07.23.2007
On the 2006 Economic Projections of the Federal Reserve
Last week, Federal Reserve Board Chairman Bernanke made his midyear appearance on Capitol Hill to testify about the economy and monetary policy. Among his duties at these semiannual events is the release of the economic projections, the most recent vintage of which can be found here.
Gene Epstein of Barron’s offered his perspective on the projections (available by subscription here). He argues that, in retrospect, Chairman Bernanke’s outlook for 2006 was fairly wide of the mark in all three of the major variables—real growth, the rate of unemployment, and core inflation. According to Mr. Epstein, Chairman Bernanke overestimated real growth while at the same time was a little too pessimistic about the rate of unemployment. And core inflation? Well, that was higher than projected. But these were fortuitous errors, says Mr. Epstein, because had the Chairman seen that growth was coming in at 3.1 percent rather than the 3.5 percent projected, he might have projected unemployment to be higher and inflation to be lower, weakening “his resolve to keep hiking the interest rate,” which would have been “ill-advised.”
I cannot speculate about how policy would have unfolded had the economic projections of 2006 been different. I simply don’t have a clue. But the Federal Reserve Bank of Cleveland is a contributor to these projections, and a few observations on them might be useful.
First, the central tendency of the semiannual economic projections reported by Gene Epstein is not necessarily the Chairman’s personal view. All Reserve Bank presidents and Fed governors offer a projection, and where the Chairman’s view lies in that set of projections cannot be determined. Further, the magnitude of the “misses” in the central tendency of the Federal Reserve’s 2006 projections was pretty small given their track record, a track record that compares favorably to private forecasts. (Don’t take my word for it. See this piece by my former colleague Bill Gavin, who, with Rachel Mandal, has evaluated the accuracy of the Federal Reserve’s projections.)
But there’s one article of particular interest—also authored by Bill Gavin–that is worth repeating here. Gavin demonstrates that the midpoint taken from the full range of policymakers opinions is a better predictor of future outcomes than the group’s central tendency (which is calculated by Federal Reserve Board staff by excluding projection outliers.) While Gavin didn’t speculate on why the midpoint of the full range of Federal Reserve forecasts is superior to the group’s central tendency, I will.
Economic forecasting is a hard business and no thoughtful forecaster enters the forecasting arena without a healthy skepticism about his or her ability to see the future—a future full of unknowns. And these unknowns—risks—can sometimes be revealed by the range of opinions coming from people who see and judge them differently. In the case of the 2006 outlook, the outcomes were within the full range of Federal Reserve officials opinions for unemployment and core inflation, and just marginally under the bottom end of the range of opinions for real GDP growth (3.1 percent vs. 3.25 percent).
Perhaps the real usefulness of economic forecasts is not their ability to see what will come next, but rather their ability to identify the risks that stand the best chance of causing the projection to go amiss. And these risks are also articulated in the Chairman’s testimony. In his February 2006 testimony, Chairman Bernanke described three such risks in the 2006 outlook. First, there was a potential for slower real growth due to problems associated with housing; second, high rates of resource utilization could add to inflation pressure; and third, “[a]dditional steep increases in the price of energy [c]ould intensify cost pressures and weigh on economic activity.”
Well, we can quibble over the accuracy of last year’s economic projections, but it’s hard to find fault with the assessment of the risks we faced.
07.20.2007
Core Values
The June CPI report, well described by my colleague David Altig in Macroblog, continued to affirm that the underlying inflation trend seems to have moderated some. Indeed, we cut the CPI data many different ways here in the research department of the Federal Reserve Bank of Cleveland and both the 16% trimmed-mean CPI and the median CPI appear to be giving back some of the acceleration they exhibited last spring.
While I may think the recent readings on inflation have been promising, it’s wise to echo the caution made yesterday by Chairman Bernanke in his Senate testimony that “month-to-month movements in inflation are subject to considerable noise, and some of the recent improvement could also be the result of transitory influences.”
That caution in mind, I think the evidence is pretty compelling that when trying to interpret the monthly, or even quarterly patterns in the price data, core measures help us dial in the signal of future CPI trends (and I back that claim up most recently in this piece.) This is potentially valuable information when the actions of the central bank take a frustratingly long time to influence the aggregate price data.
Highlighting the more moderate behavior of the core CPI measures while the headline CPI is moving sharply in the other direction is a public relations nightmare. It’s not easy to sell an improving inflation picture when the prices people pay at the gas pump continually deteriorate. And that’s the biggest risk of these core indictors. If the discrepancy between the headline CPI—the actual costs influencing household budgets—and the forward-looking core measures is large and persistent, the public’s confidence in the central bank’s resolve to achieve price stability could fade. And that would be a bad thing. Here’s the risk as described in the recently released minutes of the June FOMC meeting, when (May) headline inflation was topping 8 percent (annualized) while the core measures were softening:
“…while core consumer price inflation had moderated of late, total consumer price inflation had moved substantially higher, boosted by rising energy and food prices. While total inflation was expected to slow toward the pace of core inflation over time, a number of participants noted that recent elevated readings posed some risk of a deterioration in inflation expectations.”
I’ve looked into the relationship between the inflation measures and household inflation expectations as reported in survey data. What I’ve found is that households seem to look at the data in much the same way as economists. Specifically, when looking over the immediate future (the next 12 months), households look at all prices—core or not. But when setting their long-term expectations (over the next 5 to ten years), households tend to discount the headline inflation measure and attach greater significance to the behavior of the core measures. That bit of evidence might give you some comfort that the inflation expectations of households will also be seeing some favorable patterns in the price data, despite the continued pressures coming from energy markets.