When VCs Invest Together, Does Ability or Affinity Matter?

I woke up to a bunch of VC related things in my twitter stream this morning. I had a nice digital sabbath yesterday so I was a little surprised by how much there was. I tried cranking out a #tweetstorm of them using Little Pork Chop but I found the tweetstream experience to be very unsatisfying and very inauthentic feeling. The links are good, so here they are if you want to get in the headspace for what I really want to talk about.

1/11 Things I Read About VC This Morning I Think You Should Care About In A Compact Little Tweetstorm
2/11 Start with @fredwilson thinking about tweetstorms – http://avc.com/2014/06/tweetstorming/
3/11 Then @msuster on why VC is so much more compelling now – http://bit.ly/1mvIE5C
4/11 and @pmarca on why the IPO is not what it used to be – http://bit.ly/1ljhzlV
5/11 and congrats to @jeff on raising his new fund – http://bit.ly/1m0h6cD
6/11 thx @joshelman to the pointer to the @yoapp hackathon – http://bit.ly/1x0MhbQ
7/11 the #premoney conference recordings will be online soon – http://www.livestream.com/500startups/folder
8/11 the 2nd seed round trend @Mattermark by @DanielleMorrill – http://bit.ly/1iQTCI2
9/11 I end with Haiku
10/11 Tweetstorms perplex me a lot
11/11 Do you enjoy them

The response to 11/11 was generally “no” although a few people suggested that tweetstorming while a soccer game was going wasn’t a particularly useful test.

After I thought I was done I ran across a really interesting set of articles which didn’t make it into the tweetstorm. The first article, In Venture Capital, Birds of a Feather Lose Money Together, was a summary that let to the second article, The Cost of Friendship, which led to the actual article behind the annoying SSRN paywall. After reading the abstract, I decided to buy and read the article, especially since Paul Gompers, one of the great academic researchers on the VC industry, was the lead author.

I was once a Ph.D. student at MIT Sloan School studying innovation. Specifically, my doctoral advisor was Eric von Hippel. Eric was very kind to me, but I was a horrible Ph.D. student because I was also running a company at the time and had no interest in being an academic. Eventually I got kicked out well before I got my Ph.D.

Nonetheless, I learned how to more or less read an academic paper and some social science rubbed off on me. Actually, a lot rubbed off on me – enough for me to know that the headlines written about academic papers and studies rarely capture the essence of what is going on in the paper. Instead, reading the abstract and the carefully reading the non-analysis part of the paper, with a goal of putting yourself in the researchers’ shoes to understand what they are trying to figure out, will help you understand the punch line.

So when I read the first article, it was easy to conclude “VCs who are like each other do less well investing together.” Or, “VCs who like each other perform more poorly when investing together than those who don’t like each other.” This is consistent with the callout from the first article which says “The more affinity there is between two VCs investing in a firm, the less likely the firm will succeed, according to research by Paul Gompers, Yuhai Xuan and Vladimir Mukharlyamov.”

I read the summary, which is kind of the “PR piece” for the article, but I didn’t find it satisfying. It generalized too quickly and I kept wondering how affinity was defined. The hint was that it had to do with ethnicity, educational background, and employment history, which wasn’t how I was defining affinity when reacting to the title “In Venture Capital, Birds of a Feather Lose Money Together.”

Next, I read the executive summary of the paper. This was clear and felt fine to me. It separated affinity and ability. The punch line of the paper is:

“Collaborating for ability-based characteristics enhances investment performance. But collaborating due to shared affinities dramatically reduces the probability of investment success.”

Much different than the marketing piece about the paper that I read first. Basically, if you choose your co-investor because you think she is a great investor, that’s good, but if you choose your co-investor because you like him, that’s bad. But that felt too simple to me – no way that’s the basis for a HBS academic study. So I bought and read the paper, which was pretty easy until I got stuck in analysis stew on p.22. I hung in there and got through it, but once again was reminded of another reason I was a shitty Ph.D. student – I dislike reading academic papers.

I learned that affinity was narrowly and precisely defined, but not in the way I thought it was. Affinity to me meant that the two VCs liked each other, or had an “affinity” for one another, but instead affinity was based on biographic data, specifically gender, ethnicity, educational background, and employment history.

