Comments on Monetary Policy by Governor Jeremy C. Stein

Thank you very much. It's a pleasure for me to be here at the Council on Foreign Relations, and I look forward to our conversation. To get things going, I thought I would start with some brief remarks on the current state of play in monetary policy. As you know, at the Federal Open Market Committee (FOMC) meeting last week, we opted to keep our asset purchase program running at the rate of $85 billion per month. But there has been much discussion about recent changes in our communication, both in the formal FOMC statement, as well as in Chairman Bernanke's post-meeting press conference. I'd like to offer my take on these changes, as well as my thoughts on where we might go from here. But before doing so, let me note that I am speaking for myself, and that my views are not necessarily shared by my colleagues on the FOMC.

It's useful to start by discussing the initial design and conception of this round of asset purchases. Two features of the program are noteworthy. The first is its flow-based, state-contingent nature — the notion that we intend to continue with purchases until the outlook for the labor market has improved substantially in a context of price stability. The second is the fact that — in contrast to our forward guidance for the federal funds rate — we chose at the outset of the program not to articulate what "substantial improvement" means with a specific numerical threshold. So while the program is meant to be data-dependent, we did not spell out the nature of this data-dependence in a formulaic way.

To be clear, I think that this choice made a lot of sense, particularly at the outset of the program. Back in September it would have been hard to predict how long it might take to reach any fixed labor market milestone, and hence how large a balance sheet we would have accumulated along the way to that milestone. Given the uncertainty regarding the costs of an expanding balance sheet, it seemed prudent to preserve some flexibility. Of course, the flip side of this flexibility is that it entailed providing less-concrete information to market participants about our reaction function for asset purchases.

Where do we stand now, nine months into the program? With respect to the economic fundamentals, both the current state of the labor market, as well as the outlook, have improved since September 2012. Back then, the unemployment rate was 8.1percent and nonfarm payrolls were reported to have increased at a monthly rate of 97,000 over the prior six months; today, those figures are 7.6 percent and 194,000, respectively. Back then, FOMC participants were forecasting unemployment rates around 7-3/4 percent and 7 percent for year-end 2013 and 2014, respectively, in our Summary of Economic Projections; as of the June 2013 round, these forecasts have been revised down roughly 1/2 percentage point each. While it is difficult to determine precisely, I believe that our asset purchases since September have supported this improvement. For example, some of the brightest spots in recent months have been sectors that traditionally respond to monetary accommodation, such as housing and autos. Although asset purchases also bring with them various costs and risks — and I have been particularly concerned about risks relating to financial stability — thus far I would judge that they have passed the cost-benefit test.

However, this very progress has brought communications challenges to the fore, since the further down the road we get, the more information the market demands about the conditions that would lead us to reduce and eventually end our purchases. This imperative for clarity provides the backdrop against which our current messaging should be interpreted. In particular, I view Chairman Bernanke's remarks at his press conference — in which he suggested that if the economy progresses generally as we anticipate then the asset purchase program might be expected to wrap up when unemployment falls to the 7 percent range — as an effort to put more specificity around the heretofore less well-defined notion of "substantial progress."

It is important to stress that this added clarity is not a statement of unconditional optimism, nor does it represent a departure from the basic data-dependent philosophy of the asset purchase program. Rather, it involves a subtler change in how data-dependence is implemented — a greater willingness to spell out what the Committee is looking for, as opposed to a "we'll know it when we see it" approach. As time passes and we make progress toward our objectives, the balance of the tradeoff between flexibility and specificity in articulating these objectives shifts. It would have been difficult for the Committee to put forward a 7 percent unemployment goal when the current program started and unemployment was 8.1 percent; this would have involved a lot of uncertainty about the magnitude of asset purchases required to reach this goal. However, as we get closer to our goals, the balance sheet uncertainty becomes more manageable — at the same time that the market's demand for specificity goes up.

In addition to guidance about the ultimate completion of the program, market participants are also eager to know about the conditions that will govern interim adjustments to the pace of purchases. Here too, it makes sense for decisions to be data-dependent. However, a key point is that as we approach an FOMC meeting where an adjustment decision looms, it is appropriate to give relatively heavy weight to the accumulated stock of progress toward our labor market objective and to not be excessively sensitive to the sort of near-term momentum captured by, for example, the last payroll number that comes in just before the meeting.