Election-Year Preliminary GDP Estimates Are Overly Optimistic
Friday’s advance estimate of real GDP growth in the third quarter of 2016 came in at an encouraging 2.9 percent. This represents a significant uptick from the Bureau of Economic Analysis’ GDP figures for the previous periods: 0.8 percent growth in the first quarter and 1.4 percent in the second quarter, making an average of 1.7 percent for 2016.
Before celebrating the good news, keep in mind that the preliminary estimate is, well, preliminary. BEA will release two more estimates before it gives us the final figure. The 2.9 percent number could be revised up or down—perhaps significantly. For instance, in the second quarter the advance GDP growth estimate was 1.2 percent, but that was later revised up to 1.4 percent. Much larger misses can occur: in the first quarter of 2015, the initial figure was 0.2 percent, but the final estimate was 2.0 percent.
Third quarters in presidential election years tend to be optimistic, with 58 percent of advance estimates in these quarters revised down by more than half a percentage point. Two-thirds of all third-quarter estimates in election years were revised down. It appears to be a bipartisan phenomenon.
How common are such misses in general? The Federal Reserve Bank of Philadelphia archives the preliminary GDP estimates for each quarter going back to 1965, along with the corresponding final GDP figures. For the past half-century, statisticians have underestimated actual GDP growth in the preliminary estimate by an average of 0.4 percentage points.
However, the absolute errors are much larger. The preliminary estimate has missed the mark by an average of 1.6 percentage points since 1965. These errors have gone down over time as survey methods have improved: looking at just the last five years, the absolute error was only 0.7 percentage points.
It is reasonable, though not certain, to expect that the third quarter’s preliminary estimate will be revised. Which direction, though, is still a mystery. Historically, preliminary figures have underestimated the final number—in other words, the final GDP figures are, more often than not, better than initial estimates suggest.
That may have changed, though. In the current decade, the final GDP numbers have been more than 0.5 percentage points higher than preliminary estimates in 27 percent of quarters. Similarly, final GDP numbers have been more than 0.5 percentage points lower than preliminary estimates in 27 percent of quarters. In slightly less than half of quarters, the preliminary estimate has not missed by more than 0.5 percentage points in either direction—quite an improvement in accuracy over past decades.
It is also worth noting that historically, third-quarter preliminary estimates have been more optimistic. Forty-five percent of third-quarter advance estimates were revised down by more than 0.5 percentage points—a higher proportion than the first quarter (31 percent), the second quarter (18 percent), and the fourth quarter (29 percent). High initial estimates are also a bit more likely to be revised down.
Oddly enough, third quarters of presidential election years tend to see more optimistic preliminary figures, with 67 percent revised down. Fifty-eight percent of advance estimates were eventually revised down by more than 0.5 percentage points. While it may be tempting to attribute this to poll-obsessed presidents nudging the numbers in their favor, keep in mind that we only have 12 elections to look at—not a great sample size.
A very simple model with all these variables—quarter, historical time trend, and advance estimate—predicts third-quarter GDP growth will be revised down 0.9 percentage points, from 2.9 percent to 2.0 percent. I wouldn’t bet on it, though: these variables only explain 11 percent of the variation in historical GDP growth revisions. The most recent figure, though it seems optimistic, could easily be revised up instead of down. In life, nothing is certain. That includes statistics.
Preston Cooper is a fellow at the Manhattan Institute. You can follow him on Twitter here.
Interested in real economic insights? Want to stay ahead of the competition? Each weekday morning, E21 delivers a short email that includes E21 exclusive commentaries and the latest market news and updates from Washington. Sign up for the E21 Morning Ebrief.