In 1988, ophthalmologist Patricia Bath was granted US patent 4,744,360 for a laser-based “Apparatus for ablating and removing cataract lenses.” Cataracts have disproportionately burdened women, and Bath had long been concerned about differences she saw among patient populations. She is not alone among innovators in focusing on solving problems that they see around them. But as one of the few female inventors of her generation, she had a rare vantage point. On page 1345 of this issue, Koning et al. (1) document a similar pattern on a large scale. They show that female inventors are more likely to produce patents to solve problems that specifically or disproportionately affect women (such as menopause and fibromyalgia). Beyond recognizing the loss of human talent that arises when women are underrepresented in innovation, this finding highlights the types of problems (and solutions) that are overlooked in the current system with its support of a homogeneous group of inventors.

The idea that people with distinctive life experiences—because of their gender, race, socioeconomic background, education, or nationality (to name a few)—will be attracted to explore different problems might not be a surprise. Diverse inventors “see” the world differently. They will explore a solution space differently, making unusual connections linking otherwise disparate insights (2). What Koning et al. highlight is that diverse inventors also identify with different problems or missions—overcoming strategic blind spots that will otherwise occur if inventors are drawn from very similar backgrounds. Their approach, applied to over 440,000 US patents, is an entirely new way to map the problem spaces (i.e., applications) explored by inventors. This mapping convincingly shows the connection between inventor identity and invention domain across the life sciences (constituting over 10% of US patenting activity from 1976 to 2010). The authors show that all-female inventor teams are 35% more likely to innovate in areas of women’s health than all-male teams (and female-majority teams are 18% more likely to have a women’s focus than male-majority teams).

It is important to celebrate the particular problem focus and distinctive solutions that diverse innovators bring to the economy. Yet, to put these results into a wider perspective, Koning et al.‘s analysis finds that women are represented on only 25% of US life sciences patents during a period when women have come to make up almost 50% of US PhDs in the life sciences. Such low levels of representation in the life sciences—from invention and patenting to start-up founding and fundraising to corporate board membership—illustrate the work that remains to overcome systemic barriers to diversity and inclusion in the life sciences (3).

These patterns of participation hold across the wider US innovation economy: Only 10% of patent inventors are women (4). Further along the innovation pipeline, less than 20% of tech start-up companies have a female co-founder, and female chief executive officers receive less than 7% of venture capital (5). At the same time, just as Koning et al. illustrate how female innovators are more likely to patent in female-focused areas, women are also more likely to found companies in some subfields such as educational technology. This “sorting” across the economy is echoed at the Massachusetts Institute of Technology, where we find that women are more likely to focus on mission- or purpose-driven sectors: health security, food security, and financial inclusion.

The concentration of women in certain fields provides new insights into the continued underrepresentation of women and minorities in innovation (6) but also raises new questions about the drivers that lead innovators into particular problem domains and missions. It is only when evaluating these dynamics at a system-wide level that greater diversity and inclusion in innovation can be effectively supported and the associated welfare losses in the economy stopped.

To understand the wider system context, it is necessary to consider two parts of the innovation system: One encompasses the domains prioritized (and funded) across the innovation landscape (and the mechanisms that go into defining those priorities). Another comprises the individuals selected as inventors and innovators to deliver those priorities (and the range of mechanisms that go into selecting key projects and people). At the intersection of these two parts of the innovation economy is a system in which the allocation of diverse talent to key domains may not simply be a preference for solving certain problems. Instead, it plausibly arises as system-wide beliefs about the innovation priorities for the economy and beliefs about the competence of innovators across fields, combine and misallocate talent.

Consider what funders are searching for—how problems and missions are scoped and prioritized both in early-stage science and in later-stage translation. It is notable that women’s health has, until recently, been viewed as a marginal area of academic or commercial investment where few venture funders prioritized women’s health or markets for female-oriented products.

