nep-dcm New Economics Papers
on Discrete Choice Models
Issue of 2016‒06‒09
sixteen papers chosen by
Edoardo Marcucci
Università degli studi Roma Tre

  1. Welfare Analysis for Climate Risk Reductions: Are Current Treatments of Outcome Uncertainty Sufficient? By Makriyannis, Christos; Johnston, Robert
  2. Systematic Non-Response in Stated Preference Choice Experiments: Implications for the Valuation of Climate Risk Reductions By Abdulrahman, Abdulallah S; Johnston, Robert J
  3. Soultion and Estimation of Dynamic Discrete Choice Structural Models Using Euler Equations By Victor Aguirregabiria; Arvind Magesan
  4. The Effects of Honesty Oath and Consequentiality in Choice Experiments By Kemper, Nathan; Nayga, Rodolfo M. Jr.; Popp, Jennie; Bazzani, Claudia
  5. Accommodating satisficing behavior in stated choice experiments By Sandorf, Erlend Dancke; Campbell, Danny
  6. Retail Channel and Beef Preferences in Argentina: Experimental Results from Consumers in Buenos Aires By Colella, Florencia; Ortega, David
  7. Interpretation issues in heteroscedastic conditional logit models By Burton, Michael; Davis, Katrina; Kragt, Marit Ellen
  8. The effect of front-of-pack nutrition labelling formats on consumers’ food choices and decision-making: merging discrete choice experiment with an eye tracking experiment By Erdem, Seda; McCarthy, Tony
  9. Consumers’ Willingness-to-pay for Health-enhancing Attributes in Food Products: A Meta-analysis By Dolgopolova, Irina; Teuber, Ramona
  10. Flexible Mixed Logit with Posterior Analysis: Exploring Willingness-to-Pay for Grid Resilience By Laura-Lucia Richter; Melvyn Weeks
  11. Farmers’ Preferences for Supermarket Contracts in Kenya By Ochieng, Dennis O.; Veettil, Prakashan C.; Qaim, Matin
  12. A Latent Class Analysis of Public Attitudes toward Water Resources with Implications for Recreational Demand By Ehrlich, Oren; Bi, Xiang; Borisova, Tatiana; Larkin, Sherry
  13. Consumers’ Willingness to Pay for Locally Grown Produce: Comparison of New Hampshire and Massachusetts Results By Shi, Wei; Halstead, John; Huang, Ju-Chin
  14. What is the Causal Impact of Knowledge on Preferences in Stated Preference Studies? By Jacob LaRiviere; Mikolaj Czajkowski; Nick Hanley; Katherine Simpson
  15. Water-based Recreation and Water Quality Indices: A Revealed Preference Approach By ji, yongjie; keiser, david
  16. Modelling of Bt Corn Production under Choice of Abatement Specification By Shi, Xiangdong; Hurley, Terrance

  1. By: Makriyannis, Christos; Johnston, Robert
    Abstract: Discrete choice experiments often include attributes subject to outcome uncertainty (OU), defined as uncertainty regarding actual attribute levels that will occur. Most choice experiments that incorporate OU do so using scenarios that allow for only two possible outcomes distinguished by a single probability. Because few environmental phenomena are characterized by two possible outcomes, characterizing scenarios in this way requires analysts to reframe actual conditions, discretizing the underlying continuous probability density function into two intervals. The implications of this reframing for welfare estimation are almost universally unknown. This article evaluates the convergent validity of welfare estimates from a more complex and accurate treatment of OU, compared to the traditional two-outcome approach. Methods and results are illustrated using an application to coastal flood adaptation in Connecticut, USA. Results show that a higher-resolution, multiple outcome treatment of OU provides additional information on risk preferences and willingness to pay (WTP), but also suggest that multiple outcome treatments increase task complexity. These tradeoffs highlight challenges facing the valuation of outcomes subject to OU.
    Keywords: Discrete choice experiments, outcome uncertainty, choice scenario, generalized multinomial logit model, climate change adaptation, Environmental Economics and Policy, Risk and Uncertainty,
    Date: 2016
  2. By: Abdulrahman, Abdulallah S; Johnston, Robert J
    Abstract: Discrete choice experiments (DCEs) addressing adaptation to climate-related risks may be subject to response biases associated with variations in risk exposure across sampled populations. Systematic adjustments for such biases are hindered by the absence of rigorous, standardized selection-correction models for multinomial DCEs, together with a lack of information on non-respondents. This paper illustrates a systematic approach to accommodate risk-related non-response bias in DCEs, where variations in risk exposure may be linked to observable landscape characteristics. The approach adapts reduced form response-propensity models to correct for survey non-response, capitalizing on the fact that indicators of risk exposure may be linked to the geocoded locations of respondents and non-respondents. An application to coastal flood adaptation in Connecticut, USA illustrates implications for welfare estimation. Results demonstrate that the proposed approach can reveal otherwise invisible, systematic effects of survey response patterns on estimated WTP.
