Via Cos­mic Vari­ance, we learn of the com­plex sta­tis­ti­cal analy­sis pre­sent­ed here, orig­i­nal­ly from the Wall Street Jour­nal and based on work from the Amer­i­can Enter­prise Institute.

The curve labeled the “Laf­fer curve” is not meant to make you “laff,” it is actu­al­ly an exam­ple of the psy­cho­log­i­cal phe­nom­e­non known as parei­do­lia. More famous exam­ples include the abil­i­ty to see Jesus in burnt tor­tillas, or the name of Allah in the swirls of an ice cream cone. In this case, the AEI has dis­cov­ered their own myth­i­cal ide­al in the data on tax rev­enue and tax rates. 

The Laf­fer curve is the the­o­ret­i­cal under­pin­ning of sup­ply-side eco­nom­ics. It shows tax rev­enue ris­ing with tax rates until some point at which tax­es become so oner­ous that peo­ple pre­fer not to make mon­ey rather than pay more in taxes. 

Econ­o­mists agree that some such rela­tion­ship must exist – it would be point­less to do any work at all if you were taxed at 100%, after all. The premise of sup­ply siders is that the rate at which tax rev­enue reach­es its max­i­mum is some­where pret­ty close to cur­rent tax rates, while real econ­o­mists tend to think the max­i­mum would occur some­where far beyond cur­rent tax rates, and that it is sil­ly to think (as the Bush admin­is­tra­tion has claimed many times) that cut­ting tax rates would raise tax revenue. 

As Sean points out at Cos­mic Vari­ance, the Laf­fer curve here appears to have been drawn using the sophis­ti­cat­ed sta­tis­ti­cal tools pro­vid­ed by M$ Paint. A sim­ple lin­ear regres­sion through the data shows an upward trend, with no basis for fit­ting qua­drat­ic mod­els. Appar­ent­ly Nor­way is also in the wrong place, and in the right place, would sim­ply make the upward trend stronger (Nor­way has high tax­es on oil and gas extrac­tion, a fact includ­ed in the tax rev­enue axis but not in the tax rates axis. Real­ly.).

Humans have a remark­able abil­i­ty to find pat­terns where they don’t exist, and indeed to force pat­terns that we see to meet those expec­ta­tions. The tools of sta­tis­tics and the sci­en­tif­ic process are pow­er­ful checks on the excess­es of “com­mon sense,” and the Wall Street Jour­nal edi­to­r­i­al board would do well to rely more on the lat­ter and less on the former.