nep-dcm New Economics Papers
on Discrete Choice Models
Issue of 2014‒12‒03
sixteen papers chosen by
Edoardo Marcucci
Università degli studi Roma Tre

  1. Consumer Preferences, Ecolabels, and the Effects of Negative Environmental Information By Chen, Xianwen; Alfnes, Frode; Rickertsen, Kyrre
  2. Income taxation, labour supply and housework: a discrete choice model for French couples By Jan Kabatek; Arthur Van Soest; Elena Stancanelli
  3. “Using Experiments to Address Attribute Non-attendance in Consumer Food Choices” By Caputo, Vincenzina; Loo, Ellen J. Van; Scarpa, Riccardo; Nayga, Rodolfo M. Jr; Verbeke, Wim
  4. ANALYSING FARMERS’ USE OF PRICE HEDGING IN-STRUMENTS: AN EXPERIMENTAL APPROACH By Anastassiadis, Friederike; Feil, Jan-Henning; Mußhoff, Oliver; Schilling, Philipp
  5. Identifying which ecosystem services coastal residents actually value: A choice experiment survey of the Eastern Shore of Virginia regarding climate change adaptation By Yue, Ian T.; Swallow, Stephen K.
  6. The Effect of Forced Choice with Constant Choice Experiment Complexity By Penn, Jerrod; Hu, Wuyang; Cox, Linda J.
  7. Parameter Estimation in Multivariate Logit models with Many Binary Choices By Bel, K.; Fok, D.; Paap, R.
  8. Distribution-free Methods for Estimation of Willingness to Pay Models Using Discrete Response Valuation Data By Zapata, Samuel D.; Carpio, Carlos E.
  9. Water Demand Elasticities: Selection Effects in Meta-Analysis By Hoehn, John
  10. Adoption of Information and Communication Technologies and New Organizational Practices in the Tunisian Manufacturing Sector By Adel Ben Youssef; Walid Hadhri; Hatem Mhenni
  11. Can Explicit Price Reminders Mitigate Hypothetical Bias in Online Choice Experiment? By Lim, Kar Ho; Hu, Wuyang
  12. The effect of unconditional cash transfers on adult labour supply: A unitary discrete choice model for the case of Ecuador By Mideros A.; O'Donoghue C.
  13. Examination of the Transitions of Households into and out of Poverty in Turkey By Aysenur Acar; Cem Baslevent
  14. Wages Anatomy. Labor supply of nurses and a comparison with physicians. By Andreassen, Leif; Di Tommaso, Maria Laura; Strøm, Steinar
  15. Modelling Changing Rural Land Use in New Zealand 1997 to 2008 Using a Multinomial Logit Approach By Zack Dorner; Dean Hyslop
  16. Willingness to Pay to Reduce Wild Fire Risk in Wild land-Urban Interface: A Comparative Analysis of Public Programs and Private Actions By Christman, Laine; Kimberly, Rollins; Micheal, Taylor; Sohan, Agbonlahor

  1. By: Chen, Xianwen; Alfnes, Frode; Rickertsen, Kyrre
    Abstract: Consumers prefer ecolabeled products. However, little is known about the effects of ecolabels when consumers are simultaneously exposed to negative environmental information. We conducted a stated choice experiment in France with eight fish products that were either ecolabeled or unlabeled. Four types of negative information concerning the potential negative environmental consequences of catching wild fish or producing farmed fish were randomly administered to the participants. The data were analyzed by a mixed logit model. Several results emerged. First, there are positive ecolabeling effects on the willingness to pay (WTP) for fish. Second, ecolabeling cannot fully mitigate the negative effects on WTP of negative environmental information. Third, there is a positive effect on the WTP for substitute fish produced with the same production technology as the type of fish that receives negative environmental information.
    Keywords: environmental information, spillover effect, eco-label, fish, discrete choice, stated preferences, Agribusiness, Consumer/Household Economics, Food Consumption/Nutrition/Food Safety, Q13, Q51,
    Date: 2014
  2. By: Jan Kabatek (Tilburg University [Tilburg] - Netspar); Arthur Van Soest (Tilburg University [Tilburg] - Netspar); Elena Stancanelli (CES - Centre d'économie de la Sorbonne - CNRS : UMR8174 - Université Paris I - Panthéon-Sorbonne, EEP-PSE - Ecole d'Économie de Paris - Paris School of Economics - Ecole d'Économie de Paris)
    Abstract: Earlier studies suggest that income taxation may affect not only labour supply but also domestic work. Here we investigate the impact of income taxation on partners' labour supply and housework, using data for France that taxes incomes of married couples jointly. We estimate a household utility model in which the marginal utilities of leisure and housework of both partners are modelled as random coefficients, depending on observed and unobserved characteristics. We conclude that both partners' market and housework hours are responsive to changes in the tax system. A policy simulation suggests that replacing joint taxation of married spouses' incomes with separate taxation would increase the husband's housework hours by 1.3% and reduce his labour supply by 0.8%. The wife's market hours would increase by 3.7%, and her housework hours would fall by 2.0%.