“The education dummy variables Top College, Top Business School, Top Graduate School, and Top School equal one if a venture capitalist holds, respectively, an undergraduate, business, graduate, or any degree from a top university and zero otherwise. Ethnic Minority takes the value of one if a venture capitalist is East Asian, Indian, Jewish or Middle Eastern. Dummy variables East Asian, Indian, Jewish and Middle Eastern pin down a venture capitalist’s ethnicity; the dummy variable Female identifies an individual’s gender.”

Also, success was defined as a company having an IPO (the data range for the study was 1975 – 2003). Now, I’m not going to argue the performance variable, but as someone who has had a lot of financial success with exits that were not IPOs, I’d be curious what happens when the analysis is done where success is defined by “at least 10x return on capital for the VC.”

The big reveal is buried in the middle of p.18.

“On one hand, people display greater inclination to work with similar others.  Similarities may be in terms of ability (e.g., whether individuals hold degrees from top academic institutions) or affinity (e.g., whether individuals share the same ethnic background). On the other hand, these two sets of pairwise characteristics affect performance in opposite ways. Teams with more able participants are more likely to result in a successful investment outcome. On the contrary, investments are more likely to fail when groups are formed based upon similarities between members along characteristics having nothing to do with ability.”

Go read that again. If you pair up two people based on ability, they have better results than if you pair them up on affinity, where affinity is defined by “each went to the same school, each are the same ethnic minority (including Jewish), or each worked together in a previous company.”

Unless I missed something (and it’s entirely possible that I did), the message is “choose to work with people who have ability.”

I kind of feel like this applies to life in general!

It’ll be interesting to see how this paper gets interpreted, or misinterpreted over the next few weeks, assuming anyone else goes beyond the summary and reads the paper, no thanks to SSRN.

Just another reminder to look beyond the headlines. And don’t co-invest with someone who has no ability just because you went to the same school, are the same ethnicity, or once worked together.

  • James Khabushani

    very insightful as always. thank you.

  • James Khabushani

    very insightful as always. thank you.

  • http://jorgetorres.com/ Jorge M. Torres

    I read the HBR piece and left it at that. Thanks for taking the deep dive and posting about it. A couple of observations:

    (1) I have to say that it’s dispiriting that the ethnicity variable did not include African Americans, Latinos, or Native Americans. I understand that the omission may have been driven by the practicalities of the venture industry, but, if so, what a bummer.

    (2) Re: ability, devil is in the details. How do we know who truly has investing ability? Are the schools and firms on an investor’s LinkedIn reliable indicators of who has the best cash-on-cash returns, IRRs, and the like? Maybe sometimes, but I suspect most of the time, not. Persistence seems to be the measure LPs use, and this has been recently validated as a predictor of fund performance for venture firms. The rub is that publicly available performance data is in short supply. If you have it, great. If not, then you end up relying on less robust proxies for ability.

    • http://www.feld.com bfeld

      Strongly agree on #2. I was going to go a lot deeper on this in my post but eventually I just ran out of steam.

  • http://jorgetorres.com/ Jorge M. Torres

    I read the HBR piece and left it at that. Thanks for taking the deep dive and posting about it. A couple of observations:

    (1) I have to say that it’s dispiriting that the ethnicity variable did not include African Americans, Latinos, or Native Americans. I understand that the omission may have been driven by the practicalities of the venture industry, but, if so, what a bummer.

    (2) Re: ability, devil is in the details. How do we know who truly has investing ability? Are the schools and firms on an investor’s LinkedIn reliable indicators of who has the best cash-on-cash returns, IRRs, and the like? Maybe sometimes, but I suspect most of the time, not. Persistence seems to be the measure LPs use, and this has been recently validated as a predictor of fund performance for venture firms. The rub is that publicly available performance data is in short supply. If you have it, great. If not, then you end up relying on less robust proxies for ability.

    • http://www.feld.com bfeld

      Strongly agree on #2. I was going to go a lot deeper on this in my post but eventually I just ran out of steam.

