HICKS: Why economic forecasting is inexact science

January 1, 2011

New Year’s is a time to reflect upon the passing of a year and what the coming one might bring. For an economist of a certain stripe, that means thoughts of forecasts. I would like to explain how and why economists forecast the economy.

Forecasts are primarily used as a tool to begin, not end, conversations about business and government matters. It is easier to craft a budget or plan production from a forecast than to argue about a starting point. So a forecast can be useful and is worthy of serious effort.

It should come as no surprise that most economists have nothing to do with forecasts. Like any discipline, the field is too broad to develop technical expertise in everything.

To begin with, forecasting the economy is very different from estimating the effect of a policy, such as a change in tax rates or regulations. Economists have a much better grasp on these matters for a variety of reasons, not least of which is a relative abundance of data and experience. There have been four recessions in the past 30 years, and about a bazillion tax-rate changes at federal, state and local levels. It is literally akin to the problem of forecasting a hurricane season versus rain this afternoon.

Most forecasts today use a mix of two approaches. Both involve constructing a set of mathematical equations for which data must be available. One tactic to construct equations allows a forecaster to create relationships from observing recurring economic events, without the need to know anything about causation. The second method involves linking known relationships between different types of economic activity. For example, as gas prices rise, fewer cars are sold. In both cases, the strength of the relationships is based on historical evidence, with more recent experience carrying heavier weight. The models economists use are series of these types of equations, which are solved to come up with a prediction of the future.

These forecasts are pretty straightforward, and most of us who dabble in this dark science actually use a trick of forecasting past events as a tool for assessing how good the model actually works. To no surprise, there’s a lot of competition in forecasting. Anyone who can build a model to predict a major recession, given what was known before it started, will become rich.

These approaches have some critics. Better data would yield better models of course (that is what experimental economists are busy creating). Also, economists have long been criticized for assuming that people are rational. But, one need attend only one faculty meeting to know that economists have an unusual definition of rationality. Economists are happy if people respond to price changes given the information that is available to them. This is a low bar for social interactions, but works well for forecasts.

So, the ways economists predict growth is at least a little less murky. Now, if only it were more accurate.•


Hicks is director of the Center for Business and Economic Research at Ball State University. His column appears weekly. He can be reached at cber@bsu.edu.


Recent Articles by Mike Hicks

Comments powered by Disqus