Wednesday, July 3, 2013

Stein, Thoma, and Hamilton on Fed Communication

Mark Thoma describes four ways the Fed is creating harmful uncertainty that could block the economic recovery. A June 28 speech by Jeremy Stein, a member of the Board of Governors, describes Stein's opinions on how the Fed should address two out of the four.

Thoma says that the Fed is creating harmful uncertainty through a botched communication policy concerning when quantitative easing will begin tapering off. "The Fed does not seem to understand how anxious it has appeared to return policy to normal over the last several years," he writes. "Any sign of `Green Shoots,' such as the supposed sighting in 2009 by Chairman Bernanke, prompts the Fed to announce that it is developing an exit strategy. With that background, combined with the public perception that the Fed is itching to reverse course and get back to normal, members of the Fed should not be surprised when markets “misread” talk about when the Fed is planning to reverse course. When Fed communication is adding uncertainty instead of reducing it, that’s a big problem."

Another, related way the Fed is creating uncertainty, Thoma notes, is through its consistently overly-optimistic forecasts: "When things turn out much worse than forecast, the Fed has to reverse itself. It has eased policy after these episodes in some cases and that creates quite a bit of uncertainty over Fed policy."

Stein asserts that both the labor market and the general economic outlook have improved since the start of this round of asset purchases nine months ago, but adds that "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." In other words, "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." This seems roughly true, although I would add that it is not balance sheet uncertainty per se that is the most impactful. I think market participants, in contrast to the consensus among Fed officials, generally care more about the flow of asset purchases than the stock on the balance sheet. In other words, while the Fed believes that tapering is not tightening, since it still expands the balance sheet, if at a slower rate, the markets interpret the slower rate as tightening. So the fact that "balance sheet uncertainty" becomes more manageable does not do much to alleviate uncertainty in the financial market.

Stein makes a suggestion that might go some ways towards addressing Thoma's concern about consistently overly-optimistic forecasts:
"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...Not only do FOMC actions shape market expectations, but the converse is true as well: Market expectations influence FOMC actions. It is difficult for the Committee to take an action at any meeting that is wholly unanticipated because we don’t want to create undue market volatility. However, when there is a two-way feedback between financial conditions and FOMC actions, an initial perception that noisy recent data play a central role in the policy process can become somewhat self-fulfilling and can itself be the cause of extraneous volatility in asset prices. 
Thus both in an effort to make reliable judgments about the state of the economy, as well as to reduce the possibility of an undesirable feedback loop, the best approach is for the Committee to be clear that in making a decision in, say, September, it will give primary weight to the large stock of news that has accumulated since the inception of the program and will not be unduly influenced by whatever data releases arrive in the few weeks before the meeting – as salient as these releases may appear to be to market participants. I should emphasize that this would not mean abandoning the premise that the program as a whole should be both data-dependent and forward looking. Even if a data release from early September does not exert a strong influence on the decision to make an adjustment at the September meeting, that release will remain relevant for future decisions. If the news is bad, and it is confirmed by further bad news in October and November, this would suggest that the 7 percent unemployment goal is likely to be further away, and the remainder of the program would be extended accordingly."
The question is whether such a commitment would be credible. Market participants might believe that the Fed would treat September near-term news asymmetrically. Participants might (reasonably) believe that a data release from early September may not exert a strong influence on September decisionmaking if it were bad news, whereas if it were good news, the Fed would choose to use it to justify a return to normal.

While Stein recognizes the need to improve the Fed's communication strategy, he also points out that "there are limits to how much even good communication can do to limit market volatility, especially at times like these." In this respect his view is quite similar to that of James Hamilton, who questions the omnipotence sometimes ascribed to the Fed's communication, reminding us that the Fed's communication strategy is not the only determinant of long-term yields.
"Prior to the Great Recession, I thought we had all agreed on the practical limits on the Fed's capabilities. We understood that to some extent the Fed could control the short-term interest rate by changing the supply of reserves available to the banking system. But we also understood that the Fed's influence over longer-term interest rates was much less immediate and direct. The Fed can communicate its long-run inflation objectives, and certainly the 10-year inflation rate is a very important determinant of long-term yields. But regardless of what the Fed may say about its 10-year inflation goals, the market would form its own view of whether the Fed could or would achieve those. Other determinants of long-term yields, such as the term premium, long-run economic growth rate, and global saving and investment decisions were understood to be even farther beyond those things that the Fed can hope to control."
Stein puts it very similarly:
"At best, we can help market participants to understand how we will make decisions about the policy fundamentals that the FOMC controls – the path of future short-term policy rates and the total stock of long-term securities that we ultimately plan to accumulate via our asset purchases. Yet as research has repeatedly demonstrated, these sorts of fundamentals only explain a small part of the variation in the prices of assets such as equities, long-term Treasury securities, and corporate bonds. The bulk of the variation comes from what finance academics call “changes in discount rates,” which is a fancy way of saying the non-fundamental stuff that we don’t understand very well – and which can include changes in either investor sentiment or risk aversion, price movements due to forced selling by either levered investors or convexity hedgers, and a variety of other effects that fall under the broad heading of internal market dynamics...So while we have seen very significant increases in long-term Treasury yields since the FOMC meeting, I think it is a mistake to infer from these movements that there must have been an equivalently big change in monetary policy fundamentals."

2 comments:

  1. Once again this post pays homage to
    the traders and volitility arbitrageurs
    Its the expectations of actual real economy agents that deserve attention

    Why should the noise flux of paper markets significantly impact real agents ...except to play it safe and hold back
    Something they are doing anyway given
    Chronic demand constraints

    ReplyDelete
  2. Interesting that Stein mentions changes in risk aversion, since economists (as opposed to practitioners) usually avoid ascribing price changes to changes in the price of risk. (Although, see the recent papers by Ross and by Carr & Yu on disentangling expectations from the price of risk using option prices.)

    But neither Stein nor you address the fundamental problem that the date of Fed tightening as implied by futures prices has advance a long way forward. Tough to attribute that to "volatility".

    ReplyDelete

Comments appreciated!