The economic data has gotten confusing again. For two months running, we have had very strong jobs reports, showing accelerated job growth and wage growth. On the other hand, consumer sentiment nosedived in January, and now we have learned seasonally adjusted retail sales unexpectedly fell in December, the key Christmas shopping month.
A report from the Census Bureau said retail sales fell 1.2 percent in December, after adjusting for seasonal variation, the fastest rate of decline since 2009. Some of the decline was due to falling gas prices (when gas is cheaper, people spend less at gas stations, for obvious reasons) — but even after adjusting out gasoline sales, the drop was 0.9 percent.
What’s going on here? Is the economy getting stronger or weaker?
“I find it mystifying,” says Jason Furman, a professor of economic policy at Harvard’s Kennedy School of Government and former top economic adviser to Barack Obama.
Furman says we’re seeing signs of slowing economic growth at the same time as robust employment growth. That could be explained by low productivity growth: More people are working, but they’re not getting a lot better at what they’re doing. But that would be inconsistent with the recent wage data — wages are rising faster than prices, which tends to happen when productivity growth is strong.
As for consumer sentiment, which fell in January to its lowest level since Trump was elected, Furman says the government shutdown was a likely factor. If so, we may see improvement in the February survey. Furman also said he found, when looking at past shutdown and debt-limit disputes, these events appeared to affect measures of consumer sentiment but did not affect economic activity in a way that could be measured at the economy-wide level. So, this drop in sentiment may not tell us much about the real economy.
Ernie Tedeschi, a former Treasury Department economist who now produces economic forecasts for the investment advisory firm Evercore, emphasizes that a slowdown in the rate of economic growth is consistent with forecasts from the Federal Reserve and others. He says there was a “sugar high” from the expectation and enactment of the tax cut bill, and as that wears off, you should expect the economy to grow more slowly — which is not to say you should expect a recession, in which the economy actually shrinks.
“We’re for the most part talking about magnitudes of positive growth,” he said. “We’re debating whether or not different measures are as good as we expect them to be. We’re still not in a world where — I think, I hope — we are talking about things being, in absolute terms, bad.”
Okay, but what about the falling retail sales number? That’s not a slowdown in growth. That’s an actual decline. But, it is only one month of preliminary data. We may get a later revision that makes the figure look better. Tedeschi also notes the survey on retail sales was conducted during the government shutdown, which could be an added factor influencing its accuracy.
“The Census Bureau swears that even though there was the government shutdown, that they were still getting timely data, complete data, response rates were actually a little higher than normal,” he told me. “I trust them, but we can’t rule out that there’s some tainting of the data because of the government shutdown.”
Furman is also more inclined to look at the (positive) jobs numbers, rather than the (worrying) figures on economic output like retail sales.
“As a general rule, I would place two-thirds weight on employment data and one-third weight on output data in signal extraction,” he told me, regarding which economic indicators tell you most about where the economy is headed, based on research he did at the White House Council of Economic Advisers. “And in this case, since it is not the full output data, would put even less than one-third weight on it.”
Doug Holtz-Eakin, the president of the conservative American Action Forum and a former top economic adviser to George W. Bush and John McCain, told me he hasn’t yet come up with a story to reconcile the contradictory data because it’s early yet. The sentiment number and the retail number both look bad — but in each case, it’s just one month of data.
“Those can both be pretty noisy, so I’m just waiting for more info,” he said.