nep-dcm New Economics Papers
on Discrete Choice Models
Issue of 2012‒10‒13
fourteen papers chosen by
Philip Yu
Hong Kong University

  1. A Simple Estimator for Binary Choice Models With Endogenous Regressors By Yingying Dong; Arthur Lewbel
  2. Household Interaction and the Labor Supply of Married Women By Zvi Eckstein
  3. Labor supply as a discrete choice among latent jobs By John K. Dagsvik and Zhiyang Jia
  4. Is China climbing up the quality ladder? By Pula, Gabor; Daniel Santabarbara, Daniel
  5. Joint knowledge generation in European R&D networks: Results from a discrete choice modelling perspective By Florian Reinold; Manfred Paier; Manfred M. Fischer
  6. Housing preferences and attribute importance among Dutch older adults: a conjoint choice experiment By Petra de Jong; Jan Rouwendal; Aleid Brouwer
  7. Landscape Valuation: Choice Experiments or Contingent Valuation By Burgess, Diane; Finney, Graham; Matthews, Dave; Patton, Myles
  8. Using Choice Experiments to Measure Consumers’ Preferences for Functional Dairy Products in Germany: Are Willingness-To-Pay Estimates Affected by Starting Point Bias? By Bechtold, Kai-Brit; Abdulai, Awudu
  9. Do Preferential Trade Agreements Increase Members’ Agri-food Trade? By Ul Haq, Zahoor; Meilke, Karl D.; Orden, David
  10. Food Quality Verifications and Consumer Trust By Hobbs, Jill E.; Innes, Brian G.; Uzea, Adrian D.; Zhang, Jing
  11. Using Attribute Importance Rankings within Discrete Choice Experiments: an application to Valuing Bread Attributes. By Balcombe, Kelvin George; Bitzios, Michael; Fraser, Iain; Haddock-Fraser, Janet
  12. A Choice Experiments Application in Transport Infrastructure: A Case Study on Travel Time Savings, Accidents and Pollution Reduction By Koundouri, Phoebe; Kountouris, Yannis; Stithou, Mavra
  13. Multivariate Choice and Identification of Social Interactions By Cohen-Cole, Ethan; Liu, Xiaodong; Zenou, Yves
  14. The Marginal Willingness-to-Pay for Attributes of a Hypothetical HIV Vaccine By Michael P. Cameron; Peter A. Newman; Surachet Roungprakhon; Riccardo Scarpa

  1. By: Yingying Dong (California State University, Irvine); Arthur Lewbel (Boston College)
    Abstract: This paper provides a few variants of a simple estimator for binary choice models with endogenous or mismeasured regressors, or with heteroskedastic errors, or with panel fixed effects. Unlike control function methods, which are generally only valid when endogenous regressors are continuous, the estimators proposed here can be used with limited, censored, continuous, or discrete endogenous regressors, and they allow for latent errors having heteroskedasticity of unknown form, including random coefficients. The variants of special regressor based estimators we provide are numerically trivial to implement. We illustrate these methods with an empirical application estimating migration probabilities within the US.
    Keywords: Binary choice, Binomial response, Endogeneity, Measurement error, Heteroskedasticity, Discrete endogenous regressor, Censored regressor, Random coefficients, Identification, Latent variable model. Classification-JEL: C25, C26
    Date: 2012–06–15
  2. By: Zvi Eckstein (Tel Aviv University)
    Abstract: The major increase in the employment rate of married women while that of men remained almost unchanged is one of the most dramatic socioeconomic changes to have taken place during the last century. In this paper, we argue that shifts in social norms regarding household interaction in determining a married couple’s labor supply can provide an explanation. Specifically, we formulate and estimate a dynamic discrete-choice labor supply model, assuming that there are two types of households – Classical and Modern. The Classical household follows a Stackelberg leader game in which the wife’s labor supply decision follows her husband’s already-known employment outcome. The Modern family is characterized by a symmetric and simultaneous game that determines their joint labor supply and has a Nash equilibrium. The family type – Modern or Classical – is exogenously determined when the couple gets married but is not observable for estimation. The model is estimated using the Simulated Moments Method (SMM) and data from the Panel Study of Income Dynamics (PSID) survey for the years 1983-93. The estimated model accurately predicts employment rates and produces a good fit of mean wages to the data. We estimate that 38 percent of families are Modern and that the participation rate of women in those households is almost 80 percent. The employment rate of women in Classical families is 10 percent lower than that while the employment rates of men is almost identical in the two household types. These results support our hypothesis that part of the increase in labor supply of married women may be due to an increase in the share of Modern families in the population.
