Welcome to the Jungle
New York City, 1968. Busy street. A group of people have stopped on the sidewalk to look up at a window on the 6th floor of the adjacent office building. No one is in plain sight. No one is trying to make a jump. Nothing. A crowd is forming on the sidewalk. In fact, a crowd is forming at a surprisingly high speed.
When a group of 15 people or more stops to look up like this, 40% of passersby will join in even without knowing “what we’re looking at.” Psychologists, Milgram, Birkman, and Berkowitz conducted this experiment (“Note on the drawing power of crowds of different size” in Journal of Personal and Social Psychology, 1969) nearly 50 years ago, and its results have been confirmed several times since. Our propensity to conform and follow the crowd increases with the size of the crowd.
A JOKE (?) Two men are in the jungle when they see a lion running towards them. Frantically, one of the men starts putting on his running shoes. Surprised, the other man says “What are you thinking, you can’t outrun a lion!” The other replies: “I don’t have to outrun the lion… I just have to outrun you.”
As cynical as it may seem, being a part of a herd is a brilliant survival tactic for any prey animal who cannot outrun its predator. We can even take it one step further; if you are quicker on your feet than the one running next to you, the best strategy is not even to divert from the herd and try to out-manoeuvre your predator, rather, you should follow the herd and trust that the predator will follow too.
This herd mentality runs deep in all of us. Especially, in the face of danger and uncertainty we follow troop. Even when what’s at stake is far less critical than in the joke above. When you leave a restaurant because it’s empty to stand in line for the one next door, you are tapping into your herd mentality: There must be a reason this restaurant is empty; someone else must know something; if I mimic the actions of others I don’t have to know why this decision is the right one! Of course, herd mentality exposes us as well, and our history is full of instances where collective actions are horrifying and downright inhuman (?)
Finance and crowds
In financial markets we don’t always see the actions of others, but we do see their impact on prices. The concept of financial speculation (such as day-trading) is built on the premise that you can profit from following trends as long as you are quicker on your feet than the rest of the followers. This can in extreme cases result in what economists call speculative bubbles; prices that get inflated and in turn burst because of the huge disconnect between price level and the underlying economics. This is equivalent to the point where the crowd on the sidewalk in New York suddenly decides to leave… it happens instantly because everyone has been waiting for someone else to make the first move.
In context of (investment) crowdfunding, the concern about herd behavior is particularly grave. Startup investing (and private placements in general) are risky, and the due diligence is more or less left to the investor. For that reason, herd behavior, speculation, and manipulation are central to the discussion.
In “Crowdfunding: Geography, Social Networks, and the Timing of Investment Decisions“, professors Agrawal, Catalini, and Goldfarb find that funding propensity increases with funding levels; i.e. a campaign becomes attractive when others have found it attractive. This is not to say that it is attractive because others have found it attractive, but simply that the timing is such that you as a fundraiser will have a simpler task once you have traction.
There are a couple of important questions to ask in this context, so I reached out to Prof. Christian Catalini, from the MIT Sloan School of Management, to learn more about how herd behavior impacts investment decisions.
Christian: In our paper, we focus on the role different types of information play in crowdfunding investment decisions. The capital accumulated by a project is a signal to other investors that the project might be of higher quality and therefore might be more likely to succeed.
In relying on the amount raised, investors are using a “shortcut” instead of performing their own due diligence. Whenever the amount is a good proxy for quality, then herding might not be irrational at all.
The key issue is to understand how much information is really contained in the capital a particular project has accumulated.
Kevin: Are there any ways where we can learn whether funder propensity is in fact driven by a herd mentality?
Christian: In the data, we noticed that local investors were more likely to invest early, and did not exhibit the herding pattern of distant investors. As it turns out, the difference was driven by individuals who likely had access to offline information about the entrepreneurs: i.e., their family and friends.
If investment by family and friends increases with the quality of a project, then the crowd is making good investment decisions, as it is “learning” from people who are more informed. At the same time, it is easy to imagine situations where family and friends invest because they want (or have) to show support. Moreover, under such a scenario, founders who have social networks that can mobilize less capital would be at a disadvantage.
As online platforms evolve, their market design is also improving, allowing them to surface information more effectively. This, in turn, makes herding less likely and makes capital allocation more efficient.
Kevin: The fact that we sometimes follow the lead of others, simply because we believe in their decisions, can of course have disastrous consequences. But couldn’t we think of situations where such behavior is beneficial? — At the individual level and/or as a society as a whole?
Christian: If crowdfunding platforms create the right incentives for individuals who possess private information to reveal it, and for individuals who have the right skills to perform due diligence to conduct it, then the crowd can rely on these “experts” and avoid duplicating their effort. Experts could be rewarded with equity, other types of rewards or by establishing a reputation system that they can then use for other purposes (e.g., to get a job).
Anyone who has attended a pitch event e.g. hosted by an angel group will have seen how investors influence each other; tech savvy investors are followed if the particular offering calls for her expertise etc. Is this herd behavior, bystander-effect, information cascading or something completely different?
Herding is common also among sophisticated investors, such as VCs and business angels, but it would be wrong to consider it always irrational. Every time you ask other investors to follow you in a deal, your reputation is at stake, and investing is a repeated game. Early-investment might even generate a self- fulfilling prophecy: let’s imagine there are two startups of identical quality, A and B, raising a round to enter a hot new market. Let’s also assume that by pure chance A receives more attention, and a first group of investors decides to invest in it, building momentum around A’s fundraising campaign. Once the deal is subscribed by a substantial amount, it becomes optimal even for investors that would have chosen B to converge around A, as A is now more likely to receive funding, and receive it earlier. If fundraising momentum generates path-dependency, you could even imagine a case where the worse company gets funded, and it would be very hard to tell it even happened, as the other company might not get funded at all.
Kevin: Do you see any opportunity for crowdfunding platforms to mimic “angel group meetings,” or would that require face-to-face interactions, reputation in place, and other mechanisms?
Christian: Yes, and this is exactly where we have been focusing our research more recently. Online syndication, as implemented by platforms such as AngelList, creates the right incentives for a syndicate lead to evaluate startups, perform due diligence, and mentor founders in exchange for a part of the upside. This allows the online crowd to make investment decisions based on the reputation and past deal performance of established investors, the syndicate leads. Incentives are aligned as leads invest on the same terms of the crowd and put a tangible amount of their own capital, i.e. “have skin in the game”. Unlike a VC, online syndicates does not profit from a recurring management fee, but only from a successful exit.
As platforms develop, matching between capital and great investment opportunities is likely to improve, and specialization is likely to increase: whereas a small crowd of experts could screen deals and contribute to the success of the startups involved, the rest of the crowd could provide the bulk of the capital.
Kevin Berg Grell, PhD is CEO at APEN Designs, a next generation presentation tool for startups and SMEs. Responsibilities include investor and industry relations, financials, and strategy development. Besides this venture, he consults for private enterprises on the implications of crowd dynamics, online investing (incl. crowdfunding and P2P), and Web 2.0 driven business models. Kevin has a PhD in Finance and MSc in Mathematics and Economics. He also practices capoeira and archery, but rarely simultaneously.