Principal investigator: Omar Barroso Khodr
Authors: Xuan Zhanga and Huihua Nieb
Original title: Public Health Insurance and Pharmaceutical Innovation: Evidence from China
Location of the Intervention: China
Sample Size: 64 patients
Primary Variable of Interest: Measures of the quantity and quality of disease-related patents
Type of Intervention: Intervention in public policies aimed at the health sector.
Methodology: Differences-in-differences (DID); Fixed Effects in Poisson; and log-linear models.
Summary
Developing countries often exhibit low levels of pharmaceutical innovation. According to the authors, a possible explanation for this phenomenon is the small effective size of their markets, not due to the total population, but because of low income levels and limited health insurance coverage. Therefore, this study investigates a natural experiment resulting from the implementation of a public health insurance program for rural residents in China—the New Cooperative Medical Plan (NCMS)—to examine whether the pharmaceutical industry intensifies its innovation efforts regarding diseases covered by the NCMS and prevalent in rural areas. To this end, the authors analyze patent data between 1993 and 2009, evaluating the evolution of pharmaceutical innovation in the country. They observe that diseases with a 10% higher proportion of rural patients registered a 12,4% increase in relevant national pharmaceutical patent applications, as well as a modest improvement in patent quality after the introduction of the NCMS. Finally, the results suggest that by offering public health insurance to low-income populations, governments in developing countries can stimulate innovation.
- Policy Problem
According to the authors, the central issue lies in the global imbalance in research and development (R&D) in the pharmaceutical field. Despite being home to more than 80% of the world's population, developing countries receive only a small fraction of global R&D investment, leading to serious neglect in the treatment of diseases that predominantly affect their low-income populations. A fundamental question in economic policy is how to correct this market failure and encourage innovation for these "neglected diseases".
Thus, the study directly investigates a possible solution: whether public health insurance contributes as a powerful policy to incentivize innovation. Although these programs are primarily implemented to improve access to healthcare and financial protection, their potential to expand market size and thus create incentives for companies to innovate has been a theoretical, yet little-tested, possibility in the context of developing countries. The research tests the hypothesis that, by making medicines more accessible to a large poor population, health insurance can financially enable investment in R&D relevant to companies.
Additionally, the study addresses skepticism surrounding this mechanism in developing economies. The study explicitly confronts two main counter-arguments: first, that weak intellectual property (IP) protection undermines commercial incentives for innovation, and second, that a low level of technological capability prevents companies from responding effectively. By studying China, a developing economy but with relatively robust IP laws, the authors assess whether these barriers can be overcome.
Finally, the findings have profound implications for policymaking. The study suggests that demand-driven policies, such as health insurance, are not only tools for social welfare but can be integral components of a national innovation strategy. It implies that governments in developing countries can strategically use health financing to direct technological progress toward their specific health priorities, creating a virtuous cycle where improved access to healthcare also stimulates the national pharmaceutical industry. This provides a crucial, evidence-based alternative to relying solely on traditional supply-driven measures such as research grants.
- Policy Implementation Context
According to the authors, three main factors challenged the implementation of public health insurance policies. First, the NCMS emerged as a direct response to a serious failure in public health. Following the collapse of the old Rural Health Cooperative System in the late 1970s, more than 80% of rural residents remained uninsured for decades. The consequences were severe: approximately 38% of sick individuals in these areas failed to seek medical care, and millions of families were pushed into poverty annually due to debt from medical expenses. The NCMS, therefore, did not represent a mere technical adjustment, but rather a decisive intervention to address a systemic failure in access to rural health.
Secondly, the implementation of the NCMS was inherently complex and highly decentralized. Although largely funded by increasing government subsidies, the program was voluntary for individuals—even though it required participation from all households as a way to avoid adverse selection. Crucially, its administration was delegated to the county level. This meant that, while the central government set general guidelines and provided resources, some 3.000 rural municipalities had to design and manage their own specific programs. The result was a veritable patchwork of local implementations, rather than a uniform national policy, which added a layer of complexity for businesses and innovators interested in understanding the new market landscape.
