Discover more from Uncorrelated Interests
Venture Capital and Structural Bias
We stifle innovation by disadvantaging female innovators at every level.
Last week in The Inequity of Venture Funding, we established two foundational truths:
There is a gaping gender gap in venture capital funding.
Closing this gap (in venture and beyond) would fundamentally and positively alter US GDP for the next century.
This was a fact-based exploration - the “what” of the gender gap hypothesis. What I didn’t cover was why such a gargantuan gap still exists in 2023, nor did I offer any solutions. The latter point must wait one more week, but the former is the thrust of this piece - an examination of why gender equity is so poor in the funding of innovation (in venture).
Important as it may be, it’s a challenging piece to write, so let’s very briefly talk about the approach and structure. I am very obviously a white male, which of course means that I am the very combination of immutable demographics that created or at least enabled this funding gap in the first place.
Given this, the piece is structured as a series of hypotheses formulated to explain pieces of the gap, coupled with a review of what the academic literature believes to be true today about each of these hypotheses.
As is often the case in this newsletter, complex topics yield complex answers, and though it would be easiest, or perhaps simplest, if the answer was “male venture capitalists are uniformly sexist and routinely make biased decisions that harm women”…well, I’ll leave that conclusion for the end.
Let’s jump in.
Hypothesis 1: The gender gap reflects the present distribution of available female talent.
We start with a fairly obvious question - what if the current rates of VC investment simply reflect the current rates of entrepreneurial talent available in the market? I already established that women currently represent just a quarter of founders seeking angel investment - what if this represents the current max capacity for female founders? Unfortunate as this would be, if our goal is to close the gap we need to establish our (relatively) objective baselines.
To do this, we’ll walk through several pieces of research that compare both venture Investor and Founder rates against participation in similar industries, to track the “relative” rate of talent presently available to fund.
We can see that the share of post-grad degrees for women has climbed substantially in last 30 years, with all indications that women will likely achieve parity with men in the next decade if they haven’t already (given these data are now a decade old).
In spite of the growth in degree attainment (“credentialing”), women have seen their workforce share in the traditional “startup fields” fall over the last couple decades:
The researchers observe:
“This decline reflects not fewer women entering the software industry (the absolute number of women entering software has actually increased), but the rate of increase for women has been slower than the rate of increase for men.1”
When we compare the rates of startup founding and investing for women against other “top professions”, we see just how stark the last 30 years have been:
Women have reached gender parity in a number of traditionally high-value domains yet remain pinned to absurdly low rates of participation in the startup disciplines.
Management consulting and investment banking are the most closely associated disciplines given their levels of achievement, technical skills, and similar problem sets, so they provide the ideal equity baseline for our examination here:
Looking solely at participation rates since 2010, the gap is obvious: huge swaths of our most educated, most ambitious, most talented women are NOT making the transition into startup land2.
As the researchers articulate:
Women entrepreneurs are about one-third to one-fourth of the fraction that they are in science and engineering degrees or professions in software/computer manufacturing. Clearly, the low level of females in entrepreneurship cannot be explained by a lack of trained and skilled women.
Now, I can already hear the objections - the typical “Silicon Valley and Wall Street are fundamentally different ecosystems” type. Certainly we should agree that investment banking and venture investing are comparable, and in fact venture as a whole would likely benefit from a greater influx of Wall Street financial acumen.
And though the West Coast bias against the management consulting field is overwrought, I’ll at least concede that consulting and founding are very much not the same discipline, even if the two share very similar traits. If we instead look at STEM participation, a more concerning reality emerges:
Although women now represent 55% of all STEM jobs, their participation in the typical Silicon Valley fields of interest - here represented as “Computer” or “Engineering” jobs - the rate of participation looks much more similar to the rates of founders seeking Angel funding. More disconcerting is the trend over time:
As mentioned above, these rates do not indicate that fewer absolute women are entering these fields - just that men have done so at significantly higher rates.
In summary, this hypothesis requires a bit of nuance:
On the investor side, there is an incredible supply of female talent that can and should elevate participation far beyond the current 10% participation rate.
On the founder side, there is, like with investors, a substantial talent pool of capable business-focused founders, but the pool of female technical talent is quite similar to the female founder rates.
So we can reject the hypothesis for investors but cannot reject the hypothesis for founders. That said, we would still expect substantially greater volumes of female founders than we currently see given the adjacent talent pools.
