nep-dcm New Economics Papers
on Discrete Choice Models
Issue of 2022‒09‒19
ten papers chosen by
Edoardo Marcucci
Università degli studi Roma Tre

  1. Deep Learning for Choice Modeling By Zhongze Cai; Hanzhao Wang; Kalyan Talluri; Xiaocheng Li
  2. Marginal stochastic choice By Yaron Azrieli; John Rehbeck
  3. Ask a local: Improving the public pricing of land titles in urban Tanzania By Tanner Regan; Martina Manara
  5. Gender preference at birth: A new measure for son preference based on stated preferences and observed measures of parents' fertility decisions By Mehwish Ghulam Ali; Ashton De Silva; Sarah Sinclair; Ankita Mishra
  6. Endogeneity in Weakly Separable Models without Monotonicity By Songnian Chen; Shakeeb Khan; Xun Tang
  7. On mechanism design with expressive preferences: an aspect of the social choice of Brexit By Anindya Bhattacharya; Debapriya Sen
  8. Consumer Demand with Social Influences: Evidence from an E-Commerce Platform By El Hadi Caoui; Chiara Farronato; John J. Horton; Robert Schultz
  9. Bayesian Estimation of Large-Scale Simulation Models with Gaussian Process Regression Surrogates By Sylvain Barde
  10. An approach to generalizing some impossibility theorems in social choice By Wesley H. Holliday; Eric Pacuit; Saam Zahedian