    Keywords: Response bias, Response propensity, Coastal adaptation, Flood risk, Sea level rise., Environmental Economics and Policy,
    Date: 2016
  3. By: Victor Aguirregabiria; Arvind Magesan (University of Calgary)
    Abstract: This paper extends the Euler Equation (EE) representation of dynamic decision problems to a general class of discrete choice models and shows that the advantages of this approach apply not only to the estimation of structural parameters but also to the computation of a solution and to the evaluation of counterfactual experiments. We use a choice probabilities representation of the discrete decision problem to derive marginal conditions of optimality with the same features as the standard EEs in continuous decision problems. These EEs imply a Â…fixed point mapping in the space of conditional choice values, that we denote the Euler equation-value (EE-value) operator. We show that, in contrast to Euler equation operators in continuous decision models, this operator is a contraction. We present numerical examples that illustrate how solving the model by iterating in the EE-value mapping implies substantial computational savings relative to iterating in the Bellman equation (that requires a much larger number of iterations) or in the policy function (that involves a costly valuation step). We deÂ…fine a sample version of the EE-value operator and use it to construct a sequence of consistent estimators of the structural parameters, and to evaluate counterfactual experiments. The computational cost of evaluating this sample-based EE-value operator increases linearly with sample size, and provides an unbiased (in fiÂ…nite samples) and consistent estimator the counterfactual. As such there is no curse of dimensionality in the consistent estimation of the model and in the evaluation of counterfactual experiments. We illustrate the computational gains of our methods using several Monte Carlo experiments.
    Date: 2016–05–24
  4. By: Kemper, Nathan; Nayga, Rodolfo M. Jr.; Popp, Jennie; Bazzani, Claudia
    Abstract: Choice experiments are now one of the most popular stated preference methods used by economists. A highly documented limitation of stated preference methods is the formation of hypothetical bias in the estimation of consumers’ willingness-to-pay (WTP) for a good or a service. Honesty oaths and consequentiality scripts are two ex ante approaches that show promise in their ability to reduce or eliminate hypothetical bias. We examine these approaches independently and together and measure their effectiveness by comparing the resulting WTP values. We also explore a potential connection between consequentiality, honesty oaths, and attribute non-attendance (ANA). We infer patterns of ANA resulting from our various treatments (i.e., consequentiality script only, honesty oath only, combined script and oath, inconsequential, and control) and examine the differences. Our results suggest that the combined ex ante approach of consequentiality script and honesty oath provided significantly lower WTP values than all other experimental treatments. Conditioning our data for both consequentiality and ANA resulted in significant improvements in model fit across all treatments. Results indicate that not accounting for ANA has important implications for welfare estimates. While we cannot fully explain the connection, the combination of the consequentiality script, honesty oath, and inferred ANA allowed us to better see the differences between respondents’ attending attributes and those ignoring.
    Keywords: consequentiality, honesty oath, attribute non-attendance, choice experiment, Marketing, Research Methods/ Statistical Methods,
    Date: 2016
  5. By: Sandorf, Erlend Dancke; Campbell, Danny
    Abstract: Accumulating evidence suggests that many respondents in stated choice experiments use simplifying strategies and heuristics. Such behavior is a deviation from random utility theory and can lead to biased estimates if not appropriately considered. This paper is a first attempt to systematically explore the use of the satisficing heuristic (Simon, 1955) in the context of a stated choice experiment. We consider 944 possible satisficing rules and allow respondents to revise the rules adopted throughout the choice sequence. While only a small proportion of respondents used the same satisficing rule across the entire sequence, allowing for changes in behavior at different stages reveals evidence that the use of the heuristic follows a learning and fatigue path. Furthermore, considering respondents satisficing leads to improved model fits and different marginal willingness-to-pay estimates.
    Keywords: random utility maximization, satisficing, stated choice experiments., Environmental Economics and Policy, Food Consumption/Nutrition/Food Safety, Institutional and Behavioral Economics, Research Methods/ Statistical Methods,
    Date: 2016
  6. By: Colella, Florencia; Ortega, David
    Abstract: Traditionally-produced beef from Argentina is recognized and demanded internationally. Locally, consumers are often unable to afford these certified beef products, and may rely on external cues to determine beef quality. Are Argentinean consumers willing to pay for sustainably or organically produced beef? Are these consumers using alternative sources of information, other than product labels? Determining demand for beef attributes may require an understanding of consumers’ product purchasing strategies, which involves retailer choice. We develop a framework utilizing latent class analysis to identify consumer groups with different retailer preferences, and separately estimate their demand for beef product attributes. This framework accounts for the interrelationship between consumers’ choice of retail outlets and beef product preferences. Our analysis identifies two groups of consumers, a convenience- (67%) and a service- (33%) oriented group. We uncover significant differences in demand for beef attributes across these groups, and find that the service oriented group, while unable to pay for credence attributes, relies on a service-providing retailer (butcher) as a source of product quality assurance. We note that failing to account for the interrelationship between retailer and product preferences can lead to an overestimation of preferences and demand for beef attributes.
    Keywords: Latent class analysis, demand heterogeneity, random parameters logit, consumer behavior, butcher, decision modelling, choice experiment, Agribusiness, Consumer/Household Economics, Livestock Production/Industries, Marketing,
    Date: 2016
  7. By: Burton, Michael; Davis, Katrina; Kragt, Marit Ellen
    Abstract: This paper identifies an issue with interpretation of significance within heteroscedastic conditional logit models, due to sensitivity of reported results to arbitrary variable-normalization decisions. We advocate the use of willingness-to-pay (WTP) space models to avoid this, as estimates of WTP do not exhibit this effect. However, in cases where error variance may be high, we question whether it is correct to infer that all respondents should be considered to hold a common WTP.
    Keywords: Scale, Willingness to Pay, Error Variance, Discrete Choice Models, Aggregate Welfare, Part-worths, Environmental Economics and Policy, Research Methods/ Statistical Methods, Resource /Energy Economics and Policy, C10, C18, C51, Q51,
    Date: 2016–05–15
  8. By: Erdem, Seda; McCarthy, Tony
    Keywords: Food Consumption/Nutrition/Food Safety, Marketing,
    Date: 2016–05
  9. By: Dolgopolova, Irina; Teuber, Ramona
    Abstract: Assessing potential demand for functional or healthy foods is crucial from several perspectives. First, foods with functional attributes in many cases require more expensive production process than traditional foods, for example, when the functional attribute is provided by enhancing or enriching the products with additional substances. It is necessary, then, to estimate potential demand for functional foods prior to the delivering product to the consumers. Second, the promotion of healthier food options is related to the fact that an unhealthy diet is among the four main behavioral risk factors of non-communicable diseases (NCD) and are mostly spread in low- and middle-income countries. It has been demonstrated that prices can be a barrier for healthy food consumption, especially among low-income groups of the population. From this perspective, it needs to be clearly stated if consumers indeed are ready to pay price premiums for foods aimed at improving their health. Third, market introduction of functional foods and foods with health benefits has not always been successful. Despite the importance of a healthy diet in the prevention of some diseases and sustaining well-being in general, economists and marketing researchers observe some uncertainty in consumers’ perception and acceptance of foods with health benefits. This paper investigates the body of research that has been performed so far on consumers’ valuations of healthy attributes in food products by means of a meta-analysis. It explores if variation in willingness to pay (WTP) for healthy attributes in foods that have been reported in scientific papers on the topic can be attributed to common factors related to the choice of the methodology, the place and time of data collection, the choice of the carrier product and the health benefit specified. Thus, our study contributes to the existing literature on health-enhancing foods by (i) reviewing the existing empirical evidence on consumer valuations of different healthy attributes, (ii) identifying the major underlying drivers of differences in WTP estimates via meta-analysis and (iii) deriving directions of research to be taken into account for the future developments in the field. Literature search resulted in 28 studies which provided 175 WTP estimates. Stata meta-regression command specifically designed for meta-analyses was employed. This command allows analyzing study-level data and estimates the between-study variance and the coefficients by weighted least squares when the outcome variable is continuous. The results of the meta-regression imply that the elicitation method, the carrier product, the specific health benefit, and the place of the study significantly influence variations in WTP estimates across studies. First, hypothetical methods of willingness to pay elicitation produce higher valuations compared to non-hypothetical methods like experimental auction and real purchase data. Second, with respect to the base product the results indicate that in case of dairy products (milk, yogurt, cream cheese, cheese, butter and ice cream) and fruits and vegetables the WTP estimates for a specific health attribute are significantly lower than for all other product categories included. Third, according to our results the specific health attribute “Cholesterol lowering” leads to significantly higher WTP estimates than any other health/nutrition claim. The valuations of this attribute varied from 0% to 200% with the highest values referring to the spread for lowering cholesterol. Finally, the place where the data was collected influences WTP estimates. Our results indicate that there are no significant differences between studies conducted in Europe, the United States and other regions. However, studies conducted with Canadian consumers report statistically significant lower WTP values. In general, it can be noticed that despite an established connection between diet and the development of non-communicable diseases, economics and marketing research so far fails to provide systematic view on the consumer valuations of different healthy attributes in food and, consequently, on the perspective demand for these products. Studies reviewed reported very different valuations of healthy attributes in foods. Studies also differ greatly in basically all parameters of the research: data collection, methodology, and analysis of the results. Although it seems rather difficult to draw general conclusions about consumers’ willingness to pay for healthy attributes in foods, this research summarizes the efforts performed so far and may be employed to determine the directions for future analysis.