Hypothesis 2: Potential female founders are self-selecting out of the entrepreneurial path.
Though we haven’t been able to fully reject the first hypothesis, we have shown at least that the rates of female founders (and especially investors) should be considerably higher than presently observed given available talent. We thus must consider the possibility that these “potentials” are choosing (i.e., self-selecting) not to enter Silicon Valley.
Stanford researchers found that women are generally far less competitive than men, so an environment like VC structured around competition will, on its face, advantage men. But note that the very fields of comparison we’re making here - management consulting and investment banking - are themselves highly competitive, yet show 3-12x higher rates of participation than startup founding.
In a separate paper, researchers found that highly qualified women from MIT are less likely to become an entrepreneur if their idea lacks intellectual protection. However, women who decide to pursue their idea-full time experience no disadvantage at raising VC than their male counterparts. This is a small/specific sample but provides some evidence that women are more likely to “rationalize” their way out of pursuing entrepreneurship even when they are eminently qualified.
This is a more challenging hypothesis as it requires counterfactual leaps, and there is understandably less literature covering this topic. There is some evidence of gender effects, and though not strong enough to wholly reject the hypothesis, there’s enough here to dive in a bit further.
Perhaps there is a more subtle effect at play here.
Hypothesis 3: Female entrepreneurs are harmed by lower levels of social capital.
As we’ve discussed, the gender gap is actually starker on the investor side, as women represent just 10% to 16% of general partners at venture firms, just 5% of funds are female-managed, and only 26-30% of angels (HERE, HERE) are female.
Taken further, nearly three-quarters of firms have never hired a female GP.
Between the 3-5x differential in founders and 9x differential in venture capitalists, one could hypothesize that potential female founders subtly self-select out of the profession due to something closer to René Girard’s Mimetic Theory:
“Man3 is the creature who does not know what to desire, and he turns to others in order to make up his mind. We desire what others desire because we imitate their desires”
We might see this specifically manifest in both the size and composition of social networks in which each gender participates. This is especially true in venture capital, a cottage industry which to this day is primarily conducted via warm introductions and face-to-face meetings. Female disadvantages in social composition would necessarily then decrease the rate of founding.
We do (unfortunately) find some support for this hypothesis:
“Evidence suggests that women entrepreneurs have more restricted networks than do men, and that these networks are comprised disproportionately of members of the same sex (Aldrich, 1989; Brush, 1992, 1997).”
More specifically, the “heavy reliance on trusted referrals may also privilege those who are more connected to investors”. This paper uses Harvard Business School’s (HBS) New Venture Competition (NVC) to test whether the randomly assigned panels of judges differentially affects male vs female startup teams.
Narrator: Yes, yes it does:
“Exposure to more VCs substantially benefits male participants, while this relationship is at most only marginally positive for women…Regression estimates find that relative to the effect among women, each additional VC judge on a panel increases the likelihood of men starting a VC-backed company after HBS by 25 percent.”
Given these results, the authors sent a survey to better understand WHY males might be differentially benefitting from more VCs on these panels.
“The survey responses reveal that male participants are nearly twice as likely as women to proactively reach out to VC investors after the NVC…Conditional on reaching out, however, men and women report no difference in the rate at which VCs respond to their outreach or any difference in the degree to which VCs proactively reach out to them.”
So the higher rate of funding is solely due to more assertive followups, not on responses to those followups. And a greater number of VCs on the panel just increases total odds.
“Consistent with this, qualitative responses in the survey also suggest that because follow-up with judges at the NVC was not explicitly encouraged, women were more likely to have some reservation about leveraging the connection to discuss fundraising.”
This gets into the more implicit structural biases, or at least perceived biases, within venture capital. Men are far less likely to fear blowback or reputational harm by just reaching out, whereas women are more concerned about violating unwritten (implicit) rules and thus don’t play the game as aggressively.
“Men and women may have different beliefs about appropriate networking norms. There may also be homophily in networking, where individuals might feel more comfortable networking with others of the same gender. Since most VCs are men, this would lead to lower rates of networking with VCs among women. Furthermore, women may not reach out if they anticipate discrimination or harassment on the part of investors.”
Translating from academic speak - independent of the salience of the given idea being pitched, male founders are far more likely to play the venture networking game - and probably play it better - than their female competitors, and this yields positive returns.
