Biased decisions in Venture Capital — smart or crazy?

by Eva Gfrerer and Tanja Jänicke

In our last articles we referred to several biases in VC decision making. Then people asked: Are those biases always bad? Excellent question — let’s have a look.

Decision biases in a nutshell 🥥

First, let’s level the playing field by establishing a common understanding of what we actually mean by biases.

Decision biases are systematic deviations from rational decision behavior. For a long time economists have been modeling decision makers as rational, utility maximizing individuals. In the 1950s, psychologists and behavioral scientists began to fundamentally dispute these assumptions.

By analyzing the behavior and decision making in economic experiments, researchers found that biases and heuristics are not occasional. Individuals act systematically biased. Skeptics of Behavioral Economics argued that decision errors are randomly distributed with a mean of zero and therefore cancel out. That is not the case (Thaler, 2016). Tversky & Kahneman (1974) and many researchers after them showed in numerous experiments that those errors are systematic and predictable and therefore must be accounted for in human decision behavior.

Do biases always lead to bad decisions? 📉

A large part of the literature says: yes. Biases and heuristics are bad and impede our decision making behavior. Other researchers, such as Gerd Gigerenzer or Robert Auman, promote that biases and heuristics are also good. They are the result of evolution and make our lives easier, reduce complexity, save us from overthinking and decision paralysis and lead to better decisions after all.

So biases in Venture Capital are smart, right? Well, unfortunately no.

How do biases impede VC decision making? 💵

One of the most obvious biases in venture capital is the similarity bias.It describes that decision makers in general prefer individuals who are similar to them. Empirical research shows that venture capital investors have a strong preference for start-up founders similar to them, e.g. regarding education, professional background or even the way they think (Franke et al. 2006, Murnienks et al. 2011) .

This makes sense in part because the fact that another person went to the same school or had a similar job gives us an information edge. We know the school and the job, so we know what people learn there. As an economist would put it: choosing people similar to us reduces information asymmetry.

On top of that, it appears that also personality-wise there are unmissable similarities between VC investors and most of their portfolio companies. So we were curious and conducted a personality study with exciting results confirming our hypotheses: The results suggest that investors prefer founders with a similar personality structure. Does that make sense?

It does. It is deeply human to look for people who are similar because it is easier to get along. Now, is “getting along” relevant for venture capital decisions? It seems like — yes.

Human connection and chemistry ❤️

“The last, and arguably most important, thing is the real human connection and chemistry you share with this person.” This quote does not describe how to find a partner to fall in love and raise a family with. Here, a VC investor gives advice about how founders should select a VC to invest in them. And similar advice is given vice versa to investors. This is odd because no one in a bank would promote the idea to allocate loans based on the question whether there is a “real human connection”. Or to deny a client to buy car insurance because “there was no chemistry”. Venture Capital is the only field in financial services where it is about emotions and network, ski trips and friendship. But this is not an objective metric to select investments because then similarity bias easily kicks in.

All eggs in the same basket? 🥚

With regard to VC decisions the similarity bias is harmful because it extremely narrows our set of choices. It leads to homogeneous portfolios which contradicts one of the most important rules of financial investment: don’t put all your eggs in one basket.

In reality, VC investors put nearly all their investment eggs in the baskets of young white men from top universities because they are mostly white men from top universities themselves. The baskets of female founders, non-white founders and founders who did not attend a top school stay empty. One common reaction from VC side is to point fingers to the pipeline or funnel problem. It describes that fewer women seek VC funding and therefore investors do not have enough female talent in their investment funnel. Empirical data shows that this is not true (Ackerman, 2021). Interestingly, female VC partners invest in 2x more female founding teams (Kauffman Fellows, 2020). That shows there is enough female talent in the funnel.

From a purely financial point of view the “one-basket-only approach” means that venture capitalists miss out on a lot of opportunities and therefore leave potential returns on the table.

They also deny opportunities to a large group of individuals due to their subjective evaluations. The gatekeeping function of VCs prevents large segments of the population from realizing their entrepreneurial potential (Antretter et al. 2020, PlanBeyond 2020).

This is not fair and also not smart because more diverse investments lead to better performing funds (Deszö & Ross, 2012, Peni et al., 2014). As various studies have shown, investors who are capable of controlling for their biases achieve significantly higher portfolio returns, compared to those that are not (Antretter et al. 2020).

At MorphAIs we believe that great founders make great decisions. And that greatness cannot be merely measured by subjective metrics, or “gut-feeling”. As described above, this quickly backfires. So we want to allocate funds based on entrepreneurial talent not similarities. We use technology to debias our decisions to make more accurate investments, achieve stronger overall fund performance, and by doing so, create a more inclusive start-up ecosystem. Something that is long overdue.