  • http://www.startupmanagement.org/ William Mougayar

    Diversity in ethnicity is kind of a bullshit metric. And so old-school in today’s globalization realities.

    I once worked for a CEO who had Indian parents, but was born in Kenya. His wife is Brazilian, of Japanese descent. They have lived in the US for a long time, and have kids born in the US. What ethnicity do you give their children?

    I prefer diversity in Ability. period.

    • http://www.feld.com bfeld

      Yeah – I struggled with that also, especially the use of Jewish as a dummy variables. I’m Jewish, but an atheist, and I can’t think of a situation where I’ve cared what religion, or ethnicity, someone was.

      • ObjectMethodology.com

        Are you sure you’re an atheist? Or are you just wanting to abstain from religious membership?

    • ObjectMethodology.com

      “I prefer diversity in Ability. period.”
      .
      Well put. In the US you have to be politically correct or else you’ll be attacked. It’s funny how people will exhibit behavior worse then the behavior they are acting against. I think that’s called being a hypocrite?

      • http://www.startupmanagement.org/ William Mougayar

        But on the whole, the US/Canada are pretty open minded cultures that favor meritocracy over ‘where you come from’.
        If you want to see the flip side of that, Europe is far more discriminatory in that regard.

  • http://www.startupmanagement.org/ William Mougayar

    Diversity in ethnicity is kind of a bullshit metric. And so old-school in today’s globalization realities.

    I once worked for a CEO who had Indian parents, but was born in Kenya. His wife is Brazilian, of Japanese descent. They have lived in the US for a long time, and have kids born in the US. What ethnicity do you give their children?

    I prefer diversity in Ability. period.

    • http://www.feld.com bfeld

      Yeah – I struggled with that also, especially the use of Jewish as a dummy variables. I’m Jewish, but an atheist, and I can’t think of a situation where I’ve cared what religion, or ethnicity, someone was.

      • ObjectMethodology.com

        Are you sure you’re an atheist? Or are you just wanting to abstain from religious membership?

    • ObjectMethodology.com

      “I prefer diversity in Ability. period.”
      .
      Well put. In the US you have to be politically correct or else you’ll be attacked. It’s funny how people will exhibit behavior worse then the behavior they are acting against. I think that’s called being a hypocrite?

      • http://www.startupmanagement.org/ William Mougayar

        But on the whole, the US/Canada are pretty open minded cultures that favor meritocracy over ‘where you come from’.
        If you want to see the flip side of that, Europe is far more discriminatory in that regard.

  • http://www.thekitchensync.co Laura Yecies

    Thanks for raising this topic. I had read the initial “Birds of a Feather” HBR article as I’m on their email list. While I agree there are some weaknesses in deeper research that you point out and the choice of the word “affinity” is misleading, we shouldn’t lose sight of the major point in this article – we suffer from dramatic homophily in the startup and venture capital world. We have been wringing our hands about women and minorities being excluded – that it is wrong for moral reasons. This article says it’s also wrong for performance reasons and is confirming research that has occurred for public company boards linking diversity with positive performance.

    • http://www.feld.com bfeld

      Yup – we suffer from homophily as a society. But that’s not new information. We’ve spent a ton of time with this issue at NCWIT (http://www.ncwit.org) and have focused on root causes, like unconscious bias, to address the issue.

      There is a lot of much better research that shows that gender diversity is positively correlated with performance. I’m not aware of the research showing that ethnic diversity is positively correlated with performance – do you know of any examples of this?

      • http://www.thekitchensync.co Laura Yecies

        I can’t find anything specific to cite on ethnic diversity and performance unfortunately. Battling homophily is tough – that comfort of being with similar people is clearly a strong innate preference for many people. In addition to focusing on root cause I’m interested in motivating people to fight that preference/bias – seeing a financial and performance advantage from getting out of the “Birds of a Feather” comfort zone could motivate people to insist on diversity.

