I am an Assistant Professor of Economics at Tsinghua University, School of Economics and Management. My research interests are in health economics, insurance markets, industrial organization and applied microeconomics more broadly. I received my Ph.D. in Business (Risk and Insurance) from the University of Wisconsin at Madison and my Bachelor's degree in Economics from the University of Hong Kong.
[Download CV] Email: liuchy3 [at] sem [dot] tsinghua [dot] edu [dot] cn
How can developing countries foster pharmaceutical innovation when drug price negotiations risk eroding firms' incentives? China's National Drug Price Negotiation policy addresses this challenge by pairing substantial price cuts with expanded insurance coverage for innovative drugs, enlarging the effective market size. We examine its impact on innovation using a difference-in-differences design that compares clinical trials of new drugs (treated) with those of vaccines (untreated) in each disease group. The policy increases the number of trials by 0.56 per disease per year. The effects are particularly pronounced for more novel innovations and similar for domestic and foreign firms. We also find that the policy induces more pharmaceutical R&D firms, especially those of small size, to enter the market. Finally, we document increased collaboration and outsourcing across firms.
Health insurance markets often offer multiple plans, yet consumers frequently choose strictly dominated options. Using two natural experiments in China's private health insurance market, we study how choice set size affects choice quality. An insurer exogenously expanded renewal menus from two to three plans and later reduced them back to two. Exploiting strictly dominated plans and a regression discontinuity design, we find that adding one option increases dominated choices by 8.4 percentage points (52%), while removing one reduces them by 11.0 points (49%). Evidence points to choice overload rather than changes in average plan quality as the main reason.
This paper investigates the distortionary effects of community rating on relative prices in insurance markets with multiple products. We show that common community-rating approaches may generate counterintuitive distortions in the price differentials between plan options that run in the opposite direction as the distortions in overall price levels. We document that partial age-based community rating leads to this pattern in private health insurance exchanges in the U.S. Older individuals are subsidized overall but face marginal prices for the most generous coverage significantly above the marginal cost of that additional coverage. The distortions in these markets are large enough that older individuals often face and sometimes choose dominated options. We discuss and present simulations of the impacts of different community rating approaches on both the efficiency of plan choices and distributional outcomes.
Policy proposals have called for limiting the number of plan options in health insurance markets due to concerns that choice overload may impair individual decision-making. We analyze private health insurance markets in the federally administered Healthcare.gov exchange and document substantial variation across counties in the number of options enrollees face. We then exploit a sudden policy shock that led to a sharp relative price increase for a subset of health plans to examine whether people respond differently to price shocks in areas with larger choice sets. Contrary to the simplest predictions of choice overload, we find that enrollees in areas with very large choice sets responded somewhat more strongly to this relative price change. At the same time, we find that subsets of consumers who arguably should not have reacted to the price change nonetheless did so more in markets with a larger number of options. Our findings suggest that choice overload is not limiting consumers' sensitivity to plan features, but consumers are also not making clearly better choices in areas with larger choice sets.
Should public health insurance be administered at the national or sub-national level? This paper examines the issue in China's public health insurance drug formulary design. Before 2019, the central government allowed provinces to design their own public insurance drug lists. We find that provincial governments favor local firms, adding these firms' drugs disproportionately more to insurance coverage, while holding local demand fixed. We illustrate that a unified national formulary could eliminate such distortions, but may induce welfare losses due to the central government's incomplete information and disregard for the heterogeneity of local demand.
We study deductible choices in Switzerland's mandatory health insurance market—a large-scale, transparent setting where benefits are standardized and the entire enrollee population is observed. Comparing new enrollees, who must make an active choice, with existing enrollees, who may default to their prior plan, we document large and persistent inertia: existing enrollees are 14 percentage points more likely to choose dominated deductibles and exhibit substantially higher foregone savings. Critically, we show that the age gradient in consumer inertia is driven by cumulative market tenure rather than biological age, indicating that prolonged default exposure—rather than age-related cognitive decline—is a key mechanism underlying persistent inaction. Financial incentives cannot reliably overcome this inertia: switching rates plateau at roughly 10% even when potential savings are large. Counterfactual analysis reveals that implementing a smart default would save consumers approximately CHF 2.3 billion annually, but eliminating this inertia requires a 3–4% base premium increase to offset the loss of systemic cross-subsidies.
Horizontal mergers are often associated with product reshuffling, which may have important anti-trust consequences. This article shows that hospitals merging with local competitors reposition service lines after the merger. We use hospital-level service lists data from American Hospital Association 2002 - 2012. To avoid endogenous selection into mergers, we estimate difference-in-differences models comparing hospitals merged later to those merged earlier. We find that merging hospitals eliminate duplicate services without reducing patient volumes. Hospitals within a system become more differentiated in service positioning after the merger. However, there is limited evidence that repositioning leads to significant cost reductions.
Many developing countries face high pharmaceutical prices, even though multiple producers of generic versions of drugs often exist. One reason is that drug firms spend excessively on sales and marketing efforts. In this paper, we explore the effect of centralized procurement of drugs on limiting firms' excessive marketing efforts and reducing drug prices by leveraging a policy experiment in China. Under the policy, certain generic drugs are procured via centralized auctions in pilot cities. The winning firms can directly capture large market shares without incurring high marketing costs. We find that the centralized auction effectively reduces the sales costs and prices, and the winning firms' sales costs and marketing-related labor demand decrease significantly.
Health insurance plans often differ in coverage levels and the combinations of cost-sharing attributes to achieve that level. In this paper, I show that the proliferation of plan designs can result from distortion under asymmetric information. Though optimal risk protection requires concentrating coverage in large loss states (i.e., straight-deductible plans), low-risk types signal by sorting into plans with more coverage for smaller losses. Standardizing plans to vary only along a single dimension may exacerbate welfare loss from asymmetric information. Consistent with the model, I show that a large variation in plan designs exists in the ACA federal exchange and that straight-deductible plans attract individuals with significantly higher ex-post medical spending and ex-ante risk scores. I calibrate the potential welfare effects of standardizing plan designs in the ACA when asymmetric information and consumer confusion exist.
Prior research documents that there are sometimes dominated options in health plan menus, but is that common? We analyze Kaiser Family Foundation data on health plans that firms offer to their employees. For firms offering both a high-deductible and lower-deductible health plan, 62% of the time the high-deductible option has lower maximum spending risk for the employee. We estimate that the high-deductible plan dominates at roughly half of firms. We discuss potential mechanisms behind these surprising patterns and find support both for two explanations: widespread adverse selection pricing and some employers also differentially favoring high-deductible plans.
The capitated payment model has been used to address the high cost of health care. Under capitation, physicians are compensated with a fixed amount per patient, regardless of the services generated. We provide new evidence on how the capitation payment model changes physicians' behaviors by studying the treatment of lower back pain, as this type of treatment provides substantial scope for physicians' discretion. We use data from 2003 to 2006 from a large database of employer-sponsored health insurance claims and leverage capitation variation within the plan and physician to mitigate selection concerns. The results show that the treatment intensity—primarily derived from therapy and diagnostic testing—of patients under a capitation system is 7–12% lower than that of similar patients in a non-capitated plan. Furthermore, we find no evidence of increased relapse rates for patients in a capitated plan.