Unobserved Preference Research Articles

The Reflection Effect and Fourfold Pattern of Risk Attitudes: A Structural Econometric Analysis

The reflection effect and fourfold pattern of risk attitudes are often thought to be stylized facts about decision making under risk in lab settings. We take a closer look at this received wisdom, using 5 existing data sets from pairwise lottery choice experiments with real monetary incentives. We apply a structural econometric model that allows for within-individual correlation in risk preferences across gain and loss domains, as well as unobserved preference heterogeneity across individuals. The estimates of utility and probability weighting functions do not support the received wisdom.

Why Do Previous Choices Matter for Hospital Demand? Decomposing Switching Costs from Unobserved Preferences

Using data on women’s choice of hospital for childbirth in Florida, we find that women return to the same hospital approximately 70% of the time. We separate explanations of switching costs and unobserved preference heterogeneity using a panel data fixed effects estimator and find that switching costs account for approximately 40% of the demand effects of a lagged dependent variable. The welfare effects of excluding a hospital from a payer’s network are smaller in the short run but higher in the long run, given our estimates of switching costs, and the dynamic effects of entry on competition are significantly smaller.

Travel Data Sequence from Multi-Source Recommendation System

Due to different sort of preferences and restrictions of a trip such as time source limitation and every tourist’s destination points the travel based recommendation has become a challenging task. Most importantly the data generated by the geo-tagged social channel from the geo based tag tweets, snapshots of credentials. Due to examining this, extended data allows us to invent the profiles, daily mobility patterns, and results of the user’s. To resolve the issues and challenges of capacity providing their personalized and sequential travel to make package recommendation to a topical package model and to take using social media info in which mechanically mine person travel interest with another quality like time, cost, and period of wayfaring. Here, we had a proposal that a travel data sequence after a multi source recommendation system. We implemented a location recommendation system that derives personal preferences while accounting for restraints irremissibly by road capacity in order to change the demand of travel. We first infer unobserved preferences using a machine learning technique from data mining records. It extends our method to provide personalized suggestions based on user geo co-ordinates points. By utilizing the tree based hierarchal graphs (TBHG), location histories of the multiple users’ have been modeled. In order to collect the selected places interest level and travel knowledge of user’s, the HITS model had developed based on TBHG. Finally, hybrid filtering approach based on HITS is utilized to get the global positioning system (GPS) based personalized recommendation system. And for image based search similar images with the tag information are retrieved for the query image users.

Public transport: One mode or several?

This paper develops a methodology for testing and implementing differences in preferences for a set of public transport modes, relating to observed and unobserved attributes, in state-of-practice large-scale travel demand models. Results of a case study for commuters in the Stockholm public transport system suggest that there are preference differences among public transport modes. We found that the value of time for train is lower than for bus and metro, and that it is higher for auxiliary modes than for the main mode. Surprisingly, we found no evidence for differences proportional to the in-vehicle time between bus and metro, suggesting that characteristics of in-vehicle time in these two modes are valued equally by the travellers. Nevertheless, unobserved preference for metro is higher than the preference for bus. Regarding the existence of a rail factor, we find evidence to support the hypothesis that rail-based modes have in fact a smaller time parameter (train) or higher alternative specific constant (metro), indicating that rail modes are preferable to bus, ceteris paribus.

A comparison of regret-based and utility-based discrete choice modelling – an empirical illustration with hospital bed choice

ABSTRACTThere is some concern that the unobserved preference heterogeneity in random utility maximization theory-based discrete choice experiment modelling is an important source of error variability. The randomness in utility is often interpreted as interpersonal preference heterogeneity but it can also be intrapersonal random variation in preferences. We compare utility maximization and regret minimization-based choice models’ sensitivity to individual heterogeneity, examine differences between two consecrated models and validate with empirical illustrations. We use frequency of category (public, semi-private, and private) of bed chosen from Swiss cross-sectional datasets (2007–2012) to compare two approaches – utility maximization and regret minimization by applying multinomial logit (MNL) models in regard to the variances in utility (regret) function, goodness-of-fit and predicted marginal effects (pseudo-elasticity) of additional payment. We find parameters with the same sign and estimates with almost same order of magnitude in both the approaches. The statistical significance of attribute effects is consistent in all variants of utility -based MNL models while effects of different attributes are significant only in heteroskedastic extreme value (HEV) variant of regret-based MNL models. This empirical illustration suggests that HEV variant of regret-based models perform better in capturing attribute effects in choice behaviour.

Public preferences for improved urban waste management: a choice experiment

AbstractA discrete choice experiment, aiming to elicit public preferences for improvements in solid waste services, is carefully administered across socioeconomic zones in the city of Hawassa, Ethiopia. Observed and unobserved preference heterogeneity are analyzed using mixed logit choice models. The results show that there exists substantial willingness to pay to increase collection frequency and separate recyclable waste. A new issue is the focus on child labor in the waste management sector. Significant gender effects are found: women are more interested than men in increasing waste collection frequency and value the abolishment of child labor more highly, as do higher income households. As expected, respondents living in wealthier neighborhoods are more likely to pay higher service charges. Education indirectly influences preferences for waste separation. The study provides important insight into the social benefits of public investment decisions to improve the quality of solid waste management services in large cities in Ethiopia.

Comparison of parametric and semiparametric representations of unobserved preference heterogeneity in logit models

The logit-mixed logit (LML) model is a very recent advancement in semiparametric discrete choice models. LML represents the mixing distribution of a logit kernel as a sieve function (polynomials, step functions, and splines, among many other variants). In the first part of this paper, we conduct Monte-Carlo studies to analyze the number of required parameters (e.g., polynomial order) in three LML variants to recover the true population distributions, and also compare the performance (in terms of accuracy, precision, estimation time, and model fit) of LML and a mixed multinomial logit with normal heterogeneity (MMNL-N). Our results indicate that adding too many parameters in LML may not be the best strategy to retrieve underlying taste heterogeneity; in fact, overspecified models generally perform worst in terms of BIC. We recommend to use neither minimum-BIC nor the most flexible specification, but we rather suggest to start with the same number of parameters as a parametric model (such as MMNL-N) while checking changes in the derived histogram of the mixing distribution. As expected, LML was able to recover bimodal-normal, lognormal, and uniform distributions much better than the misspecified MMNL-N. Computational efficiency makes LML advantageous in the process of searching for the final specification. In the second part of the paper, we estimate the willingness-to-pay (WTP) estimates of German consumers for different vehicle attributes when making alternative-fuel-car purchase choices. LML was able to capture the bimodal nature of WTP for vehicle attributes, which was not possible to retrieve using standard parametric specifications.

URBAN TRANSPORTATION MODE CHOICE AND TRIP COMPLEXITY: BICYCLISTS STICK TO THEIR GEARS

Commuters' lives have become more complicated with rising income. In my model, transportation mode choices are made simultaneously with the choice of whether to make multiple stops. Using travel behavior data, I estimate the model using an error components logit (ECL) specification to account for commuters' unobserved preferences for particular modes and find that omitting unobserved preferences underestimates value of travel time relative to the crossing‐components ECL. The estimated model predicts that increased trip complexity causes substitution away from public transit to automobiles, with the exception that bicyclists transition only to more complex trips and do not change transportation modes. (JEL R41, C25)

The Effects of the Timing of Childbirth on Female Labour Supply: An Analysis Using the Sequential Matching Approach

In this study, we estimate the effects of childbirth on female labour supply by using Japanese data. The novel contributions of our study are twofold. Firstly, we include the effects of unobserved preferences on female labour supply. Secondly, we apply a dynamic version of the sequential matching approach to analyse the causal effects of childbirth on female labour market outcomes. The estimated results show that childbirth decreases current employment outcomes (participation in regular and non-regular work) and that this decrease is larger for regular employees than for non-regular employees. On the timing of childbirth, while the negative effects of childbirth on regular work increase by delaying the age at childbirth, these negative effects on non-regular employment slightly decrease by delaying the age at childbirth. On future employment outcomes, childbirth does not affect the probability of choosing non-regular work in the next period regardless of childbearing age. By contrast, delayed childbirth decreases the probability of choosing regular work in the next period significantly.

Understanding How Angler Characteristics and Context Influence Angler Preferences for Fishing Sites

AbstractUnderstanding angler heterogeneity is critical for fisheries managers to be able to develop management approaches that minimize stress to fish and aquatic ecosystems while maximizing benefits to anglers. We used a stated‐preference choice model of British Columbia, Canada, anglers to predict their behavioral intentions to fish at lakes described by catch‐related and non‐catch‐related attributes, such as expected catch and travel distance. We investigated how different means of accounting for preference heterogeneity and decision context affected conclusions about angler preferences for fishing sites. Our preferred model of fishing site choice accounted for both observable and unobservable (to the researcher) preference heterogeneity for site attributes, angler characteristics (i.e., recreation specialization and residence), and context (i.e., trip duration and target species). On average, anglers’ preferences conformed to expectations, but preferences for different site attributes varied greatly among anglers and contexts. For example, highly specialized anglers were more influenced by catch rates, fish size, and bag limits and were less deterred by travel than were less‐specialized anglers. Anglers facing multiple‐day trip contexts were influenced more by fish size and bag limits than were anglers considering day fishing trips. This latter result is especially important, as researchers often estimate models of angler behaviors by excluding anglers that undertake multiple‐day trips, who might be the most sensitive to changes in regulations and fishing quality characteristics.Received June 19, 2017; accepted September 18, 2017Published online November 10, 2017

Unobserved Preference Heterogeneity in Demand Using Generalized Random Coefficients

We model unobserved preference heterogeneity in demand systems via random Barten scales in utility functions. These Barten scales appear as random coefficients multiplying prices in demand functions. Consumer demands are nonlinear in prices and may have unknown functional structure. We therefore prove identification of Generalized Random Coefficients models, defined as nonparametric regressions where each regressor is multiplied by an unobserved random coefficient having an unknown distribution. Using Canadian data, we estimate energy demand functions with and without random coefficient Barten scales. We find that not accounting for this unobserved preference heterogeneity substantially biases estimated consumer-surplus costs of an energy tax. JEL codes: C14 D12 D13 C21 Keywords: unobserved heterogeneity, nonseparable errors, random utility parameters, random coefficients, equivalence scales, consumer surplus, welfare calculations. The authors wish to thank Jinyong Hahn, Richard Blundell, Stefan Hoderlein, Rosa Matzkin, and anonymous referees for helpful discussions and suggestions. Earlier versions of this paper were circulated under the

Urban Transportation Mode Choice and Trip Complexity: Bicyclists Stick to Their Gears

The benefits of public policies designed to alleviate traffic congestion, reduce pollution, and improve public health rest heavily on the value that commuters place on time and how they make transportation mode decisions. However, over the last several years, commuter lives have become more complicated with rising income. This has impacted transportation decisions. In my model, transportation mode choices are made simultaneously with the choice of whether or not to make multiple stops (termed travel complexity). Unlike other studies, I account for commuters' unobserved preferences for particular transportation modes. Using travel behavior data collected by the Oregon Department of Transportation, I estimate the model using an error components logit (ECL) specification and find that not allowing for unobserved commuter preferences for particular transportation modes underestimates value of travel time (VOT) and over-predicts complex trip choice relative to the crossing-components ECL. The estimated model predicts that increased trip complexity causes substitution away from public transit to automobiles, with the exception that bicyclists transition only to more complex trips and do not change transportation modes.

Comparison of Parametric and Seminonparametric Representations of Unobserved Taste Heterogeneity in Discrete Choice

The recently derived Logit-Mixed Logit (LML) model captures unobserved preference heterogeneity seminonparametrically as a function of polynomials, step functions, and splines. This study contributes twofold in exploring and extending the usefulness of the LML specification. First, we conduct a Monte-Carlo study to analyze the number of required LML parameters to recover the true distributions, and also compare LML performance --in terms of accuracy, precision, estimation time, and model fit-- with a Mixed Multinomial Logit specification with Normal heterogeneity (MMNL-N). As expected, LML is able to retrieve the underlying Bimodal-Normal (mixture of two Normal distributions), Lognormal, and Uniform distributions much better than the MMNL-N model. In an empirical case study, we also estimate the willingness to pay (WTP) of German consumers for different vehicle attributes when making choices for alternative-fuel cars. LML is able to capture the bimodal nature of WTP for vehicle attributes, which was not possible to retrieve using the MMNL-N model. LML also appears to be more useful for panel data because the computation time is not affected by the number of choice situations. Second, whereas the original LML formulation assumes all utility parameters to be random, we extend the model to a combination of fixed and random parameters (LML-FR). We further show that the gradient of the LML-FR likelihood loses its convenient properties, leading to a much higher estimation time than that of the original LML specification. In an empirical case study about preferences for alternative-fuel vehicles in China, estimation time increased by a factor of 15-20 when introducing fixed parameters to the LML model. Thus, LML is computationally efficient and better than MMNL-N in retrieving the true distribution of the random taste heterogeneity, but loses its computational efficiency feature if any parameter is assumed to be fixed.

Asymmetric triangular mixing densities for mixed logit models

A novel method is proposed to estimate random parameter logit models using the asymmetric triangular distribution to describe unobserved preference heterogeneity in the population of interest. The asymmetric triangular mixing density has the potential to overcome behavioural limitations associated with the most frequently applied mixing densities like the normal and log-normal distribution. With only three parameters it remains parsimonious whilst its bounded support can easily be brought in line with behavioural intuitions. The triangular mixing density is not associated with an incredibly large upper (or lower) bound and it can accommodate varying degrees of skewness in unobserved preference heterogeneity. The proposed estimation procedure is based on the principle of mixture densities and circumvents additional simulation chatter arising when applying the inverse cumulative density function method to generate draws from the mixing density.

Disability benefit generosity and labor force withdrawal

A key component for estimating the optimal size and structure of disability insurance (DI) programs is the elasticity of DI claiming with respect to benefit generosity. Yet, in many countries, including the United States, all workers face identical benefit schedules, which are a function of one's labor market history, making it difficult to separate the effect of the benefit level from the effect of unobserved preferences for work on individuals' claiming decisions. To circumvent this problem, we exploit exogenous variation in DI benefits in Austria arising from several reforms to its DI and old age pension system in the 1990s and 2000s. We use comprehensive administrative social security records data on the universe of Austrian workers to compute benefit levels under six different regimes, allowing us to identify and precisely estimate the elasticity of DI claiming with respect to benefit generosity. We find that, over this time period, a 1% increase in potential DI benefits was associated with a 1.2% increase in DI claiming.

Same concerns, same responses? A Bayesian quantile regression analysis of the determinants for supporting nuclear power generation in Japan

Using internet survey data from 6,500 individuals, this study examines the determinants for supporting the restart of nuclear power plants operation in Japan. As in previous studies, the variable of interest is a categorical and ordered variable that measures the level of support, for which ordered logit or ordered probit is commonly estimated. This study departs from the literature by using Bayesian ordinal quantile regression recently proposed by Rahman (2015). This approach allowed us to explore whether covariates have differential effects at various conditional quantiles of the latent response variable, which can be interpreted as the willingness to support the restart. The results show that for most of the covariates we examined, including concerns about meltdowns and the storage of nuclear material and concerns about global warming, the effects differ across conditional quantiles. In other words, the covariate effects depend on individuals’ unobserved preferences for the restart (conditional on observables). The results also show that for some covariates, the effects differ considerably across gender.

Variance-component-based nested logit specifications: Improved formulation, and practical microsimulation of random disturbance terms

The initial motivation leading to the results in this paper was a problem most choice modeling researchers may have not considered: how to simulate random disturbance terms from nested logit (NL) models. We develop an approach using results from Cardell (1997), who proved the existence of a probability distribution (C(λ)) that could be used to formulate NL models based on statistically independent variance components. These components can be interpreted as unobserved preference heterogeneity for the choice ‘dimensions’ used to define NL tree structures. Simulation aside, we consider this formulation to have other practical advantages for empirical work, but it does not appear to have penetrated the literature (possibly due to notational obstacles). We use notation from Daly (2001) to implement an equivalent representation, which also establishes mathematical equivalence between Cardell (1997) and other important results in the NL literature.

Effect of Smoking and Other Economic Variables on Wages in the Euro Area

The smoking effects on wages has been examined in this work using different econometric methodologies with the use of European Community Household Panel (ECHP) datasets. We employ econometric tools like Instrumental Variable technique, Heckman correction factor, Endogenous Switching and Matching estimates. The initial results from regression estimates (OLS and IV methods) revealed that the wage gap between smokers and non-smokers ranges between 1% to 22.7%. Moreover, endogenous switching and matching estimator also showed a negative average treatment effects of approximately 47% and 4.3% to 6.9% respectively. Thus smokers observed less wage effects are explained in part by real effects on their health status and a measure of unobserved preferences.

Marital Matching, Economies of Scale and Intrahousehold Allocations

We propose a novel structural method to empirically identify economies of scale in household consumption. We assume collective households with consumption technologies that define the public and private nature of expenditures through Barten scales. Our method recovers the technology by solely exploiting preference information revealed by households' consumption behavior. The method imposes no parametric structure on household decision processes, accounts for unobserved preference heterogeneity across individuals in different households, and requires only a single consumption observation per household. Our main identifying assumption is that the observed marital matchings are stable. We apply our method to data drawn from the US Panel Study of Income Dynamics (PSID), for which we assume that similar households (in terms of observed characteristics like age or region of residence) operate on the same marriage market and are characterized by a homogeneous consumption technology. This application shows that our method yields informative results on the nature of scale economies and intrahousehold allocation patterns. In addition, it allows us to define individual compensation schemes required to preserve the same consumption level in case of marriage dissolution or spousal death.

A Dynamic Model of Demand for Houses and Neighborhoods

This paper develops a dynamic model of neighborhood choice along with a computationally light multi-step estimator. The proposed empirical framework captures observed and unobserved preference heterogeneity across households and locations in a flexible way. We estimate the model using a newly assembled data set that matches demographic information from mortgage applications to the universe of housing transactions in the San Francisco Bay Area from 1994 to 2004. The results provide the first estimates of the marginal willingness to pay for several non-marketed amenities—neighborhood air pollution, violent crime, and racial composition—in a dynamic framework. Comparing these estimates with those from a static version of the model highlights several important biases that arise when dynamic considerations are ignored.