S4 Ep7: Patent pools for generic drugs

Episode Summary

Executive Summary: The episode examines whether the Medicines Patent Pool speeds access to life-saving drugs in low- and middle-income countries. Mark Shankerman explains that while MPP dramatically increases licensing, the effect on actual drug launches, sales, and prices is much smaller, implying that bundled geographic licenses boost paper access more than real-world diffusion. The discussion closes with reform ideas such as unbundling, pricing licenses by country, and broader use beyond drugs.

Main Topics: What patent pools are and how MPP differs (Priority: 5/5): Shankerman explains patent pools as centralized licensing platforms that aggregate patents for downstream licensing, and contrasts traditional technology pools with the Medicines Patent Pool’s geographic, drug-specific model. Why drug diffusion matters in low- and middle-income countries (Priority: 5/5): The conversation opens with evidence that important drugs can take many years to reach poorer countries, creating large health losses that patent pools aim to reduce. Data limitations and empirical strategy (Priority: 5/5): Because licensing and launch data are usually proprietary, the study combines MPP records, IQVIA market data, and patent sources, then uses difference-in-differences, matched negotiation controls, and fuzzy regression discontinuity to identify causal effects. Impact of MPP on licensing (Priority: 5/5): The MPP strongly raises the likelihood of licensing for product-country pairs, with effects several times larger than baseline, but much of this reflects geographic bundling rather than actual market entry. Impact on launches, sales, volume, and prices (Priority: 5/5): Shankerman finds much weaker effects on actual launches than on licensing, little effect on sales, higher volume, and lower prices, suggesting diffusion gains exist but are modest relative to headline licensing figures. Who benefits most from the MPP (Priority: 4/5): The largest licensing effects occur in smaller, poorer countries with moderate HIV incidence, while large potentially profitable markets are less likely to be included because firms may prefer bilateral licensing there. Policy reforms and future applications (Priority: 4/5): Possible improvements include unbundling country licenses, charging differentiated fees, reducing coordination problems among generic entrants, and extending the model to diagnostics or green technologies.

Key Arguments: Patent pools can reduce transaction costs by centralizing licensing, but the MPP is unusual because it is geographic and focused on access rather than royalty revenue. Drug diffusion is slow enough in many countries to justify policy intervention; a seven-year delay is presented as typical rather than exceptional. Empirical evidence on patent pools is scarce because licensing terms and downstream outcomes are hard to observe; this study addresses that gap using proprietary and administrative data. The MPP increases the probability of licensing dramatically, but bundled licensing inflates the count of licenses relative to actual launches. Licensing gains do not fully translate into diffusion: many licensed products never reach market entry due to manufacturing, distribution, and WHO quality-certification hurdles. The MPP’s largest effects are in smaller, less profitable markets, consistent with the idea that it fills niches that bilateral licensing leaves underserved. Price and volume effects are offsetting: higher volume is accompanied by lower prices, so sales change little even when access improves. Policy changes like unbundling or pricing licenses by market attractiveness could better align licenses with actual launches and competition. Some of the observed benefits likely come from centralized licensing transaction-cost savings rather than simply from suspending patents. The model may be useful beyond drugs, especially for technologies that are socially valuable but commercially unattractive in poorer markets.

Data Points: Time lag for etravirine availability: 7 years - Approved in the US and Europe by 2008, but only available in about 25% of Eastern and Central European countries by 2015. Availability of etravirine in Eastern/Central Europe: ~25% of countries - Illustrates slow diffusion of a life-saving HIV drug. MPP establishment year: 2010 - Created by the United Nations to improve access to HIV-related medicines. MPP remit expansion: 2015 - Widened to include tuberculosis and hepatitis C. Study product count: 173 distinct products - Includes different dosages, formulations, and packaging variants covered by the analysis. Study country coverage: Up to 133 countries - Observed across countries covered by the data from 2005 to 2018. Study period: 2005 to 2018 - Time window used for the empirical analysis. Licensing effect: ~500% increase - Probability that a product-country pair is licensed in at least one country. Alternative licensing effect: ~700% increase - Reported in another sample/specification. Launch effect: ~40% increase - Effect of MPP inclusion on actual product launches, much smaller than licensing effects. Volume effect: ~45% increase - Average increase in volume after launch in MPP-covered settings. Sales effect: Approximately zero - Sales rise little because lower prices offset higher volumes. Royalty level: Zero or under 5% - Typical MPP licenses are zero royalty or very low royalty. Income-country classification: World Bank low, lower-middle, upper-middle income - Defines MPP-eligible countries; upper-income countries are excluded. Sample window for fuzzy RDD: $500 to $2,000 either side of income cutoff - Used to compare countries just below and above World Bank thresholds.

Pivotal Quotes: "The MPP has a really large impact on the probability a license occurs for a given product country." — Mark Shankerman: Summarizing the core empirical result on licensing. "Most of the licenses never get translated into launches." — Mark Shankerman: Explaining why licensing effects overstate actual diffusion outcomes. "We should be thinking about whether these geographically pooled, not necessarily bundled, patent pools might be applicable for other kinds of biomedical, maybe diagnostic instruments or certain green technologies." — Mark Shankerman: Describing future policy and research directions.

Implications: Patent pools like the MPP can sharply increase licensing, but real access gains depend on launch and distribution. Policymakers should focus on design features—especially bundling and pricing—to turn legal access into actual treatment availability.