Unobserved Preference Research Articles

The Role of Intermediaries in Selection Markets: Evidence from Mortgage Lending

Abstract We study the role of brokers in selection markets. We find broker-clients in the Canadian mortgage market are observationally different from branch-clients. They finance larger loans with more leverage and longer amortization. We build and estimate a model of mortgage demand to disentangle three possible explanations for these riskier product choices: (1) selection on observables, (2) unobserved borrower preferences for riskier loans, and (3) a causal effect of brokers. Although we find that brokers influence product choices, the main reason borrowers choose high-leverage products is unobserved preferences. Borrowers prefer larger loans and brokers facilitate qualification for them. (JEL G21, D12)

Growing Attention

We model a boundedly rational agent who exhibits limited and heterogeneous attention. The agents' attention can be naturally sorted in an ascending order. We provide a characterization result using two simple behavioral postulates. Unlike earlier models, consideration sets' composition and frequency are uniquely identified. Within our framework, we accommodate unobserved heterogeneity in attention and unobserved preferences. Furthermore, we explore various extensions of our model and provide insights into the comparative statics of attentiveness levels in two probabilistic choices.

Augmented degree correction for bipartite networks with applications to recommender systems

In recommender systems, users rate items, and are subsequently served other product recommendations based on these ratings. Even though users usually rate a tiny percentage of the available items, the system tries to estimate unobserved preferences by finding similarities across users and across items. In this work, we treat the observed ratings data as partially observed, dense, weighted, bipartite networks. For a class of systems without outside information, we adapt an approach developed for dense, weighted networks to account for unobserved edges and the bipartite nature of the problem. The approach begins with clustering both users and items into communities, and locally estimates the patterns of ratings within each subnetwork induced by restricting attention to one community of users and one community of items community. The local fitting procedure relies on estimating local sociability parameters for every user and item, and selecting the function that determines the degree correction contours which best models the underlying data. We compare the performance of our proposed approach to existing methods on a simulated data set, as well as on a data set of joke ratings, examining model performance in both cases at differing levels of sparsity. On the joke ratings data set, our proposed model performs better than existing alternatives in relatively sparse settings, though other approaches achieve better results when more data is available. Collectively, the results indicate that despite struggling to pick up subtler signals, the proposed approach’s recovery of large scale, coarse patterns may still be useful in practical settings where high sparsity is typical.

Identifying Heterogeneous Willingness to Pay for New Energy Vehicles Attributes: A Discrete Choice Experiment in China

New energy vehicles (NEVs) have emerged as a promising solution to reduce carbon emissions and address environmental concerns in the transportation sector. In order to effectively accelerate market acceptance, it is crucial to prioritize the heterogeneity of consumer preferences for NEV attributes. This study employs the multinomial logit model (MNL) and latent class model (LCM) to investigate both observed and unobserved preference heterogeneity based on stated preferences obtained from a discrete choice experiment conducted across seven cities in China. Results from the MNL model indicate that all attributes significantly influence alternative utility. In particular, there are differences in the willingness to pay (WTP) for attributes of battery electric vehicles (BEVs) and plug-in hybrid electric vehicles (PHEVs). Analysis of MNL subgroups reveals observed heterogeneity in WTP for identical attributes among consumers from regions with different latitudes and markets with different NEV penetration rates. Furthermore, the LCM model uncovers unobserved preference heterogeneity by classifying respondents into four distinct classes and identifies specific socioeconomic variables associated with each class. The recognition of heterogeneous WTP for NEV attributes across vehicle types, regions, markets, and consumer classes provides important implications for formulating targeted policies that promote the sustainable development of the NEV industry.

A maximum likelihood bunching estimator of the elasticity of taxable income

SummaryThis paper develops a maximum likelihood (ML) bunching estimator of the elasticity of taxable income (ETI). Our structural approach provides a natural framework to simultaneously account for unobserved preference heterogeneity and optimization errors and for measuring their relative importance. We characterize the conditions under which the parameters of the model are identified and show that the ML estimator performs well in terms of bias and precision. The paper also contains an empirical application using Swedish data, showing that both the ETI and the standard deviation of the optimization friction are precisely estimated, albeit relatively small.

