As you know from the Greenbook and as is shown in the upper left panel of exhibit 4, the unemployment rate is currently just a tenth of a percentage point above our estimate of the NAIRU. To put the matter delicately, however, the most impressive aspect of our estimate of the NAIRU may be the imprecision of it: According to one representative econometric equation, a 70 percent confidence interval around the point estimate of the NAIRU extends 2 full percentage points from top to bottom. Accordingly, one might reasonably ask whether any evidence can be marshaled to corroborate our view that a smidgeon of slack still remains in the labor market.
The upper right panel presents one such bit of evidence. As shown by the black line, the participation rate has declined a percentage point over the past five years. To be sure, the model that we use to analyze movements in the participation rate writes off part of that decline as reflecting trend demographic developments, summarized by the red line. But the larger part of the overall decline, according to the model, reflects a swing from a participation rate that was above its trend level through the late 1990s to one that has been below its trend level the last few years. On our interpretation, conditions in the labor market were sufficiently uninviting to cause some workers to take themselves out of the workforce. But as job opportunities continue to improve, these workers will be available to rejoin the labor force—and will do so in great enough numbers to hold the participation rate steady despite a declining trend.
The middle left panel shows a broad measure of conditions in the labor market— total hours worked relative to trend. This measure combines the effects of the unemployment gap, the participation rate gap, and the gap in the workweek relative to its trend, not shown but currently about zero. As you can see from the chart, total hours worked remain a little below the estimated trend despite the recent gradual improvement in labor market conditions.
The idea that some labor market slack remains is corroborated by a few other indicators. For example, as shown in the middle right panel, individuals still perceive jobs as somewhat harder to get than they were in the first half of 1997, when labor markets seemed roughly to be in equilibrium. Similarly, as shown in the lower left panel, the NFIB [National Federation of Independent Business] reports that small businesses still are not encountering quite as much difficulty filling positions as they did in the first half of 1997. I should also note that a few other indicators suggest that the labor market is at least as tight as it was in early 1997. For example, as shown in the lower right panel, the percent of employed people working part-time for economic reasons has fallen below its average over the first half of 1997.
That said, we still see the balance of evidence as suggesting that a little slack remains in the labor market, but that the margin is narrow, and narrower today than it was six months or a year ago. Moreover, such slack as may still exist will, on our projection, essentially be taken up over the course of the next year and a half.
Other things equal, in the past, labor market slack has restrained the growth of nominal compensation and thus has helped to keep price pressures in check as well. And yet, as shown in the upper left panel of exhibit 5, compensation per hour as measured in the BLS’s [Bureau of Labor Statistics] Productivity and Cost [P&C] system soared more than 10 percent in the fourth quarter of last year and posted another hefty increase in the first quarter of this year. As shown by the black line to the right, even on a four-quarter basis, the growth of compensation per hour has picked up sharply in the last two quarters, raising the question as to whether underlying compensation pressures are building and whether they may be feeding price inflation.
We think not, for two main reasons. First, as shown by the blue line in the upper right panel, the other main indicator of compensation trends, the ECI [Employment Cost Index], has actually been decelerating slightly since mid-2003, suggesting that the compensation situation may be a good deal less alarming than one might conclude based on the P&C measure alone. Second, we diagnose the recent bulge in compensation per hour as having been influenced by a spike in stock option exercises, which we think have little bearing on firms’ pricing decisions. Indeed, the last time a bulge of this sort occurred in comp per hour, in 2000, core PCE price inflation—the red line in the upper right panel—responded little if at all.
The middle left panel reviews some of the considerations that lead us to implicate option exercises in this matter. For now, in the interest of time, I will only touch on the first and most obvious factor, namely, that option exercises are included in P&C compensation per hour but not in the ECI. The attribution of a sizable portion of the spike in comp per hour to option exercises is important, because growth in that form of compensation is not likely to persist at its fourth-quarter pace.
Indeed, to that point, the middle right panel summarizes our outlook for the growth of compensation per hour. As shown on line 1 of the table, we have the four- quarter change in P&C comp per hour coming back down sharply this year, as about half of last year’s fourth-quarter spike is unwound. As shown on line 2, we have ECI compensation per hour moving up along a smoother trajectory, pushed up by the gradual tightening of labor markets, the continued pass-through of increases in structural productivity to real wages, and the lagged effects of higher consumer price inflation last year and this year. Overall, on our numbers, the growth of compensation will be moderate enough to allow a gradual deceleration in core consumer prices from here forward while still permitting firms to maintain the markup of prices over unit labor costs (not shown) noticeably above its average over the past 35 years.
But we could be wrong about all of that, and the bottom two panels illustrate the potential consequences of a misjudgment on our part. As outlined in the lower left panel, we sketched an alternative scenario in the Greenbook in which hourly compensation increases 1 percentage point faster than in the baseline over the remainder of this year and next. In addition, we assumed that firms have sufficient pricing power to protect their profit margins, and are able—by the end of the scenario —to return the markup to its baseline value. As shown by the dashed red line to the right, the resulting boost to core PCE price inflation is substantial: By the end of 2006, core inflation is running at a 3 percent annual rate and headed further north from there.
While we view a compensation-led spiral as a risk to the outlook, we do not see compensation as having been a major source of the recent upward creep in core inflation. But if not compensation, then what?
