Rationalising Choice with Multi-self Models

Abstract

This article shows that when multiple selves' preferences are aggregated into a decision, even if the researcher has a fully specified theory of how preferences get aggregated, there are typically no testable implications without restricting the number of selves. This points to the importance of collecting reliable information on the number of selves in interpersonal and intrapersonal decision -making contexts. We establish the result through a linear relationship between the number of selves and the set of choice functions an aggregator is guaranteed to rationalise. The latter relates to a choice function's number of independence of irrelevant alternatives violations, a new measure of irrationality.Since the seminal work of May (1954), a growing number of papers have proposed models of multi-self decision making, with the primary motivation of accommodating context-dependent behaviour (thereby relaxing the axiom of independence of irrelevant alternatives (IIA), which requires that if an alternative is chosen from a set, it is also chosen from any subset in which it is contained). 1,2 Formal models of multi-self decision making include among others -

State dependent choice

“…Similar observations hold for the recent works by Green and Hojman [24] and Ambrus and Rozen [3]. Green and Hojman describe an agent as a probability distribution over all possible preference orderings.…”

“…We look at a long-standing debate on labour supply responses by workers who (like taxi-drivers) can choose their supply in a non-lumpy way. 3 Even a very stripped down 3 Some of the relevant literature is cited in section 5. version of the model (that postulates 'target-driven' behaviour as suggested in the empirical literature) can accommodate the 'anomaly' of negative wage elasticities. The new key insight we provide lies in the asymmetry between income targets, which can be 'unrealistic', and leisure targets, which have a physical bound (there are so many hours in a day).…”

Collective Dynamic Choice: The Necessity of Time Inconsistency

“…10 Some notable models with multiple personalities, preferences, or motives of agents include, among others: Thaler and Shefrin (1981) Ambrus and Rozen (2009);and Cherepanov, Feddersen, and Sandroni (2009). For a simple non-unitary discount model in a two-motive, CRRA utility setting, see Hori and Futagami (2012).…”

Partial knowledge restrictions on the two-stage threshold model of choice

“…One approach has been to weaken the consistency properties that preferences are ordinarily assumed to possess. 1 Another has been to study relationships between preference and choice other than straightforward maximization. 2 And a third has been to permit additional, non-preference-related factors-as well as multiple preferences-to influence decision making in some way.…”