Thank you, Mr. Chairman. A lot has changed since the June Greenbook forecast was put to bed. Unfortunately, a lot has changed since the August Greenbook was put to bed, too. [Laughter] In particular, in the June Greenbook, we forecasted that payroll employment gains would average about 300,000 per month in the second and third quarters of this year against a backdrop of GDP growth averaging about 5 percent at an annual rate. By the end of this year, we expected that the GDP gap would have narrowed to just ½ percent. By the time we published the August Greenbook last week, we had tempered our growth assessment substantially and had marked up our expectation for the GDP gap at the end of this year to a full 1 percent. But we were still interpreting the June employment report essentially as a one-time hiccup and were looking for employment gains going forward only slightly smaller than we had projected in the June Greenbook. Then came last Friday’s employment report, with its news of only a 32,000 advance in private nonfarm payrolls, a downward revision to June employment of 51,000, and only a 0.1 hour increase in the average workweek. Our misery over coming up short nearly 300,000 jobs was only slightly alleviated by having so much company.
So, as I said, a lot has changed. On the other hand, I should also underscore that a lot has stayed the same. Even in the wake of Friday’s report, we still see overall macroeconomic conditions as likely to support a return to above-trend growth—in other words, growth that is sufficiently rapid to narrow the gap in resource utilization over the next six quarters, even if only slowly and incompletely. While we have taken quite a bit out of our forecasts for payroll employment increases in August and September, we still have those increases tracking into the neighborhood of 300,000 in the fourth quarter of this year. And in response to the employment report, we trimmed our forecast for third-quarter GDP growth only 0.6 percentage point, from 3.9 percent in the Greenbook to 3.3 percent currently, on the theory that the deficit in hours relative to our expectation informed us mainly about productivity in the current quarter rather than output.
In this connection, I should note there was one bright spot in Friday’s report, namely that employment in the manufacturing sector looked, if anything, a little better than we had expected. This is particularly encouraging in light of the fact that manufacturing over the decades has been a fairly sensitive and reliable bellwether of cyclical conditions. If a pronounced cyclical softening truly were under way, we would have expected to see it in manufacturing employment, but it simply was not evident in the numbers for July.
Our outlook for inflation also falls in the category of “not much changed since the June Greenbook.” The most recent CPI report was quite reassuring that the type of moderation we were looking for seems to be in train. Indeed, the news was good enough to cause us to shave our projection for core CPI inflation by a tenth both this year and next. Because the BEA penciled in a faster pace of inflation in the so-called non-market-based prices, we did not revise down our forecast for core PCE inflation in line with the core CPI. Even so, however, we have core PCE inflation slowing from 1¾ percent this year to 1½ percent next year, reflecting smaller contributions from import, energy, and other commodity prices and now a slightly greater amount of restraint from slack in resource utilization.
But while a lot is still the same, enough has changed to make it more than fair to ask, What on earth is going on? Lest I spark any hope on your part that I’m about to produce a fully satisfying explanation, let me assure you to the contrary. In point of fact, we do not claim to suddenly understand everything that has happened in the last few months; if we were that smart now, we probably would have been a little smarter last week; and if we had been that smart last week, we certainly wouldn’t have hid it so well from the Committee. [Laughter] Unfortunately, the large-scale econometric model that we often consult for insight on such issues shares our incomplete understanding of the current situation. Indeed, that model, left to its own devices, would have wanted to see an additional 2 percentage points of real GDP growth in the second quarter, with much of it in PCE. Evidently, in order to provide a full accounting of the recent weakness in aggregate demand, we will have to appeal to some factor not captured by our conventional models. In light of the weakness of consumer spending, factors bearing on the household sector will have special standing. Let me list some of the possibilities.