Estimating the economic value of hydropeaking externalities in regulated rivers

Hydropower is a flexible form of electricity generation providing both baseload and balancing power to accommodate intermittent renewables in the energy mix. However, hydropower also generates various externalities. This study investigates individuals' preferences for policies aiming to reduce short-term regulations (i.e., hydropeaking in regulated rivers) while accounting for associated externalities with a discrete choice experiment. This is the first valuation study focusing on hydropeaking that considers both negative and positive externalities. The results imply that most individuals prefer stronger restrictions on short-term regulations to mitigate local environmental impacts. Individuals especially value improvements in recreational use, fish stocks, and the ecological state. On the other hand, potential increases in CO2 emissions are linked with a clear disutility. The estimated benefits obtained from an improved state of the river environment due to such restrictions exceed the disutility caused by increased CO2 emissions. The results also reveal unobserved preference heterogeneity among individuals, which should be accounted for in the willingness-to-pay (WTP) estimation using a model specification with correlated utility coefficients. Overall, the findings can inform policy-makers and environmental managers on the economic value of hydropeaking externalities and further guide the sustainable management of rivers regulated for hydropower generation.

Pre-peak fare discount policy for managing morning peak demand of interregional bus travel: a case study in Seoul metropolitan area

ABSTRACT Pre-peak fare discount policies have become increasingly common in urban transit services and show potential to spread morning peak-hour demand. This study analyzes the departure time choice behavior of interregional bus passengers during morning hours. A mixed logit model is applied to estimate the probability of shifting departure time as a function of fare discount rate, in-vehicle congestion, and other influential factors, accommodating heterogeneity of passenger preferences. The analysis data are obtained through a stated preference survey, which is conducted to passengers who regularly use the interregional bus in Seoul metropolitan area. Choice models are segmented by departure time periods (peak, pre-peak, and post-peak) and occupation types (fixed time worker, flexible time worker, and student). Marginal utilities of the unobserved preferences are estimated to suggest policy implications. From the experimental analysis, target passengers for shifting departure time during peak hour are suggested as flexible time workers and students.

Are californians willing to use shared automated vehicles (SAV) & renounce existing vehicles? An empirical analysis of factors determining SAV use & household vehicle ownership

Future smart transportation systems are anticipated to integrate automation and sharing of vehicles. Responding to the expected changes in shared mobility services, this study used representative data from over 4100 households in California to examine consumers' affinity to use shared automated vehicles (SAV) and their willingness to renounce existing vehicle(s) in the presence of SAVs - two interlinked factors determining the long-term success of SAVs. Results of the study showed that around 50 % of the households reported willingness to use SAVs but unwillingness to renounce their current vehicle(s). As innovators, another 9.1 % stated their affinity to use SAVs and renounce existing vehicle(s) but with heterogeneity across vehicle ownership levels. Random parameter logit models revealed that households using existing vehicles to work for ride hailing companies were significantly less likely to use SAVs and renounce their current vehicle(s). Households who had high awareness about self-driving cars, possessed full electric vehicles, used sustainable work travel modes (bike, e-bike, bikeshare), had greater number of leased vehicles and ridesharing trips were more likely to use SAVs and renounce current vehicle(s). Significant unobserved preference heterogeneity was recorded in the effects of behavioral and sociodemographic correlates. The study contributes by shedding new light on the behavioral determinants of the vehicle ownership impacts of SAVs. We discuss policy implications of the key findings.

Labor Force Transitions of Older Married Couples in Germany

This study investigates the labor force behavior of older married couples in Germany. Monthly observations from the first eleven waves of the German Socio-Economic Panel (GSOEP) are used to describe and analyze the relationship between the labor force behavior of husbands and wives. The empirical model is a discrete time, competing risks hazard model of transitions among labor force states defined by the employment status of both spouses. The analysis indicates that the probability of one spouse exiting employment is much larger if the other spouse is not employed than if the other spouse is employed. Similarly, one member of a couple is much more likely to enter employment if the spouse is employed than if the spouse is not employed. Observed covariates including wages and retirement benefits help to explain these patterns, but unobserved preferences for shared leisure also appear to play an important role.

Classmates and friends matter! Peer effects on cognitive ability formation

In this paper, we investigate the impact of peer effects on cognitive ability formation at two different levels – class peers and close friends simultaneously. We use random class assignments in the China Education Panel Survey (CEPS) to deal with ability sorting and self-selection into classroom when estimating class peer effects. To identify close friend peer effects, we include initial human capital to control for time-invariant unobservables, as stable friendship implies that unobserved preference based on which students make friends is likely to stay unchanged, especially within the one-year frame of the data. We find significant positive peer effects on students' cognitive ability formation at both levels. Peer effects are heterogeneous across student ability. Peer effects work through two channels – peer conformity and peer complementarity. We find both channels generate positive peer effects and jointly determine the size of overall estimated peer effects.

On Assignment to Classes in Latent Class Logit Models

Random parameter logit models address unobserved preference heterogeneity in discrete choice analysis. The latent class logit model assumes a discrete heterogeneity distribution, by combining a conditional logit model of economic choices with a multinomial logit (MNL) for stochastic assignment to classes. Whereas point estimation of latent class logit models is widely applied in practice, stochastic assignment of individuals to classes needs further analysis. In this paper we analyze the statistical behavior of six competing class assignment strategies, namely: maximum prior MNL probabilities, class drawn from prior MNL probabilities, maximum posterior assignment, drawn posterior assignment, conditional individual-specific estimates, and conditional individual estimates combined with the Krinsky–Robb method to account for uncertainty. Using both a Monte Carlo study and two empirical case studies, we show that assigning individuals to classes based on maximum MNL probabilities behaves better than randomly drawn classes in market share predictions. However, randomly drawn classes have higher accuracy in predicted class shares. Finally, class assignment based on individual-level conditional estimates that account for the sampling distribution of the assignment parameters shows superior behavior for a larger number of choice occasions per individual.

Matching and weighting in stated preferences for health care

There is an increasing interest in the use of stated preference methods to understand individuals' preferences for health and healthcare. There is also a growing interest in understanding heterogeneity in individuals' preferences. Consequently, stated preference studies frequently consider models that capture either or both observed and unobserved preference heterogeneity. A popular preliminary investigation into heterogeneity involves split-sample analysis to compare subgroups' preferences e.g., comparing patients with clinicians, or older patients with younger. In fixed-effects models, the constant variables (the individuals’ characteristics) remain stable across choice sets and therefore only enter the choice model when interacted with various attributes and/or levels. However, subgroups of respondents may differ on multiple variables that may not easily be implemented with interaction terms because of complexity and a lack of power thus only one, or a few, variables are typically taken into account in each subgroup model. This paper presents an overview of methods for matching and balancing samples to weight individuals with different characteristics in subgroup analysis and an example of how unweighted comparisons may produce erroneous conclusions regarding the degree of heterogeneity in preferences. We illustrate the issue with synthetic and empirical datasets to explore methods for matching subgroups before and within simple choice models. Our results show that entropy balancing and propensity score matching could be more appropriate than analyses using unmatched preference data when heterogeneity is driven by multiple factors. The paper concludes with a discussion of when matching and weighting may and may not be useful in healthcare decision-making.

