Poole reflects on 10 years at the St. Louis Fed

Outgoing St. Louis Fed president William Poole gave a speech to the St. Louis NABE today. I linked to the Reuters story in the previous post. Now, I would like to post some excerpts from the speech that didn't make the wires. Most of the speech has to do with central bank communications.

My general approach has been to speak primarily about the policy process rather than the specific situation facing the FOMC at its next meeting. I try to think of myself as speaking to portfolio managers who have a medium-term horizon rather than to traders who have a horizon measured in hours or a few days. I do not disparage traders—they perform an important function. Obviously, I have had internal information that would be of interest to traders but it would be entirely inappropriate—indeed illegal—to disclose confidential FOMC information.

Traders, portfolio managers and many others always want to know my forecast of what will happen at the upcoming FOMC meeting. My standard answer is that I do not forecast monetary policy decisions—my job is to participate in making those decisions. I confess that, initially, this response was something of a dodge, because I usually had a pretty good idea weeks in advance of what my own position at a meeting would be. However, over the years I have become impressed by how often my own position would change even in the days just before a meeting as a consequence of the arrival of new information, including staff analysis and sound arguments by my FOMC colleagues. It is not that my views are pushed this way and that by arrival of the latest economic data reports. What happens is that, from time to time, compelling new information does arrive. I hope that my policy outlook was stable even as my view on the appropriate policy action might change in the light of incoming data.

Thinking through the matter led me to a bit of research. Working with Bob Rasche, the St. Louis Fed research director, I had already studied the accuracy of market expectations about FOMC decisions using data from the federal funds futures market the day before each FOMC meeting. Given that those futures market forecasts have proven to be quite accurate, an obvious question was forecast accuracy longer in advance. Bob and I studied futures market predictions of the fed funds rate three and six months in advance and found that the accuracy was pretty low. The reason these forecasts have not been very good is that new information arrives that calls for a changed expectation on the monetary policy setting, both for the markets and for the FOMC. I not only became more aware of the need for me to keep my mind open but also thought it important to explain to my audiences how the policy process worked to be responsive to new information.

...

I also became troubled by the following argument. If current economic conditions were such to suggest a high probability that future economic conditions would justify a future increase in the funds rate target, why not just raise the rate at the current meeting? Given lags in the effects of policy actions, the current policy had to be based on the future outlook. On the other hand, if the probability were low, would it serve the cause of good communication to state a bias? Wouldn’t it be more helpful to work harder to articulate the conditions under which the committee might change the target—to explain in more detail the nature of the policy rule or response function?

Another problem with forward policy guidance was that a slow accumulation of information sometimes made the prior balance-of-risks language out of date, but it was not easy to take it out of the statement without sending a message, or seeming to, that a future policy adjustment in the other direction was contemplated. This problem arose in 2006. In August 2006, the committee kept the funds rate target unchanged, after increasing it by 25 basis points at each of its previous 17 meetings. However, the statement indicated a bias toward a further increase by saying this: “Nonetheless, the Committee judges that some inflation risks remain. The extent and timing of any additional firming that may be needed to address these risks will depend on the evolution of the outlook for both inflation and economic growth, as implied by incoming information.” Just ahead of the August meeting, the market had assigned a probability of about 0.8 on an FOMC target fed funds rate of 5.25 percent and a probability of about 0.2 on a target rate of 5.5 percent.

Just after the August 2006 FOMC meeting, the market assigned these probabilities to the FOMC decision at its forthcoming September meeting: a target of 5.25 percent had a probability of about 0.78, a target of 5.5 percent had a probability of about 0.2 percent, and a target of 5.75 percent had a probability of about 0.02 percent. Although the FOMC retained the language that “firming may be needed” at subsequent meetings, over time the market lowered its probability that the FOMC would in fact raise the target rate. Ahead of the FOMC meeting of Jan. 30-31, 2007 the market placed essentially zero probability on any target rate above the prevailing rate of 5.25 percent.

At its meeting of March 20-21, 2007, the committee dropped the language referring to possible firming. Doing so made little difference given that the market had discounted the possibility of firming for some time.

Here's the key paragraph of the whole thing...

I have recounted several of these episodes in some detail to illustrate the general issue. As a consequence of observing this process for 10 years, I have concluded that an FOMC attempt to provide forward guidance in the policy statement causes more communications difficulties than it solves. A key reason is that the economy is subject to more shocks and reversals than one might think. These shocks sometimes require more frequent policy actions than I would have thought likely when I came to St. Louis. At a minimum, changing economic conditions change the likelihood that the FOMC will want to adjust the fed funds target in the direction previously thought. Directional language tends to remain in the FOMC policy statement beyond the time it applies and removing the language creates the possibility of miscommunication. Every change in the policy statement leads naturally to market questions as to what the change means and whether the change is meant to provide a hint about the future direction of policy. To my mind, every time new language is inserted into the policy statement, there needs to be as much thought given as to how to exit from the language as to the rationale for inserting it. (Emphasis mine)

Oh, how true. I blogged about this over two years ago. At the time, I was hopeful that this was a solvable problem. The intervening two years have made me less optimistic--apparently, I'm not alone.

Now, some might be surprised or taken aback by his talk of "hunches" in the next paragraph, but I think most veteran Fed watchers know what he means. I know I do.

I know that market participants are hungry for insight into the FOMC’s thinking and into the likelihood of future adjustments in the target federal funds rate. My judgment is that, most of the time, the committee cannot provide what the market wants because the committee itself is not clairvoyant. No one knows how the economy is going to evolve and how events will change the appropriate setting of the federal funds target rate. Most of the time over the past 10 years I had hunches about the policy direction I would be advocating at the next FOMC meeting, but “hunches” really is the right word. I had hunches and not settled convictions. Furthermore, the more I reflected and the more experience I accumulated, the more I realized how frequently surprise changes in conditions required that I change my hunches. I should not be misinterpreted as saying that I necessarily changed my view on the appropriate setting of the fed funds rate target. But when the information on which my prior hunch was based changed significantly, I had to start over, in a sense, to figure out whether the new information required a change in the policy stance.

This one's going on the reading list.