    Keywords: Time use; Taxation; Discrete choice models
    Date: 2014–04
  3. By: Caputo, Vincenzina; Loo, Ellen J. Van; Scarpa, Riccardo; Nayga, Rodolfo M. Jr; Verbeke, Wim
    Abstract: A number of choice experiments (CEs) studies have shown that survey respondents employ heuristics such as attribute non-attendance (ANA) while evaluating food. This paper addresses a set of methodological questions. First, it explores if ANA is an issue to take into account in food valuation studies. Second, it assesses if there is any difference in terms of welfare estimates between the two common ways of collecting self-reported stated ANA (serial and choice task). Next, it validates the statements of ANA behavior provided by the respondents across serial and choice task self-reported ANA. Lastly, it explores the issue of concordance between the stated ANA and inferred ANA methods. We estimated a set of choice models including inferred or observed ANA information. Our results show a clear winner between the two stated approaches, choice task, which also better matches the inferred ANA data.
    Keywords: serial stated attribute non-attendance, choice task attribute non-attendance, </dc:subject><dc:subject>inferred attribute non-attendance, food choice experiments, </dc:subject><dc:subject>food labeling, Consumer/Household Economics, Food Consumption/Nutrition/Food Safety, Research Methods/ Statistical Methods,
    Date: 2014–07
  4. By: Anastassiadis, Friederike; Feil, Jan-Henning; Mußhoff, Oliver; Schilling, Philipp
    Abstract: This paper analyses the influencing factors of farmers’ use of price hedging instruments (PHIs) based upon a discrete choice experiment with German grain farmers. A mixed logit model is used to determine whether farmers’ choices of PHIs against cash sales are influenced by their price expectation, their risk attitude and their available storage capacities. The results show that farmers with a price expectation below the actual price level have a higher preference for using PHIs against cash sales in general and that the individual degree of risk aversion can have a significant impact on farmers’ choices of a specific PHI. A generally lower preference of farmers with available storage capacities for using PHIs as assumed in many theoretical contributions in the literature, however, cannot be confirmed.
    Keywords: Price hedging instruments, grain marketing, discrete choice experiment, mixed logit, Demand and Price Analysis, Farm Management,
    Date: 2014–09
  5. By: Yue, Ian T.; Swallow, Stephen K.
    Keywords: Environmental Economics and Policy, Land Economics/Use,
    Date: 2014
  6. By: Penn, Jerrod; Hu, Wuyang; Cox, Linda J.
    Abstract: In this analysis we compare WTP estimates of responses separately modelling choice sets of forced choices from those of unforced choices in order to measure the effect of forced choice in Choice experiment. This comparison is done while still maintain constant task complexity and evaluated in WTP space. We find evidence to suggest that individual WTP are different in forced and unforced choice sets and in joint tests for parameter equality.
    Keywords: Choice Experiment, Environmental Economics and Policy,
    Date: 2014
  7. By: Bel, K.; Fok, D.; Paap, R.
    Abstract: __Abstract__ The multivariate choice problem with correlated binary choices is investigated. The Multivariate Logit [MVL] model is a convenient model to describe such choices as it provides a closed-form likelihood function. The disadvantage of the MVL model is that the computation time required for the calculation of choice probabilities increases exponentially with the number of binary choices under consideration. This makes maximum likelihood-based estimation infeasible in case there are many binary choices. To solve this issue we propose three novel estimation methods which are much easier to obtain, show little loss in efficiency and still perform similar to the standard Maximum Likelihood approach in terms of small sample bias. These three methods are based on (i) stratified importance sampling, (ii) composite conditional likelihood, and (iii) generalized method of moments. Monte Carlo results show that the gain in computation time in the Composite Conditional Likelihood estimation approach is large and convincingly outweighs the limited loss in efficiency. This estimation approach makes it feasible to straightforwardly apply the MVL model in practical cases where the number of studied binary choices is large.
