nep-dcm New Economics Papers
on Discrete Choice Models
Issue of 2023‒05‒08
twelve papers chosen by
Edoardo Marcucci
Università degli studi Roma Tre

  1. Mismatch in Preferences for Working from Home – Evidence from Discrete Choice Experiments with Workers and Employers By Lewandowski, Piotr; Lipowska, Katarzyna; Smoter, Mateusz
  2. Torch-Choice: A PyTorch Package for Large-Scale Choice Modelling with Python By Tianyu Du; Ayush Kanodia; Susan Athey
  3. Testing and Identifying Substitution and Complementarity Patterns By Rui Wang
  4. The Devil You Know: Rational Inattention to Discrete Choices when Prior Information Matters By Bruno Pellegrino
  5. Incorporating Search and Sales Information in Demand Estimation By Ali Hortacsu; Olivia R. Natan; Hayden Parsley; Timothy Schwieg; Kevin R. Williams
  6. Do Consumers Acquire Information Optimally? Experimental Evidence from Energy Efficiency By Andrea La Nauze; Erica Myers
  7. Which transport modes do people use for travelling to coworking spaces (CWSs)? By Stéphanie Souche-Le Corvec
  8. Complexity and Time By Benjamin Enke; Thomas Graeber; Ryan Oprea
  9. On Graphical Methods in Stochastic Choice By Christopher Turansick
  10. Wardrop Equilibrium Can Be Boundedly Rational: A New Behavioral Theory of Route Choice By Jiayang Li; Zhaoran Wang; Yu Marco Nie
  11. A Note on Quasi-Maximum-Likelihood Estimation in Hidden Markov Models with Covariate-Dependent Transition Probabilities By Demian Pouzo; Zacharias Psaradakis; Martin Sola
  12. Extremum Monte Carlo Filters: Real-Time Signal Extraction via Simulation and Regression By Francisco Blasques; Siem Jan Koopman; Karim Moussa

  1. By: Lewandowski, Piotr (Institute for Structural Research (IBS)); Lipowska, Katarzyna (Institute for Structural Research (IBS)); Smoter, Mateusz (Institute for Structural Research (IBS))
    Abstract: We study workers' and employers' preferences for remote work, estimating the willingness to pay for working from home (WFH) using discrete choice experiments with more than 10, 000 workers and more than 1, 500 employers in Poland. We selected occupations that can be done remotely and randomised wage differences between otherwise identical home- and office-based jobs, and between otherwise identical job candidates, respectively. We find that demand for remote work was substantially higher among workers than among employers. On average, workers would sacrifice 2.9% of their earnings for the option of remote work, especially hybrid WFH for 2-3 days a week (5.1%) rather than five days a week (0.6%). However, employers, on average, expect a wage cut of 21.0% from candidates who want to work remotely. This 18 pp gap in the valuations of WFH reflects employers' assessments of productivity loss associated with WFH (14 pp), and the additional effort required to manage remote workers (4 pp). Employers' and workers' valuations of WFH align only in 25-36% of firms with managers who think that WFH is as productive as on-site work.
    Keywords: working from home, remote work, discrete choice experiment, willingness to pay
    JEL: J21 J31 J81
    Date: 2023–03
  2. By: Tianyu Du; Ayush Kanodia; Susan Athey
    Abstract: The $\texttt{torch-choice}$ is an open-source library for flexible, fast choice modeling with Python and PyTorch. $\texttt{torch-choice}$ provides a $\texttt{ChoiceDataset}$ data structure to manage databases flexibly and memory-efficiently. The paper demonstrates constructing a $\texttt{ChoiceDataset}$ from databases of various formats and functionalities of $\texttt{ChoiceDataset}$. The package implements two widely used models, namely the multinomial logit and nested logit models, and supports regularization during model estimation. The package incorporates the option to take advantage of GPUs for estimation, allowing it to scale to massive datasets while being computationally efficient. Models can be initialized using either R-style formula strings or Python dictionaries. We conclude with a comparison of the computational efficiencies of $\texttt{torch-choice}$ and $\texttt{mlogit}$ in R as (1) the number of observations increases, (2) the number of covariates increases, and (3) the expansion of item sets. Finally, we demonstrate the scalability of $\texttt{torch-choice}$ on large-scale datasets.
    Date: 2023–04
  3. By: Rui Wang
    Abstract: This paper studies semiparametric identification of substitution and complementarity patterns between two goods using a panel multinomial choice model with bundles. The model allows the two goods to be either substitutes or complements and admits heterogeneous complementarity through observed characteristics. I first provide testable implications for the complementarity relationship between goods. I then characterize the sharp identified set for the model parameters and provide sufficient conditions for point identification. The identification analysis accommodates endogenous covariates through flexible dependence structures between observed characteristics and fixed effects while placing no distributional assumptions on unobserved preference shocks. My method is shown to perform more robustly than the parametric method through Monte Carlo simulations. As an extension, I allow for unobserved heterogeneity in the complementarity, investigate scenarios involving more than two goods, and study a class of nonseparable utility functions.
