The Upshot

STATISTICAL PUZZLE

Shrinking Economy? No, but It’s Not Surging Either

MAY 29, 2015

Justin Wolfers

The government reckons that the American economy shrank over the winter, but no one really believes it.

Specifically, the Bureau of Economic Analysis on Friday revised the nation’s gross domestic product to a new estimate that it contracted by 0.7 percent in the first three months of the year from its initial guess that the economy grew over the winter at an annual rate of 0.2 percent.

There seem to be as many asterisks on this report as might apply to Barry Bonds’s baseball records. A better reading is that the economy continued to grow through the winter, albeit at a slower rate than in late 2014. While it is unlikely that growth was actually negative, it was still disappointing.

As I reported before the initial release of these numbers, many economists believe that there’s a glitch in how government statisticians calculate these numbers. At issue is problems with how they adjust for the economy’s usual seasonal ups and downs. These problems — which the government statisticians have acknowledged and promise to fix — mean that the official numbers have year after year understated economic growth in the winter quarter.

Another snowstorm in New York City, on March 5. A tough winter in heavily populated areas helped slow economic activity in the first quarter. CreditDamon Winter/The New York Times

A reasonable guess is that this problem subtracted one or two percentage points from the headline measure. So this glitch alone explains why the numbers recorded a shrinking economy even when it probably grew.

Fortunately, there’s a more reliable measure of economic growth, although it remains scandalously underappreciated. The headline that attracts all the attention is based on adding up each dollar of spending across the whole of the economy. But because every dollar you spend is a dollar of income to someone else, you can also measure the size of the economy by adding up people’s incomes. In theory, the two measures should be identical. In practice, they differ because they’re based on different surveys. An impressive body of research has demonstrated that the income-based measure is more reliable.

This alternative measure, called gross domestic income, grew at an annual rate of 1.4 percent in the winter, after growing at an impressive 3.7 percent the previous quarter. This alternative measure bears special emphasis right now because it appears to be unaffected by the seasonal adjustment problems that are muddling the spending-based estimate.

Here at The Upshot we have been emphasizing this alternative measure for some time. It’s a point that the Bureau of Economic Analysis has recently endorsed, announcing that as of July it will publish a new measure of gross domestic product: a simple average of the spending- and income-based measures. Economic commentary will be enormously improved if this becomes the new headline that analysts focus on. Right now, it suggests that the economy grew a healthy 3.1 percent over the past year.

The other asterisks on this report are less about problems with measurement and more about what the measurements represent. For instance, this winter was especially snowy in heavily populated areas, and international trade was disrupted by a port strike. Jason Furman, the White House economist, estimates that the snow alone “reduced annualized growth by about a full percentage point this quarter.” The port strike might be worth another quarter or perhaps half a percentage point. But the snow has melted, and the ports are open again, so this is the sort of setback that should not cause too much worry.

Moreover, much of our economic weakness occurred in types of spending where weakness is often fleeting. To get a sense of the underlying momentum of business and consumer spending, economists often focus on those components that tend to predict future momentum. In particular, the Bureau of Economic Analysis will soon emphasize a measure called “final sales to private domestic purchasers.”

It’s a dreadful mouthful, but it’s basically a measure that discards the volatile parts of G.D.P. — inventories, international trade and government spending. That measure grew by 1.2 percent in the winter quarter. Even as the headline estimate of economic growth was revised downward, this measure of underlying momentum was revised up a tick, suggesting that all of the bad news in this report occurred in volatile measures that do not much speak to the economy’s prospects.

All of this suggests that the economy is growing faster than this extremely pessimistic report suggests. This isn’t an argument for optimism, though, so much as reduced pessimism.

However you view the numbers, economic growth looks to have slowed through the winter to a disappointing pace. More worryingly, the one-off factors that slowed the economy in winter should cause faster growth in the spring, as the economy snaps back on track, but this has yet to materialize.

Given the statistical fog, the best forecast I can muster is to expect uncertainty to continue to cloud the outlook.