Don’t Believe That GDP Number. Or Any Other Number.

(Bloomberg Opinion) -- Gross domestic product grew at an inflation-adjusted annual pace of 4.2 percent in the second quarter, we learned this week. Great, right? Not so fast:

GDwhat? GDI is gross domestic income, yet another metric of economic activity produced by the GDP estimators at the Bureau of Economic Analysis that’s released with a little more lag time than GDP. That is, the “advance” estimate of second-quarter GDP that was released in late July contained no information on second-quarter GDI. The “second estimate” of GDP that came out Wednesday did.

Now you might think that Jason Furman, a professor at Harvard’s Kennedy School of Government who was the final chairman of President Barack Obama’s Council of Economic Advisers, was just emphasizing this to make the current presidential administration look bad. But that’s really not fair, given that Furman wrote an op-ed in the Wall Street Journal in early July that wielded the GDP-GDI average to argue that the economy had really grown at a 2.8 percent pace in the first quarter, not the 2 percent reflected in GDP growth alone (those numbers have since been revised to 3.1 percent and 2.2 percent).

Also, it’s not just Furman and not just this year. The BEA started reporting the GDI-GDP average in 2015. And in 2013, the Federal Reserve Bank of Philadelphia began publishing its own “alternative measure of real U.S. output growth” called “GDPplus,” which combines GDP and GDI using a statistical smoothing technique known as a Kalman filter. That delivered a 2.1 economic growth rate for the second quarter, down from 3.3 percent in the first quarter.

GDP and GDI are both estimates of the size of the economy, one focused on expenditure and the other on income. They should in theory add up to the same amount, but they apparently never do, even after multiple rounds of revisions. The smallest gap on record is currently $200 million in 2012 dollars, recorded in the first quarter of 1974. In that quarter GDP was bigger than GDI, which it has been about two-thirds of the time since 1947. But during the current economic expansion, GDI has been bigger for 31 out of 36 quarters.

I’m not aware of any comprehensible explanation for this, but in a 2010 paper published by the Brookings Institution, economist Jeremy Nalewaik — then at the Federal Reserve Board, previously at the BEA and now at Morgan Stanley — did argue that GDI was “the more accurate measure of output growth.” This was because, he wrote, initial GDI estimates:

  1. Appeared to predict subsequent revisions to GDP growth.
  2. Were a “better predictor of a wide variety of business cycle indicators that should be correlated with true output growth,” and
  3. Had identified the onset of each of the last few cyclical downturns more quickly than initial GDP estimates had.

“Henceforth it will be impossible for macroeconomic analyses to proceed comfortably simply using GDP,” declared prominent University of Pennsylvania economist Francis X. Diebold, one of the commenters on the paper. Diebold then collaborated with Nalewaik and several other economists to devise the Philadelphia Fed’s GDPplus.

Before one concludes from all this that the economy really only grew at just a 2.1 percent pace in the second quarter, though, it’s important to note that the GDI fans have yet to convert all of their colleagues in the economics profession. The other commenter on that 2010 paper, long-time BEA director J. Steven Landefeld (he retired in 2014), said that while he appreciated Nalewaik’s efforts, “the evidence suggests that GDP and GDI provide roughly the same picture of economic activity over the business cycle and that a review of the source data and performance of the two measures favors GDP rather than GDI.” And just this May, three Federal Reserve Bank of San Francisco economists studied the record of GDP, GDI and GDPplus in forecasting subsequent changes in the unemployment rate and the Chicago Fed National Activity Index, a measure that sums up 85 different economic data series, and found that GDP did the best job.

Meanwhile, in his time on the Council of Economic Advisers and since, Furman has also touted another measure that the BEA started reporting in 2015, final sales to private domestic purchasers, as a more reliable gauge of “the underlying trend in the economy” because it strips out exports and inventory fluctuations and with them presumably some of the volatility of GDP. In the second quarter, though, real final sales to private domestic purchasers rose at a 4.3 percent annualized pace — faster than GDP.

Don’t Believe That GDP Number. Or Any Other Number.

Perhaps the most useful lesson to take from all this is that accounting for national income and output is an imprecise business, and that the economy was probably growing at an annual rate somewhere between 2.1 percent and 4.3 percent in the second quarter. Which is pretty good.

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

Justin Fox is a Bloomberg Opinion columnist covering business. He was the editorial director of Harvard Business Review and wrote for Time, Fortune and American Banker. He is the author of “The Myth of the Rational Market.”

©2018 Bloomberg L.P.