nep-dcm New Economics Papers
on Discrete Choice Models
Issue of 2023‒02‒20
sixteen papers chosen by
Edoardo Marcucci
Università degli studi Roma Tre

  1. Revisiting Panel Data Discrete Choice Models with Lagged Dependent Variables By Christopher R. Dobronyi; Fu Ouyang; Thomas Tao Yang
  2. Drivers of intentions and drivers of actions: willingness to participate versus actual participation in fire management in Sardinia, Italy By G. Concu; C. Detotto; M. Vannini
  3. Do Household Tax Credits Increase the Demand for Legally Provided Services? By Lilith Burgstaller; Annabelle Doerr; Sarah Necker
  4. Identification in a Binary Choice Panel Data Model with a Predetermined Covariate By St\'ephane Bonhomme; Kevin Dano; Bryan S. Graham
  5. The Effect of Credit Constraints on Housing Prices: (Further) Evidence from a Survey Experiment By Tom Cusbert
  6. The refugee mobility puzzle: Why do refugees move to cities with high unemployment rates once residence restrictions are lifted? By Wiedner, Jonas; Schaeffer, Merlin
  7. Comparing the impact of subsidies and health prompts on choice process variables and food choice: The case of dietary fiber By Gustafson, Christopher R.
  8. Personalized information and willingness to pay for non-financial risk prevention : an experiment By Yves Arrighi; David Crainich; Véronique Flambard; Sophie Massin
  9. Retirement Decision of Belgian Couples and the Impact of the Social Security System By Cetin, Sefane; Jousten, Alain
  10. ddml: Double/debiased machine learning in Stata By Achim Ahrens; Christian B. Hansen; Mark E. Schaffer; Thomas Wiemann
  11. Customer Responses to (Im)Moral Behavior of Service Robots Online Experiments in a Retail Setting By Kegel, Mona Mareen; Stock-Homburg, Ruth
  12. Statistical inference for the logarithmic spatial heteroskedasticity model with exogenous variables By Bing Su; Fukang Zhu; Ke Zhu
  13. Recovering utility By Christopher P. Chambers; Federico Echenique; Nicolas S. Lambert
  14. Expectile hidden Markov regression models for analyzing cryptocurrency returns By Beatrice Foroni; Luca Merlo; Lea Petrella
  15. Rationalizable Learning By Andrew Caplin; Daniel J. Martin; Philip Marx
  16. Revisiting Conduct Parameter Estimation in Homogeneous Goods Markets: At Least, Linear Model is Valid By Yuri Matsumura; Suguru Otani

  1. By: Christopher R. Dobronyi; Fu Ouyang; Thomas Tao Yang
    Abstract: This paper revisits the identification and estimation of a class of semiparametric (distribution-free) panel data binary choice models with lagged dependent variables, exogenous covariates, and entity fixed effects. Using an "identification at infinity" argument, we show that the model is point identified in the presence of a free-varying continuous covariate. In contrast with the celebrated Honore and Kyriazidou (2000), our method permits time trends of any form and does not suffer from the "curse of dimensionality". We propose an easily implementable conditional maximum score estimator. The asymptotic properties of the proposed estimator are fully characterized. A small-scale Monte Carlo study demonstrates that our approach performs satisfactorily in finite samples. We illustrate the usefulness of our method by presenting an empirical application to enrollment in private hospital insurance using the HILDA survey data.
    Date: 2023–01
  2. By: G. Concu; C. Detotto; M. Vannini
    Abstract: Changing wildfire regimes coupled with budget cuts are spurring increased involvement of communities and citizens in fire management programs. Policy making faces the task of understanding citizens' willingness to participate and mobilizing will into actions. As there is no reason to expect that the same factors affect willingness to participate and actual participation in the same direction, policy making would require information both on citizens' preferences over management programs and on drivers and barriers to participation. In this paper we compare data on preferences from a Discrete Choice Experiment (DCE) with data on adoption of fire prevention and mitigation measures. The objective is to test if the same factors explain actual participation and willingness to participate in fire management programs. Results suggest that sufficient information for policy design cannot be gained exclusively from the DCE or the analysis of actual behavioural data as the sets of explanatory factors do not entirely overlap. However, two variables – knowledge of fire regulations and community's capacity – can be used to influence both the adoption of prevention and mitigation measures and citizens' willingness to participate in fire management. Policy makers can directly control these factors to nudge the public towards greater involvement in fire prevention and mitigation.
