In Sit-Down Restaurants, an Economic Indicator

Have you been eating more at restaurants with waiters rather than fast-food joints?

If so, you are not alone, and that in fact is an indication that the American economy is improving.

Over the 12 months through January, sales at what the government calls full-service restaurants were 8.7 percent higher than in the previous 12 months. That was the fastest pace of growth since the late 1990s, when the economy was booming. Moreover, as is seen in the accompanying charts, that rate was much greater than the rate of growth in sales at limited-service restaurants.

Since those numbers became available 20 years ago, that difference has been a reliable indicator of how the economy is going. In tough times, people may still eat out, but they cut back.

Full-service restaurants may or may not be expensive. Le Bernardin in Manhattan qualifies, but so does Red Lobster. The range at limited-service places is not nearly as wide.

Americans now spend about $220 billion a year at full-service restaurants, and $211 billion at the limited-service places. (They also spend $21 billion at what the government calls “drinking places,” also known as bars. Bar sales are now rising slower than at either type of restaurant, although history does not indicate that has any particular significance for the economy.)

To my knowledge, no previous article has pointed to the relative sales-growth figures at different types of restaurants as an economic indicator. The lack of attention from economists may be partly because of the fact that the government releases the restaurant figures on a delayed basis and does not offer seasonal adjustments for the sectors. The Census Bureau reported the January numbers this week as it reported overall retail sales for February, and most attention was focused on the fact that overall sales improved in February and that figures for the previous two months were revised upward.

As can be seen from the charts, over time the relative sales trends of the different types of restaurants have generally coincided with changes in the gross domestic product numbers. But recently, that relationship seems to have broken down, with the economy growing much more slowly than the restaurant numbers would indicate.

There is a much better-known economic relationship, known as Okun’s Law, that also seems to have stopped working. That law, propounded by the late Arthur Okun, ties G.D.P. growth to changes in the unemployment rate. In late 2011, the rapid decline in the jobless rate should have been associated with strong G.D.P. growth, not the slow growth that was reported. Some commentators have suggested the law is no longer applicable and looked for reasons.

Perhaps, however, the problem lies not with Okun’s Law but with the uncertainty of G.D.P. figures.

Back in 2009, there was a similar discussion of the apparent failure of Okun’s Law, but in the opposite direction. Then, the unemployment rate had risen to an extent that could not be explained by the G.D.P. figures.

It turned out most of the gap was caused by bad G.D.P. numbers. At the time, the government said the economy had declined by less than 1 percent from the fourth quarter of 2007 through the fourth quarter of 2008. Now, after all the revisions to the G.D.P. — revisions reflected in the accompanying chart — the economy is believed to have shrunk by 3.3 percent in 2008.

Could it be that Okun’s Law is still working, at least to a greater extent than now appears, and that both the decline in unemployment and the change in restaurant sales are providing more accurate indications of the state of the economy? If so, the benchmark revisions for the 2011 G.D.P. numbers, expected in July, could provide evidence of that.