Identifying Industry Margins with Price Constraints: Structural Estimation on Pharmaceuticals
with Pierre Dubois
American Economic Review 2018, 108(12): 3685–3724 https://doi.org/10.1257/aer.20140202
This paper develops a structural model to investigate the effects of pharmaceutical price regulation on demand and on manufacturers' price-setting behavior in France. We estimate price-cost margins in a regulated market with price constraints and infer whether these constraints are binding. Once the shape of demand is recovered, the identification strategy exploits cost restrictions across drugs, which come from observing the same drugs in potentially price-constrained markets (France) and in markets where prices are unregulated (the US and Germany). Using data on the anti-ulcer market from 2003 to 2013, we find that some drugs are significantly constrained by regulation. Counterfactual simulations suggest that price constraints generated some modest savings (approximately 2% of total expenses) and increased consumer surplus, relative to a free pricing scenario. These effects come from shifting consumption towards branded drugs at the expense of generics. However, a policy defining price caps based on prices in other countries (external reference pricing) would increase generic penetration and generate additional savings and consumer surplus by reducing all prices, particularly those of generics.
Collusion in the US Generic Drug Market
with Rob Clark and Chris Fabiilli
International Journal of Industrial Organization 2022, 85, 102878 https://doi.org/10.1016/j.ijindorg.2022.102878
We study cartels that operated in the US generic drug industry, leveraging quarterly Medicaid data from 2011-2018 and a difference-in-differences approach comparing the evolution of prices of allegedly collusive drugs with a group of competitive control drugs. Our analysis highlights (i) the difficulty of establishing a suitable control group when collusion is pervasive, (ii) the importance of accounting for market structure changes when defining the control period, and (ii) the existence of across- and within-drug heterogeneity. We focus on six drug markets that were part of the expanded initial complaint and where there was no entry in the same class during the collusive period, permitting a clean measure of the causal impact of collusion on prices. Our most conservative estimates suggest that collusion led to price increases of between 0% and 166% for each of the six drugs, and damages of between $0 and $3 million for the Medicaid market.
Biased Beliefs and Stigma as Barriers to Treatment and Innovation Adoption
with Laura Grigolon
Lung cancer is associated with smoking and is characterized by low treatment rates and research funds. We estimate a model of treatment choice where patients internalize societally biased beliefs on the effectiveness of treatment and stigma, basing their treatment decision on the treatment decisions of their reference group. Identification rests on the exogenous variation in the treatment propensity of physicians. Placing all patients in a neighborhood characterized by low social discrimination increases treatment rates by 4% and the use of innovative therapies by 3%. Social effects account for around 2% of the gap in research funding for lung cancer.
Delisting of Pharmaceuticals from Insurance Coverage: Effects on Consumption, Pricing and Expenditures in France
Working paper
This work structurally estimates the impact on demand and supply of public insurance coverage of pharmaceuticals and its removal (so-called delisting). The analysis focuses on the 2008 delisting of oral phlebotonics (drugs to treat venous circulation disorders) in France, where insurance coverage is tied to price regulation: when a drug is delisted, its price becomes unregulated and the manufacturer can set it freely. This regulatory change and the fact that some drugs were never covered before 2008 provide the variation needed to identify price-cost margins and simulate the counterfactual pricing equilibrium without coverage and price regulation. Results suggest that insurance coverage with price regulation stabilizes prices and guarantees demand for some drugs, like generics, which would sell much less in the absence of coverage. Without insurance coverage and price regulation, increased competition and higher price elasticity would result in lower average prices and reduce demand for most drugs on the market, in line with what is observed after 2008 at the delisting or oral phlebotonics.
Strategic Tier Design in Health Insurance: The Case of Medicare Part D
with Alessandro Iaria and Léa Bignon
CEPR Discussion Paper 17937
We study the role of tier design in Medicare Part D. In the period 2013-2017, plans expanded the number of tiers in their formularies from three/four to five and system- atically shifted generics to higher tiers subject to higher cost sharing. The systematic tier upgrading caused significant increases in the out-of-pocket costs, up to 6 times for some generics. This resulted in additional average per-enrollee spending on generics of $76 in 2017, totalling $1.5 billion for the Part D population, and increased mortality by 5.4% due to reduced utilization of generics with documented mortality benefits.
Prescription of Generics: Evidence from France
Working paper
This work investigates the factors that enter the prescription decision between brand-name and generic drugs, to verify whether physicians adjust the choice of the brand type to patients preferences and characteristics. The aim is to shed some light on the still limited generic prescription in France and on the documented increase in sales for brand-name drugs subject to reference pricing. The analysis uses CEGEDIM data on a representative sample of 326 French General Practitioners and up to 7,500 patients, receiving prescriptions between 2000 and 2008 for drugs in five therapeutic classes (anti-ulcer, anti-diabetes, anti-hypertensive, anti-cholesterol and antidepressant drugs). Results suggest that physicians and patients characteristics play a major role in explaining the decision on the brand type: patients that are fully reimbursed show a higher probability of receiving a brand-name prescription. In addition, the probability of generic prescription drops significantly once a drug is subject to reference pricing, but out of pocket expenses for drugs with a price higher than the reference price discourage brand-name prescription.
Geographic Disparities in Access to Cancer Treatment and Health Outcomes in Ontario
with Laura Grigolon and William Evans