    Date: 2023–04
  4. By: Bruno Pellegrino
    Abstract: In the seminal rational inattention model of Matĕjka and McKay (2015), logit demand arises from the discrete choice of agents who are uncertain about choice payoffs and have access to a flexible, costly information acquisition technology (RI-logit). A notable limitation of this powerful framework is the lack of known general closed-form solutions that allow the decision maker’s prior information to be asymmetric across choices. In this paper, I solve the RI-logit model analytically for a large family of priors known as multivariate Tempered Stable (TS) distributions. In my analytical formulation, decision makers can be biased, display aversion to prior uncertainty, and thus tend to select choices that are familiar (i.e. for which they hold a less disperse prior). My result extends the applicability of the RI-logit model to a new range of settings where prior information matters. I provide one such application, by showing how it can be used to model the behavior of risk-averse investors who select risky projects in an environment characterized by epistemic uncertainty (risk-adjusted expected returns are unknown, but can be learnt at a cost).
    Keywords: rational inattention, discrete choice, uncertainty
    JEL: D11 D81 D83
    Date: 2023
  5. By: Ali Hortacsu (University of Chicago and NBER); Olivia R. Natan (University of California, Berkeley); Hayden Parsley (University of Texas, Austin); Timothy Schwieg (University of Chicago, Booth); Kevin R. Williams (Cowles Foundation, Yale University)
    Abstract: We propose a demand estimation method that allows for a large number of zero sale observations, rich unobserved heterogeneity, and endogenous prices. We do so by modeling small market sizes through Poisson arrivals. Each of these arriving consumers solves a standard discrete choice problem. We present a Bayesian IV estimation approach that addresses sampling error in product shares and scales well to rich data environments. The data requirements are traditional market-level data as well as a measure of market sizes or consumer arrivals. After presenting simulation studies, we demonstrate the method in an empirical application of air travel demand.
    Keywords: Discrete Choice Modeling, Demand Estimation, Zero-Sale Observations, Bayesian Methods, Airline Markets
    JEL: C10 C11 C13 C18 L93
    Date: 2021–11
  6. By: Andrea La Nauze; Erica Myers
    Abstract: We use an experiment to test whether consumers optimally acquire information on energy costs in appliance markets where, like many contexts, consumers are poorly informed and make mistakes despite freely available information. To test for optimal information acquisition we compare the average utility gain from improved decision making due to information with willingness to pay for information. We find that consumers acquire information suboptimally. We then compare two behavioral policies: a conventional subsidy for energy-efficient products and a non-traditional subsidy paying consumers to acquire information on energy costs. The welfare effects of each policy depend on the benefits of improved decisions versus the losses of mental effort (from the information subsidy) or distorted choices (from the product subsidy). In our context, information subsidies dominate product subsidies. In a variety of settings where decisions are made and information is delivered online, paying for attention could more effectively target welfare improvements.
    Keywords: endogenous information acquisition, behavioral bias, information interventions, energy efficiency
    JEL: D91 D12 D83 Q41
    Date: 2023
  7. By: Stéphanie Souche-Le Corvec (LAET - Laboratoire Aménagement Économie Transports - UL2 - Université Lumière - Lyon 2 - ENTPE - École Nationale des Travaux Publics de l'État - CNRS - Centre National de la Recherche Scientifique)
    Abstract: There are many different ways of practicing coworking, and many different forms of CoWorking Spaces (CWSs). In this paper, we define CWSs in economic and spatial terms, and we provide some explanations on the transport mode used to travel to. Is the transport mode used to get to CWSs the same as that which is usually used for travel for work purposes? Does the spatial location of the CWS (in large city - medium-sized town - rural community) have an impact on the transport mode choice? The key issue is determining whether these new working spaces favor a shift in travel behavior toward practices less centered on the car. We use a survey of coworkers conducted in 2019 in the Auvergne-Rhône-Alpes Region (AURA), from which data is processed using a binomial logit model. The estimation results show that the mode choice for traveling to CWSs does not allow us to identify a characteristic that is fundamentally different from mode choice for work purpose. However, the availability of a parking place in CWSs is identified as a possible public policy level.