    Keywords: Citizens' participation;Discrete Choice Experiment;Drivers of preparedness;Latent class;Willingness to participate
    Date: 2023
  3. By: Lilith Burgstaller; Annabelle Doerr; Sarah Necker
    Abstract: We study the causal effects of household tax credits on the willingness to demand legally provided services using two survey experiments with 1.974 German homeowners. Participants choose between hypothetical offers of service providers and are randomly assigned to a policy scenario 1) without a tax credit, 2) a tax credit households can claim through the annual tax return, or 3) a tax credit granted by the seller at source. We also vary the refund rate of the tax credit (20/30%) and whether the price including the tax reduction is displayed. All tax credits increase the willingness to pay for offers with invoice as well as the probability to select an offer with invoice. The effectiveness of the tax credit is significantly higher when two attractive features (at source+30%) are combined or when the reduction is made salient. We estimate that about two thirds of respondents who would use the tax credit would have demanded an offer without invoice also without the tax credit.
    Keywords: tax credit, financial rewards for compliance, tax evasion, tax compliance, third-party reporting, survey experiment, discrete choice experiment
    JEL: H26 C93 E26 J22 O17
    Date: 2023
  4. By: St\'ephane Bonhomme; Kevin Dano; Bryan S. Graham
    Abstract: We study identification in a binary choice panel data model with a single \emph{predetermined} binary covariate (i.e., a covariate sequentially exogenous conditional on lagged outcomes and covariates). The choice model is indexed by a scalar parameter $\theta$, whereas the distribution of unit-specific heterogeneity, as well as the feedback process that maps lagged outcomes into future covariate realizations, are left unrestricted. We provide a simple condition under which $\theta$ is never point-identified, no matter the number of time periods available. This condition is satisfied in most models, including the logit one. We also characterize the identified set of $\theta$ and show how to compute it using linear programming techniques. While $\theta$ is not generally point-identified, its identified set is informative in the examples we analyze numerically, suggesting that meaningful learning about $\theta$ is possible even in short panels with feedback.
    Date: 2023–01
  5. By: Tom Cusbert (Reserve Bank of Australia)
    Abstract: The response of housing prices to financing conditions is determined by the effect on the marginal buyer, not the average household. I use heterogeneous willingness to pay (WTP) data from a stated preference experiment in Fuster and Zafar (2021) to estimate the effects of changes in mortgage rates and collateral constraints on housing prices by analysing the structure of housing demand curves. This work builds on their research, which focused on average changes in WTP. Relaxing down payment constraints has a large average effect on WTP, but the effect on price is less than half as large. Financially constrained households tend to respond more to relaxed constraints, but those households often have WTPs that are too low to affect market prices. Changing the mortgage rate has the same average effect on WTPs and on market prices, because there is no systematic relationship between a household's response to mortgage rates and their location on the demand curve. I use a heterogeneous user cost model of individual WTPs to understand how household heterogeneity determines the structure of overall housing demand. An empirical model using observable household characteristics allows the experimental findings to be applied to other household survey data to simulate the effects of credit conditions. The simulated effects of easing collateral constraints in Australia are fairly stable over the past 20 years, and show a similar pattern to the US results.
