Interview with Professor Gilles Chemla

The following interview discusses the paper entitled “Learning through crowdfunding” by Chemla G. & Tinn K. (2020) published in Management Science, Vol 66, pages 1783-1801.

1. Your model shows that firms could learn from crowdfunding. How do firms actually benefit from crowdfunding?

There are several types of crowdfunding, such as reward-based crowdfunding, which is the most famous type with the platforms such as Kickstarter and Indiegogo, security-based crowdfunding, lending based platform and equity-based platform. We distinguish between two. Reward based crowdfunding is a way for firms to learn about demand of products. Let’s say you come out with the prototype, you have a brand-new product, but you are not quite sure how popular it is going to be. On reward-based crowdfunding, essentially future consumers can pre-order the goods typically at a discount. If you see a lot of people pre-order your prototype, it means it will be a lot of demand; then only when you have a lot of demand, you can start the production process. If you do not have a lot of demand, you know that it is not working. On the other hand, it fits well with the literature on entrepreneurship as experimentation. It is very worthwhile to have entrepreneurship because people can learn with new projects and then we figure out some of the project will be very successful and some will fail. With reward-based crowdfunding, you could actually experiment and learn about consumer demand without even having to spend on investment. You only have to spend on production process if there is enough demand. So, there is a real option argument. However, when you have security-based crowdfunding, people invest in equity instruments. You do not learn so much about consumer products but only on your securities. It can lead to booms, bust and bubbles, so we claim that it is much less stable.

2. We are aware that crowdfunding is subject to moral-hazard problem. What is your finding with this regard?

In fact, entrepreneurs can actually crowdfund a project, get the money and take the money and run. But if you look at Kickstarter, more than 97% of the project that are successfully funded end up delivering the products. So, in equilibrium, that does not seem to be a big problem with money diversion because they are genuinely attempt to produce the products. With the 97% success rate means we do well to eliminate more moral hazards. In fact, crowdfunding platforms are not legally liable in delivering the product, everybody has the best effort and duty but we cannot really commit. So, how does it work? It seems that the way that crowdfunding campaigns work do well to eliminate moral hazards. Essentially, we claim that it starts from “All-or-nothing” scheme where you only get the money as entrepreneur if the campaign succeeds. First, you announce that target, for example 50K dollars and only if you raise at least 50K dollars that you get the money. If you get 45k dollars with the crowdfunding campaigns then the money returns to the backers. So, this 50k is not going to be just enough for the NPV of the project to be positive but also enough so that the NPV of continuation for the entrepreneur is higher than the value of taking the money and run. You actually place a target that is higher than what you need to break-even. Sometimes, entrepreneurs still completed the project even though they did not get enough funding because they were really trying to test the demand.

3. As most successfully projects are followed by the successful campaigns, what is the relationship between the crowdfunding outcomes with the sample size?

Indeed, it brings the broader question on how important it is to have large sample.  There is a tendency to have big data. Data is a new goal and everybody loves to have big data. First, there is a number of researches show that if you just use the big data without thinking the valid meaning, you can get things completely wrong, even in the causal inference problem. In this paper, what we see is that you do not want your sample to be so large, what you want is actually enough demand left after the crowdfunding campaign, so that the value of continuation is sufficient enough. If you have a fixed number of consumers overall, you want to moderate sample sizes during the crowdfunding campaign so that there will be big enough consumer demand after the crowdfunding campaign. For Kickstarter, they decided to cut the period from 3 months to 1 month. In fact, there are big companies, such as Sony, Apple and large tech companies to crowdfund on their own platforms. But there could be credibility and commitment problems that they could always decide to select the crowdfunding campaign as they do not have big enough demand. We state that third party platforms do have more credibility. In fact, some large established companies that are cash cows, do crowdfund new devices and new products. For example, ZTE crowdfunds new cell phone on Kickstarter. So, it is very consistent with the idea that you do crowdfund project on Kickstarter to find out what is the consumer demand.

4. Is your model consistent with existing empirical findings?

It is consistent with the observation that Kickstarter shortened the campaigns. It is also consistent with the idea of another working paper by Ethan Mollick from Wharton that surveys entrepreneur from Kickstarter. He asked the entrepreneurs what are the main reasons for them to crowdfund projects and learning about consumer demand. It is also very consistent with the outcomes that the crowdfunding prices in which you pre-order products are lower than the retail prices. But then you can also see from new empirical predictions that 61% to 65% of the projects successfully crowdfunded on Kickstarter are innovative consumer products. These are typical products where the most important problem is to find out the demand. It may be not so difficult to come out with the prototype, but the main question is whether it is profitable as a project to market, and then form the platform event.

5. In your opinion, can crowdfunding bakers make wiser decisions than those professional investors?

In the wisdom of crowd arguments, population can make wiser decisions than professional investors. Of course, it goes back for the democracy, the people make wiser decisions than head of government. In consumer demand, consumers make their decision based only on their own value, so they do not need to form belief about other consumer values. In security-based crowdfunding, your decision will depend on others’ belief. This is much unstable as the crowds will make decision subjects to informational cascades and will probably not make the best decision. From the lending-based platform literature, indeed, bakers can predict credit score better than professional. So those backers can really make very sound decisions. But recent working paper from Mohammadi finds that actually they underperform professional investors. Investor banks have large teams that specialize in valuation, there are professional finance investors and teams of experts that working together, so it is very hard to outperform those teams. Even for venture specialists, they are professional investors but sometimes cannot breakeven on most of the project they invested. It is very difficult to outperform these professional investors.

6. Any other important implications from your paper?

In our paper, we do not focus on lending-based crowdfunding and the equity-based crowdfunding platforms. We also have a survey article coming out from the Palgrave-MacMillan Handbook on Alternative Finance to look at the investor-based crowdfunding. What we know is that the market where you have a huge booming platform, there were regulatory backlash. It seems to be less stability on security-based crowdfunding than in reward-based crowdfunding. That could be informational cascades and firm’s behaviour problem. On security-based crowdfunding you would have more bubbles, what we expect from this line of research is that when security-based crowdfunding is going to be more unstable with boom and bust, some regulatory oversight is inevitable. Reward-based crowdfunding is actually going to be more stable.

Gilles Chemla*, Lee Wan Ling

* Professor Gilles Chemla is a Professor of Finance at Imperial College Business School and co-director of Imperial's Centre for Financial Technology. He is also a research fellow at Centre National de la Recherche Scientifique, a research fellow at Centre for Economic Policy Research (CEPR), a senior research fellow at the Rimini Centre for Economic Analysis (RCEA), and a member of the American Finance Association, American Economic Association, Western Finance Association, and European Finance Association. He currently serves as an Associate Editor at the Journal of Empirical Finance.
Close Menu