|Institution:||University of New South Wales|
|Keywords:||Commitment decisions; Elicitation methods; Cognitive ability|
|Full text PDF:||http://handle.unsw.edu.au/1959.4/54358|
This thesis reports on an experimental study of time preferences and commitment decisions. Time preferences and associated discounting of future utility payoffs play a fundamental role in intertemporal decisions, and they have been the centre of much attention and controversy in the recent experimental literature. Since Strotz (1955), commitment has been recognised as a method of improving decision-making in the context of inconsistent decisions over time. Such inconsistencies can occur due to changes in the time discount over different periods in the future. The study compares three different elicitation methods used in time preference research: the Time Trade-off (TTO) method (Attema et al. 2010), the Equivalent Delay (ED) method (Takeuchi, 2010) and the Intertemporal Choice Reversal (ICR) method. These methods classify people as time-consistent or -inconsistent, and further classify time-inconsistent people as present-biased or future-biased, free from the shape of the utility function. Our experimental results show that the TTO, ED, and ICR methods produce different classifications. Compared to the ICR method, the TTO method is more likely to have a present-biased classification; the ED method is more likely to have a future-biased classification, and less likely to have a time-consistent classification. Using the classification of people into these types, we study whether a person’s elicited time preferences and cognitive ability predict his commitment decisions. Since only sophisticated people who are aware of their time-inconsistency would commit, assuming the cognitive score of a person to be a proxy for his sophistication, we focus on the impact of the interaction of time-inconsistency and cognitive score on people’s decision to commit. Our experimental results show that there is no strong evidence that commitment decisions are consistent with elicited time preferences and cognitive ability. The reasons could be that subjects are uncertain about future income streams in their real life, the cognitive score is not a good proxy for sophistication, or sophistication does not play an important role in commitment decisions. Additionally, we find the design of commitments matters. Subjects are sensitive to the degree of binding power as well as cost/benefit of a commitment option.