    Keywords: Multivariate Logit model, Stratified Importance Sampling, Composite Likelihood, Generalized Method of Moments
    Date: 2014–10–20
  8. By: Zapata, Samuel D.; Carpio, Carlos E.
    Abstract: The Turnbull method is the standard approach used in contingent valuation studies to estimate willingness to pay (WTP) models using discrete responses without making assumptions about the distribution of the data. However, this approach has several limitations. The purpose of this study is to develop alternative distribution-free methods for the estimation of WTP models using nonparametric conditional imputation and local regression procedures. The proposed approaches encompass the recovery of the individuals’ WTP values using an iterated conditional expectation procedure and subsequent estimation of the mean WTP using linear and nonparametric additive models. In contrast to the Turnbull approach, the proposed estimation methods allow the inclusion of covariates in the modeling of WTP estimates, as well as the complete recovery of its underlying probability distribution. Monte Carlo simulations are employed to compare the performance of the proposed estimators with that of the Turnbull estimator. We also illustrate the use of the proposed estimation techniques using a real data set.
    Keywords: Additive models, double-bounded elicitation, kernel functions, iterated conditional expectation, non-parametric regression, Turnbull estimator., Research Methods/ Statistical Methods,
    Date: 2014–05
  9. By: Hoehn, John
    Abstract: A meta-analysis of reported water demand elasticities shows that publication selection bias results in water demand estimates that overstate price elasticities. The analysis uses panel data estimators to measure and adjust for publication bias. A random effects estimator is used to derive water demand elasticity estimates that are efficient and unbiased by selection effects. The sample mean elasticity for reported elasticities is -.37%. Once publication bias is removed, the mean elasticity estimate drops to -.08%. Indoor water demand is almost perfectly inelastic. Water demands and their elasticities are heterogeneous in specific uses and locations. Water demand elasticities vary by location, water use type, econometric approach and publication bias. Water use elasticities range up -1.48% for domestic irrigation demand in the southwestern US. Estimated elasticities are also higher when derived with discrete choice data and estimators.
    Keywords: water, demand, elasticity, metaanalysis, Meta-analysis, meta analysis, panel data, random effects, fixed effects, Demand and Price Analysis, Resource /Energy Economics and Policy, D12, Q25,
    Date: 2014–07–27
  10. By: Adel Ben Youssef (University of Nice Sophia Antipolis, France; GREDEG CNRS); Walid Hadhri (Université de Tunis); Hatem Mhenni (Ecole Supérieure de Commerce de Tunis; Université de la Manouba)
    Abstract: The aim of this paper is to explore the relationship between the adoption of Information Technologies (IT) and the adoption of New Organizational Practices (NOP) in the context of an emerging country (Tunisia). Based on face-to-face questionnaire, to a random sample of 175 Tunisian manufactures, and using an ordered logit model, our empirical results show a significant link between IT adoption and NOP. We show that the complementarity is strengthened when the technology evolves. Adoption and usage of latest technologies are pushed by the prior adoption of NOP.
    Keywords: ICT adoption, ICT usage, Ordred Logit Model, New Organizational Practices
    JEL: L21 O31 O33
    Date: 2014–10
  11. By: Lim, Kar Ho; Hu, Wuyang
    Keywords: Agribusiness, Consumer/Household Economics, Food Consumption/Nutrition/Food Safety, Marketing, Research Methods/ Statistical Methods,
    Date: 2014
  12. By: Mideros A.; O'Donoghue C. (UNU-MERIT)
    Abstract: We examine the effect of unconditional cash transfers by a unitary discrete labour supply model. We argue that there is no negative income effect of social transfers in the case of poor adults because leisure could not be assumed to be a normal good under such conditions. Using data from the national employment survey of Ecuador ENEMDUR we estimate the effect of the Bono de Desarrollo Humano BDH. Results show that cash transfers, unconditional in labour, do not produce labour disincentives in the case of household heads, but may be paying for housework and childcare provided by partners and single adults. However, labour market and care work gender inequality must be addressed by complementary policies.
    Keywords: Personal Income and Other Nonbusiness Taxes and Subsidies; includes inheritance and gift taxes; Fiscal Policies and Behavior of Economic Agents: General; Time Allocation and Labor Supply; Microeconomic Analyses of Economic Development;
    JEL: H24 H30 J22 O12
    Date: 2014
  13. By: Aysenur Acar (Bahcesehir University Center for Economic and Social Research); Cem Baslevent (Istanbul Bilgi University)
    Abstract: Using a balanced panel drawn from Turkstat’s Survey of Income and Living Conditions (SILC), we aim to identify the main determinants of Turkish households’ entry into and exit from poverty. During the 4 year period (2007-2010) examined, the relative income poverty rate declined moderately, implying that households were more likely to exit than enter poverty. In addition to a descriptive analysis where poor, non-poor, entrant, and exitor households are compared in terms of basic household and household head characteristics, the empirical work involves the estimation of binary choice models that analyze the relative importance of these factors. Our models reveal that the employment status and schooling of the household head and household size are closely associated with poverty status changes. The probability of entry into poverty, for instance, is higher for larger households with many inactive/dependent members. However, model specifications that produce the best fit are the ones that take into account the changes in household composition and the amounts of income types received.
