This past week the National Institute for Labor Relations Research (NILRR) issued its latest analysis showing that, according to a host of economic indicators, having a state Right to Work law on the books is positively correlated with faster growth and higher employee earnings and overall incomes, when regional differences in the cost of living are taken into account. (See the first link below.)
For the average concerned citizen, the fact that both total private-sector employment and total real employee compensation grew more than twice as fast in Right to Work states as in forced-unionism states from 2002 to 2012 is significant evidence that ought to be considered in states that are now debating whether or not to adopt their own laws prohibiting compulsory union dues and fees. So is the fact that over the same period the aggregate young-adult (aged 25-34) population of Right to Work state grews nearly 4.5 times as rapidly as the young-adult population of forced-unionism states.
However, professional apologists for monopolistic unionism such as University of Oregon labor-studies professor Gordon Lafer, who is also affiliated with the Big Labor-funded Economic Policy Institute (EPI), insist the public ought to ignore the positive correlations between Right to Work and job and income growth, as well as income levels, documented by NILRR because NILRR does not attempt to incorporate the data it collects into an economic model.
Lafer’s dismissiveness is unwarranted. As a large and growing number of professional economists acknowledge, in their field the use of “models” often leads to a “false sense of precision and understanding,” as George Mason University’s Russ Roberts put it in a January 2011 contribution to the Cafe Hayek blog. (See the second link below.)
Roberts, who earned his PhD in economics from the University of Chicago in 1981, frankly describes his field as “not a very predictive or very precise science or whatever you want to call it.”
While Roberts commends economics as an excellent tool for “organizing your thinking,” he argues that theoretical economic models rarely have any predictive value:
My claim is simply that we should recognize the limits of reason in analyzing complex systems with millions of decision-makers, numerous feedback loops, institutional features (synthetic CDOs, the repo market, the willingness of the Fed to bail out bondholders) that are difficult to model in tandem with the outcomes we care about. Finally, there are important variables that we cannot observe directly such as expectations, anxiety, confidence, overconfidence and so on.
On the other hand, even though they cannot measure causes and effects “with great precision,” economists can “understand something and sometimes a lot about the direction that things are likely to go if something changes and nothing else does.” And “armchair reasoning” drawing largely on “correlations in the data that we might view as sufficiently close to natural experiments” is the basis of much of what is known today in economics.
Theoretical models “of varying degrees of complexity” can be helpful, but models that purport to assess the economic impact of policy changes “with numerical precision,” as do the EPI’s anti-Right to Work analyses and, sadly, some analyses made by academic Right Work proponents, are “scientism and intellectually bankrupt.”
In economics, correlations in the data are often the best evidence available. They do not enable you to predict the economic future, but neither do models that only pretend to be able to do so.