Differentiated Excise Taxation in the Beer Market

This paper studies excise tax policy that applies different tax rates to differentiated products. We propose a general method, from the revealed preferences perspective, that allows the identification of the policy maker’s unobserved preferences when setting excise tax rates. We use it to evaluate excise beer tax policy in the Canadian province of Ontario, which differentially taxes local craft beer and large manufacturers’ beer. We identify the government’s preferences over the total surplus and net externalities associated with the production and consumption of different types of beer products. The results show that the government believes the positive externality associated with the production of local craft beer, outweighs the negative externality associated with its consumption, whereas it goes the other way around for large manufacturers’ beer. This means that the government’s subjective per-litre cost of reducing the equilibrium consumption through local craft beer is significantly higher than through the large manufacturers’ beer. The observed gap in excise taxes is aligned with the discrepancy in the percentage decreases in equilibrium quantities of two types of beer with respect to a dollar increase in own excise tax, respectively, given different beer market sizes. We also identify the increasing share of local craft beer consumption with respect to household income as a unique source of the regressiveness of the differentiated excise taxes. The regressiveness does not seem large in terms of the extra tax payment imposed on lower-income households even though the regressiveness of the effective tax rate itself is not insignificant.

Exploring the Influence of Multidimensional Tourist Satisfaction on Preferences for Wetland Ecotourism: a Case Study in Zhalong National Nature Reserve, China

Tourist satisfaction, as a principal element for sustainable development of wetland ecotourism and destination competitiveness, has received extensive attention in the research on its impact on behavioral intention. This paper employs choice experiment to study tourist satisfaction as an important motivation for behavioral intention of willingness to pay and preference on wetland ecotourism improvement policy, and explores the potential causal mechanism. A 15-item scale is designed to measure tourist satisfaction and four constructs are defined. Under the framework of discrete choice modeling, unidimensional and multidimensional data are incorporated as the standard for population segmentation and explanatory variables. The empirical results estimated by three methods basically confirms the proposed hypotheses and indicate that tourist satisfaction has a certain effect on tourist preference of wetland ecotourism and there is significant preference heterogeneity among groups with different satisfaction. Besides, different dimensions of satisfaction have a different effect on tourist preference of wetland ecotourism. This study enriches the studies on non-market value evaluation and preference motivation in the field of ecotourism, because tourist satisfaction may explain some unobserved preference heterogeneity. More effective wetland ecotourism policies can be designed when data on satisfaction are included.

The effect of unobserved preferences and race on vaccination hesitancy for COVID-19 vaccines: implications for health disparities.

BACKGROUND: Reducing the extra burden COVID-19 has on people already facing disparities is among the main national priorities for the COVID-19 vaccine rollout. Early reports from states releasing vaccination data by race show that White residents are being vaccinated at significantly higher rates than Black residents. Public health efforts are being targeted to address vaccine hesitancy among Black and other minority populations. However, health care interventions intended to reduce health disparities that do not reflect the underlying values of individuals in underrepresented populations are unlikely to be successful. OBJECTIVE: To identify key factors underlying the disparities in COVID-19 vaccination. METHODS: Primary data were collected from an online survey of a representative sample of the populations of the 4 largest US states (New York, California, Texas, and Florida) between August 10 and September 3, 2020. Using latent class analysis, we built a model identifying key factors underlying the disparities in COVID-19 vaccination. RESULTS: We found that individuals who identify as Black had lower rates of vaccine hesitancy than those who identify as White. This was true overall, by latent class and within latent class. This suggests that, contrary to what is currently being reported, Black individuals are not universally more vaccine hesitant. Combining the respondents who would not consider a vaccine (17%) with those who would consider one but ultimately choose not to vaccinate (11%), our findings indicate that more than 1 in 4 (28%) persons will not be willing to vaccinate. The no-vaccine rate is highest in White individuals and lowest in Black individuals. CONCLUSIONS: Results suggest that other factors, potentially institutional, are driving the vaccination rates for these groups. Our model results help point the way to more effective differentiated policies. DISCLOSURES: No funding was received for this study. The authors have nothing to disclose.