    Keywords: poverty transitions, relative income poverty, binary choice models
    Date: 2014–11
  14. By: Andreassen, Leif; Di Tommaso, Maria Laura; Strøm, Steinar (University of Turin)
    Abstract: We estimate a dynamic discrete choice model of Registered Nurses’ labor supply with random terms. A distinguished feature of our model is that random terms are correlated over time and jobs (habit persistence). Past options and not only the past optimal choices matter for the current choices. Given observed incentives, we find that nurses are mobile when they are young (less than physicians), but there is also a weak tendency of higher mobility again when they are approaching retirement age. Wage increases have a modest impact on labor supply. The overall elasticity for nurses are close to zero (like for physicians). These low elasticities shadow for stronger responses, shifting labor away from part time jobs in the public and private sector towards full time jobs in the private sector. A change in taxation away from the progressive tax system towards a flat tax of 28% gives Registered Nurses a very modest incentive to shift their job to private hospitals. For physicians the impact is stronger.
    Date: 2014–10
  15. By: Zack Dorner (Monash University); Dean Hyslop (Victoria University of Wellington)
    Abstract: Rural land use in New Zealand is an important driver of economic activity and has clear implications for the environment, including for biodiversity, climate change emissions and water quality. The spatial distribution and aggregate shares of rural land use is always changing, but change occurs slowly. To better understand the drivers of rural land use change, this paper addresses three questions using the popular multinomial logit modelling approach. First, do recent commodity prices have any predictive power on land use conversions? Second, is recently sold land more likely to change use? Third, does land which is marginal between uses have identifiable characteristics? The data used consists of the New Zealand Landcover Database version 3 (LCDB3), with observations in 1997, 2002 and 2008; 6 year average profitability data for dairy, sheep and beef and forestry; QVNZ land sales data; and I control for land quality and Maori tenure. In answering the first and third questions, I evaluate the predictive power of a spatially explicit land use share multinomial logit model, estimated from 2002 cross-sectional variation. To supplement the land use share multinomial logit for questions one and three, and to address question two, I use a land use transition multinomial logit, estimating the likelihood of transition from a single starting land use between 1997 and 2002, similar to Lubowski et al. (2008). Finally, I compare the two modelling approaches.
    Keywords: land use; rural; conversion; multinomial logit; land sales; marginal land
    JEL: Q15
    Date: 2014–11
  16. By: Christman, Laine; Kimberly, Rollins; Micheal, Taylor; Sohan, Agbonlahor
    Abstract: Wildfire threat is an intrinsic part of life in rural communities in the arid Western U.S. Western U.S. continues to experience a rapid population growth of Wildland-Urban Interface (WUI) areas, an increase in the frequency of wildfire due to increase human-caused ignition (Cardille et al. 2001), and overall fire severity due to over-accumulation of vegetation density on the surrounding public lands. Due to these trends, wildfire tends to have significantly higher economic cost associated with communities located in the WUI relative to wildfires on more remote public lands since WUI residences have an increased chance of damage (NIFC 2004). Fire protection services within these rural areas are scarce however; fire protection can be leveraged by wildfire risk mitigation actions done to both private properties and public lands adjacent to WUI areas. Private mitigation actions include creating a defensible space zone and use of fire resistant construction materials on the house, such as concrete siding or tiled roofing. Defensible space reduces the chance a wildfire will move from the public lands and damage or destroy a house by eliminating the combustible material and vegetation on the landscape within 30 feet of the home and promotes the use of fire resistant landscaping, i.e., well-irrigated lawns, deciduous shrubs and trees, xeriscaping, etc. Fire resistant construction materials reduce the chance a house will catch fire, given it ignites on the property. Public fuel reduction treatments remove strips of the native vegetation around the community creating a barrier between the public lands and the WUI which fire protection services can use to prevent fire from threatening homes. These public actions can be targeted to high priority areas within the community. Private and public provisions to reduce risk can impose considerable costs to residents and the individual risk of damage to a residence from a wildfire is rather low, therefore decisions about investment in these provisions can be seen as a gamble due to the uncertainty regarding the outcome. In one prospect, the resident can pay nothing and face uncertainty about a loss incurred from a wildfire. In another prospect, the resident can reduce the uncertainty by investing in risk mitigating provisions. Financial incentives for WUI residents in areas of elevated fire risk can be imposed in the form of policies and cost-sharing programs such as block grants. Such incentives can relieve some of the burden placed on residents and increase of overall level of defensible space thereby decreasing the level of risk facing WUI residents. Other incentives include educational components that help residents understand the risk they face. We