Transcripts of the monetary policymaking body of the Federal Reserve from 2002–2008.

I will be referring to the exhibits that follow the green nonfinancial cover page. The indicators that we have received since the last FOMC meeting suggest that real activity has been contracting rapidly, and as Dan has described, financial conditions have continued to deteriorate. Starting with the labor market, private payroll employment (the inset box in the top panel) plunged 540,000 in November, and the figures for September and October were revised down noticeably. All told, November’s drop brought the three-month decline in private employment (the black line in the top panel) to an annual rate of 4.2 percent—a much steeper pace of job loss than we were expecting in the previous Greenbook. The retrenchment in November was pervasive across both goods- and service-producing industries. As shown in the middle left panel, initial claims for unemployment insurance have continued their steep climb since the November survey week, consistent with a further large drop in employment in December.

Meanwhile, in the household sector, nominal sales at the retail control grouping of stores (the inset box in the middle right panel) declined 1.5 percent in November, which given a large energy-driven price decline, translates into an increase in real spending for the month. We had been expecting real outlays in this category to fall further. As shown by the rightmost blue bar, with the latest retail sales number, we now think that real PCE on goods other than motor vehicles is on track to fall at an annual rate of 5¼ percent, not as bad as the 9¼ percent decline we projected in the Greenbook but still very weak. Near-term indicators of consumer spending point to further weakness in the coming months. Notably, the Reuters/University of Michigan survey index of consumer sentiment, plotted in the bottom left panel, remained at recessionary low levels in the first part of December. Sales of light vehicles (the bottom right panel), which slumped in October, fell further in November, to an annual rate of 10.1 million units, a much slower pace than we had anticipated. Anecdotal reports suggest that sales are soft again this month. Given the weak fundamentals for demand and tight credit, we project vehicle sales to remain depressed over the next several months.

As shown in the top left panel of exhibit 2, residential construction has continued to slide. Single family starts fell further, to 441,000 units—somewhat lower than our expectations—and permits continued to move down last month. In the business sector, new orders for nondefense capital goods excluding aircraft (the red line in the top right panel) dropped for a third straight month in October; and with orders falling below shipments for a second consecutive month, the backlog of unfilled orders shrank further. Yesterday, the Board released data on industrial production in November. Total IP (the inset box in the middle panel) fell 0.6 percent in November. Excluding the effects of rebounds from the Boeing strike and September hurricanes, IP moved down 1.6 percent last month. The black line in the panel plots the three- month diffusion index of manufacturing IP, a measure reflecting the net fraction of industries that experienced an increase in production. As you can see, over the past few months that number has plummeted to 21, indicating that the contraction in industrial activity has been remarkably widespread. Overall, the incoming data led us to steepen significantly the contraction in real GDP that we are forecasting for the current quarter and the first quarter of next year. In these two quarters, we expect output to decline at an average annual rate of nearly 5 percent. In both quarters, a significant chunk of the revisions has come in private domestic final purchases (lines 3 and 4 of the table). In addition, the available indicators suggest that firms are acting aggressively to limit unwanted increases in their inventories. As shown in line 5, firms have been drawing down inventories at a moderate pace in the second half, and we expect even faster liquidation next quarter in response to the sizable contraction under way in final sales.

Your next exhibit focuses on the medium-term outlook, starting with some of the key background factors that have influenced our thinking about the outlook since the October Greenbook. On the downside, as noted earlier, financial conditions have deteriorated further in recent weeks. As can be seen in the top left panel, the index of financial stress that we track continued to rise sharply. One important component of that increase in stress has been the further widening of corporate bond spreads for investment-grade issues (shown by the gray shaded area in the top right panel). In addition, equity prices are lower than we projected, taking an even bigger bite out of household resources. Meanwhile, in the external sector, the path of the dollar (shown in the middle right panel) is somewhat stronger than in our October forecast, and the outlook for foreign economic activity has weakened further. Shaghil Ahmed will have more to say about these developments shortly.

As you know from reading Part 1 of the Greenbook, in light of the intensification of recessionary forces that emerged over the intermeeting period, we based this forecast on the assumption of a lower level of the federal funds rate than in our previous projection (not shown). In addition, we now assume that $500 billion in new fiscal stimulus actions will be enacted early next year, on top of the roughly $165 billion in the stimulus package enacted earlier this year. These assumed fiscal actions include permanent tax cuts for most individuals, higher transfer payments, grants to state and local governments, and support for housing. As shown in the bottom left panel, these new federal programs boost the impetus to GDP growth from fiscal policy considerably relative to our assumptions in the October Greenbook. Two other factors also provide more support to real activity in this forecast. First, mortgage rates (the bottom right panel) have fallen about ½ percentage point since the October Greenbook. With the spread over the 10-year Treasury yield still quite wide, we project that mortgage rates will fall further over the projection period. Second, oil prices (not shown) have fallen more in recent weeks than we anticipated in the October Greenbook; the lagged effects of these declines provide a greater lift to spending in 2009.

The top left panel of your next exhibit summarizes our medium-term projection. On balance, we expect that the factors restraining activity will far outweigh the supportive influences, with real GDP falling at an annual rate of 3 percent in the first half of next year. The decline is led by a steep drop in business fixed investment (lines 5 and 6). In the second half of the year, real activity begins a slow recovery as PCE (line 3) picks up and residential investment (line 4) begins to stabilize. In 2010, the recovery is projected to gain momentum as household spending strengthens further and business purchases of equipment and software begin to rise. As is typical, the recovery in nonresidential investment is expected to lag.