    Keywords: credit; housing; collateral constraints
    JEL: G21 G51 R21 R38
    Date: 2023–01
  6. By: Wiedner, Jonas; Schaeffer, Merlin (WZB Berlin Social Science Center)
    Abstract: Social science research demonstrates that dispersal policies and restrictions on the freedom of residence have inhibited refugees’ socio-economic integration. The dominant explanation is that such policies prevent refugees from moving to places where they can employ their skills most fruitfully. However, previous studies of refugees’ actual residential choices in Europe provide little evidence that refugees move to places with good employment prospects. The combination of negative effects of residence restrictions and emerging evidence of disadvantaging secondary migration forms what we call the ‘refugee mobility puzzle’. In this study, we address this puzzle and ask: What attracts refugees to deprived areas, and can their seemingly unfortunate residential choices be understood as moves to labor market opportunity after all? Empirically, we draw on the IAB-BAMF-SOEP Survey of Refugees, track the location of more than 2, 500 refugee respondents, and estimate discrete choice models across all German counties and postcodes. Our results confirm the existence of the refugee-mobility puzzle and complicate recent critiques of dispersal policies and restrictions by suggesting that refugees’ need for affordable housing and their desire to be close to (co-ethnic) friends and family may turn into an unintended lock-in factor in the mid- and long-run.
    Date: 2023–01–23
  7. By: Gustafson, Christopher R. (University of Nebraska-Lincoln)
    Abstract: Fiscal tools—taxes and/or subsidies—are increasingly used to address diet-related health problems. However, some studies have found that these tools are markedly more effective if attention is draw to the tax or subsidy, suggesting that the price change alone may go unnoticed in the complex food environments that consumers face. Interventions that prompt individuals to consider health during choice show promise for promoting healthy food choices in both simple laboratory settings and complex, real-world markets. In this pre-registered study, I examine the impact of dietary fiber health prompts and/or dietary fiber subsidies on the per-serving fiber content of foods chosen, the documented set of products considered, and (self-reported) nutrition information use by participants in an online supermarket setting. Participants were randomized to one of four conditions: 1) control, 2) subsidy, 3) fiber prompt, and 4) fiber prompt + subsidy. Results show that both the prompt and prompt + subsidy conditions significantly increase fiber content of foods chosen (with the latter having a larger effect). While all three interventions influence the probability of using nutrition information during food choice and affect the set of products that respondents consider relative to the control condition, the effects were larger for the prompt and prompt + subsidy conditions. A multiple mediation analysis illustrates that both direct and indirect (through the set of products considered and the use of fiber information during choice) pathways lead to the significant overall increase in fiber content of selected foods.
    Date: 2023–01–27
  8. By: Yves Arrighi (LEM - Lille économie management - UMR 9221 - UA - Université d'Artois - UCL - Université catholique de Lille - Université de Lille - CNRS - Centre National de la Recherche Scientifique); David Crainich (LEM - Lille économie management - UMR 9221 - UA - Université d'Artois - UCL - Université catholique de Lille - Université de Lille - CNRS - Centre National de la Recherche Scientifique); Véronique Flambard (LEM - Lille économie management - UMR 9221 - UA - Université d'Artois - UCL - Université catholique de Lille - Université de Lille - CNRS - Centre National de la Recherche Scientifique); Sophie Massin (LEM - Lille économie management - UMR 9221 - UA - Université d'Artois - UCL - Université catholique de Lille - Université de Lille - CNRS - Centre National de la Recherche Scientifique)
    Date: 2022
  9. By: Cetin, Sefane (Université catholique de Louvain, LIDAM/CORE, Belgium); Jousten, Alain (Université de Liège)
    Abstract: This paper investigates the retirement patterns of married couples in Belgium. To forecast retirement behavior, we use administrative Social Security data from 2003 to 2017 and a discrete choice random utility model. In particular, we concentrate on the spousal bonus of pension payments to comprehend how financial incentives resulting from the social security system’s structural design affect both partners’ retirement decisions. We simulate the effect of the elimination of the spousal bonus and find that a small portion of women delay their retirement whereas the rest substitute into alternative social security benefits. Our results not only highlight the significance of cross-program spillovers between various Social Security benefits, but also the heterogeneity in preferences for retirement and asymmetry of retirement behavior between husbands and wives.
    Keywords: Old-Age Labor Supply ; Retirement Incentives ; Spousal Bonus ; Pension Reforms
    JEL: D10 H55 J26
    Date: 2022–11–16
  10. By: Achim Ahrens; Christian B. Hansen; Mark E. Schaffer; Thomas Wiemann
    Abstract: We introduce the package ddml for Double/Debiased Machine Learning (DDML) in Stata. Estimators of causal parameters for five different econometric models are supported, allowing for flexible estimation of causal effects of endogenous variables in settings with unknown functional forms and/or many exogenous variables. ddml is compatible with many existing supervised machine learning programs in Stata. We recommend using DDML in combination with stacking estimation which combines multiple machine learners into a final predictor. We provide Monte Carlo evidence to support our recommendation.