    Keywords: Transport modes, Mode choice, Coworking, Binomial logit model
    Date: 2023–06
  8. By: Benjamin Enke; Thomas Graeber; Ryan Oprea
    Abstract: We provide experimental evidence that core intertemporal choice anomalies – including extreme short-run impatience, structural estimates of present bias, hyperbolicity and transitivity violations – are driven by complexity rather than time or risk preferences. First, all anomalies also arise in structurally similar atemporal decision problems involving valuation of iteratively discounted (but immediately paid) rewards. These computational errors are strongly predictive of intertemporal decisions. Second, intertemporal choice anomalies are highly correlated with indices of complexity responses including cognitive uncertainty and choice inconsistency. We show that model misspecification resulting from ignoring behavioral responses to complexity severely inflates structural estimates of present bias.
    Keywords: complexity, hyperbolic discounting, present bias, bounded rationality, noise, cognitive uncertainty
    JEL: C91 D91 G00
    Date: 2023
  9. By: Christopher Turansick
    Abstract: In recent years there has been an influx of papers which use graph theoretic tools to study stochastic choice. Fiorini (2004) serves as a base for this literature by providing a graphical representation of choice probabilities and showing that every interior node of this graph must satisfy inflow equals outflow. We show that this inflow equals outflow property is almost characteristic of choice probabilities. In doing so, we characterize choice probabilities through graph theoretic tools. As an application of this result, we provide a novel characterization of stochastic rationality on an incomplete domain.
    Date: 2023–03
  10. By: Jiayang Li; Zhaoran Wang; Yu Marco Nie
    Abstract: As one of the most fundamental concepts in transportation science, Wardrop equilibrium (WE) has always had a relatively weak behavioral underpinning. To strengthen this foundation, one must reckon with bounded rationality in human decision-making processes, such as the lack of accurate information, limited computing power, and sub-optimal choices. This retreat from behavioral perfectionism in the literature, however, was typically accompanied by a conceptual modification of WE. Here we show that giving up perfect rationality need not force a departure from WE. On the contrary, WE can be reached with global stability in a routing game played by boundedly rational travelers. We achieve this result by developing a day-to-day (DTD) dynamical model that mimics how travelers gradually adjust their route valuations, hence choice probabilities, based on past experiences. Our model, called cumulative logit (CULO), resembles the classical DTD models but makes a crucial change: whereas the classical models assume routes are valued based on the cost averaged over historical data, ours values the routes based on the cost accumulated. To describe route choice behaviors, the CULO model only uses two parameters, one accounting for the rate at which the future route cost is discounted in the valuation relative to the past ones and the other describing the sensitivity of route choice probabilities to valuation differences. We prove that the CULO model always converges to WE, regardless of the initial point, as long as the behavioral parameters satisfy certain mild conditions. Our theory thus upholds WE's role as a benchmark in transportation systems analysis. It also resolves the theoretical challenge posed by Harsanyi's instability problem by explaining why equally good routes at WE are selected with different probabilities.
    Date: 2023–04
  11. By: Demian Pouzo (University of California); Zacharias Psaradakis (University of London); Martin Sola (Universidad Torcuato di Tella)
    Abstract: We consider hidden Markov models with a discrete-valued regime sequence whose transition probabilities are covariate-dependent. We show that consistent estimation of the parameters of the conditional distribution of the observable variables is possible via quasi-maximum-likelihood based on a (misspecified) mixture model without Markov dependence. Some related numerical results are also discussed.
    Keywords: Consistency; covariate-dependent transition probabilities; hidden Markov model; mixture model; quasi-maximum-likelihood; misspecified model
    Date: 2023–04
  12. By: Francisco Blasques (Vrije Universiteit Amsterdam); Siem Jan Koopman (Vrije Universiteit Amsterdam); Karim Moussa (Vrije Universiteit Amsterdam)
    Abstract: This paper introduces a novel simulation-based filtering method for general state space models. It allows for the computation of time-varying conditional means, quantiles, and modes, but also for the prediction of latent variables in general. The method relies on generating artificial samples of data from the joint distribution implied by the model and on estimating the conditional quantities of interest via extremum estimation. We call this procedure Extremum Monte Carlo and define a corresponding class of filters for signal extraction. The method can be applied to any model from which data can be simulated and is not liable to the curse of dimensionality. Furthermore, the use of extremum estimation allows for a wide range of conditioning sets, including data with missing entries and unequal spacing. The filtering method also places the computational burden predominantly in the off-line phase, which makes it particularly suitable for real-time applications. We present illustrations for some challenging problems characterized by nonlinearity, high-dimensionality, and intractable density functions.
    Keywords: Nonlinear non-Gaussian state space models, Least squares Monte Carlo, Real-time filtering, Intractable densities, Curse of dimensionality
    Date: 2023–03–24

This nep-dcm issue is ©2023 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.