Tuesday, November 25, 2014

Study That

Here’s a line from this report in the New York Times; I add emphasis to that part of the line that is most germane to this blog post:

Studies show that simply raising the price of an alcoholic beverage by 10 percent reduces alcohol consumption by 7 percent, suggesting that higher taxes on alcohol could make a significant dent in excessive drinking.

Is anyone surprised by this finding?  I’m not.  I have no doubt that any increase in the cost to consumers of acquiring alcoholic beverages reduces the consumption of those beverages from what that consumption would have been had the cost of alcoholic beverages not risen.  I’m much less certain about the seven-percent finding.  It could be accurate for the time period that was studied; this figure sounds plausible enough.  But the actual figure might be higher or it might be lower.  That question – the question of the magnitude of the decline in alcohol consumption – isn’t of interest to me here.  I am, however, quite certain about the direction of the change in alcohol consumption that is caused by a change in the consumers’ costs of acquiring alcohol: higher cost, lower demand.

I’m also quite certain that intrepid-enough researchers could produce honest and impressive-looking empirical studies that find that raising the price of alcohol by 10 percent does not result in any reduction in the consumption of alcohol.  I’m pretty sure, however, that no one has done such studies.  The reason is that no one – at least no respectable economist – doubts for a moment that the higher the price of alcohol, ceteris paribus, the lower is the quantity demanded of some well-specified unit of alcohol.  Therefore, when study after study turns up the expected relationship between price and quantity demanded for alcohol, it makes sense.  No one is inspired to set about to add different control variables, to segregate the data into finer categories, or to otherwise fiddle with the empirical studies in order to see if it’s possible to find empirical evidence in favor of the proposition that, at least for small increases in the retail price of alcohol, higher alcohol prices do not decrease the quantities of alcohol demanded.

And if some researchers did undertake such a project, when they would then try to publish their findings in a scientific journal their paper(s) would be rejected.  The referees of the paper would, quite correctly no doubt, point out any number of flaws in the specification of the model, the manner of collecting data, the classification of the data, and the interpretation of the empirical findings.  The findings would just not make sense absent some paradigm-shifting theoretical breakthrough to explain such findings.

And yet, such studies are frequently done for one item: low-skilled labor.  While many studies continue to find – and confirm – what is predicted by foundational economics (namely, that - ceteris paribus – the higher the minimum wage, the fewer are the quantity of hours of low-skilled labor demanded by employers), many other studies find the opposite.

The reason for the flood of the ‘pro-minimum-wage’ studies is not, I firmly believe, that anyone intentionally sets out to disprove the standard, ‘anti-minimum-wage’ theory and findings.  Rather, the reason is that there are people whose priors have them invested in believing that somehow, for some reason, low-skilled workers are the exception to the rule that is at the very core of economic science – that rule being the law of demand.  Believing, pre-scientifically, ‘in’ the goodness of the minimum wage, scholars are led to explore the data intrepidly until they find what they know must be true – namely, that at least small hikes in the minimum wage do not reduce the quantity demanded of low-skilled labor.

And the enormous complexity of economic reality renders finding such confirmation in appropriately parsed, diced, and sliced data easy enough to do.  Again, one produces such findings for the effect of alcohol taxes on alcohol consumption (or the effect of carbon taxes on carbon emissions, or the effect of penalties for rape and the frequency of rape or …. you name it).  That no one does these other studies is explained by the fact that too few people are invested in believing that the law of demand does not apply in these other situations.  There are, however, many people – including (I believe sadly) some economists – who are invested in believing that minimum-wage legislation simply must help the people it is ostensibly meant to help.  And, again, finding empirical confirmation for such a prior in this incredibly complex and ever-changing world of ours is really rather simple.

No comments: