Monday, May 28, 2012

Believe in yourself, for my benefit.


In his Stanford CS183 class on tech startups, Peter Thiel asks the age-old question: Do entrepreneurs succeed because of skill, or because of luck?

He also gives the correct answer: Since every new company is different, we can't do statistical analysis to predict a startup's success or failure from past data (in econometrics jargon, startup success is non-ergodic). We know for sure that some skill is involved in the mix; we also know for sure that some luck is involved in the mix. How much skill and how much luck, though, is something we'll never really know.

So here's the question: Should a prospective entrepreneur believe in skill, or luck? And my answer is: Depends on who's asking! From an individual standpoint, it could go either way. But there are reasons to think that even if belief in skill is wrong, it might benefit society! Here's why.

If you believe in skill, you're more likely to be an entrepreneur, because you'll be more convinced that you can beat the market. Which is to say, a person who believes she is more skilled than the average will only bet on that belief if she thinks skill matters. If you think entrepreneurial success is mostly just down to luck, then what you should do (as Thiel points out) is to diversify your portfolio and manage your risk. Go get a nice stable high-income job and put your savings into index funds. I.e., don't be an entrepreneur at all.

This fits with economic theory. Economist Frank Knight famously described a difference between "risk" and "uncertainty". Risk is when you don't know the outcome, but you know the probabilities of the outcomes (e.g. flipping a coin). Uncertainty is when you don't even know the probabilities of success and failure (e.g. starting the first social-networking company). Entrepreneurs, he said, must accept not just risk, but also true uncertainty, since no one really knows how likely a new business is to succeed. This is especially true of tech entrepreneurs.

If you believe in skill (and if you have skill), you are more likely to accept true uncertainty, because you probably think that your guess regarding your probability of success is better than other people's guess. In other words, you believe that your knowledge of your own skill is inside information, so instead of diversifying your portfolio (i.e. insuring against risk), you concentrate your investments in one big bet on your inside info.

So will believing in skill be good or bad for you? It's tough to know. In some eras, there are lots of good opportunities just out there for the taking (e.g. when the internet was just being invented), but in other eras, the available ideas are more scarce, and you just never know which era you're in (it also differs by sector).

But some people argue that even when believing in your own skill is a bad bet from an individual standpoint, it might be good for society to have lots of people who believe (even irrationally!) in their own skill. The reason, essentially, is this: There is always new stuff out there, and if society is going to progress, somebody needs to discover it. If nobody is confident in their own skill, nobody will try any new stuff; each person will limit her exposure to true uncertainty, but society will suffer. In other words, there is a positive externality to exposing oneself to true uncertainty.

Blogger and CTO Matt Sherman puts it this way:
The process of going from nothing to something...is inherently irrational...To embark on it is to leave the world of economic modeling...[P]rogress requires madness, that is, the freedom to pursue choices whose rationality can’t be measured.
Economists Antonio Bernardo and Ivo Welch actually have a model that describes how this could happen. From the abstract to their paper:
By ignoring the herd, the actions of overconfident individuals ("entrepreneurs") convey their private information. However, entrepreneurs make mistakes and thus die more frequently. The socially optimal proportion of entrepreneurs trades off the positive information externality against high attrition rates of entrepreneurs[.]
Same basic idea!

So there are theoretical reasons to believe that overconfident entrepreneurs are more likely to benefit society than they are to benefit themselves. If you try to be the next Mark Zuckerberg, chances are you're taking on a lot more risk than you would take if you weren't such an overconfident nut. But without overconfident nuts like you, our economy will be poorer and our society will be more stagnant.

So, by all means, young entrepreneur, believe in skill! Believe that you, not the person sitting next to you, will be the next tech billionaire. You're probably wrong. But the chance that you might be right is enough to make your quixotic madcap adventure worth it for the rest of us.

26 comments:

  1. Anonymous3:41 PM

    As an overconfident nut with lots of ideas , I fully expect that you and your readers will send me lots of money.

    You know , for the benefit of society and all.

    ReplyDelete
  2. http://www.hks.harvard.edu/fs/drodrik/Research%20papers/selfdisc.pdf

    Dani Rodrik says something similar, but argues just subsidising new products is the way to go rather than encouraging a culture change or asking people to be "irrational":

    In the presence of uncertainty about what a country can be good at producing, there can be great social value to discovering costs of domestic activities because such discoveries can be easily imitated. We develop a general-equilibrium framework for a small open economy to clarify the analytical and normative issues. We highlight two failures of the laissez-faire outcome: there is too little investment and entrepreneurship ex ante, and too much production diversification ex post. Optimal policy consists of counteracting these distortions: to encourage investments in the modern sector ex ante, but to rationalize production ex post. We provide some informal evidence on the building blocks of our model.

    I think being slightly delusional is quite common among entrepreneurs anyway, in a good way.

    ReplyDelete
  3. Anonymous8:03 AM

    In some ways, the risk associated with betting on yourself is inherently lower than the risk associated with betting on others.