  • laurayecies

    Thanks for raising this topic. I had read the initial “Birds of a Feather” HBR article as I’m on their email list. While I agree there are some weaknesses in deeper research that you point out and the choice of the word “affinity” is misleading, we shouldn’t lose sight of the major point in this article – we suffer from dramatic homophily in the startup and venture capital world. We have been wringing our hands about women and minorities being excluded – that it is wrong for moral reasons. This article says it’s also wrong for performance reasons and is confirming research that has occurred for public company boards linking diversity with positive performance.

    • http://www.feld.com bfeld

      Yup – we suffer from homophily as a society. But that’s not new information. We’ve spent a ton of time with this issue at NCWIT (http://www.ncwit.org) and have focused on root causes, like unconscious bias, to address the issue.

      There is a lot of much better research that shows that gender diversity is positively correlated with performance. I’m not aware of the research showing that ethnic diversity is positively correlated with performance – do you know of any examples of this?

      • http://www.thekitchensync.co Laura Yecies

        I can’t find anything specific to cite on ethnic diversity and performance unfortunately. Battling homophily is tough – that comfort of being with similar people is clearly a strong innate preference for many people. In addition to focusing on root cause I’m interested in motivating people to fight that preference/bias – seeing a financial and performance advantage from getting out of the “Birds of a Feather” comfort zone could motivate people to insist on diversity.

  • ObjectMethodology.com

    “choose to work with people who have ability.”
    .
    Yes. Work with people who have the ability to get things done. If someone doesn’t put their abilities to work they aren’t of value.

  • ObjectMethodology.com

    “choose to work with people who have ability.”
    .
    Yes. Work with people who have the ability to get things done. If someone doesn’t put their abilities to work they aren’t of value.

  • http://www.preacquaint.com/ Mario Cantin

    I’m not a VC and so I should probably have kept my pie hole shut on this one, but when I read the Carmen Nobel article yesterday after you tweeted about it, I personally felt it was largely bullshit. I couldn’t help it, it was the little voice in the back of my head talking very so softly — but firmly ;)

    I like how you’ve dissected the research paper and made some sense of it.

    At a glance the main problem with the research seems that they can’t even define “affinity”.

    True affinity is based on shared values and understanding, not on superficial social fluff.

    Anyone’s who needs “research”to be told that they must drop their inept school buddy in favor of someone more capable is doomed anyway.

    Personally, it’s academic pedantry like that that makes me not regret having dropped out of college.

    • http://www.feld.com bfeld

      Harsh but sharply appropriate. As I’ve reflected more on the paper, the definition of success (IPO) is also distressingly bad, especially for the 1998 – 2003 time frame.

      • http://www.preacquaint.com/ Mario Cantin

        I don’t typically say harsh things in general, and this would hold true about a paper coming out of Harvard, Stanford or whatever, but to me, this research begged to be challenged.

        The word “affinity” may have been selected for lack of a better word, but the particular definition which can be made to fit the context is too uncommon — and this could be misleading for many readers.

        And the premise seems flawed. I understand the point being made, but what about “chemistry”?

        Try to imagine Pink Floyd with a more “capable” drummer — one with less “affinity” (different school, geography, ethnicity, etc.)

        I don’t think so either.

        And there are other issues as you’ve pointed out (definition of success, etc.)

        Anyway, enough said :)

        I apologize for ranting on your blog.

        • http://www.feld.com bfeld

          Rant accepted! Very appropriate.

        • ObjectMethodology.com

          “Try to imagine Pink Floyd with a more “capable” drummer — one with
          less “affinity” (different school, geography, ethnicity, etc.)
          .
          I don’t think so either.”
          .
          I think the most important part is what drummer can sit down and *make music with the rest of the band*? What makes someone the best drummer is not what they did yesterday! But what they’ll do today and into the future. If someone played country yesterday but decided they want to play rock-n-roll instead they might be the best to choose.
          .
          Yesterday is gone. Today is what exists. Tomorrow may not come.

          • http://www.preacquaint.com/ Mario Cantin

            Agreed. If the drummer quit or needed to be replaced they’d have to move on and it would behoove them to look for one outside of the boundaries of “affinity”, as it is being defined in the article. I’m with you there.

            I was making a point about chemistry though.