    Date: 2012
  3. By: John K. Dagsvik and Zhiyang Jia (Statistics Norway)
    Abstract: This paper discusses a modeling framework in which workers are assumed to choose their preferred job from latent worker-specific choice sets. This point of departure yields a framework that formalizes the widely used ad hoc approaches (fixed cost of working and dummies at peak hours) in the literature of discrete labor supply models. We discuss the conditions under which the preferences and job opportunity restrictions can be separated using conventional data on hours and wages only. Subsequently, we show that the framework is consistent with stochastic choice sets and a relaxation of the IIA assumption. An empirical model version for married/cohabiting couples is estimated using Norwegian micro data. Based on the empirical model, we discuss further important empirical issues, such as functional form, prediction performance and simulation of counterfactual policy reforms.
    Keywords: Labor supply; non-pecuniary job attributes; latent choice sets; random utility models.
    JEL: J22 C51
    Date: 2012–10
  4. By: Pula, Gabor (BOFIT); Daniel Santabarbara, Daniel (BOFIT)
    Abstract: There is an ongoing debate in the literature about the quality content of Chinese exports and to what extent China imposes a threat to the market positions of advanced economies. While China’s export structure is very similar to that of the advanced world, its export unit values are well below the level of developed economies. Building on the assumption that unit values reflect quality the prevailing view of the literature is that China exports low quality varieties of the same products than its advanced competitors. This paper challenges this view by relaxing the assumption that unit values reflect quality. We derive the quality of Chinese exports to the European Union by estimating disaggregated demand functions from a discrete choice model. The paper has three major findings. First, China’s share on the European Union market is larger than would be justified only by its low average prices, implying that the quality of Chinese exports is high compared to many competitors. Second, China has gained quality relative to other competitors since 1995, indicating that China is climbing up the quality ladder. Finally, our analysis on the supply side determinants reveals that the relatively high quality of Chinese exports is related to processing trade and the increasing role of global production networks in China.
    Keywords: Chinese exports; vertical product differentiation; quality ladder; global production networks; discrete choice model; COMEXT database
    JEL: F10 F12 F14 F15 F23
    Date: 2012–10–02
  5. By: Florian Reinold; Manfred Paier; Manfred M. Fischer
    Abstract: The objective of this study is to explore the determinants of joint knowledge generation within European networks of R&D collaboration. This study distinguishes between two types of joint knowledge generation: scientific and commercially relevant knowledge generation. Joint generation of scientific knowledge is measured by co-authored scientific publications, while joint commercially relevant knowledge is measured by co-owned patents and artefacts. Unit of analysis are dyads of organisations jointly participating in projects of the 5th EU Framework Programme (FP5). The data for carrying out this study is taken from a survey among FP5 participants and the EUPRO database. 23 EU member countries (Bulgaria, Cyprus, Malta and Rumania are excluded) plus Switzerland and Norway are included. Regression methods for discrete choice (logit and probit) are employed to meet the objective. The independent variables taken into consideration encompass the types of organisations involved in the dyad, geographical and cultural obstacles, relational factors and project characteristics. Results show that dyads involving universities have the highest probability not only to jointly generate scientific knowledge but also to jointly generate commercially relevant knowledge, whereas the involvement of an industry organisation results in a low probability for both types of knowledge generation. Perhaps, this can be attributed to the fact that joint knowledge generation entails disclosure of own knowledge, which is actually a task of universities but is problematic for industry organisations. Another important result is that crossing national border has a significant positive rather than negative effect on joint scientific knowledge generation, which is essentially a consequence of how the Framework Programmes had been set up. Similarly, crossing EU-15 external border has a positive effect on joint knowledge generation, indicating that the FPs work well in achieving their aim of supporting the catching up process of CEE countries. But, joint generation of commercially relevant knowledge is negatively influenced by language borders. This can be explained by the fact that the co-development of patentable knowledge or artefacts requires more intensive and complex interactions than to co-author a scientific publication where English is the lingua franca anyway. Results on relational factors and project characteristics satisfy expectations: Duration of collaboration and the existence of previous collaboration have a positive effect on joint knowledge generation, whereas the project size, measured by number of participants, affects joint knowledge generation negatively.