Local buyers’ exclusion through price on the farmland market: a potential outcome approach

ABSTRACT The influx of non-local buyers into the land market is commonly held responsible for the exclusion of local buyers through price. We study the seaside farmland market in Corsica. A massive price gap between non-local and local buyers is observed. To explore this gap, we rely on a treatment effect approach. We first estimate hedonic price models, while controlling for omitted variable bias using an innovative method. We go further and estimate a general potential outcome model which allows capturing observable and unobservable preference heterogeneity between buyers. We find that the price gap indeed reflects preference heterogeneity, suggesting market segmentation at work.

Imperfect Learning and Neglected Long Run Economic Constraints

This paper shows that an economy with unobservable preference shocks can be temporarily explosiveif adaptive learning agents focus on short run optimization but ignore long run constraints. However,when agents slowly learn that the economy will not be sustainable under the explosive path, they canrestore the economy to a stationary process by changing their current consumption decisions. Withoutany structural changes, these temporarily explosive dynamics lead to more volatile and longer fluctuationsof the consumption to capital ratio compared to those of real business cycle models focusing on highfrequency economic fluctuations. Since these fluctuations are not optimal responses to structural changes,this paper provides policy suggests to prevent these sub-optimal fluctuations of the consumption to capitalratio.

The Effect of Unobserved preferences and Race on Vaccination Hesitancy for COVID-19 Vaccines: Implications for Health Disparities

Background: Reducing the extra burden COVID-19 is having on people already facing disparities is among the main national priorities for the COVID-19 vaccine rollout. Early reports from states releasing vaccination data by race show that White residents are being vaccinated at significantly higher rates than Black residents. Public health efforts are being targeted to address vaccine hesitancy among Blacks and other minority populations. However, health care interventions intended to reduce health disparities that do not reflect the underlying values of individuals in underrepresented populations are unlikely to be successful. Objective: To identify key factors underlying the disparities in COVID-19 vaccination. Data sources: Primary data were collected from an online survey of a representative sample of the population of the four largest U.S. states (New York; California; Texas; Florida) between August 10 through September 3rd, 2020. Study Design: Using latent class analysis, we built a model identifying key factors underlying the disparities in COVID-19 vaccination.Principal findings: We found that subgroups among Black residents are not hesitant at all.Conclusions: Results suggest that other factors, potentially institutional, are driving the vaccination rates for these groups. Our model results help point the way to more effective differentiated policies.

Nonparametric Welfare Analysis for Discrete Choice: Levels and Differences of Individual and Social Welfare

Empirical welfare analyses often impose stringent parametric assumptions on individuals’ preferences and neglect unobserved preference heterogeneity. In this paper, we develop a framework to conduct individual and social welfare analysis for discrete choice that does not suffer from these drawbacks. We first adapt the broad class of individual welfare measures introduced by Fleurbaey (2009) to settings where individual choice is discrete. Allowing for unrestricted, unobserved preference heterogeneity, these measures become random variables. We then show that the distribution of these objects can be derived from choice probabilities, which can be estimated nonparametrically from cross-sectional data. In addition, we derive nonparametric results for the joint distribution of welfare and welfare differences, as well as for social welfare. The former is an important tool in determining whether those who benefit from a price change belong disproportionately to those who were initially well-off. An empirical application illustrates the methods.

Unobserved preferences and dynamic platform pricing under positive network externality

We study the dynamic optimization of platform pricing in industries with positive direct network externalities. The utility of the network for the consumer is modeled as a function of three components. Platform price and participation rate affect the consumer’s decision to join the platform. The platform operator is assumed to know the consumer’s sensitivities with respect to these components. In addition, the consumer’s utility is a function of other attributes, such as network privacy policies and environmental effects of the service. We assume that the distribution of these unobserved preferences in the potential customer base is known to the platform operator. We show analytically how the unobserved preferences affect the dynamic platform price design. Both static and rational expectations with respect to the platform participation are presented. We simulate an electricity market demand side management service application and show that the platform operator sets low prices in the launch phase. The platform operator can set higher launching prices if it can affect customers’ preferences, expectations or adjustment friction.