Remember again that in this study, the actual scores from judges were roughly equivalent between male and female founders. The results of the competition yielded gender gaps not from the value of the pitches but from the networking abilities - and really, the absence of concern for implicit networking norms - of male founders.
“While our analysis is based on a sample that is quite particular, we believe the results have broader significance. If networking frictions matter in this already highly ambitious and well-connected sample of potential entrepreneurs, it seems likely that they will matter more generally in VC and thereby perhaps contribute to the overall gender gap in VC-backed entrepreneurship… Structural solutions that focus only on providing female entrepreneurs more exposure to VCs may not always be sufficient to eliminate such networking frictions.”
This is a really interesting point! The sample already includes some of the best networked and most-privileged set of women - these are Harvard undergrads, after all - and gender differences emerge even here. Exposure to venture is clearly important, but insufficient, as men will play the game in a manner that women either don’t or feel like they can’t.
“This reliance on networks may privilege those who are more connected or those who are most comfortable forming connections with investors”
At present, investors are biased toward the best networked rather than the absolute best ideas which, given our overall thesis of maximizing global innovation, places networking frictions as a key area of future innovation. The researchers drive home this point:
“More generally, since the individuals behind ideas are intricately tied to the ideas themselves at a venture’s earliest stages, and the distribution of good ideas is not perfectly correlated to the distribution of good access to VC, our results suggest that promising ideas may go unfunded because of systematic variations in VC access rather than because of the inherent quality of the idea.”
Here we can more definitively confirm the hypothesis that female founders are significantly disadvantaged compared to their equivalently competent male peers due to fundamental differences in the approach and quality of networking.
Hypothesis 4: Homophilic bias significanly disadvantages female entrepreneurs.
I included a bit of a throwaway line from the prior section:
“Men and women may have different beliefs about appropriate networking norms. There may also be homophily in networking”
Let’s unpack this further, given the 9x discrepancy in venture investor gender.
Homophily is the tendency to seek out and partner with people that more closely resemble us. Small groups are both more likely to be homophilous and more likely to see biases aggregate into expressed decision-making.
Most venture firms are quite small, typically with fewer than 5 investors in each, and this structure itself makes venture investing more likely to exhibit homophily.
The phenomenon is not purely theoretical. From here, co-investment patterns in venture capital are driven by social similarities, in which venture capitalists who are more similar in terms of gender, ethnicity, school background, and work history are more likely to collaborate. More specifically, the likelihood of two investors investing together increases by:
16.3% when both hold a degree from a top university
34.4% when both graduated from the same university
39.2% when both are of the same ethnic minority
Investors are more likely to make homophilic investments, but what about the other side of the negotiating table, the founding teams themselves?
Ethnicity and gender are the two strongest homophilic forces in social networks, with individuals 25% more likely to form groups with people of the same ethnicity or gender relative to randomly matching within a set of students who choose the groups that they work with on real micro-businesses.
Important here is that the effects are not equivalent within a given demographic:
Homophily in education and working experience is stronger among males than females.
Homophily is (again) often stronger among smaller groups.
So we’ve shown that, for both investors and founders:
Homophily overall is a strong factor in partnerships
Homophily is strongest for demographic similarities, especially amongst males.
Small groups (of which both founders and investors belong) exacerbate homophily.
Given all of this, we would expect the homogeneity (90% male) of venture investors to manifest as homophilic investment in startups, which is in fact what we see.
Female-led startups experience significantly more difficulty garnering interest and raising capital from male investors compared to observably similar male-led startups. In particular, women are less successful with male investors, even controlling for a battery of startup/founder characteristics that encompass much of the information that was available to investors online when making the decisions we are studying.
However, this bias is reduced with more experienced investors. What’s not clear is the mechanism by which more experienced investors would be less homophilic:
Have they pushed down their biases due to successes in female funding in the past?
Or is there a survivorship bias, whereby the more experienced (successful) investors were themselves less biased to begin with, invested in females more regularly, and from there saw better returns?
In conclusion, we find broad support for this hypothesis:
Individuals of the same gender are more likely to work together on both the investor and founder sides.
Investors tend to favor founders of the same gender.
This gender bias is stronger for males.
Because 90% of venture investors are male, homophily has a broadly negative impact on female founders.
Hypothesis 5: Investors hold female founders to a different standard.
Given the significant homophilic effects I just discussed, we might naturally expect that female founders are literally treated differently in pitch meetings or any further discussions with investors. This perspective is widely supported in the literature.