Hypothesis #1: Maybe it’s just noise. This hypothesis actually comes in two variants. The first variant takes the view that the data may simply be sending the wrong signal, not truly representative of macroeconomic reality. I certainly cannot rule out this variant, and indeed, we had discounted the first disappointing report on payroll employment, given the relatively encouraging recent readings on initial claims for unemployment insurance and improving household perceptions of labor market conditions. On the other hand, this variant seems less plausible in light of a second consecutive disappointing labor market report as well as the June readings on retail sales, motor vehicle sales, and industrial production, all suggesting that something real happened toward the end of the second quarter. A second variant of the noise hypothesis accepts the accuracy of the incoming data but questions the persistence of the softness: Maybe we are just on a temporary detour away from above-trend growth and shrinking margins of resource slack. We put some stock in this variant as well, in part because we have been impressed by a number of the other relatively encouraging indicators of demand in the third quarter, including the rebound in motor vehicle sales, the strength in orders and shipments of capital goods, and the relatively upbeat anecdotal information that we continue to receive from the ISM and other business surveys, from the Beige Book, and from our own business contacts. In addition, we continue to view the fundamental determinants of growth, especially the stance of monetary policy—and, in the near term, the stance of fiscal policy—as consistent with above-trend growth. On the other hand, we are operating in an environment of considerable uncertainty, and a key possibility that we must entertain is that something more fundamental is going on than just a temporary blip— something we don’t understand and that may persist.
Hypothesis #2: Maybe it’s energy prices. Since June, the spot price of West Texas intermediate has increased about $6 per barrel. Moreover, prices on long-dated futures have undergone a stunning nearly vertical takeoff in the last few weeks. Certainly the hypothesis that higher energy prices are having a much more pronounced restraining effect on activity than we had anticipated has to figure prominently on our list of possibilities. Maybe the second surge in gasoline prices in May and June caused households to reassess the outlook for energy prices and suddenly to take on board a much larger hit to their purchasing power.
To be sure, the timing of the most recent energy price spike is incriminating. But how much of the recent shortfall in aggregate activity can we really lay at its doorstep? In a memo that we delivered to the Committee just before the June meeting, we estimated that, all else being equal, the run-up in oil prices between December and June might lop ½ percentage point off the growth of GDP this year and a further tenth off growth next year. At that time, the spot price had risen about $10 per barrel from its late-December level. Since June the spot price has increased another $6 or so, suggesting that the total oil-related deduction from the growth of GDP this year might amount to about ¾ percentage point—or roughly two-thirds of the total downward revision to our GDP projection over that time.
But while the price of oil is crucial for understanding the revision since the December Greenbook to our projection for GDP growth this year, can it account for the steep downward revision to GDP growth since the June Greenbook? Here the answer is murkier. By June, roughly two-thirds of the increase in oil prices had already occurred and so should have been taken on board in the forecast we delivered to the FOMC just before your two-day meeting. That suggests that the increase in oil prices since then might account for a downward revision to our estimate of growth this year of only about ¼ percentage point—only a fraction of the nearly 1 percentage point revision that we have put through since the June Greenbook. On the other hand, a number of factors counsel against drawing sharp conclusions. For one, our ability to pinpoint the timing of an adjustment to higher oil prices probably is very limited. For another, we may not have sufficiently adjusted the June Greenbook projection to the oil prices then already in view. And for a third, there is little doubt that our formal econometric models omit potential channels of influence from oil to real activity, and whether we have adequately compensated for those omissions in our judgmental projection is not clear.
All that said, the coincidence between the spike in oil prices and the air pocket in spending seems too striking to write off to mere chance. So at this point, we are thinking that energy prices almost surely are part of the explanation; we’re just not sure how big a part.
Hypothesis #3: Faltering fiscal stimulus. In the baseline projection, we have been assuming that the 2001 and 2003 tax cuts are still supporting the growth of consumer spending this year but by a gradually diminishing amount. On our assumptions, these tax cuts boosted the growth of real PCE ¾ percentage point in the first half of this year but will lift the growth of PCE in the second half by less than ½ percentage point and will be essentially neutral for PCE growth in 2005. But here, too, we have only limited ability to pinpoint the timing of adjustment, and this hypothesis suggests the possibility that we should have attributed a larger fraction of the strength in consumer spending earlier this year to short-lived tax-induced stimulus and a smaller fraction to an underlying fundamentals-based thrust.
In sum, we think these hypotheses contain some of the seeds of an explanation. Whether they constitute a full or only a partial explanation, whether we need to add to our list, or whether we need to subtract from it, all remain to be seen. But as usual, we will have to remain alert in coming months for clues that may shed further light on the nature of the recent step-down in the growth of real activity. Karen will now continue our presentation.