Results 1 
6 of
6
Heterogeneity and the nonparametric analysis of consumer choice: conditions for invertibility”, cemmap Working Papers
, 2005
"... This paper considers structural nonparametric random utility models for continuous choice variables with unobserved heterogeneity. We provide sufficient conditions on random preferences to yield reducedform systems of nonparametric stochastic demand functions that allow global invertibility betwe ..."
Abstract

Cited by 12 (3 self)
 Add to MetaCart
This paper considers structural nonparametric random utility models for continuous choice variables with unobserved heterogeneity. We provide sufficient conditions on random preferences to yield reducedform systems of nonparametric stochastic demand functions that allow global invertibility between demands and nonseparable unobserved heterogeneity. We distinguish between new classes of models in which heterogeneity is separable and nonseparable in the marginal rates of substitution, respectively. Invertibility is essential for global identification of structural consumer demand models, for the existence of wellspecified probability models of choice and for the nonparametric analysis of revealed stochastic preference.
Estimation of Stochastic Preferences: An Empirical Analysis of Demand for Internet Services
, 2000
"... The rapid inc rease in demand for Internet servic es and new, bandwidth and timeintensive applic ations require high quality ac c to the Internet. Servic quality may be assured through e#c ient allo coj on of Internetac ss cj ac y. E#c ntc apac ity allo cj tionc an be ac hieved through nonlinear p ..."
Abstract

Cited by 7 (4 self)
 Add to MetaCart
The rapid inc rease in demand for Internet servic es and new, bandwidth and timeintensive applic ations require high quality ac c to the Internet. Servic quality may be assured through e#c ient allo coj on of Internetac ss cj ac y. E#c ntc apac ity allo cj tionc an be ac hieved through nonlinear pric es. Nonlinear pric ing is motivated through preferenc heterogeneity. The objec tive of this paper is to develop anec onometric model of Internet users' preferenc s over servic e attributes thatcj tes unobserved preferenc e heterogeneity. To this end, a stoc hastic p referenc model is proposed and estimated on data from the U.C. Berkeley Internet Demand Experiment (INDEX). A struc tural ec onometric framework is developed within whic h ac onsumer learns in the proc ess ofservic ec onsumption and randomness inc hoic es,c onditional on pric s and expenditure, arises from unobserved heterogeneity in thec onsumer's preferenc es. Heterogeneity in preferenc es is estimated using a simulationassisted estimation methodology. The empiric al analysis of Internet user data shows thatc onsiderable heterogeneity in preferenc es exists, among di#erent users and foreac h user over the observation horizon. Moreover, users appear to learn in the c nsumption pro c ss, deviating in their online valuations from their ex antec onsumption plans. A user's variation in online valuations also typic ally exc eds the variation in ex ante valuations. For the purpose of demand management, the estimated model appears to quite ac c urately predic t the distribution of thec ontinuousc hoic e variables,c onditional on disc rete servic ec hoic e, andc an be used to explore di#erent pric ng sc enarios. KEYWORDS: Internet Demand , Preference Heterogeneity, Bayesian Learning, DiscreteContinuous Choices, M...
Unobserved Preference Heterogeneity in Demand Using Generalized Random Coefficients
, 2011
"... We model unobserved preference heterogeneity in demand systems via random Barten scales in utility functions. These Barten scales appear as random coefficients multiplying prices in demand functions. Consumer demands are nonlinear in prices and may have unknown functional structure. We therefore pro ..."
Abstract

Cited by 4 (0 self)
 Add to MetaCart
We model unobserved preference heterogeneity in demand systems via random Barten scales in utility functions. These Barten scales appear as random coefficients multiplying prices in demand functions. Consumer demands are nonlinear in prices and may have unknown functional structure. We therefore prove identification of Generalized Random Coefficients models, defined as nonparametric regressions where each regressor is multiplied by an unobserved random coefficient having an unknown distribution. Using Canadian data, we estimate energy demand functions with and without random coefficient Barten scales. We find that not accounting for this unobserved preference heterogeneity substantially biases estimated consumersurplus costs of an energy tax.
Strategic choice during economic crisis: Domestic market position, . . .
 JOURNAL OF WORLD BUSINESS
, 2009
"... ..."
c ○ 2008 The Review of Economic Studies Limited Heterogeneity and the NonParametric Analysis of Consumer Choice: Conditions for
, 2005
"... This paper considers structural nonparametric random utility models for continuous choice variables with unobserved heterogeneity. We provide sufficient conditions on random preferences to yield reducedform systems of nonparametric stochastic demand functions that allow global invertibility betwe ..."
Abstract
 Add to MetaCart
This paper considers structural nonparametric random utility models for continuous choice variables with unobserved heterogeneity. We provide sufficient conditions on random preferences to yield reducedform systems of nonparametric stochastic demand functions that allow global invertibility between demands and nonseparable unobserved heterogeneity. Invertibility is essential for global identification of structural consumer demand models, for the existence of wellspecified probability models of choice and for the nonparametric analysis of revealed stochastic preference. We distinguish between new classes of models in which heterogeneity is separable and nonseparable in the marginal rates of substitution, respectively. 1.
Generalized Random Coefficients With Equivalence Scale Applications
, 2011
"... We propose a generalization of random coefficients models, in which the regression model is an unknown function of a vector of regressors, each of which is multiplied by an unobserved error. We also investigate a more restrictive model which is additive (or additive with interactions) in unknown fun ..."
Abstract
 Add to MetaCart
We propose a generalization of random coefficients models, in which the regression model is an unknown function of a vector of regressors, each of which is multiplied by an unobserved error. We also investigate a more restrictive model which is additive (or additive with interactions) in unknown functions of each regressor multiplied by its error. We show nonparametric identification of these models. In addition to providing a natural generalization of random coefficients, we provide economic motivations for the model based on demand system estimation. In these applications, the random coefficients can be interpreted as random utility parameters that take the form of Engel scales or Barten scales, which in the past were estimated as deterministic preference heterogeneity or household technology parameters. We apply these results to consumer surplus and related welfare calculations.