  1. By: Zhongze Cai; Hanzhao Wang; Kalyan Talluri; Xiaocheng Li
    Abstract: Choice modeling has been a central topic in the study of individual preference or utility across many fields including economics, marketing, operations research, and psychology. While the vast majority of the literature on choice models has been devoted to the analytical properties that lead to managerial and policy-making insights, the existing methods to learn a choice model from empirical data are often either computationally intractable or sample inefficient. In this paper, we develop deep learning-based choice models under two settings of choice modeling: (i) feature-free and (ii) feature-based. Our model captures both the intrinsic utility for each candidate choice and the effect that the assortment has on the choice probability. Synthetic and real data experiments demonstrate the performances of proposed models in terms of the recovery of the existing choice models, sample complexity, assortment effect, architecture design, and model interpretation.
    Date: 2022–08
  2. By: Yaron Azrieli; John Rehbeck
    Abstract: Models of stochastic choice typically use conditional choice probabilities given menus as the primitive for analysis, but in the field these are often hard to observe. Moreover, studying preferences over menus is not possible with this data. We assume that an analyst can observe marginal frequencies of choice and availability, but not conditional choice frequencies, and study the testable implications of some prominent models of stochastic choice for this dataset. We also analyze whether parameters of these models can be identified. Finally, we characterize the marginal distributions that can arise under two-stage models in the spirit of Gul and Pesendorfer [2001] and of kreps [1979] where agents select the menu before choosing an alternative.
    Date: 2022–08
  3. By: Tanner Regan (George Washington University); Martina Manara (London School of Economics)
    Abstract: Information on willingness-to-pay is key for public pricing and allocation of services but not easily collected. Studying land titles in Dar-es-Salaam, we ask whether local leaders know and will reveal plot owners' willingness-to-pay. We randomly assign leaders to predict under different settings then elicit owners' actual willingness-to-pay. Demand is substantial, but below exorbitant fees. Leaders can predict the aggregate demand curve and distinguish variation across owners. Predictions worsen when used to target subsidies, but adding cash incentives mitigates this. We demonstrate that leader-elicited information can improve the public pricing of title deeds, raising uptake while maintaining public funds.
    Keywords: property rights; willingness-to-pay; public pricing; local publicly provided goods
    JEL: O17 H40 R21 D80
    Date: 2022–07
  4. By: Vadim A. Petrovsky (National Research University Higher School of Economics)
    Abstract: This paper interprets the hard-to-explain discrepancies between the results of empirical research into achievement motivation (with free choice of the levels of task challenge, as in Hoppe’s experiments) and the predictions of the risk-taking model by Atkinson that is based on a combination of three variables: motives for achieving success, avoiding failure, and the probability of success. This model predicts that when the motive of success dominates the motive of failure, subjects choose tasks of an average level of difficulty, but this (and some other consequences of the model) is not confirmed by empirical data. Unlike most works that introducing additional variables into the classical Atkinson model to account for these, this paper proposes a different solution, based on the development of a model of readiness for a bipolar choice by Lefebvre. Lefebvre’s original model contains the perceived “pressure of the environment” (a1) which impels the choice of a positive pole; the image of the pressure of the environment (a2); subjective intentions (a3); and the objective readiness to make a choice (A). The variables are interconnected by an operator of material implication, A = ((a3?a2)?a1 and its continuous counterparts. Just like the classical risk-taking model by Atkinson, the model of readiness for a bipolar choice (defined here as the “reflexive model of risk-taking”) includes the three variables of the subjective probability of success, the motive of success, and the motive of avoiding failure. As a combination they are considered as a predictor of preferences for tasks of various difficulty levels. It is possible to adjust Atkinson’s model to account for experimental data without increasing the number of variables
    Keywords: achievement motivation, model of risk-taking, model of reflexive choice, intentional choice, congruence, choice strategies
    JEL: Z
    Date: 2022
  5. By: Mehwish Ghulam Ali; Ashton De Silva; Sarah Sinclair; Ankita Mishra
    Abstract: Investigating preference for sons is a continuing focal area of development economics and demographic research. Son preference presents a challenge in achieving the United Nations Sustainable Development Goals of 'no poverty', 'good health and wellbeing', and 'gender equality' by 2030. It is thus important to investigate son preference to inform policy-makers of the potential challenges in achieving these goals. Inaccurate interpretation of the mechanisms of son preference could misinform policy analysis and result in unintended consequences.
    Keywords: Human fertility, Family planning, Welfare, Wellbeing
    Date: 2022
  6. By: Songnian Chen; Shakeeb Khan; Xun Tang
    Abstract: We identify and estimate treatment effects when potential outcomes are weakly separable with a binary endogenous treatment. Vytlacil and Yildiz (2007) proposed an identification strategy that exploits the mean of observed outcomes, but their approach requires a monotonicity condition. In comparison, we exploit full information in the entire outcome distribution, instead of just its mean. As a result, our method does not require monotonicity and is also applicable to general settings with multiple indices. We provide examples where our approach can identify treatment effect parameters of interest whereas existing methods would fail. These include models where potential outcomes depend on multiple unobserved disturbance terms, such as a Roy model, a multinomial choice model, as well as a model with endogenous random coefficients. We establish consistency and asymptotic normality of our estimators.
    Date: 2022–08
  7. By: Anindya Bhattacharya; Debapriya Sen
    Abstract: We study some problems of collective choice when individuals can have expressive preferences, that is, where a decision-maker may care not only about the material benefit from choosing an action but also about some intrinsic morality of the action or whether the action conforms to some identity-marker of the decision-maker. We construct a simple framework for analyzing mechanism design problems with such preferences and present some results focussing on the phenomenon we call "Brexit anomaly". The main findings are that while deterministic mechanisms are quite susceptible to Brexit anomaly, even with stringent domain restriction, random mechanisms assure more positive results.
    Date: 2022–08
  8. By: El Hadi Caoui; Chiara Farronato; John J. Horton; Robert Schultz
    Abstract: For some kinds of goods, rarity itself is valued. "Fashionable'" goods are demanded in part because they are unique. In this paper, we explore the economics of rare goods using auctions of limited-edition shoes held by an e-commerce platform. We model endogenous entry and bidding in multi-unit auctions and construct demand curves from realized bids. We find that doubling inventory reduces willingness to pay by 7-15%. From the producer perspective, ignoring the value of rarity leads to substantial overproduction: auctioned quantities are 82% above the profit-maximizing level. From the consumer perspective however, the negative spillovers of restricting quantity more than offset the benefits of rarer goods.
    JEL: D12 D44 L81
    Date: 2022–08
  9. By: Sylvain Barde
    Abstract: Large scale, computationally expensive simulation models pose a particular challenge when it comes to estimating their parameters from empirical data. Most simulation models do not possess closed form expressions for their likelihood function, requiring the use of simulation-based inference, such as simulated method of moments, indirect inference or approximate Bayesian computation. However, given the high computational requirements of large-scale models, it is often difficult to run these estimation methods, as they require more simulated runs that can feasibly be carried out. This paper aims to address the problem by providing a full Bayesian estimation framework where the true but intractable likelihood function of the simulation model is replaced by one generated by a surrogate model. This is provided by a sparse variational Gaussian process, chosen for its desirable convergence and consistency properties. The effectiveness of the approach is tested using both a Monte Carlo analysis on a known data generating process, and an empirical application in which the free parameters of a computationally demanding agent-based model are estimated on US macroeconomic data.
    Keywords: Bayesian estimation; surrogate methods; Gaussian process; simulation models
    JEL: C14 C15 C52 C63
    Date: 2022–08
  10. By: Wesley H. Holliday; Eric Pacuit; Saam Zahedian
    Abstract: In social choice theory, voting methods can be classified by invariance properties: a voting method is said to be C1 if it selects the same winners for any two profiles of voter preferences that produce the same majority graph on the set of candidates; a voting method is said to be pairwise if it selects the same winners for any two preference profiles that produce the same weighted majority graph on the set of candidates; and other intermediate classifications are possible. As there are far fewer majority graphs or weighted majority graphs than there are preference profiles (for a bounded number of candidates and voters), computer-aided techniques such as satisfiability solving become practical for proving results about C1 and pairwise methods. In this paper, we develop an approach to generalizing impossibility theorems proved for C1 or pairwise voting methods to impossibility theorems covering all voting methods. We apply this approach to impossibility theorems involving "variable candidate" axioms--in particular, social choice versions of Sen's well-known $\gamma$ and $\alpha$ axioms for individual choice--which concern what happens when a candidate is added or removed from an election. A key tool is a construction of preference profiles from majority graphs and weighted majority graphs that differs from the classic constructions of McGarvey and Debord, especially in better commutative behavior with respect to other operations on profiles.
    Date: 2022–08

This nep-dcm issue is ©2022 by Edoardo Marcucci. It is provided as is without any express or implied warranty. It may be freely redistributed in whole or in part for any purpose. If distributed in part, please include this notice.
General information on the NEP project can be found at For comments please write to the director of NEP, Marco Novarese at <>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.