A key system-level driver of these choices has been women’s underrepresentation in the highest levels of decision-making (where markets or diseases are prioritized for investment) and in grant and investment making (where specific choices are made). Being absent shapes attention, or lack thereof, to these critical problems, leading to what Criado-Perez has called “invisible women” (7). More broadly, a lack of women in decision-making roles has likely shaped decisions— by predominantly male investors—to overlook key problems and markets. The same system dynamics are true for underrepresented minorities. Not having diverse voices at the table defining problem spaces perpetuates a system that narrows the projects and problems that are prioritized.

If female innovators identify particularly strongly with certain missions, or if they have specific solutions to these problems, then the lack of funding priority accorded to these areas will contribute to the low levels of women in the innovation economy. What funders search for—which investments they fund as priorities—shapes who is attracted into innovation. This system can only be overcome by expanding who sits at the table to set priorities.

Another aspect of the innovation system that must be attended to in driving diversity is more familiar: identifying, evaluating and selecting people to pursue innovation priorities. Many studies show that women (and other underrepresented groups) often receive lower evaluation scores when they submit research proposals, resumes, or business pitches—not because of a lack of quality, but because of assumptions being made about their capabilities (8), or because of the questions they are asked (9). To overcome these system-wide issues, it is necessary to consider the processes by which projects are selected: how solicitations are written, how venture pipelines are sourced, who is at the table during the evaluation process, what criteria are used, and how conversations are structured. Research shows that a wide range of voices needs to be heard in the selection process (10).

The criteria for selection are an especially critical part of how a system shapes diversity. In grant-making settings, decision-makers frequently rely on measures that are generally considered to be objective indicators of quality, e.g., citations, but are, in fact, biased in favor of over- not underrepresented groups (2). Thus, even signals of “merit” that are used to support innovation selections perpetuate a system that builds cumulative advantage for some groups over others. For example, if women receive smaller grants or smaller funding rounds early in their careers (11). This will have a cumulative long-term impact on productivity (in science or in commercialization), plausibly explaining the lower rates of invention observed by Koning et al. and across the economy. A “quick fix,” e.g., training aimed at unconscious bias, is not enough. Considering who is at the table making selection decisions and the criteria they use are essential to system change.

The interaction between what is prioritized and who is selected has important, unintended, cumulative, system-wide consequences that magnify inequality at every stage—from grant funding, to patenting, to translation and venture funding. Women have lower levels of confidence when they consider entering some domains (12), causing them to shy away from certain fields—a dynamic that is magnified if they do not observe role models (13) or decision-makers with whom they identify. Similarly, if women believe that they have a disadvantage when pursuing funding in traditionally male-focused fields—whether those are male-focused diseases in life sciences or traditional sectors of the economy—they may be less likely to apply in the first place from “lack of fit” (14). These beliefs are not speculative; evidence from the start-up community shows that women’s proposals were valued on average at only two-thirds the value of equivalent proposals by males in traditionally male-focused business sectors. Men faced no such “valuation discount” in female-focused sectors (15). The outcome of such dynamics reinforces the representation patterns seen in life sciences patenting with women selecting more heavily into women’s health.

When overlaying upon this system of selection a context in which diverse problem domains and missions are often underfunded or overlooked (for lack of effective diversity within priority setting bodies), then the drivers of underrepresentation on many dimensions in the innovation economy become clearer—albeit more daunting.

The logic of this analysis suggests a system in which the allocation of diverse talent to key domains may not simply be a preference for solving certain problems. As well, it arises as the result of system-wide beliefs about innovation priorities and beliefs about the competence of innovators across fields. If these beliefs are manifest early on in the innovation process, they are likely replicated at each subsequent stage, leading to cumulative disadvantage for underrepresented groups of women and other minorities and considerable welfare losses.

We must celebrate the positive impact of attracting some of the most talented women life scientists into critical and previously under-explored research areas such as women’s health. Similarly, in areas of mission-driven innovation, the sense of purpose that drives underrepresented groups is a powerful force for change. But simply using traditional approaches to prioritize and select projects and people without appreciating the subtle dynamics at play across the entire system stifles the allocation of diverse talent across the entire innovation economy and does not bring the best and brightest to bear on the most pressing challenges.