et al. study seem to make a fairly strong case for an even bigger reduction in the threshold, perhaps to as low as 5.5%. But our central case is now that the FOMC will reduce the threshold from 6.5% to 6% at the March 2014 FOMC meeting, alongside the first tapering of QE; however, a move as early as the December 2013 meeting is possible, and if so, this might also increase the probability of an earlier tapering of QE.
In other words... Goldman is expecting Taper to be announced at the December FOMC and believs the Fed need to cut its threshold on rate hikes to as low as 5.5% to balance the potential 'tightening' implicit in the Taper announcement that the market is likely be unhappy about... because - it would seem - Goldman (like many others) still do not understand that it is all about the "flow" not the stock.