    For example, I am 100% certain that I will not run off with my own money. I don't know that the same hods for you.

    ReplyDelete
    Replies
    1. OTOH, what are the relevant probabilities on you blowing the money by screwing it up, vs others?

      Delete
  4. "The process of going from nothing to something...is inherently irrational...To embark on it is to leave the world of economic modeling...[P]rogress requires madness, that is, the freedom to pursue choices whose rationality can’t be measured."

    This is interesting when - take this quote from Knight:

    "To begin with a general abstract answer, it will be evident to anyone with a rudimentary understanding of economic processes and analysis that profit (always in the sense of pure profit) would be absent under the conditions of equilibrium with “perfect competition,” (which may be defined in more than one way). The”tendency” of the competitive processes of buying and selling and the control of production is to impute the whole product to the productive agencies which create it, leaving nothing for entrepreneurship as a distinct function (except for monopoly gain, referred to below). This means that under the conditions of ideal equilibrium (stationary or moving) the function of entrepreneurship itself is entirely absent from the economy."

    Could it be possible that economic models effectively assume away entrepreneurship - why would a rational utility maximising individual start a firm under uncertainty?

    ReplyDelete
    Replies
    1. Could it be possible that economic models effectively assume away entrepreneurship

      Yes they most certainly do! People have tried to build entrepreneurship into models, and no one really paid attention...just like no one pays attention to most of the interesting features that people throw into models from time to time...

      Delete
  5. "He also gives the correct answer: Since every new company is different, we can't do statistical analysis to predict a startup's success or failure from past data (in econometrics jargon, startup success is non-ergodic). We know for sure that some skill is involved in the mix; we also know for sure that some luck is involved in the mix. How much skill and how much luck, though, is something we'll never really know."

    Noah, this is a big bunch of not-much. Technically, *everything* is different from everything else. You entering a Ph.D. program is different from anybody else, even in the same program.

    The question is how much different?

    ReplyDelete
    Replies
    1. The point is, skill and luck play differing roles in differing activities at different times, and in any activity, at any time, the relative contributions of skill and luck (and their interaction!), not to mention different kinds of skill and different kinds of luck, will be immeasurable.

      Delete
    2. It is a shame you Americans are not familiar with cricket. Anybody who is, can't have problems being aware of how both skill and luck affect outcomes massively in outcomes where positive feedback effects dominate. The very best batsman will often score less than the very worst in any given innings, but will score on average many times as much. Now imagine there is only ONE innings.

      Delete
    3. Noah: "The point is, skill and luck play differing roles in differing activities at different times, and in any activity, at any time, the relative contributions of skill and luck (and their interaction!), not to mention different kinds of skill and different kinds of luck, will be immeasurable."

      Again, this applies to a whole lot of stuff. And if you took this seriously, it seems to me that it'd chop economics down by a lot (much deserved, perhaps).

      IIRC, this was been looked at in the Silicon Valley side, with many VC's preferring people who've started a business, even if it failed.

      Delete
    4. IIRC, this was been looked at in the Silicon Valley side, with many VC's preferring people who've started a business, even if it failed.

      That could militate in favor of either skill or luck, couldn't it! ;-)

      Delete
    5. This comment has been removed by the author.

      Delete
  6. Phil Koop8:43 AM

    I don't think you need to encourage entrepreneurship, whether by economic or social means (dulce et decorum est ...) Humans are naturally overconfident nuts. It is sufficient to remove obstacles and reduce the attraction of alternatives. So, no stigma to failure, social safety net, crack down on rent-seeking whenever possible = innovation.

    ReplyDelete
    Replies
    1. There is something missing from this discussion. The fact that "Austrian" economists (or rather their followers) trumpet BOTH entreneurship AND moral hazard. Surely they can't have it both ways once you introduce the concept of uncertainty. Their economic heroes must be accept heavy punishment for failure, but risk failure anyway. This is taking irrationality to new heights.

      Delete
    2. "crack down on rent-seeking whenever possible = innovation"

      I'm not sure that this is not a contradiction. Surely rent of some sort is exactly the reward that risk taking entrepeneurs are looking for. But on the other hand, locking up intellectual property rights reduces the raw material of new ideas that entrepeneurs have to work with. This needs some original thinking - only Dean Baker seems to take this major issue seriously.

      Delete
  7. Let's not exaggerate though. A lot of times entrepreneurs have good -- or very good -- evidence that their skill and specialized knowledge in an area is far better than average, or the relative comparison group. And they may go in with a great deal of evidence that their idea will work, relative to the payoff from working.

    ReplyDelete
  8. Anonymous, where do I send the check?

    ReplyDelete
  9. Larry Headlund3:46 PM

    Why are the only two factors in entrepreneurial success skill and luck? Are these two terms being used in a special way I don't know about?

    To point to an example from current events one of Mitt Romney's sons has started a hedge fund (entrepreneur). He started with a $10 million investment from his mother and has attracted a number of other investors. He says that investors were attracted by his proprietary investing strategy (skill).

