National News

Why is the Government Always Revising Data?

by Mark Knold, Chief Economist

Why does the federal government seemingly always revise economic data? Because it is part of a plot to frustrate and keep you off balance. So one may think. But actually, the reason the government often revises data is because you and I are relatively impatient. We want whatever information the government can gather as quickly as possible, even if not all of the data has yet been gathered. Time is of the essence, even more so than quality.

If you haven’t surmised by now, economic data comes into the federal government’s domain at a somewhat measured pace. Typical government, right? Functions at the pace of a snail. In many cases, it is not the government’s bureaucratic fault, but is instead, the nature of the programs through which the government collects data.

For example, the U.S. Bureau of Labor Statistics (BLS) is charged with quantifying the nation’s job count. There is a great conduit in place in each state for BLS to accomplish this comprehensive employment count. It is each state’s unemployment insurance program—the one that provides laid-off workers with unemployment benefits. This is a program run by the states, not the federal government.

It works like this. At the end of a calendar quarter—say January, February, and March—in early April the states send out forms to all their employers who hire workers, who then report back the earnings of those workers for the unemployment insurance program. It is a wonderfully thorough way to count people in jobs. Just tabulate the job counts that the businesses send back. But it takes time to get all those forms returned and tabulated. Businesses have better things to do than to fill out the government form as soon as they get it. Because of how comprehensive this can be, it can take until July or August to get a good grasp of the employment levels for January, February, and March. But being impatient as we are, this delay will not do. So what to do?

Do a survey in the meantime—and that BLS does. The ongoing BLS monthly employment and unemployment report, eagerly awaited by the media and the investment community (Wall Street), is nothing more than the results of a survey. But surveys are only built to get within the ballpark. It is the more comprehensive employment counts yet to come through the unemployment insurance programs that will be the real measure of the nation’s employment.

So where do the revisions come in? They arise when the original survey estimations are replaced by the more comprehensive unemployment insurance counts further down the road. If the survey did a good job, the revisions are minimal. But surveys come with an inherent error range, and sometimes the revisions can be noticeable.

Why not just forget surveys and wait for the real data? Because of the nature of people—we want it all and we want it now. Since that is the case, we will have to live with revisions.

Government data elements that see revisions:

• Employment
• Unemployment
• Gross Domestic Product
• Unemployment Insurance Claims
• Retail Sales
• Home Sales
• Corporate Profits
• Exports
• Personal Income
• Consumer Expenditures
• Industrial Productivity