By postwar standards, we are projecting a deep, prolonged recession and a very sluggish recovery. The middle left panel provides some perspective. The black line shows the level of real GDP, indexed to its own peak, in the second quarter of 2008. By way of comparison, the red and blue lines show the paths of real GDP during the recessions that started in November 1973 and July 1981, respectively. As can be seen, the projected contraction in real GDP in the current episode is about in line with that experienced during those two earlier “big” recessions, but our projected recovery is noticeably more prolonged. The box to the right summarizes some of the important factors that impede the projected recovery. First, we think that the economy will continue to face significant (albeit moderating) financial headwinds over the next two years, with elevated risk premiums, restrictive lending conditions, and general uncertainty restraining real activity for some time to come. Second, tight monetary policy did not help generate the current recession, and we don’t see monetary ease as likely to generate as much impetus to recovery. In fact, as the third bullet point notes, the federal funds rate is already close to the zero lower bound, greatly limiting the scope for conventional monetary policy to provide further stimulus to real activity.

The importance of this last factor is illustrated by the bottom set of panels. The exercise is similar to the simulations presented in the Bluebook, except that this one allows the federal funds rate to turn negative. The green line in the bottom left panel shows the simulated federal funds rate path. If unconstrained, the optimal funds rate would fall below zero in early 2009 and drop to negative 5½ percent in the third quarter of 2010. In this scenario, the unemployment rate, shown in the middle panel, peaks at 7¾ percent in 2009—about ½ percentage point lower and a year earlier than in the Greenbook projection. The four-quarter change in core PCE prices, shown at the right, bottoms out at just over 1¼ percent in mid-2010 and then turns up, in contrast with the continued deceleration in the extended staff forecast. In the context of the zero bound, achieving an outcome for real activity and inflation consistent with the unconstrained optimal control exercise would likely require some combination of greater fiscal stimulus and nontraditional monetary actions than we have built into the current projection.

Your final exhibit summarizes the outlook for inflation. As shown in line 1 of the top left panel, we now project that total PCE price inflation will slow to about 1 percent in 2010, while the core rate (line 7) falls to 0.8 percent. The decline in inflation reflects a combination of widening slack in resource utilization, reduced energy and materials prices, and a net decline in core import prices; these factors also push down long-run inflation expectations. I should note that we received data on the November CPI this morning. The total CPI fell 1.7 percent, driven by a sharp drop in energy prices. The core CPI was unchanged last month. We had been expecting an increase of 0.1 percent. One measure of resource utilization, the unemployment rate, is projected to reach 8¼ percent in 2010. In our forecast, the wide unemployment rate gap puts substantial downward pressure on costs and prices. However, as suggested in the middle left panel, we may be overstating the downward pressure on inflation caused by slack. In particular, the current cycle has been associated with especially large employment declines in several industries, notably construction and finance. If these declines are leading to an unusual amount of employment reallocation across industries, structural unemployment would increase, which in turn would raise the NAIRU.

The remaining panels present some analysis of the issue using a measure of sectoral reallocation and a set of Beveridge curves. The middle right panel depicts an index of sectoral employment reallocation across industries. The index is based on the growth rates of employment in 15 industries relative to the growth rate of total employment, adjusted at the industry level for typical cyclical movements in employment shares and is similar in spirit to a measure constructed at the Chicago Fed. As can be seen, even after removing the typical cyclical behavior, increases in sectoral reallocation often occur around recessions, which is not surprising since each business cycle produces a unique set of imbalances. Indeed, the amount of sectoral reallocation indicated by this measure has been rising over the past year or so. However, to date it remains low relative to previous spikes in reallocation. Another way to determine whether there has been a rise in structural unemployment is through the so-called Beveridge curve, two versions of which are shown in the bottom panels. The version at the bottom left plots the unemployment rate, adjusted for the Emergency Unemployment Compensation program, on the horizontal axis against the job openings rate, measured by the Job Openings and Labor Turnover Survey (JOLTS), on the vertical axis. The bottom right panel shows a Beveridge curve calculated using the Help-Wanted Index as the measure of job openings. If there has been a significant increase in structural unemployment, then one would expect that for a given level of the job openings rate, the unemployment rate would be unusually high—that is, to the right of the plotted Beveridge curve. This might occur, for example, if many of the job openings were for nurses but a disproportionate number of the unemployed were bond traders, who are not qualified for the job openings. [Laughter]

So, what do these Beveridge curves say about structural unemployment? The blue and red circles in the two panels show the data points for the third quarter of 2008 and for the most recent months available. The latest readings from the JOLTS do stand to the right of the estimated curve, while the latest readings from the Help- Wanted Index are closely in line with past experience. At this point we are reluctant to draw strong conclusions from just a couple of observations from either measure, and we read the evidence as consistent with at most a small increase in structural unemployment thus far. Of course, these data do not tell us what will happen in the coming quarters, when we anticipate further job losses in financial services and continued weakness in construction and manufacturing. Shaghil will continue our presentation.

Keyboard shortcuts

j previous speech k next speech