    Is the only alternative to his tale of skill is that after starting the company he just got lucky?

    ReplyDelete
    Replies
    1. His luck was in having a mother with $10 million to spare and a rolodex full of those in similar station.

      Tagg started Solamere with Spencer Zwick, national finance director for Mitt's 2008 campaign. Who do you suppose invested in his fund?

      Neither had significant experience in the word of high finance.

      "In their prospectus, they said they expected to reinvest the money in an array of equity funds, each of them also managed by Romney political supporters."

      http://articles.boston.com/2011-10-07/news/30254612_1_spencer-zwick-mitt-romney-equity

      A fund to invest in other funds -- Damn, those guys must have some gonzo skills!

      JzB

      Delete
  10. Ken Galbraith, back in the 1950s, called to over-optimism of entrepreneurs the "convenient xocial vitrue." Man could coin a phrase...

    ReplyDelete
  11. One of my college roommates always envisioned being a successful entrepreneur (and probably will be one day). His natural way of thinking led to frequent innovative ideas as potential business opportunities. My natural way of thinking is to focus on the operational aspects of any idea. What is the market? How costly is the product to produce? What are reasonable expectations for profit margins? This combination of personalities led to a constant back and forth of creative ideas versus practicality of actually running that business.

    In my opinion (hopefully his too), our debates were ultimately very beneficial to both parties. He taught me not to get bogged down in the details and I helped him to probably avoid wasting time on profit-less ideas. The point is, between my roommate and I, he is far more likely to believe in his skill as an entrepreneur and take those risks. While there is certainly a role for the micro-thinkers, our society clearly benefits from those individuals willing to pursue out-of-the-box ideas. For all of our sakes, I hope that he and many others will give entrepreneurship a run (although I obviously hope my old roommate is the next Zuckerberg).

    http://bubblesandbusts.blogspot.com/2012/05/noah-smith-believe-in-yourself-for-my.html

    ReplyDelete
  12. Anonymous11:47 PM

    We know that societies benefit from various forms of irrational "exuberance". Investment and entrepreneurship are risky on the level of the individual but are vital to group success. It's actually best if most individuals are somewhat ignorant of the risks that they are taking.

    However, at the current time the risks associated with entrepreneurship and investment have become so disproportionately great that both have stalled. Companies see no good opportunities for expansion so rather than investing in capital goods, etc. they simply sit on mountains of cash. Health care costs are so unaffordable that starting a business could very well result in premature deaths in one's family. Myriad huge government subsidies to established businesses make competition from new entrants impossible.

    ReplyDelete
  13. Noah - I'm not sure how it could relate to luck. But assuming there is a learning curve in starting a business, it probably relates to a particular skill. That is why I wanted to talk about cricket. In cricket the probability of getting out, decreases the longer you have been batting (hence positive feedback). Lots of very small scores, a few very big scores. Entrepeneurship is a bit like that. Economists are used to dealing strictly with negative feedback systems, because they are used to assuming equilibrium or tendencies to equilibrium.

    ReplyDelete
  14. To me, there's a difference between thinking I can beat the market as an investor and being an entrepreneur, unless I'm entering to compete in an already-established business (I'd argue Facebook wasn't one of those, it was creating a new idea).
    In most entrepreneurs' case they see a product or service that is in demand but not being provided. They know there are people interested in buying that product or they wouldn't start. They then step in to fill that void. Now, HOW LONG their profits last can very much be more a matter of luck than people tend to write books about, I agree. They may get lucky and not have someone eager to step in and compete. They may get lucky and be the first to file a patent for their product and thus reap a much longer benefit. But the initial impulse to provide something the market currently is not doesn't seem so random.

    ReplyDelete
  15. "If you try to be the next Mark Zuckerberg, chances are you're taking on a lot more risk than you would take if you weren't such an overconfident nut."

    ... however:
    1) Building Facebook required an awesome accumulation of skill (even if Zuckerberg's triumph was 95% luck). Even if FB flopped, he'd have an meaningful story to tell potential employers and a powerful set of skills that any tech manager would value. In contrast to Zuckerberg, most undergrads spend their time day-dreaming, drinking... or worse, prepping for GRE exams so they can get into a good grad school. (Case in point: One of my best friends founded an internet start-up, flopped, but was thereafter immediately hired for a fat 6-figure salary.)

    2) You seem to be assuming that "success" is defined primarily by compensation. However, most tech entrepreneurs seem to have very different preferences than the rest of us e.g. a preference to build a disruptive/society-changing technology OR a preference to become famous (which most moderately rich people are not) OR a preference for equity-like compensation (in lieu of a fixed salary).

    But thank you for your blog Noah! Awesome selection of topics and analysis.

    Dave R,
    University of Michigan
    Masters in Applied Economics Candidate

    ReplyDelete
  16. I discovered your web site via Google while looking for a related subject, lucky for me your web site came up, its a great website. I have bookmarked it in my Google bookmarks. You really are a phenomenal person with a brilliant mind!

    ReplyDelete