The top left panel of exhibit 6 steps back and reviews the scope of the recent run- up in core PCE price inflation. The red line in this panel shows the evolution over the past 18 months of our forecast for core PCE inflation during 2004. As shown by the leftmost point on that line, back in December 2003 we forecasted that core PCE prices would increase about 1.1 percent over the course of 2004. As shown by the peak in the red line, by August of last year, we had marked our forecast up, on net, about ¾ percentage point, to 1.8 percent. Thus far, in any event, that August forecast has proved to be a little too pessimistic; as shown by the rightmost point on the red line, our latest estimate—which now simply tracks the published figure for 2004 from the BEA—is 1.6 percent.
The blue line similarly tracks the evolution of the forecast for 2005. There, as you can see, the upward revision in our forecast has been more pronounced. As shown by the rightmost point on the blue line, the current Greenbook calls for core PCE inflation to come in this year at 2.1 percent. As shown by the green line, we did not begin providing a forecast of inflation in 2006 until last September’s Greenbook, but since then, that forecast has followed a similar upward trajectory. So what has been going on?
The panel to the right provides half of the answer: As you are well aware, oil prices have surprised both us and the rest of the world greatly to the upside. As you can gauge from the vertical distance between the dotted blue line at the bottom and the dashed black line at the top, the forecast implicit in oil futures contracts back in December 2003 for oil today was too low by about $30 per barrel.
But other developments have worked in the same adverse direction. As shown in the panel to the middle left, the PPI for core intermediate materials prices also has risen considerably faster than we expected as of a year and a half ago, and is expected to continue along its current higher trajectory. Similarly, import prices (shown to the right) have risen faster, on net, than we expected in December 2003, even though events have unfolded a little more favorably than we had expected as of December 2004.
The table in the lower left panel traces through the implications of these developments for our inflation outlook. As shown on line 1, between December 2003 and last week’s Greenbook, we’ve revised up our estimate of core PCE inflation in 2004 by 0.5 percentage point, on net, and have marked up our forecast for inflation this year 1 percentage point. Lines 2 and 3 show that the deterioration in the oil outlook can account for about half of the upward revision in both years, while import and materials prices can account for nearly all of the rest, leaving only a small amount, line 4, to be explained by other factors. We take some comfort from this result because we expect the upward impetus from the sources represented on lines 2 and 3 to prove transitory.
As shown in the table to the lower right, as those transitory influences wane, we expect the inflation picture to improve. For this year, as you know, we have topline inflation (line 1) coming in at 2½ percent again, little different from last year, before dropping back to 1¾ percent next year, under the influence of the swing in energy prices. And, as shown on line 4, we have core PCE inflation stepping up ½ percentage point this year before moderating slightly next year.
Given that oil prices have surprised so much to the upside, one might have expected our projections for the growth of real GDP to have been marked down substantially of late. But, as shown in the upper left panel of exhibit 7, that has been only partly true.
This panel does for the growth of real GDP what the one from the preceding exhibit did for core PCE inflation. Thus, the red line shows the evolution of our projection for real GDP growth over the four quarters of 2004. As shown by the steadily sinking portion of that line, which paralleled the feeling we had in our stomachs at the time, we marked down our projection for the growth of real activity substantially as the bad news came in during the first two-thirds of 2004. By contrast, our projections for 2005 and 2006 have been remarkably stable. How could this have been so?
The panel to the upper right shows that the increases in oil prices relative to what we had assumed in the December 2003 Greenbook in fact were, on our estimates, a major negative for the growth of real GDP both last year and this. As shown by the red line, we estimate that oil price increases trimmed half a percentage point from the growth of real GDP in 2004. And, as shown by the blue line, they will cut nearly a percentage point from the growth of real GDP in 2005.
Clearly, oil price effects can be only part of the explanation, because they fail to explain the full extent of the disappointment in 2004 and they explain too much in 2005. The middle left panel fills part of the gap. As you will recall all too clearly, our estimates as to the influence of the partial expensing provision proved rather wide of the mark. As shown by the declining height of the bars marked “2004,” over the course of the past year and a half, we have progressively trimmed our estimate of fiscal impetus in 2004, explaining another 0.2 percentage point of our overall downward revision. By the same token, the bars marked “2005” show that the patching of the pothole plus some unanticipated defense spending have added to the projected growth of real GDP this year.
The table to the middle right summarizes the net effect of these two influences. As shown on lines 2 and 3, the increase in oil prices and the reduction in fiscal stimulus can explain about half of the shortfall in 2004 growth relative to our projection on the eve of the year. The remaining 0.7 percentage point was a puzzle in real time, and largely remains so today.
For 2005, the parsing-out exercise yields a tidier result: By our lights, oil prices were indeed a big negative, but that negative was almost completely offset by our miss on the partial expensing provision. Without doubt, other influences have been bearing on the growth of real GDP but, conveniently, they happen to net to zero in the arithmetic for 2005.
One implication of this exercise is that the underlying growth of demand— supported, to be sure, by low interest rates and appreciable gains in housing and stock market wealth—has been pretty robust. Looking ahead, we expect oil to remain a negative for the level of real GDP, but not to any great degree for its rate of growth going forward. Thus, if oil prices flatten out, as we assume, we would expect the relatively robust pace of underlying demand to show through to topline growth.
The table at the bottom of the page summarizes our outlook for the growth of real GDP. As shown on line 1, for this year and next, we expect real GDP growth to average about 3½ percent per year. This growth rate is close to our estimate of the rate of growth of potential, so (as shown on line 7) the output gap closes only a little from its current level of about 1 percent. Throughout the projection period, domestic final sales are the key driver of aggregate demand, supported by interest rates that remain quite low. Net exports (line 5) subtract a little less from the growth of aggregate demand this year, but resume their more substantial pace of erosion next year. Inventory investment (line 6) is expected to be about neutral for real growth this year and next.