Judging the Fed’s Inflation Forecast
Does it tell us anything?
Today the Bureau of Economic Analysis (BEA) released its May estimate for the Federal Reserve’s preferred price index, the Personal Consumption Expenditure Price Index (PCEPI)—this estimate is part of the BEA’S larger release of data on personal income and consumer spending. There is good news on the inflation front: the PCEPI rate of inflation from April to May equaled zero percent! Coupled with the similarly stagnant CPI rate of inflation, one might dare to dream that inflation is finally stabilizing.1
The May value for the PCEPI comes a week after the Fed stood pat on their target interest rate. As discussed last week by Macrosight, the Fed’s projections from that meeting revealed that the Fed turned more pessimistic on inflation and the potential for rate cuts. Yet, on the heels of the BEA’s latest data drop, we might now speculate on the increased chance of a rate cut sooner rather than later.
This good inflation news has Macrosight feeling quite optimistic. Yet, contrasted with the Fed’s pessimistic turn last week, Macrosight cannot help but wonder: what good are the Fed’s forecasts anyway?
The Fed’s Pessimism
Figure 1 displays a snapshot of the Fed’s latest forecasts, released last week (the same snapshot from last week’s Macrosight post), with the inflation forecast circled in green.
As highlighted on Figure 1, the Fed bumped their PCEPI inflation forecast for the end of 2024 from 2.4 percent (the value from March 2024’s meeting), to 2.6 percent. That is not a huge increase, but it is an increase nonetheless.
That increase coincides with the pessimistic change in the “projected appropriate policy path” for the Federal Funds rate from an expected 4.6 percent by the end of the year to 5.1 percent (shown at the bottom of the snapshot).
The Fed’s Track Record
The juxtaposition between the Fed’s recent pessimism and the recent inflation data made Macrosight wonder how the Fed has adjusted their inflation forecast the past few years. For instance, have their forecasts done a good job of anticipating actual inflation?
Table 1 displays the Fed’s forecasted value of inflation for 2024, at each of the last 14 meetings at which projections were published (from March 2021 through June 2024, available here). Alongside those forecasts are the actual values of PCEPI inflation at the conclusion of the meeting-month (measured as the percent change from a year ago).
What does this Table show?
The Fed’s forecast has been relatively stable over this time period, while the actual rate of inflation has not (the standard deviation of the former is 0.2 percent, while the latter’s standard deviation is 1.7 percent).
As the actual rate of inflation was rising rapidly, the Fed’s forecast for the PCEPI was adjusted slowly and in small increments. Over this time period the PCEPI rate of inflation averaged 4.7 percent, reached as high as 7.1 percent, yet the forecast never breached 2.5 percent until this month.
What does this Table tell us? A point-counterpoint
Point: In defense of the Fed, the forecast values represent the median value of the numerous forecasts produced within the Federal Reserve System (as explained here and here if you scroll down). Indeed, the full table of Fed projections, of which Figure 1 above only shows a segment, provides the ranges from which the median estimates are drawn. As such, one would expect a “smoothed” or stable series.
Counterpoint: In critique of the Fed, smoothed or not, if the forecast provides little insight into the nature of inflation, then what good is the forecast?
That is, what insight does the recent increase in the forecast from 2.4 percent in March to 2.6 percent in June impart? Does it tell us anything about what to expect for the rest of 2024?
Or, if the last fourteen forecasted values appear to have little correlation to the evolution of inflation over that time span, then what can we possibly hope to glean from the recent forecast values for the rest of 2024?
The contrast between the forecast since 2021 and the actual rate of inflation certainly leaves one wanting.
Counterpoint to the Counterpoint: This discussion is not meant to be a take-down of the Fed. In fact, as I tell my macro students, there is no greater collection of macroeconomic geeks, nerds and data-dorks than in the Federal Reserve System. And I say so with affection. There are dozens (if not hundreds) of statisticians, data scientists, and economists all trying to the forecast the macroeconomy using state-of-the-art statistical techniques. These macro enthusiasts are not political hacks or shills; they are experienced professionals that have devoted many years of their lives getting masters degrees and PhDs, and working for the Federal Reserve (when most if not all could earn more money in the private sector).
Rather, what Macrosight takes away from the contrast shown in Table 1 is the following: forecasting the macroeconomy is really, really, fricking difficult. Even the most experienced forecasters, using the most sophisticated techniques, can only do so much (especially in a time of relatively volatile inflation). The content of Table 1 is a manifestation of that reality.
Where does that leave us going forward? Like the Fed, the best we can do is assess a multitude of information in real-time, update our expectations accordingly, and assess our errors as we go. Oh, and read Macrosight.
One could argue that inflation has already stabilized. From June 2023 through May 2024, the PCEPI rate of inflation averaged 2.9 percent with a low of 2.5 percent and a high of 3.4 percent (as measured on a monthly basis as “percent change from a year ago”). The issue with respect to the Fed is that their target for inflation is 2.0 percent. So, we could instead say “one might dare to dream that inflation is going to soon converge to the Fed’s 2.0 percent target.”