            You’re not into Pink Floyd, are you? :)

            (Don’t get worked up, I’m just teasing you a little)

          • ObjectMethodology.com

            I like Pink Floyd. I can quote you some lyrics if you want. Also I like that you opened up with your post. Everyone today says things that seem so canned or fearful. Lacking passion and opinion. Nothing to inspire deep thought.
            .
            I guess I was just thinking that if you pick team members based on yesterday or something written on a piece of paper you’re probably missing out on some really good folks.

  • http://www.preacquaint.com/ Mario Cantin

    I’m not a VC and so I should probably have kept my pie hole shut on this one, but when I read the Carmen Nobel article yesterday after you tweeted about it, I personally felt it was largely bullshit. I couldn’t help it, it was the little voice in the back of my head talking very so softly — but firmly ;)

    I like how you’ve dissected the research paper and made some sense of it.

    At a glance the main problem with the research seems that they can’t even define “affinity”.

    True affinity is based on shared values and understanding, not on superficial social fluff.

    Anyone’s who needs “research”to be told that they must drop their inept school buddy in favor of someone more capable is doomed anyway.

    Personally, it’s academic pedantry like that that makes me not regret having dropped out of college.

    • http://www.feld.com bfeld

      Harsh but sharply appropriate. As I’ve reflected more on the paper, the definition of success (IPO) is also distressingly bad, especially for the 1998 – 2003 time frame.

      • http://www.preacquaint.com/ Mario Cantin

        I don’t typically say harsh things in general, and this would hold true about a paper coming out of Harvard, Stanford or whatever, but to me, this research begged to be challenged.

        The word “affinity” may have been selected for lack of a better word, but the particular definition which can be made to fit the context is too uncommon — and this could be misleading for many readers.

        And the premise seems flawed. I understand the point being made, but what about “chemistry”?

        Try to imagine Pink Floyd with a more “capable” drummer — one with less “affinity” (different school, geography, ethnicity, etc.)

        I don’t think so either.

        And there are other issues as you’ve pointed out (definition of success, etc.)

        Anyway, enough said :)

        I apologize for ranting on your blog.

        • http://www.feld.com bfeld

          Rant accepted! Very appropriate.

        • ObjectMethodology.com

          “Try to imagine Pink Floyd with a more “capable” drummer — one with
          less “affinity” (different school, geography, ethnicity, etc.)
          .
          I don’t think so either.”
          .
          I think the most important part is what drummer can sit down and *make music with the rest of the band*? What makes someone the best drummer is not what they did yesterday! But what they’ll do today and into the future. If someone played country yesterday but decided they want to play rock-n-roll instead they might be the best to choose.
          .
          Yesterday is gone. Today is what exists. Tomorrow may not come.

          • http://www.preacquaint.com/ Mario Cantin

            Agreed. If the drummer quit or needed to be replaced they’d have to move on and it would behoove them to look for one outside of the boundaries of “affinity”, as it is being defined in the article. I’m with you there.

            I was making a point about chemistry though.

            You’re not into Pink Floyd, are you? :)

            (Don’t get worked up, I’m just teasing you a little)

          • ObjectMethodology.com

            I like Pink Floyd. I can quote you some lyrics if you want. Also I like that you opened up with your post. Everyone today says things that seem so canned or fearful. Lacking passion and opinion. Nothing to inspire deep thought.
            .
            I guess I was just thinking that if you pick team members based on yesterday or something written on a piece of paper you’re probably missing out on some really good folks.

  • ObjectMethodology.com

    Ability is what matters. The ability to get things done!

  • http://blog.kwiqly.com/ James Ferguson @kWIQly

    Yup I think you missed something Brad

    Think Venn diagram.

    Intersection = Affinity characteristics,

    Outside Intersection = Non-redundant contribution capacity.

    Two VCs in two helpful scenarios

    a)

    VC One: “So I could introduce you to Martha ”

    VC Two: “So can I” == Chocolate tea-pot

    b)

    VC One – I can’t help with that

    VC Two – but I can == Gold dust

    Differentiation != Affinity

    Commodity != Differentiated

    => Affinity == Dumb money ( market price)

    Differentiation =>. Added value

    Give me a challenge not a bunch of cheerleaders