    Date: 2023–01
  11. By: Kegel, Mona Mareen; Stock-Homburg, Ruth
    Date: 2023–01–03
  12. By: Bing Su; Fukang Zhu; Ke Zhu
    Abstract: The spatial dependence in mean has been well studied by plenty of models in a large strand of literature, however, the investigation of spatial dependence in variance is lagging significantly behind. The existing models for the spatial dependence in variance are scarce, with neither probabilistic structure nor statistical inference procedure being explored. To circumvent this deficiency, this paper proposes a new generalized logarithmic spatial heteroscedasticity model with exogenous variables (denoted by the log-SHE model) to study the spatial dependence in variance. For the log-SHE model, its spatial near-epoch dependence (NED) property is investigated, and a systematic statistical inference procedure is provided, including the maximum likelihood and generalized method of moments estimators, the Wald, Lagrange multiplier and likelihood-ratio-type D tests for model parameter constraints, and the overidentification test for the model diagnostic checking. Using the tool of spatial NED, the asymptotics of all proposed estimators and tests are established under regular conditions. The usefulness of the proposed methodology is illustrated by simulation results and a real data example on the house selling price.
    Date: 2023–01
  13. By: Christopher P. Chambers; Federico Echenique; Nicolas S. Lambert
    Abstract: We provide sufficient conditions under which a utility function may be recovered from a finite choice experiment. Identification, as is commonly understood in decision theory, is not enough. We provide a general recoverability result that is widely applicable to modern theories of choice under uncertainty. Key is to allow for a monetary environment, in which an objective notion of monotonicity is meaningful. In such environments, we show that subjective expected utility, as well as variational preferences, and other parametrizations of utilities over uncertain acts are recoverable. We also consider utility recovery in a statistical model with noise and random deviations from utility maximization.
    Date: 2023–01
  14. By: Beatrice Foroni; Luca Merlo; Lea Petrella
    Abstract: In this paper we develop a linear expectile hidden Markov model for the analysis of cryptocurrency time series in a risk management framework. The methodology proposed allows to focus on extreme returns and describe their temporal evolution by introducing in the model time-dependent coefficients evolving according to a latent discrete homogeneous Markov chain. As it is often used in the expectile literature, estimation of the model parameters is based on the asymmetric normal distribution. Maximum likelihood estimates are obtained via an Expectation-Maximization algorithm using efficient M-step update formulas for all parameters. We evaluate the introduced method with both artificial data under several experimental settings and real data investigating the relationship between daily Bitcoin returns and major world market indices.
    Date: 2023–01
  15. By: Andrew Caplin; Daniel J. Martin; Philip Marx
    Abstract: The central question we address in this paper is: what can an analyst infer from choice data about what a decision maker has learned? The key constraint we impose, which is shared across models of Bayesian learning, is that any learning must be rationalizable. To implement this constraint, we introduce two conditions, one of which refines the mean preserving spread of Blackwell (1953) to take account for optimality, and the other of which generalizes the NIAC condition (Caplin and Dean 2015) and the NIAS condition (Caplin and Martin 2015) to allow for arbitrary learning. We apply our framework to show how identification of what was learned can be strengthened with additional assumptions on the form of Bayesian learning.
    JEL: D83 D91
    Date: 2023–01
  16. By: Yuri Matsumura; Suguru Otani
    Abstract: We revisit the conduct parameter estimation in homogeneous goods markets. In contrast to the pessimistic simulation results of linear models shown in Perloff and Shen (2012), our simulation shows that the estimation becomes accurate by properly adding demand shifters in the supply estimation and increasing the sample size. We also investigate log-linear models widely used in Industrial Organization literature and recommended by Perloff and Shen (2012) and find other estimation problems. Based on the numerical investigation, at least the linear model can achieve a proper estimation of the conduct parameter.
    Date: 2023–01

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.