|
on Discrete Choice Models |
| By: | Robin Cubitt (School of Economics, University of Nottingham); Maria Ruiz-Martos (Economics Department, University of Warwick); Chris Starmer (School of Economics, University of Nottingham) |
| Abstract: | The paper reports an experiment which tests the principle of separability, i.e. that behaviour in a dynamic choice problem is independent of history and of unreachable eventualities. Although this is a well-known principle of orthodox decision theory and central to conventional economic modelling, it has been questioned on grounds suggested by non-expected utility models of choice under risk and by the psychology of affective influences on risk-taking. Our experimental design, which provides between-subjects tests of separability using three treatments in which the history preceding a decision is manipulated, is inspired by these concerns. We expose separability to a clean and harsh test, but find no evidence that it is violated. |
| Keywords: | Separability; history-independence; non-expected utility; risk and affect |
| Date: | 2010–01 |
| URL: | http://d.repec.org/n?u=RePEc:cdx:dpaper:2010-01&r=dcm |
| By: | John M. Rose (Institute of Transport and Logistics Studies (ITLS), The University of Sydney, Australia); Lorenzo Masiero (Istituto Ricerche Economiche (IRE), Università della Svizzera Italiana, Svizzera) |
| Abstract: | The importance of willingness to pay (WTP) and willingness to accept (WTA) measures in the evaluation of policy measures has led to a constant stream of research examining survey methods and model specifications seeking to capture and explain the concept of marginal rates of substitution as much as possible. Stated choice experiments pivoted around a reference alternative allow the specification of discrete choice models to accommodate the prospect theory reference dependence assumption. This permits an investigation of theories related to loss aversion and diminishing sensitivity, and to test the discrepancy between WTP and WTA, widely documented within the literature. With more advanced classes of discrete choice models at our disposal, it is now possible to test different preference specifications that are better able to measure WTP and WTA values. One such model allowing for utility to be directly specified in WTP space has recently shown interesting qualities. This paper compares and contrasts models estimated in preference space to those estimated in WTP space allowing for asymmetry in the marginal utilities by estimating different parameters according to reference, gain and loss values. The results suggest a better model fit for the data estimated in WTP space, contradicting the findings of previous researches. The parameter estimates report significant evidence of loss aversion and diminishing sensitivities even though the symmetric specification outperforms the asymmetric ones. Finally, the analysis of the WTP and WTA measures confirms the higher degree of WTA compared to WTP, and highlights the appeal of the WTP space specification in terms of plausibility of the estimated measures. |
| Keywords: | choice experiments, willingness to pay space, preference asymmetry |
| JEL: | C25 L91 |
| Date: | 2010–03 |
| URL: | http://d.repec.org/n?u=RePEc:lug:wpaper:1006&r=dcm |
| By: | Lorenzo Masiero (Istituto Ricerche Economiche (IRE), Università della Svizzera Italiana, Svizzera); David A. Hensher (Institute of Transport and Logistics Studies (ITLS), The University of Sydney, Australia) |
| Abstract: | Choice behaviour might be determined by asymmetric preferences whether the consumers are faced with gains or losses. This paper investigates loss aversion and diminishing sensitivity, and analyzes their implications on willingness to pay and willingness to accept measures in a reference pivoted choice experiment in a freight transport framework. The results suggest a significant model fit improvement when preferences are treated as asymmetric, proving both loss aversion and diminishing sensitivity. The implications on willingness to pay and willingness to accept indicators are particular relevant showing a remarkable difference between symmetric and asymmetric model specifications. Not accounting for loss aversion and diminishing sensitivity, when present, produces misleading results and might affect significantly the policy decisions. |
| Keywords: | freight transport, choice experiments, willingness to pay, preference asymmetry |
| JEL: | C25 L91 |
| Date: | 2010–03 |
| URL: | http://d.repec.org/n?u=RePEc:lug:wpaper:1004&r=dcm |
| By: | Lorenzo Masiero (Istituto Ricerche Economiche (IRE), Università della Svizzera Italiana, Svizzera); David A. Hensher (Institute of Transport and Logistics Studies (ITLS), The University of Sydney, Australia) |
| Abstract: | It is widely recognized that individual decision making is subject to the evaluation of gains and losses around a reference point. The estimation of discrete choice models increasingly use data from stated choice experiments which are pivoted around a reference alternative. However, to date, the specification of a reference alternative in transport studies is fixed, whereas it is common to observe individuals adjusting their preferences according to a change in their reference point. This paper focuses on individual reactions, in a freight choice context, to a negative change in the reference alternative values, identifying the behavioural implications in terms of loss aversion and diminishing sensitivity. The results show a significant adjustment in the valuation of gains and losses around a shifted reference alternative. In particular, we find an average increase in loss aversion for cost and time attributes, and a substantial decrease for punctuality. These findings are translated to significant differences in the willingness to pay and willingness to accept measures, providing supporting evidence of respondents’ behavioural reaction. |
| Keywords: | Willingness to pay, gains and losses, freight choice, reference alternative |
| JEL: | C25 L91 |
| Date: | 2010–03 |
| URL: | http://d.repec.org/n?u=RePEc:lug:wpaper:1005&r=dcm |
| By: | Mimako Kobayashi (Department of Resource Economics, University of Nevada, Reno); Klaus Moeltner (Department of Resource Economics, University of Nevada, Reno); Kimberly Rollins (Department of Resource Economics, University of Nevada, Reno) |
| Abstract: | In many stated preference settings stakeholders will be uncertain as to their exact willingness-to-pay for a proposed environmental amenity. To accommodate this possibility analysts have designed elicitation formats with multiple bids and response options that allow for the expression of uncertainty. We argue that the information content flowing from such elicitation has not yet been fully and efficiently exploited in existing contributions. We introduce a Latent Thresholds Estimator that focuses on the simultaneous identification of the full set of thresholds that delineate an individual's value space in accordance with observed response categories. Our framework provides a more complete picture of the underlying value distribution, the marginal effects of regressors, and the impact of bid designs on estimation efficiency. We show that the common practice of re-coding responses to derive point estimate of willingness-to-pay leaves useful information untapped and can produce misleading results if thresholds are highly correlated. |
| Keywords: | Stated Preference; Multiple Bounded Elicitation; Polychotomous Choice; Bayesian Estimation; Value Uncertainty |
| JEL: | C11 C15 C35 C52 Q51 |
| Date: | 2010–01 |
| URL: | http://d.repec.org/n?u=RePEc:unr:wpaper:10-001&r=dcm |
| By: | Dmitriev, Mikhail |
| Abstract: | This is one of the first effort to make structural estimation of fertility in Russia.Normal distribution of random individual effect and extreme value distribution of the error are assumed. |
| Keywords: | Dynamic Discrete Choice Estimation;Fertility; Demography; Russia |
| JEL: | J13 |
| Date: | 2009–11 |
| URL: | http://d.repec.org/n?u=RePEc:pra:mprapa:21150&r=dcm |