    Date: 2012–10
  6. By: Petra de Jong; Jan Rouwendal; Aleid Brouwer
    Abstract: The “ageing†of the developed world population is a well-known and well documented phenomenon. In the year 2012 16 percent of the Dutch population is aged 65 and older. By the year 2040 this figure will rise to approximately 26 percent. The change in the number and the proportion of older adults in our society will have numerous implications. For one, the increase of older adults will place an enormous burden on existing income systems, health care systems, social services and retirement programs. It has also been estimated that, for the period 2006 to 2015, there is a shortage of 406.000 houses suitable for older adults in the Netherlands. One of the biggest challenges is to provide proper housing conditions that correspond with the diverse housing preferences of an ageing population. In order to respond accurately to this challenge we need to develop a further understanding of the housing preferences of Dutch older adults. Over the years, several modelling approaches have been proposed and applied in research to measure preferences. An emerging, and one of the most popular modelling approaches at the time, is the stated preferences approach. In this research the stated housing preferences of Dutch older adults will be analysed based on a carefully constructed survey, which is designed as conjoint choice experiment. A conjoint choice experiment is an approach that seeks to describe and predict preferences of respondents by looking at their responses to hypothetical residential profiles that can be viewed as integral descriptions of housing situations (characteristics of the house, the environment, etc.). In this study, approximately 1000 older adults are presented with 24 so-called choice sets, each existing out of three residential profiles. The older adults are asked to indicate the first and second most preferred residential profile, thereby revealing their complete orderings of the three. Subsequently, the utility function (e.g. preference) can be estimated by using logistic regression analysis. Based on these estimates it is possible to determine the relative importance of certain housing attributes. Keywords: stated preferences, conjoint analyses, older adults, housing JEL codes: C35 - Discrete Regression and Qualitative Choice Models, J14 - Economics of the Elderly, R21 - Housing Demand
    Date: 2012–10
  7. By: Burgess, Diane; Finney, Graham; Matthews, Dave; Patton, Myles
    Abstract: Landscapes represent the dynamic interaction of natural and cultural processes acting on the environment. Increasingly human impacts are dominating the natural processes resulting in landscape change and habitat loss. Due to the public good nature of landscapes, no market price exists to indicate their economic value and consequently impacts to the landscape are excluded from decision making processes. To include landscape change within the decision making process, valuation studies have been undertaken; primarily stated preference methods. In common with the valuation of many public goods, Choice Experiments (CE), have dominated the landscape valuation literature. However, CE makes the implicit assumption that the value of the good can be captured by the attributes of the the good. In CE a landscape would be described in terms of its features i.e. trees, field boundries. Drawing from psychology/cognitive research, we explore whether the spatial configuration of those landscape features has an impact on preferences. The findings of two surveys indicate that spatial configuration does have an impact on landscape preferences and therefore potentially on economic values. This would indicate that unless CE can incorporate spatial configuration, they may not be an appropriate method for valuing landscapes.
    Keywords: Landscapes, Stated Preference methods, Environmental Economics and Policy, Q24, Q51, Q57,
    Date: 2012–04
  8. By: Bechtold, Kai-Brit; Abdulai, Awudu
    Abstract: This study employs stated preference data from a choice experiment to address two issues related to consumer demand for functional dairy products: (1) Consumers’ preferences for functional dairy product attributes in Germany, and (2) are willingness-to-pay estimates obtained in the choice experiment affected by starting point bias? Based on a random parameter logit model, our results indicate that dairy products enriched with omega-3 fatty acids and bearing a health claim that is aimed at healthy blood vessels and healthy metabolism are highly valued. Furthermore, results reveal that willingness-to-pay is indeed susceptible to starting point bias. In a two-split sample approach, we find that varying the price levels displayed in the first choice set significantly affects respondents’ willingness-to-pay for functional dairy products.
    Keywords: Functional food attributes, choice experiments, preference heterogeneity, willingness-to-pay, starting point bias, Demand and Price Analysis, Health Economics and Policy, Research Methods/ Statistical Methods, C25, D12,
    Date: 2012
  9. By: Ul Haq, Zahoor; Meilke, Karl D.; Orden, David
    Abstract: This study estimates the effect of a diverse group of 30 PTAs on members’ trade of 26 agri-food products categorized into eight commodity sectors for 1990, 1995, 2000 and 2000 using disaggregated trade data for 40 countries and the Heckman selection model. Results show that whether reported zero trade-flows are considered actual or potential affects the size of the estimated PTA impacts. However, irrespective of the true nature of the zero trade-flows, the effects of PTAs are found positive and statistically significant. OLS estimates fall between the Heckman-model-derived conditional and unconditional effects of PTAs.
    Keywords: Preferential Trade Agreements, agri-food trade, selection bias, Heckman, Food Consumption/Nutrition/Food Safety, F130, C180,
    Date: 2012–09–28
  10. By: Hobbs, Jill E.; Innes, Brian G.; Uzea, Adrian D.; Zhang, Jing
    Abstract: Food markets are increasingly characterized by an array of quality assurances with respect to credence attributes, reflecting a growing interest in where food comes from and how it is produced. The provision and signalling of these credence quality attributes includes both public and private sector initiatives. How effective are quality signals in addressing the information asymmetry inherent in credence attributes? To what extent do consumers trust quality assurances from different sources, and does this trust differs across food products or across credence attributes? The paper presents a simple economic welfare analysis of the market for a credence attribute under different assumptions with respect to the strength of consumer preferences, the existence of voluntary versus mandatory standards, and the credibility of third party certification. This is followed by an empirical analysis drawing from two consumer surveys in Canada using discrete choice experiments. Food quality claims related to farm animal welfare in a meat product and to environmental sustainability in a bread product are examined. Latent Class models reveal significant heterogeneity in consumer preferences, both in terms of the value consumers place on farm animal welfare and environmentally sustainable quality assurances, and the extent to which it matters who is verifying these assurances.
