What is (not) behavioural in labour economics?

Abstract

Thomas Dohmen's survey on the impact of behavioural economicson labour economics is an impressive manifesto about the pervasive-ness of behavioural assumptions in current labour economics research.Behavioural economics is typically defined via the inverse of the tradi-tional economic agent, who is a self-interested, rational, and forward-lookinghomoeconomicus,withperfectforesight,whoseutilityfunctionis stable and well-behaved. In some surveys of behavioural economics(Berg, 2006, Kaufman, 1999), the traditional agent is a representativeagent who has noimportant personal characteristics as there is no het-erogeneity, who lives in a world where only money matters, wheresortingplays norole, and where only cognitive skills matter. Moreover,this agent has constant preferences over a long, often infinite, horizon.Looking at this description, one can without doubt paraphrase MiltonFriedman in saying “we are all behavioural economists now”.According to such a definition of an economic agent, all currentempiricallaboureconomistsqualifyasbehaviouraleconomists,becauseno one believes in homogeneous effects or preferences which are fixedacrossdifferentpolicy regimes or overlongtime periods.Gary Becker–asexemplarilyshowninhisNobellecture–stressestheexplicitmotiva-tion to go beyond the traditional economic man: “… the economicapproach I refer to does not assume that individuals are motivatedsolely by selfishness or material gain. It is a method of analysis, not anassumption about particular motivations. Along with others, I havetried to pry economists away from narrow assumptions about self-interest. Behaviour is driven by a much richer set of values and prefer-ences” (Becker, 1993, 385). Still, few would claim Becker as a founderof behavioural economics or proclaim James Heckman as a frontrunnerbecausehepioneered heterogeneoustreatmenteffectsandstressestheimportance of cognitive skills.Behavioural labour economists have made important contributionsin the past decades: It is widely recognised – and modelled – that deci-sionmakingisoftensituationalandthatreactionstopoliciesareheavilyinfluenced by circumstances. Current research recognises that inter-temporal optimization suffers from self-control problems and that it isdifficult to optimise in complex situations, e.g., discounting, formingexpectations or Bayesian learning. Behavioural aspects have enteredpolicy design and feature on policy recommendations as it is possible,perhaps only in selected cases, to nudge individuals into different –better? – decisions. Modern labour economics does not shy away fromnon-selfishpreferences;forexample,fairnesspreferenceshavebecomeastandardintheexperimentallabs.Onamoremacro-orientedway,la-boureconomistsusetheinsightsfrombehaviouralresearchtostudythereasonsforwage stickinessorrelated problemsfor wagesetting,whichmight have consequences for a micro-foundation of business cycles ormonetary theory. In short, one might say: Behavioural economicshelped us to get rid of notions where homo economicus was taken tooliterally.The integration of behavioural economics into labour issues is cer-tainlyasuccessstory.Laboureconomicsisprobablyafieldwheretheas-sumptions and methods of behavioural economists were mostwelcome. I can only speculate on reasons for that: I suppose the highlyempiricalcontentoflabour,thefactthatlaboureconomistsweremostlyconcerned with the behaviour of individuals and their readiness to usemicro data to analyse how persons (or firms) react to incentives in dif-ferentformsmighthavemadetheadaptationtoexplicitbehaviouralas-sumptions much easier.Thomas was perhaps a bit too optimistic when he counted the suc-cesses of behavioural economics and I want to mention some caveatswhich give rise to a more modest assessment of behavioural laboureconomics. As any other economic tribe, the behaviouralists use someconvenient workhorses. Workhorses are useful shortcuts for modelsand empirical setups; still it should not be forgotten that not everyhorse is appropriate for every competition.