≡ Menu

Falsifiability and Consistency

Here’s a follow-up to my long post of earlier today in which I argued that economics’s ability to enhance our understanding of observed reality is every bit as scientific and worthwhile as is another science’s ability to generate specific, falsifiable predictions about real-world events.

One of the most desirable functions served by specific, falsifiable predictions about empirical reality is that such predictions prevent the analyst – or anyone invoking the science in question – from making truth claims that are purely subjective.  The insistence upon such predictions, when such predictions are possible, prevents the mere fact that Jones feels or believes or ‘wants’  X to be true to be taken by anyone, including Jones himself, as a sufficient reason to pay heed to Jones’s claim that X is true.  A specific, falsifiable prediction about X supplies a reasonably objective criterion by which people can, with some assurance, judge without bias the validity of Jones’s claim about X.

Yet such an objective criterion for judging the validity of someone’s truth claim does not necessarily require that that someone make a specific, falsifiable prediction about the real world.  Another objective criterion is consistency.  If Jones claims that 1+1=2 on earth and on all other celestial bodies other than Mercury, Jones is obliged to explain why being on Mercury creates an exception to this rule.  His general acceptance of the rule that 1+1=2 supplies a objective criterion against which his claim that, on Mercury, 1+1≠2 can be evaluated.

Jones’s insistence that on earth and most other places 1+1=2 constrains any attempts he might make to insist that 1+1≠2 on Mercury.  Jones’s insistence that on earth and most other places 1+1=2 gives people hearing Jones not only greater reason to doubt his claim that, on Mercury, 1+1≠2, but an objective fact to point to when asking Jones to reconcile his apparently inconsistent claims.

I do not claim that the demand for consistency in theorizing (that is, the demand for consistency in story-telling about reality) provides as much objectivity, as well as much protection against the mere whims or human imperfections of analysts, as do specific, falsifiable predictions.  Indeed, I believe that they do not do so.  But when reality does not permit specific, falsifiable predictions – as it often does not – there are other means we use to promote scientific objectivity.  Consistency of theoretical explanations is one such means.


Regular Cafe patrons likely predict (!) correctly that I will end this post with the example of minimum-wage legislation.

Nearly everyone on earth understands and accepts the law of demand: if the cost to Smith of doing Y rises, Smith will do less of Y (and, of course, if the cost to Smith of doing Y falls, Smith will do more of Y).  Yet not only do all economists share in this general recognition of the validity of the law of demand, they – as economists – use this law as the basis for much of their formal theorizing.  So when some economist argues that raising employers’ cost of employing low-skilled labor empirically does not cause employers to employ less low-skilled labor, that economist makes a claim that warrants extra scrutiny and skepticism.  The reason for this extra scrutiny and skepticism is that this claim about the effects of minimum-wage legislation is objectively – observably – at odds with a general, often formal, proposition that that economist makes in almost every other situation.

This inconsistency might be only apparent: the economist can specify conditions (namely, monopsony power in the labor market as a necessary condition but not a sufficient one) that cause real-world consequences of the minimum wage to be observably different from those that are ‘predicted’ by the law of demand.  That is, it’s possible for the economist to reconcile her claim about the no-disemployment effects of the minimum wage with her general acceptance of the law of demand.  But when she makes such a claim about the minimum wage – namely, that it doesn’t reduce any workers’ employment prospects – she then must bear an extra-heavy burden of persuasion.  She bears this extra-heavy burden because in making her claim she is objectively contradicting a general proposition that she normally accepts as valid.  Pointing to some empirical studies (each one necessarily done in a world of, as Paul Krugman says, “other things” that change frequently) while repeating “monopsony power” is simply insufficient to cause good economists to accept as a general truth the claim that raising the minimum wage, even by only a little bit, will not in reality destroy jobs for some low-skilled workers.