Finally, and particularly relevant to this study, the policy was introduced into a national pharmaceutical sector that was still ill-prepared for high-level innovation. The market was dominated by generic drug manufacturers, and most innovative activity consisted of imitation, not disruptive advances. Furthermore, although China's patent system had improved since the 1990s, it still favored a "first to register" culture, which often prioritized quick, low-quality applications over substantial inventions. The central challenge faced by the authors was to demonstrate that a demand-driven policy, such as the NCMS, could effectively stimulate innovation even in an environment where productive capacities and intellectual property norms were still consolidating.
- Evaluation Details
The primary data for measuring pharmaceutical innovation in this study comes from patent information. Metrics such as the number of patent applications, granted patents, renewals, future citations, and number of claims are used to assess both the quantity and quality of innovation. The authors justify this focus on patents by noting that they are the dominant form of intellectual property protection in the pharmaceutical industry and demonstrate a positive correlation (with an elasticity of 0,37) between research and development (R&D) spending and patent applications among Chinese pharmaceutical companies. The patent data comes from the China National Intellectual Property Administration (CNIPA) and is supplemented with data from Google Patents for citations and claims. The analysis is limited to patents from 1993 to 2009 to avoid periods prior to the introduction of chemical compound protection and to exclude the period following major policy changes, including an amendment to patent law and the launch of a major new drug development project in 2009. The authors explicitly avoid using clinical trial and new drug approval data from the National Medical Products Administration due to a lack of initial data and credibility issues stemming from a high-profile corruption scandal.
For disease data, the study relies on morbidity and prevalence rates from the Chinese National Health Surveys (NHSS) and the Chinese Center for Disease Control and Prevention (China CDC). This data covers 24 specific diseases, including 11 directly from the NHSS, 10 major infectious diseases, two occupational diseases, and infertility. A crucial point of the analysis is that five of these diseases (gonorrhea, syphilis, infertility, pneumoconiosis, and chemical poisoning) are not covered by the New Cooperative Health System (NCMS), China's main public health insurance program at the time, as they fall under the exclusions for sexually transmitted diseases, infertility, and work-related accidents.
Finally, the study incorporates several other datasets to provide context. These sources include the China Health and Nutrition Surveys (CHNS) to calculate health insurance rates, the Annual Survey of Industrial Enterprises (ASIF) for sales and R&D data at the enterprise level, and various national yearbooks to track the evolution of health insurance, economic growth, and demographics.
- Method
The empirical strategy adopted by the authors consists of testing whether the introduction of China's New Cooperative Health System (NCMS), which expanded health insurance to rural populations, encouraged pharmaceutical innovation for diseases that became more profitable as a result. The main methodological choice was to employ a generalized difference-in-differences (DID) model that leverages the variation in treatment intensity among different diseases. Instead of a simple before-and-after comparison, the analysis creates a unique "treatment" variable for each disease, combining whether it was covered by the NCMS with the pre-existing proportion of its patients living in rural areas (Hausman, Hall, and Griliches, 1984; Wooldridge, 1999).
The rationale is that diseases with a higher proportion of rural patients experienced a greater expansion in their expected market size after the implementation of the NCMS, and this should lead to a stronger innovation response. To isolate this effect, the model controls for other factors that could drive innovation, such as market growth in urban areas (using the urban patient count) and underlying disease-specific trends. This approach is crucial because it allows researchers to attribute changes in innovation not only to the existence of the policy but also to the specific economic mechanism—the change in market size—that theory predicts should drive firm behavior. Furthermore, the study takes into account the potential for confounding factors arising from other policies implemented later in the period and uses a statistical model suitable for count-based innovation data, such as patent filings.
To ensure the robustness and credibility of their main conclusions, the authors resorted to an alternative statistical method—log-linear models—as a complementary verification to the primary analysis. The objective was to demonstrate that the central result, according to which the NCMS policy boosted innovation, did not stem from a specific choice of statistical modeling. Since the data on the number of patents contained many zero values (referring to illnesses or years without records), which represents a challenge for conventional log-linear models, the authors applied two distinct transformations to the data to adequately handle these zeros.
The fact that these alternative models produce results consistent with the main analysis reinforces the conclusion that the observed increase in innovation is a real effect of the policy, and not a methodological artifact. This step is fundamental to the integrity of rigorous empirical research, as it demonstrates that the conclusions remain robust under different analytical specifications.
- Main results
According to the authors, the results indicate that the NCMS had a significant positive impact on the quantity of pharmaceutical innovation, although its effect on quality was more modest. The analysis shows that, for a disease covered by the NCMS with a 10% higher proportion of rural patients, there was a 7,8% increase in total patent applications. This growth was entirely driven by domestic applicants, whose applications increased by 12,4%, while foreign applicants showed no reaction. The domestic response was immediate, beginning in 2003, shortly after the policy announcement.
However, when examining more refined quality indicators, the effects proved to be less significant.
The increase in grants and renewals of domestic patents accounted for only about half of the growth observed in applications, suggesting that many of the new patents were of inferior quality or repetitive in nature. Similarly, metrics such as the number of future citations and the volume of claims showed only modest positive impacts and, in some cases, with a time lag.
The authors emphasize that a more detailed analysis of the nature of these innovations revealed that domestic companies were the main drivers, with an increase of over 27% in their patent applications. The new patents largely concerned innovations that combined product and process—a promising indicator, given that this type of patent tends to be of higher quality than those that focus exclusively on the process. Finally, the researchers addressed possible methodological limitations and confirmed that their main conclusions regarding the positive impact of the NCMS on national innovation, although more pronounced in the quantitative dimension, remain robust.
- Lessons in Public Policy
A key lesson in public policy is that state-funded health insurance should not be viewed merely as a social welfare expenditure, but also as a powerful lever for industrial policy and innovation. The authors demonstrate that by expanding coverage to rural and low-income populations, governments in developing countries can strategically create broader and more predictable markets for specific diseases. This market expansion effect generates a direct financial incentive for pharmaceutical companies to increase their investments in research and development (R&D) in often-neglected areas—such as those that disproportionately affect the poorest.
In this sense, when planning the expansion of healthcare, policymakers can deliberately consider the desired direction of technological change and use coverage as an instrument to guide innovation toward national health priorities. Evidence suggests, however, that to fully leverage this stimulus for innovation, it will be necessary to complement such policies with efforts aimed at improving the quality of innovative outcomes. Although the study identified a significant increase in patent applications, quality indicators showed more modest progress. This indicates that, while market demand drives quantity, additional measures are essential to ensure significant and high-quality progress.
In this context, policymakers could incorporate quality-based incentives, linking, for example, tax credits or R&D subsidies not only to the number of patents registered, but also to subsequent milestones—such as the successful completion of clinical trials or the development of drugs with relevant therapeutic advantages.
Finally, the analysis highlights the importance of a more comprehensive evaluation framework for public health policies. Simply increasing innovative activity does not, in itself, guarantee an improvement in social well-being. It is crucial that policymakers support and fund long-term studies that track the entire innovation cycle—from patent to product—and conduct rigorous cost-benefit analyses. This would allow them to verify whether new medicines are effective, accessible, and ultimately generate a net benefit for society, ensuring that public resources invested in health insurance produce the greatest possible return, both in terms of health and economic development.
References
Hausman, J., Hall, B.H., Griliches, Z., 1984. Econometric Models for Count Data with an Application to the Patents-R&D Relationship. Econometrica
52, 909-38.
Wooldridge, J.M., 1999. Distribution-free Estimation of Some Nonlinear
Panel Data Models. Journal of Econometrics 90, 77–97.
