The Cost of Friendship

Interesting if perhaps not so surprising: in a new working paper called “The Cost of Friendship,” Paul Gompers, Vladimir Mukharlyamov, and Yuhai Xuan argue that even in as performance-based an industry as venture capital, people tend to collaborate with people who have similar backgrounds, often to their detriment:

This paper explores two broad questions on collaboration between individuals.  First, we investigate what personal characteristics affect people’s desire to work together.  Second, given the influence of these personal characteristics, we analyze whether this attraction enhances or detracts from performance.    Addressing these problems in the venture capital syndication setting, we show that venture capitalists exhibit strong detrimental homophily in their co-investment decisions.  We find that individual venture capitalists choose to collaborate with other venture capitalists for both ability-based characteristics (e.g., whether both individuals in a dyad obtained a degree from a top university) and affinity-based characteristics (e.g., whether individuals in a pair share the same ethnic background, attended the same school, or worked for the same employer previously).  Moreover, frequent collaborators in syndication are those venture capitalists who display a high level of mutual affinity.  We find that while collaborating for ability-based characteristics enhances investment performance, collaborating for affinity-based characteristics dramatically reduces the probability of investment success.    A variety of tests show that the cost of affinity is not driven by selection into inferior deals; the effect is most likely attributable to poor decision-making by high-affinity syndicates post investment.  Taken together, our results suggest that non-ability-based “birds-of-a-feather-flock-together” effects in collaboration can be costly.


"Ain't no use talking to me/It's just the same as talking to you." Bob Dylan.


I would assume that choosing to work with someone for "ability based" reasons, at least easily measurable ones, like graduating from a top university isn't mutually exclusive with choosing to work with someone who's your friend (or for affinity based reasons). Basically, working with a friend doesn't preclude working with someone who has good resumes or achievements. And in fact, if you're working with other venture capitalists, that is already a pretty strict screen for some kind of achievement.

So, I wouldn't think about it as choosing friendship over achievement hurts performance, which seems fairly straightforward. Instead, I would think about it as: what is it about working with your friend that makes your investments perform worse?

My first thought is that even in the competitive environment of VC, you might hold back criticism in investment decisions, which could certainly hurt performance (if not feelings). Or it could be the effect KevinM pointed out, that people who share affinity characteristics might lack a diversity of viewpoints that could lead to better investment decisions.



I would argue in the affinity based, you have a pair that is cut from the same clothe so you end up thinking a like and no one to look at a deal at a different angle or perspective. Considering if you came from the same firm before, you were trained the same way and same method of doing something. Valuating companies is an art not a science, so you definitively need multiple perspectives.

Voice of Reason

It's all abouts supply/demand. The demand (your need for an employee) stays the same, but the supply artificially decreases if you decrease the pool of applications by a filter as arbitrary as you friendship with said person.


While I applaud any effort to reduce the opacity around an industry/market like venture capital or private equity, this study doesn't look at the right variables when evaluating investment successes/failures and then goes onto mix correlation and causation.

Most fundamentally, it seems to suggest that investment outcomes (whether positive or negative) are basically wholly driven by investors. And while nobody to-date has developed an algorithm to predict startup success (at least an algorithm that works), it is clear that any such model will have to consider a great deal more than just the investors.

There is value in understanding co-investment relationships from the perspective of understanding if these relationships enable certain firms to access proprietary dealflow, i.e., fish in a better pond, but simplistically suggesting that co-investment relationship characteristics are the predictors of success is naive & misguided.

We do think tools to evaluate co-investment relationships in the VC area are important and recently looked at Sequoia, Accel and Kleiner Perkins co-investment relationships. Might be of interest to those who want to understand the venture ecosystem's web of relationships: