The Academic AdvantageGender Disparities in Patenting

A patent is a document that turns a discovery into ownership. It is how knowledge becomes an asset that is legally protected, potentially licensed, and sometimes worth millions. In 1976, women filed 2.7 percent of patents. Not twenty-seven percent. Just two point seven. This raises an important question that Sugimoto, Ni, West, and Larivière explored over four decades of data: where exactly does that gender gap narrow, and what does the answer reveal about how science and commerce treat women differently? Their dataset is comprehensive, consisting of 4.6 million utility patents issued by the United States Patent and Trademark Office between 1976 and 2013, linked to 10.8 million inventor name strings. To measure gendered contributions without inflating counts from patents with many co-inventors, the team uses fractionalized inventorship. The logic is straightforward: each inventor on a patent receives a fraction equal to one divided by the total number of inventors whose gender could be assigned. So, a patent with four inventors assigns each a quarter of an inventorship, rather than counting everyone as one. Those fractions are then aggregated by country, by sector, and by technology area to produce a clear picture of who is actually doing the inventing. The headline result: women's fractionalized share of patenting rose from 2.7 percent in 1976 to 10.8 percent by 2013—nearly four times as much. That sounds like progress—and it is. However, Sugimoto and colleagues quickly place this in context. Women account for roughly 33 percent of researchers and about 30 percent of scientific authorships. Their patenting share, even at its 2013 peak, falls far short of those benchmarks. While the pipeline of women entering science has grown, the conversion of that science into patents has not kept pace. Something in that conversion process is filtering women out. The central finding of the paper is where it gets genuinely unexpected. In every single technological area—not most, every one—women's patenting is proportionally more likely to occur at universities than at corporations or government labs. The authors call this the academic advantage, and it holds without exception across macro-disciplinary fields. Mean female participation in university-owned patents was about 11 percent across the full study period, compared with less than 8 percent in firms. By 2013, women accounted for 18 percent of university-owned patents, 10 percent of industry-owned patents, and 12 percent of patents owned by individuals. The gap is real, and it widens over time. This is not a quirk of the American system. The academic shelter appears internationally. The dataset covers inventors from more than 185 countries, with eight countries accounting for over 91 percent of all inventorships—the United States at 53 percent, Japan at 19 percent, and Germany at 7 percent. Male dominance in patenting appears in nearly every country; 42 countries in the dataset list no female inventors at all. Only five countries are female-dominated, and each of those has fewer than 35 fractionalized patents—exceptions too small to constitute a counter-pattern. Yet across this global landscape, the same regularity holds: wherever you look by technology area or by country, universities show higher female participation than firms. The finding is systemic, not local. Why would academia produce this shelter effect? The authors point to several mechanisms. The Bayh-Dole Act, passed in 1980, gave universities the right to patent federally funded research, and in response, the number of university Technology Transfer Offices increased tenfold. Those offices actively seek out academic inventors, reducing the burden on any individual to navigate the patent system alone. Sugimoto and colleagues also suggest that universities' less hierarchical structure may make it easier to build the social networks that are essential to commercialization—networks that women in corporate settings have repeatedly been found to lack access to. Academia is not equitable, but it may impose fewer of the specific structural barriers that suppress female patenting in industry. Now here is where the story turns. The academic setting that gives women a patenting advantage also produces a paradox that the data make hard to ignore. Patents involving women are more collaborative—they list more co-inventors on average. They are more interdisciplinary—they span a higher number of International Patent Classification codes, which are the system-wide tags that categorize a patent's technology domain. More co-inventors and more IPC codes sound like markers of ambitious, wide-ranging work. However, those same patents receive fewer forward citations. Citations to patents are the standard measure of technological impact, and Sugimoto and colleagues quantify this using an Average Relative Citations indicator—field-normalized, meaning each patent's citation count is divided by the average for patents issued in the same year and the same technology area. A score above one means more cited than the field average; a score below one means less. By this measure, patents with female contributors consistently underperform patents with male contributors, regardless of assignee type. The counterintuitive part is that the impact gap is narrowest in the corporate sector and widest in academia. The setting where women patent most is the setting where the citation shortfall is worst. This is not because women's academic patents do unusually poorly—their data show that female university patents and female corporate patents have similar impact. The gap widens in academia because male university patents have unusually high impact relative to male patents in other settings. Women get into academic patenting, while men extract maximum advantage from it. The paper raises possible explanations for this citation gap without claiming to resolve it. The well-documented tendency of male academics to self-cite is one possible contributor. There’s also the question of whether interdisciplinary patents are harder to locate through standard citation channels. If a patent spans many technology categories, it may fall between the reference habits of any single community. The authors acknowledge the possibility of systematic under citation of women's work. These are real possibilities. The paper does not pick a winner among them, and that honesty is worth noting. What it does establish, clearly and across multiple dimensions, is that inclusion and recognition are not the same thing. Women are patenting more—10.8 percent by 2013, up from 2.7 percent in 1976. They are more present in academic institutions than anywhere else. Their patents are more collaborative and more interdisciplinary than men's. Yet those patents are cited less, their names sometimes disappear between the related scientific article and the patent itself, and fewer women see commercialization or licensing from their inventions. The policy implications that Sugimoto and colleagues draw are specific. The academic setting is a rare case of a structural arrangement that seems to reduce—not eliminate, but reduce—the gender gap in who gets to patent. Technology Transfer Offices, by actively recruiting inventors and handling the administrative and legal complexity of patenting, may perform an equity function that nobody explicitly designed them for. Understanding that function and expanding it is a concrete lever. The less hierarchical social environment of universities, if it genuinely lowers network barriers, is another. These are institutional features, which means they are, in principle, adjustable. But the citation gap indicates that getting women into the patent count is not the end of the story. A patent that nobody cites has lower value—it generates less licensing interest, less follow-on innovation, and less recognition for the inventor and the institution. If women's patents are systematically receiving less of that downstream reward, then the academic advantage is partial. It opens a door without guaranteeing that walking through it leads to the same room. Thirty-seven years of data, 4.6 million patents, and inventors from more than 185 countries show that the trend is upward. The mechanisms are partly visible. What Sugimoto and colleagues leave you with is not pessimism, but a precise diagnosis: the innovation economy does not treat all knowledge equally, and the gap between who does the science and who owns the discovery is not shrinking fast enough on its own. The question remains whether institutions—universities in particular, which the data suggest are already doing something right—will act deliberately on what this analysis makes visible. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.