Obviously, in terms of our understanding how much productivity and efficiency gains are still there to be harvested—let’s say one-time increases—we have almost nothing upon which to gauge that. We can talk to businesses or try to undertake econometric estimates, but we have been consistently surprised over the course of the last two years by the performance of productivity. One of the biggest challenges in putting together the forecast has been to try to assess how much of what we’ve seen in productivity is going to be sustainable growth going forward. We don’t know how much was a series of one-time adjustments that firms accumulated over the late 1990s through tremendous amounts of capital investments or how technology may have offered organizational efficiencies that businesses are just being able to take advantage of.
I don’t think we have very much confidence at all in our ability to pinpoint the quarter or even half-year in which that stock of efficiencies will play out. Could that go on for another year? I think that’s certainly possible, and one of the simulations we had in the Greenbook was in essence a higher level of productivity prospectively. What that forecast produced was a weaker labor market, higher unemployment, and lower inflation than we are showing in the baseline. I think you have to give that some reasonable probability weight in your thinking about how the employment situation is likely to progress.
On the other hand, as hard as it is to believe that there are upside risks to our productivity forecast, that also remains a possibility. A year or so ago productivity turned out to be much stronger than we thought. So it may be that over the second half of last year there was in fact an improvement in underlying productivity that was being recognized by businesses and households and those entities really did step up their spending quite significantly. And it may be that what we are seeing now is just not a terribly unusual lag between that step-up in output growth and a pickup in employment and that we will soon start to see a more significant increase in jobs. Taking our models that use labor market indicators—such as initial claims, layoff announcements, and past payroll employment gains—we don’t need to see any further improvement in those indicators to get job growth on the order of the 150,000 per month that we are forecasting for the first half. Those same models have been predicting better employment gains than we’ve seen for the last six months. So there are still some very significant risks, and I think that your queasiness about the extent of this step-up is well founded because of our inability to be able to pinpoint precisely when those productivity opportunities will for the most part have played out.