Research

Directing consumers to higher-quality service providers has been considered an effective policy to improve service outcomes and consumer welfare in various contexts. However, higher-quality providers may tend to be more congested, and congestion may be detrimental to outcomes and welfare. We study this congestion-quality tradeoff and discuss its policy implications in the context of Japanese nursing homes. We find evidence that (1) within nursing homes, higher occupancy leads to poorer care outcomes but (2) between nursing homes, occupancy and outcome-based quality measures are positively correlated. To evaluate the welfare impact of patient reallocation policy, we then develop a model of demand for nursing home care where choice set is potentially constrained in an unobserved manner by providers’ rationing behavior. We find that nursing homes are less likely to admit patients at higher occupancy but no evidence that patients dislike congestion. Simulation of a reallocation policy suggests a potential gain from occupancy smoothing even though the policy sends patients to lower-quality care providers on average.

Keywords: congestion, value added, long-term care, demand estimation, choice constraint, consideration set.

We study the effect of complementing public health care with private care. Leveraging a policy at the Veterans Health Administration that generates discontinuity in private care access, we find that expanding coverage to private care increases private outpatient care by $53 (SE: 5) and decreases VA outpatient care by $20 (SE: 7), with no impact on inpatient care. The policy led to a marginally significant 0.1 p.p. (2.8%, SE: 0.04) decrease in one-year mortality, possibly because of decreased wait times and increased access to certain specialty care. Given our estimates, the benefit of access expansion significantly outweighs the increased costs.

We estimate net effects of offline stores on online spending using US internet activity and store-location data. These effects are ambiguous in sign: the effect of a retailer's offline stores on its online sales depends on opposing cannibalization and cross-channel complementarity effects. Similarly, effects of offline stores on rivals' online sales depend on opposing business-stealing and showrooming effects. We find that a consumer's spending at multichannel retailer's online store falls (1.1-3.8%, on average) when a rival adds a nearby storefront but rises (7.1-32.3%) when the retailer opens its own storefront. Offline stores often boost Amazon's sales, suggesting showrooming's relevance.

Social norms are an important determinant of behavior, but the behavioral and welfare effects of norms are not well understood. We propose and axiomatize a decision-theoretic model in which a reference point is formed by the decision maker's perceptions of which actions are admired (prescriptive norms) and which are prevalent (descriptive norms), and utility depends on the pride of exceeding the reference point or the shame of falling below it. The model is simple, yet provides a unified explanation for previous empirical findings, and is useful for welfare analysis of norm-evoking policies with a revealed preference approach.