    Keywords: health-enhancing attributes, meta-analysis, willingness to pay, Agricultural and Food Policy, Food Consumption/Nutrition/Food Safety, Q18 Agricultural Policy • Food Policy,
    Date: 2016
  10. By: Laura-Lucia Richter; Melvyn Weeks
    Abstract: This paper presents and employs an alternative approach to explore consumer preferences and willingness-to-pay (WTP) for resilience of the electricity grid. The methodological and practical relevance of this approach is demonstrated using the example of the UKs incentive regulation scheme for electricity distribution network operators (DNOs). The estimation strategy exibly accounts for preference heterogeneity in the population, allows for scale heterogeneity (i.e. heterogeneity in the randomness of choice) and exploits individual posterior distributions to improve the estimates. Since the results suggest significant parameter heterogeneity within and across DNOs, it is argued that Ofgem's current method to evaluate consumer preferences and WTP is likely to result in an inefficient level of resilience services. The welfare implications in case of public and private goods and services are discussed. The suggested approach is straightforward to implement, could improve policy evaluations and foster more nuanced and efficient incentive regulation.
    Keywords: Willingness-to-Pay, electricity service quality, mixed logit, social welfare
    JEL: L98
    Date: 2016–05–10
  11. By: Ochieng, Dennis O.; Veettil, Prakashan C.; Qaim, Matin
    Abstract: Empirical analysis of contractual arrangements between supermarkets and smallholder farmers remains scarce, yet farmers’ contract preferences influence their participation in supermarket contracts. We employ mixed logit model to analyze farmers’ preferences for contracts using discrete choice data from a sample of vegetable farmers of central Kenya, sampled through stratified random sampling procedure. Results show that farmers generally do not exhibit risk aversion to contracts and would choose them depending on their attributes. Certain farmer characteristics influence decision to contract and preferences for contract design attributes. Findings also show that group marketing could be an interesting option to reduce individual risks and transaction costs. Designing contracts that lower risks to smallholder farmers, and or with transparent risk-sharing clauses would enhance their participation in supermarket contracts. Some wider policy implications are discussed.
    Keywords: supermarkets, contracts, farmers’ preferences, choice experiment, Kenya, Agribusiness, Agricultural and Food Policy, Food Security and Poverty, Institutional and Behavioral Economics,
    Date: 2016
  12. By: Ehrlich, Oren; Bi, Xiang; Borisova, Tatiana; Larkin, Sherry
    Abstract: This study examines the extent to which heterogeneous environmental attitudes influence recreational demand in a river basin and the valuation of recreational benefits. We first employed a latent class analysis to reveal two distinct classes of respondents that differ in their environmental attitudes despite representing similar demographic characteristics. We then estimated a recreational demand model conditional on respondent’s latent class membership after controlling for the probabilistic nature of the membership classification. We found that environmental attitudes directly influence consumer recreational demand and valuation. Ignoring preference heterogeneity leads to overestimation of the recreational benefits.
