Polling and Measuring Economic Data

Anyone following electoral politics for the last couple of years will likely admit: opinion polling has definitely trended both less accurate and precise. As this article from The New York Times outlines, there are several factors underlying the decline in accuracy in public polling of election races. For the apolitical or those unconcerned with elections, this may not be a problem. But when it comes to economics, this may be a problem that should concern us all.

It is often easy to overlook data and its sources, and take the information we obtain for granted. Yet, behind the numbers on unemployment, growth and GDP, inflation, and other macro and micro variables, are strict methodologies that attempt to measure data as accurately and precisely as possible. These methodologies are of course oriented to ensure large random samples when necessary, and generally to avoid bias and ensure consistency in the variables that organizations and government agencies report on a regular basis. Some of these methods—such as seasonal adjustments—are more familiar than others. But the problem is that, if standard and tested methods are failing when it comes to measuring political sentiment, those economic variables that take in the public’s answers and opinions may be inaccurate as well.

Most economic variables are measured very consistently over time—such as unemployment, by interviewing a set number of households over a time period—and generally, since economic surveys try to capture facts on the economic situation as opposed to opinions, there should in theory be less of a concern than there is in political polling.

However, many other variables cited frequently both by the media and practicing economists—such as consumer confidence, indices of job creation, etc.—rely on the opinion of those polled (which may not be a random sample or not reflect a consistent sample across time) to less clearly defined questions (i.e. questions such as “is your company hiring?” “is the economy headed in a good direction?” “rate the strength of the economy” that might have much less objective criteria or set of answers). Moreover, the methodology to ask these questions often does not rely on the sophisticated methods used to measure other economic variables, such as the Current Population Survey.

As such, the use of these measures to gauge the strength of the economy, potentially define economic or public policy, or sway the outcomes of elections, can be highly problematic. While Economics is a rigorous science, and policy-making an intense study of all the factors in play, how we measure the economy has always been a topic of discussion, and one that we should continue to examine carefully—particularly regarding the sources and, in turn, the accuracy and precision of our measures. From the criticism of economists—including Nobel Prize laureate Joseph Stiglitz—regarding GDP and our measures of welfare, to the surprising lack of accuracy in political opinion polling running up to electoral events, how we measure the data around us should be as important (if not more) as the applications we find for that data. After all, without ascertaining the reliability of the data itself, all inferences obtained from that data should be moot points.