nep-dcm New Economics Papers
on Discrete Choice Models
Issue of 2018‒07‒23
five papers chosen by
Edoardo Marcucci
Università degli studi Roma Tre

  1. Using penalized likelihood to select parameters in a random coefficients multinomial logit model By Joel L. Horowitz; Lars Nesheim
  2. The Impact of Cigarette Smoking on Health Care Utilization among Rural Residents in China By Changle Li; Siripen Supakankunti
  3. Does Limiting Internet Use Among Adolescents Affect their Offline Social relationships? By Mora Villarrubia, Ricardo; Li, Yunrong
  4. Geography, Ties and Knowledge Flows: Evidence from Citations in Mathematics By Keith Head; Yao Amber Li; Asier Minondo
  5. Network and panel quantile effects via distribution regression By Victor Chernozhukov; Ivan Fernandez-Val; Martin Weidner

  1. By: Joel L. Horowitz (Institute for Fiscal Studies and Northwestern University); Lars Nesheim (Institute for Fiscal Studies and cemmap and UCL)
    Abstract: The multinomial logit model with random coefficients is widely used in applied research. This paper is concerned with estimating a random coefficients logit model in which the distribution of each coefficient is characterized by finitely many parameters. Some of these parameters may be zero or close to zero in a sense that is defined. We call these parameters small. The paper gives conditions under which with probability approaching 1 as the sample size approaches infinity, penalized maximum likelihood estimation (PMLE) with the adaptive LASSO (AL) penalty function distinguishes correctly between large and small parameters in a random-coefficients logit model. If one or more parameters are small, then PMLE with the AL penalty function reduces the asymptotic mean-square estimation error of any continuously differentiable function of the model’s parameters, such as a market share, the value of travel time, or an elasticity. The paper describes a method for computing the PMLE of a random-coefficients logit model. It also presents the results of Monte Carlo experiments that illustrate the numerical performance of the PMLE. Finally, it presents the results of PMLE estimation of a random-coefficients logit model of choice among brands of butter and margarine in the British groceries market.
    Date: 2018–04–26
  2. By: Changle Li (Faculty of Economics, Chulalongkorn University); Siripen Supakankunti (Faculty of Economics, Chulalongkorn University)
    Abstract: Objectives: To assess how cigarette smoking affects health care utilization among rural residents in China and to analyze the choice of health facility among current and former smokers in rural China.Methods: Fixed effect and random effect logistic models were used to assess the impact of cigarette smoking on health care utilization. The multinomial logistic model was employed to analyze the choice of health facility among current and former smokers. This study employed three waves (2010-2014) of the China Family Panel Studies, which is a nationally representative, biennially longitudinal survey in China. The data set included 10,330 adults in each wave. Results: First, according to fixed effect logistic model, the results showed that current smokers (light and heavy smokers) and former smoker (recent, moderate-term and long-term quitters) are positively (ORs>1) and significantly associated with outpatient care. Moreover, recent, moderate-term and long-term quitters are positively (ORs>1) and significantly associated with inpatient care. Second, based on random effect logistic model, the results displayed that heavy smokers and long-term quitters are positively (ORs>1) and significantly related to outpatient care, and former smokers (recent, moderate-term and long-term quitters) are positively (ORs>1) and significantly associated with inpatient care. Last, the results of multinomial logistic model uncovered that recent and long-term quitters increase the probabilities of visiting general hospital by 36% and 49% compared to visiting village health center and increase the probabilities of visiting special hospital by 90% and 79% compared to visiting clinic.Conclusions: Firstly, the present study employing fixed effect logistic model found that current and former smokers use more outpatient care than never smokers. Moreover, former smokers use more inpatient care than never smokers. Secondly, the current study showed that although former smokers use more inpatient care than never smokers, long-term quitters decrease the probability of using inpatient care compared with recent and moderate-term quitters according to the random effect logistic model. Lastly, this study discovered that recent and long-term quitters have much higher probability of visiting general hospital compared to vising village health center and clinic. Therefore, the tobacco tax policy and smoking cessation treatments should be implemented in rural China.
    Keywords: Smoking; Health Care Utilization; Rural China
    JEL: I14
    Date: 2017–07
  3. By: Mora Villarrubia, Ricardo; Li, Yunrong
    Abstract: We use longitudinal individual data from a nationally representative sample of Chinese children aged 10 to 15 to investigate whether a partial ban on internet use affects offline social interactions among adolescents. We present both IV and mixed logit estimates and control for spillover effects and unobservable heterogeneity. Although online and offline relationships are strongly positively correlated, the existence of unobservable common tastes for social interactions at household level explains an important part of this correlation. We also find that (i) most children do reduce their offline social relationships when facing limits to their use of social networking sites; (ii) there are spillover effects in the sense that these negative effects increase with the size of the online local network; (iii) the negative effects are sizable and statistically significant for older children (aged 14 and 15); and (iv), for these older children, they occur even after discounting the network effect. To sum up, we find large negative effects on offline social relationships for older children even if the limits are imposed to all children simultaneously.
    Keywords: Online social relations; complementarity; unobservable common tastes
    JEL: J14 J13
    Date: 2018–07
  4. By: Keith Head; Yao Amber Li; Asier Minondo
    Abstract: Using data on academic citations, career and educational histories of mathematicians, and disaggregated distance data for the world's top 1000 math departments, we study how geography and ties affect knowledge flows among scholars. The ties we consider are co-authorship, past colocation, advisor-mediated relationships, and alma mater relationships (holding a Ph.D. from the institution where another scholar is affiliated). Logit regressions using fixed effects that control for subject similarity, article quality, and temporal lags, show linkages are strongly associated with citation. Controlling for ties generally halves the negative impact of geographic barriers on citations. Ties matter more for less prominent and more recent papers and show no decline in importance in recent years. The impact of distance - controlling for ties - has fallen and is statistically insignificant after 2004.
    Keywords: knowledge diffusion, distance, borders, networks, academic genealogy
    JEL: F1 O3 R1
    Date: 2018–06
  5. By: Victor Chernozhukov (Institute for Fiscal Studies and MIT); Ivan Fernandez-Val (Institute for Fiscal Studies and Boston University); Martin Weidner (Institute for Fiscal Studies and cemmap and UCL)
    Abstract: This paper provides a method to construct simultaneous con fidence bands for quantile functions and quantile effects in nonlinear network and panel models with unobserved two-way effects, strictly exogenous covariates, and possibly discrete outcome variables. The method is based upon projection of simultaneous confi dence bands for distribution functions constructed from fixed effects distribution regression estimators. These fi xed effects estimators are bias corrected to deal with the incidental parameter problem. Under asymptotic sequences where both dimensions of the data set grow at the same rate, the confi dence bands for the quantile functions and effects have correct joint coverage in large samples. An empirical application to gravity models of trade illustrates the applicability of the methods to network data.
    Date: 2018–03–22

This nep-dcm issue is ©2018 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.