QRE,NSNX and the Paradox of Voting

Abstract

Levine and Palfrey's QRE account of turnout in large elections raises the broader question of how much of a departure from standard rational choice theory is justified by the considerable repertoire of rational choice anomalies that has accumulated since Downs and Olson half a century ago. An alternative but more controversial unconventional view turns on what I call NSNX motivation to account for how agents seek a balance between self-interest and social motivation. NSNX agents have irreducibly dual utility functions. QRE agents have a standard utility function but they do not maximize it. I review the situation showing why in situations where NSNX effects could be expected, QRE might mirror those effects. I show how, by varying parameters of an experiment, I can cleanly distinguish between actual QRE effects, which should bring predictions closer to the data than Nash, and NSNX effects, which should do the same.