Giving in dictator games: Experimenter demand effect or preference over the rules of the game?

Which preference underlies giving in dictator games? To date, the experimental evidence has either been interpreted as a preference over the distribution of pay-offs, or as an experimenter demand effect. We show that under strict dictator-dictator as well as strict dictator-recipient anonymity, giving in dictator games springs from a preference over the distribution of decision rights. In contrast, concerns which trigger experimenter demand (Andreoni and Bernheim 2009) are negatively correlated with dictator game giving. Our experiments cover a series of dictator game variants which have sparked the experimenter demand debate. In addition, we identify the sets of ethical ideals that dictators em-ploy to derive the 'right' course of action in a formal moral judgement test and model dictator transfers econometrically by means of dictators' actual ethical ideals. Our results explain the lion’s share of results from the literature: lower transfers when dictators earn the pie (Cherry et al. 2002); lower transfers when 'take' options are available (List 2007; Bardsley 2008); lower transfers when anonymity is lifted (Hoffmann et al. 1994); generous dictators consistently pre-ferring to avoid the game altogether, if given the option (Lazear et al. 2012), and findings that social norms and beliefs cause dictator transfers (Krupka and Weber 2013; Di Tella et al. 2015; Kimbrough and Vostroknutov 2015).