    Keywords: credence attribute, animal welfare, environmental sustainability, quality verification, discrete choice experiment, Agricultural and Food Policy, Risk and Uncertainty, Q13, Q18,
    Date: 2012
  11. By: Balcombe, Kelvin George; Bitzios, Michael; Fraser, Iain; Haddock-Fraser, Janet
    Abstract: In this paper we present results from a Choice Experiment (CE) incorporating respondent debrie…ng ranking information about attribute importance employing a modi…ed Mixed Logit using Bayesian methods. Our results indicate that a CE debrie…ng question that asks respondents to rank the importance of attributes, as opposed to simply indicating attendance or non-attendance, helps to explain the resulting choices. We also examine how mode of survey delivery (online and mail) impacts model performance and …nd that our results are not substantively a¤ected by the mode of survey delivery. We conclude that the ranking data is a complementary source of information about respondent utility functions within CE.
    Keywords: Choice Experiment, Attribute Non-Attendance, Survey Mode, Labor and Human Capital, C11, C25, L66,
    Date: 2012–04
  12. By: Koundouri, Phoebe; Kountouris, Yannis; Stithou, Mavra
    Abstract: This paper presents the results of a Choice Experiment (CE) conducted to estimate the values derived from a highway construction project in Greece. To account for preference heterogeneity conditional logit with interactions and random parameter logit models are estimated. The results indicate that individuals have significant values for travel time savings, percentage decrease in traffic accidents, percentage decrease in traffic related emissions and landscape modifications. Models where the attributes are interacted with socioeconomic variables perform better and produce lower welfare estimates compared to models without interactions with important implications for cost benefit analysis.
    Keywords: Choice experiments; transport infrastructure;travel time savings; accidents; pollution reduction
    JEL: C9 N7 C93
    Date: 2012
  13. By: Cohen-Cole, Ethan; Liu, Xiaodong; Zenou, Yves
    Abstract: In this paper, we investigate the impact of peers on own outcomes where all agents embedded in a network choose more than one activity. We develop a simple network model that illustrates these issues. We differentiate between the ‘seemingly unrelated’ simultaneous equations model where people are influenced only by others within the same activity, the ‘triangular’ simultaneous equations model, where there is some asymmetry in the peers’ cross effects, and the ‘square’ simultaneous equations model, where all possible cross-choice effects are taken into account. We develop the conditions under which each model is identified, showing that the general ‘square’ simultaneous equations model with both simultaneity effect and cross-choice peer effect cannot be identified without any exclusion restrictions. We then study the impact of peer effects on education and screen activities and show that the estimated within- and cross-choice peer effects both have non-trivial impacts on adolescent behavior. We find, in particular, that, keeping peers’ grades and screen activities fixed, watching more TV could be beneficial to a student’s grade.
    Keywords: identification; peer effects; Social networks
    JEL: C21 C3 I21 Z13
    Date: 2012–09
  14. By: Michael P. Cameron (University of Waikato); Peter A. Newman (University of Toronto); Surachet Roungprakhon (Rajamangala University of Technology Phra Nakhon); Riccardo Scarpa (University of Waikato)
    Abstract: This paper estimates the marginal willingness-to-pay for attributes of a hypothetical HIV vaccine using discrete choice modeling. We use primary data from 326 respondents from Bangkok and Chiang Mai, Thailand, in 2008-2009, selected using purposive, venue-based sampling across two strata. Participants completed a structured questionnaire and full rank discrete choice modelling task administered using computer-assisted personal interviewing. The choice experiment was used to rank eight hypothetical HIV vaccine scenarios, with each scenario comprising seven attributes (including cost) each of which had two levels. The data were analyzed in two alternative specifications: (1) best-worst; and (2) full-rank, using logit likelihood functions estimated with custom routines in Gauss matrix programming language. Knowledge of the relative importance of determinants of HIV vaccine acceptability is important to ensure the success of future vaccination programs. Future acceptability studies of hypothetical HIV vaccines should use more finely-grained biomedical attributes, and could also improve the external validity of results by including more levels of the cost attribute.
    Keywords: HIV vaccine; willingness-to-pay; conjoint analysis; discrete choice; Thailand
    JEL: I19
    Date: 2012–09–30

This nep-dcm issue is ©2012 by Philip Yu. 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.