explore how residents' value these provisions when the stakes are high but the chances of loss are low, as well as, how the value changes between public and private risk reduction provisions. Moreover, we explore the welfare loss given exogenous shifts in wildfire risk. Expected Utility Theory, a generally accepted and widely applied normative model of rational choice (Kahneman and Tversky 1979), would suggest rational individual will base investment decisions under uncertainty on comparison of expected utilities between the two prospects (Mongin 1997). The more risk adverse the individual, the more willing the individual is to forego a riskier situation for a more certain outcome. To investigate these issues, we measure the value of wildfire risk reduction to individual WUI residents using an experimental design based on a stated preference framework. We surveyed residents living in rural communities in Nevada that are representative of the different vegetation types, community designs, and population densities found within the Western U.S. We provide the homeowner two means to reduce the risk. The first method is investment in private defensible space on the property to lower the risk to the homeowner's property. The second method is investment is a community-wide fuel treatment program completed on public lands surrounding the community to lower the overall risk to the community and subsequently to the homeowners. In each case the homeowner was asked to state whether they would be attain those provisions at a given dollar amount. For the private provision the respondent was asked to pay a one-time cost. In the public provision, the cost was an annual fee. Within the survey instrument, the risk reduction from investment is explicitly stated to the homeowner. This reduction is explained as a decrease in the probability of damage to the house if provisions are taken. We employed a discrete choice, polychotomous, contingent valuation method that allowed the level of response intensity to be directly incorporated into the response option. The experiment design included 72 versions of the private provision and 24 versions of the public provision, which were randomly assigned to the respondents. For both provisions, there are multiple risk reduction categories that were randomly assigned to the respondents. Similarly, for both there were two year spans for which the risk reduction was valid, e.g, 5 or 10 years. Within the private provision there are three possible loss amounts ($50,000, $100,000, $200,000) to the homeowner, should a wildfire reach the home, which are also varied across respondents. Within the public provision, no loss amount was specified in order to view residents' value for reducing wildfire risk to the community in general. Our results indicate that willingness to pay (WTP) for private provisions of defensible space are higher than public provisions, i.e., $297 for the private provision and $47 per year for the public provision on average. When considering a 5 year time period the two WTP measures are relatively similar. When considering a 10 year period, the WTP for the public provision becomes greater however our models reveal respondents' WTP are not sensitive to the different time horizons. The similarity between the two provisions indicates that the incentives from private or public program would likely have similar outcomes in reducing risk, therefore it should be left to the policy makers to determine the priority each provision has to a given community. Our results for both provisions indicate that as risk reduction increases, so does WTP. For the private provision the average WTP for a 1% reduction is $147 and increases to $450 for a 4% reduction, while the average annual WTP for a 1% reduction is $36 and increases to $51 for a 4% reduction, suggesting people are considering the magnitude of the risk reduction when making decision about investment in fire mitigation provision across the board. From a policy perspective this would indicate people have a higher value for programs that strategically target areas where risk reduction is most effective. Furthermore, educational programs that inform residents about the how effective mitigation actions are from both private property and within the community as a whole can increase support of the provisions. Moreover, because the risk reduction is identical across the two provision but the expected value i.e., the reduction of risk and loss amount, was explicitly stated to the homeowner only for the private provision, the similarity between the two WTP estimates suggest that residents seem more concerned with the overall reduction in risk rather than the actual expected value gained from the provision. Furthermore, within the private provision our results reveal concavity in WTP over expected value which suggests the decreasing marginal value of investment. This finding is consistent with the EUT literature which assumes a diminishing marginal utility of wealth. This notion that marginal WTP incremental declines as expected value increases suggests incentives to increase private levels of defensible space only work up to a certain point. Beyond that, people begin to experience disutility with increased investment. Therefore, incentives to raise levels of defensible space in the most at-risk WUI communities must account for this lower marginal WTP and possible consider cost-sharing programs to offset this diminishing utility.
    Keywords: wildfire risk, wildland-urban interface, mitigation, contingent valuation, Community/Rural/Urban Development, Institutional and Behavioral Economics,
    Date: 2014

This nep-dcm issue is ©2014 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.