Female founders find their presentations are subject to more numerous challenges and pushback, especially with regard to technical questions. And this effect actually worsens when a prototype is readily available.
Note that the researchers here find no evidence that female investors discount female founder credentials - this is a male-only effect that manifests heavily in the types of questions they pose:
“Investors tend to ask male entrepreneurs promotion-focused questions and female entrepreneurs prevention-focused questions.”
That is, investors are more inherently attuned to the upside opportunity with male founders but flip this to downside risk mitigation when faced with female founders.
“We demonstrate that every additional prevention-focused question significantly hinders the entrepreneur’s ability to raise capital, fully mediating gender’s effect on funding.”
This should come as no surprise given that we already established in Gender Inequity that female entrepreneurs raise 20-50% less than their male counterparts, but now we have at least one specific mechanism.
“Separately, we also find that female-led startups are at less of a disadvantage with male investors when they seek low amounts of capital or operate in female-centric industries.”
Note that, again, neither of these effects occur with female investors evaluating female founders - they neither require lower levels of investment nor pigeon-hole female founders in “female industries”.
So we can indeed confirm this hypothesis, with a caveat - male investors do indeed hold female founders to a different (higher) standard than male founders.
Hypothesis 6: Male characteristics fundamentally privilege male founders over female founders.
A natural hypothesis that flows from our last finding is that there may be something inherent about male-ness that also drives these disparate outcomes. Anecdotally it’s easy to conclude that males tend to be more (over)confident, more brash, and that these exaggerated traits yield better results. But are such suppositions supported by research?
We have preliminary evidence:
“Male founders are more likely to make bold projections and assumptions in their pitches. Women, by contrast, are generally more conservative in their projections and may simply be asking for less than men.”
But it’s not simply the hyperbolic verbiage, as it appears that maleness itself drives better returns.
Investors prefer pitches from attractive men, and even just hearing a male voice - the same pitch with different timbre - garnered significantly higher likelihood of success compared to a female voice. In contrast, a female founder’s level of attractiveness does not appear to impact funding decisions.
An obvious mechanic to counteract these visual and verbal biases to the masculine persuasion is leveraging an application process4, which in theory should neuter the biases.
Survey data of angel investors platforms does show that financing proposals sent by women to angel platforms experienced no difference in funding success than male founders. Similarly, researchers conducted a large field experiment that sent cold emails from fictional founders to real investors and found a positive impact of being female.
Kapor Capital is one of the more renowned venture firms that eschews the requirement of warm intros and face-to-face initial meetings in favor of applications in order to broaden the founder pool. And yet…
Additional research demonstrates that women score lower even in blinded grant application evaluations because they tend to use more narrow words, despite having better scientific output conditional on funding. Women simply don’t boast, don’t hyperbolize, like males do, regardless of the output medium.
Application design can certainly be improved, but given what we know already, there is ample evidence to support the current hypothesis that maleness itself biases investors toward male founders.
Hypothesis 7: Overt sexism is a prime driver of poorer outcomes for female entrepreneurs.
I’ve hedged a bit in leaving this obvious hypothesis to last, and I’ll admit this was done purposefully to avoid the expected moral, ethical, and political inflammation that would necessarily impact your own reading of the remainder of the piece.
Note as well the framing of this hypothesis - I specifically included “prime driver” because there is zero argument that overt sexism is at play in an industry so heavily dominated by males. We have plenty of unfortunate examples to illustrate horrendously sexist, misogynist behavior from male investors.
That last linked piece, focused on Ellen Pao5 - the unfortunate face of and avatar for Silicon Valley misogyny - cuts specifically at a difficult area for researchers right in the title: “The Sexism You Can’t Quite Prove”.
As the author describes, so much of the sexism that permeates male-driven industries like venture capital exists in subtle, gray areas:
“Exhibit A: Pao’s performance reviews knocked her for her “sharp elbows.” There were similar negative comments in Pao’s male colleagues’ reviews, but they were nevertheless promoted. Does that demonstrate that Kleiner Perkins treated Pao differently because she was a woman? Might they have interpreted her assertiveness as “bitchiness,” and her male colleagues’ assertiveness as “strength” or “conviction”? Maybe she really did have sharp elbows, hurting her relationships with clients? Can’t women ever be criticized for being caustic?”
