|
on Discrete Choice Models |
| By: | André De Palma (ENS Cachan - École normale supérieure - Cachan, Université Paris-Saclay); Julien Monardo (ENS Cachan - École normale supérieure - Cachan, Université Paris-Saclay) |
| Abstract: | We study the general nesting logit (GNL) model for differentiated products proposed by Fosgerau and de Palma (2016) as a member of the family of generalized entropies built by Fosgerau, Melo, de Palma and Shum (2017), to estimate demand when using aggregate data. We show that the GNL model allows products to be independent, substitutable, or complementary. While Fosgerau and de Palma (2016) show that it can be transformed into a linear regression, we show that this linear regression is very similar to that of Berry (1994) for the nested logit in that it is just a regression of market shares on product characteristics and terms related to its nesting structure. We then use the Dominick's database for estimating the demand for cereals in Chicago in 1991-1992. |
| Keywords: | Demand estimation, Differentiated products, Discrete choice,Generalized entropy, Representative consumer,C26,D11,D12,L |
| Date: | 2017–07–27 |
| URL: | http://d.repec.org/n?u=RePEc:hal:wpaper:hal-01552455&r=dcm |
| By: | Marcelo Lima |
| Abstract: | This paper considers whether the answers to stated preference surveys (of the type used to monitise non-market goods) are affected by the survey's sponsoring institution. The sponsor is indicated to respondents by the logo used in the survey instrument, an online questionnaire. Survey repondents are randomly assigned to one of eight types of sponsor and whether stated willingness-to-pay for a product that reduces mortality risk is affected by the sponsor is observed. It is also considered whether sponsorship has an effect on measures of respondent engagement with the survey (survey completion rates, item response rates, time spend on the willingness to pay question and on the survey as a whole). The analysis finds that respondents that believe the survey to be sponsored by an environmental ministry or a health ministry are willing to pay significantly less for the product than those that believe that the survey is sponsored by other types of institution. There are also apparent trade-offs between the different repondent engagement measures considered. |
| Date: | 2017–08 |
| URL: | http://d.repec.org/n?u=RePEc:lsg:lsgwps:wp272&r=dcm |