When Similarity Becomes a Problem: The Dark Side of Homophily

Can Similarity Be Harmful?

The tendency to surround yourself with similar people — homophily — is one of the best-documented findings in relationship research. But this inclination has a dark side: too much similarity can undermine the quality of decisions, restrict access to new information, and foster groupthink.

A large-scale study in venture capital finance provides particularly strong evidence (Gompers, Mukharlyamov & Xuan, 2016). The authors examined over 15,000 co-investments by venture capitalists and found: investors who share the same ethnic background, attended the same university, or worked at the same employer collaborate significantly more often. Yet this similarity-based collaboration is associated with markedly worse investment outcomes. The probability of success drops by 11.5% for shared alma mater, 22.5–26.4% for former colleagues, and up to 32.2% for shared ethnic minority membership.

Echo Chambers Among Friends

The mechanism is what matters: the worse outcomes are not because similar partners select worse projects (selection effect). Even when controlling for the ex-ante quality of investments, the negative effect persists. It is a treatment effect: similar partners make worse decisions after the investment. They are prone to social conformity pressure and groupthink, while more heterogeneous teams benefit from diverse perspectives, knowledge bases, and more critical questioning (Williams & O’Reilly, 1998).

This finding transfers to personal friendships: when friendship networks are too homogeneous, they can restrict access to information. Bridging relationships with diverse people provide new perspectives and resources. Excessive homophily produces redundancy — you hear the same opinions, confirm each other, and miss alternative viewpoints. The diminishing returns of multidimensional homophily point in the same direction: similarity on too many dimensions at once delivers less added value than expected.

The Groupthink Effect

The study by Gompers et al. (2016) was published in the Journal of Financial Economics and leverages the unique data landscape of the venture capital sector: every investment has a clearly measurable outcome (IPO = success). The authors use instrumental variable approaches and counterfactual pairs to strengthen causal claims. Williams and O’Reilly (1998) provide the theoretical framework: diverse teams benefit from a variety of perspectives, improving creativity and problem-solving. The transfer to personal friendships is supported by Granovetter’s (1973) theory of the “strength of weak ties,” which shows that distant, dissimilar contacts offer the greatest information gain.

The study focuses on professional collaborations, not personal friendships. The success metric (IPO) does not transfer to friendships — friendships primarily serve emotional support and belonging, not decision optimization. Furthermore, research on perceived similarity shows that the feeling of being alike increases emotional closeness in friendships — a benefit not captured by a purely economic analysis. The negative consequences of homophily thus mainly affect the informational and decision-making level, not the emotional function of friendships.

The VC Study

The current state of research on this aspect is summarized below.

When Similarity Helps

The current state of research on this aspect is summarized below.

Frequently Asked Questions

Can a too-similar friend group be harmful?

The tendency to surround yourself with similar people — homophily — is one of the best-documented findings in relationship research.

What is an echo chamber?

A large-scale study in venture capital finance provides particularly strong evidence (Gompers, Mukharlyamov & Xuan, 2016). The authors examined over 15,000 co-investments.

When does similarity become a problem?

The study focuses on professional collaborations, not personal friendships. The success metric (IPO) does not transfer to friendships — friendships primarily serve emotional support, not decision optimization.

How can you get more diversity?

The mechanism is key: the worse outcomes are not because similar partners select worse projects (selection effect). Even controlling for ex-ante investment quality, the negative effect persists.