Graham Loomes from Warwick Business School presented "Diagnosing the Causes of Intertemporal Preference Reversals" at the FUR conference in York (25 June to 28 June 2018)

Summary from the FUR programme:

Intertemporal preference reversals arise when individuals choose a ‘smaller sooner’ (SS) option (e.g. receiving £1,600 in 18 months' time) over a ‘larger later’ (LL) option (£2,500 in 5 years' time) in a straight choice between the two, but place a higher value on the LL option when asked to state a direct present value equivalent for each option separately. Tversky et al. (1990) diagnosed these reversals as being primarily due to overvaluing the LL option in the direct valuation task, as well as undervaluing the SS option some degree relative to the values implied by their choices. They detected some degree of intransitivity but considered it to be a relatively minor contributor. We revisited this phenomenon, changing the experimental design in various ways. First, instead of relying on a single preselected present value amount to compare with each future option, we examined choices between each option and a range of present amounts so as to more accurately estimate each individual’s present value for SS and for LL when each was inferred from binary choices rather than elicited by a direct valuation question. Second, we used choice lists as well as binary choices and direct valuations to provide another means of elicitation. Third, we repeated the various binary choices, the choice lists and the direct valuation questions several times in order to get some measure of the degree of variability and imprecision in people’s responses to each type of question and in order to be able to base our diagnosis on measures of central tendency rather than on single observations which may be more volatile. Even after allowing for the considerable variability and imprecision in many people’s stated preferences, we found high levels of intertemporal preference reversals. However, we disagree with Tversky et al. about the causes of the majority of these reversals, which we ascribe to substantial levels of intransitivity in respondents’ binary choices as well as differential overpricing of both options relative to the values inferred from their choices.