The Uncertainty Of Economic Data
As though it's not difficult enough to divine what the future holds when making policy or business decisions, it's sometimes just as difficult to know what has happened in the past. This became vividly clear in early March, when it was announced that, instead of adding more than 40,000 jobs during 2011, Massachusetts added less than 10,000. In other words, our assessment of last year had to be revised to include 31,600 fewer jobs that we had thought. Our original view of the past was not pure guesswork, but rather based on official government data releases at both the state and federal levels.
It's extremely important to have an accurate and up-to-the-minute accounting of economic performance. State economic data are used in a variety of decision making, both public and private. In the public sector, economic performance is employed to help form policy making as well as politicking. Budgets are developed using "best forecasts" of the coming fiscal year. Forecasting is always, at least in part, seeing the future by looking at the past. If historical data are flawed, then forecasts of future performance will also be flawed. In the private sector, business decision makers want to see customers come through the door, and an important way to anticipate future customers is to gauge the current health of the economy.
So why are economic data subject to so much change? A stylized timeline can be constructed that applies to virtually all economic data, at the local, state, or national level. Let's use jobs data as an example. Our first indication of job growth or decline is based on a large survey of companies, hundreds of thousands at the national level, thousands within the state. The sample data are subject to revision, even after its release, as companies in the survey can be slow to submit their data and as errors in already-received survey forms are corrected. Once initially released, data are subject to this type of revision fairly quickly.
Some months after the initial data release and the first round of revision, a new data source is used to "benchmark" the jobs numbers, using a more complete jobs picture. Virtually all firms are required to register with the state's unemployment insurance (UI) programs, and the enrollment process provides a nearly complete census of jobs. So the first data release has already been subject to two revisions. Finally, the UI data are supplemented with information from organizations not enrolled in the system. This happens more than a year after the initial release of data, and provides the third revision.
For other types of information, such as unemployment and gross domestic product data, the details differ but the process is very similar. Sample data are revised as more data arrive and as more comprehensive information becomes available.
Today, there are many of us who believe that when the dust clears on 2011, job growth will be much closer to the original estimate than the current revision.
Welcome to the murky world of economic data.
Robert Nakosteen is a professor of economics at the University of Massachusetts Amherst and executive editor of MassBenchmarks, a state economic journal published by the UMass Donahue Institute.
WBJ Web Partners
Most Recent
Most Recent
0 Comments