    Keywords: Latent Class Analysis (LCA), Recreational Demand, Travel Cost Method (TCM), Non-market Valuation., Environmental Economics and Policy, Resource /Energy Economics and Policy, Q51, Q57,
    Date: 2016–07–31
  13. By: Shi, Wei; Halstead, John; Huang, Ju-Chin
    Abstract: The increasing interest in locally grown produce in the U.S. has resulted in a number of studies examining consumers’ willingness to pay for local specialty food. This paper extends the literature to investigate Massachusetts and New Hampshire survey respondents’ preferences for locally grown and other attributes of a variety of produce types. Choice experiments are used to discern the relative importance of these produce attributes. Our results show that the average premiums for locally grown green beans, cucumbers and snap peas are respectively 30.74 percent, 67.30 percent, and 32.62 percent above the prices of the non-locally grown counterparts among New Hampshire respondents. In comparison, the average premiums for locally grown green beans, cucumbers and snap peas are respectively 57.66 percent, 17.31 percent (insignificant), and 35.45 percent above the prices of the non-locally grown counterparts among Massachusetts respondents. We also find mixed results on the willingness to pay for the organic feature across different produce. Consumers are willing to pay a price premium for organically grown green beans (about 29.02 percent in New Hampshire and 32.63 percent in Massachusetts), but none for snap peas.
    Keywords: Local Agriculture, Willingness to Pay, Choice Experiments, Agribusiness, Agricultural and Food Policy, Consumer/Household Economics,
    Date: 2016
  14. By: Jacob LaRiviere (Department of Economics, University of Tennessee); Mikolaj Czajkowski (Faculty of Economic Sciences, University of Warsaw); Nick Hanley (Department of Geography and Sustainable Development, University of St. Andrews); Katherine Simpson (Economics Division, University of Stirling, Scotland)
    Abstract: This paper reports the results of a stated preference experiment designed to test for how information provided in a survey affects knowledge, and how knowledge affects preferences for a public good. A novel experimental design allows us to elicit subjects’ ex ante knowledge levels about a good’s attributes, exogenously vary how much new objective information about these attributes we provide to subjects, elicit subjects’ valuation for the good, and elicit posterior knowledge states about the same attributes. We find evidence of incomplete learning and fatigue: as subjects are told more information, their marginal learning rates decrease. We find there is no marginal impact of knowledge on the mean nor the variance of WTP for changes in the environmental good; but that ex ante knowledge does affect stated WTP. Our results are consistent with preference formation models of confirmation bias, costly search, or timing differences in learning and preference formation. Our results raise questions about the purpose and effects of providing information in stated preference studies.
    Keywords: Learning, Information, Behavioral Economics, Decision Making Under Uncertainty
    JEL: D83 D81 Q51
    Date: 2016
  15. By: ji, yongjie; keiser, david
    Abstract: Bene t estimates of water pollution control policies rely heavily on water quality in- dices. Since the 1970s, these measures of water quality have been used extensively in stated preference surveys to estimate willingness to pay for water quality that is suitable for recre- ational use. However, there is little empirical evidence of how well these indices correspond to observed recreational behavior. This paper utilizes a unique micro-dataset of individual household recreational use and water quality in a revealed preference framework to explore how well several major water quality indices explain water-based recreational use.
    Keywords: Water Recreation, Water Quality Indices, Environmental Economics and Policy, Resource /Energy Economics and Policy, Q51, Q53, Q57,
    Date: 2016–05–25
  16. By: Shi, Xiangdong; Hurley, Terrance
    Abstract: For modelling Bt corn production, we proposed a joint abatement function to capture pest life cycle. Simulation of pest adaptation to popular Bt corn varieties was developed and applied to generate random sample of abatement, where in-field parameters were adopted in favor of practical corn production. On the basis, logistic model was used to fit the sample data. The estimates were compared to those from exponential model for choice of abatement specification. At last, Cobb-Douglas production model was integrated with the pre-estimated abatement function along with farm level data and distribution of corn growers’ choices, where instrument variables estimator and delta method were adopted to solve simultaneity problem induced by farm level data. The estimated marginal net return was positive, suggesting that Bt corn as pesticide is underutilized. However, the value kept decline in recent three years. We attributed the decline of marginal net return to the widespread adoption of new Bt corn varieties with blended refuge, and the insect resistance management would benefit from it.
    Keywords: Bt corn, production, abatement, simultaneity problem, Crop Production/Industries, Environmental Economics and Policy, Production Economics, Productivity Analysis, Q16, C23,
    Date: 2016–05–25

This nep-dcm issue is ©2016 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.