As I established above, we do have evidence that this subtle form of gender bias occurs in pitch meetings in the types of questions levied against female versus male founders, or in the downplaying of female (technical) credentials in ways that rarely impact male founders.
Given the preponderance of subtler, often unintentional or emergent gender bias, it should come as little surprise that the academic literature concludes that overt, explicit sexism is not the major contributing factor in poorer outcomes for female founders.
From the prior research I used to illustrate the impacts of investors' gender composition on both willingness to enter funding competitions and on outcomes of such competitions, researchers stated:
“We do not find obvious evidence of explicit bias by male VCs against female participants in our sample. In addition to VCs responding to outreach equally by participant gender, we show that the private scores of VC judges are in fact slightly lower for male-led ventures than for women-led ventures.”
In this case at least, women actually scored slightly higher than men! Now, a college competition is not perfectly analogous to an industry that dispenses $150B+ in funding annually, and we unfortunately do not have extensive research captured within actual meetings between investors and founders. Again I return to the prior article:
“The Sexism You Can’t Quite Prove”
For this final hypothesis, I’ll tentatively conclude that there is both ample anecdotal evidence and a clear lack of scaled quantitative research to appropriately assess the relative impact of overt sexism.
I fully expect that female readers will find this conclusion too ambiguous, too hedged, and I sympathize with the feedback.
A core challenge for research of all kinds is accounting for what you haven’t actually captured. What is different about those who didn’t participate? And as much as we statisticians can use various numerical rationalizations for why a given sample is representative, the unfortunate truth is that “representative sampling” is a directional effort at best.
Regardless of the “true” pervasiveness of explicit sexism in venture, the narratives are both frequent and well accepted. Returning to my first hypothesis exploring the potential talent gap, how many hyper-talented women have chosen not to enter venture because of the perceptions of sexism? Because it’s just not worth the pain?
Such a counterfactual is both uncaptured by current research and potentially significantly underplays the impact of sexism on the funding and innovation gap. That is, the second-order effects of overt sexism may cast a far wider net than the first-order effects imply.
Here is what I believe we know based on available research today:
There is a massive supply of potential female founders and investors who are not naturally migrating into the venture/entrepreneurship track but who could unlock substantial innovation value if barriers were removed.
This talent supply would not drive gender equivalence, given the relatively low rate of female undergraduates in software-related fields, but it would significantly bridge the gap without any changes to college studies.
The exact mechanisms driving the delta between skilled workers and participation in the venture economy are not wholly clear.
We do know that homophily plays a large role in behavior, and as such, we would expect that a lack of people who look like you should negatively impact your decision to enter the industry.
Because venture investors are 90% male, and because homophily is exacerbated by the small investment teams that are the industry norm, the networks of investors will be highly skewed toward males.
We do in fact see much stronger networks from male founders more generally.
These stronger networks exacerbate the gender funding gap precisely because venture still largely operates through its personal social networks.
That is, female founders are far less likely to actually make it into the standard venture pipelines.
However, even when investors and founders do connect, females are treated differently - their credentials are downplayed and investor questions focus more on downside than upside.
Further, the inherent lack of maleness itself places female founders at a disadvantage, even when we strip out obvious gender classification via online applications.
That said, the use of online applications can significantly mitigate gender effects, and we should expect that wider adoption would increase equity.
This is an investigation only into venture capital - a massive catalyst for especially economic value capture from innovation, but not representative of the Innovation Economy writ large. I’ll explore other innovation industries in future posts, but just know that I recognize not to generalize too broadly here.
A final reminder that the purpose of this series is to better understand where we exist “objectively” in order to maximize the total scale of innovation. Whether I’ve precisely or accurately assessed the “why” questions has zero impact on the massive gender gap that presently exists in innovation. As long as this gap persists, we cannot hope to maximize innovation in this country or beyond.
So how do we (start to) close these gaps? Let’s save that for next time.
Here we get into the persistent challenge of rates (%) versus absolute (#) when discussing equity, so I’ll do my best to keep these clear throughout the piece.
Note that the entrepreneur value is a bit lower than we established in Part 1 - somewhere between 15-25% of founders seeking funding, and ~20% who actually receive funding - but the directionality here is consistent.
A tough look to include such a gendered quote, but such is life.
I'll be spending a lot more time examining methods for achieving better funding outcomes through various application processes throughout the course of this newsletter.