Survival of the Fittest

The Wall Street Journal reports that `daily deal’ sites like Groupon are dying fast, casualties of the expensive competition for new users. Groupon now spends about $8 to lure one active user.

This looks like a good example of Darwinian competition yielding an inefficient outcome — as we should expect. (See Chapter 8 of my book The Armchair Economist for more on why Darwinian competition is nothing like market competition, and far more likely to yield bad outcomes.) Vast sums are being spent in an arms race with relatively little social value. Surely consumers benefit from all this competition, but it’s highly implausible that they benefit enough to justify such high expenditures. Even after the recent Great Winnowing, about 350 of these sites remain; surely consumers (none of whom have the time to visit 350 sites a day) would be almost equally well served by 50 sites, at about 1/7 the cost.

Most of the time, markets do a spectacular job of allocating resources efficiently. Sometimes they don’t. Often that’s because the market is structured in such a way that one firm (or a very small number of firms) can supply the entire marketplace, which inspires a largely unproductive race to the top. Consumers benefit, but not enough to justify all that resource expenditure.

The market for oranges works well because producers earn rewards commensurate with the social value of the oranges they grow. The market for daily deals seems to work less well (I say “seems” because I might be overlooking something) because Groupon can earn enormous rewards for being just a tiny bit easier to navigate than LivingSocial. This inspires Groupon (and LivingSocial) to invest a lot of resources making minor improvements.

(Sometimes those minor improvements turn into major improvements that are well worth their cost — but that’s not something we can bank on.)

Economists call this the “tournament” problem. It comes up in any market that’s potentially dominated by “superstars”. Athletic competitions, for example, are prone to drain more resources than they’re worth. (See previous blog posts here and here.) Major league baseball is a good thing — it entertains the fans. But if we could randomly eliminate half the players, we’d lower the quality of competition just a little, while freeing up a lot of ex-ballplayers to start businesses, drive cabs, or write sports blogs. The market fails to find that outcome.

The fact that it’s so easy for markets to fail is part of why we should be so astonished, and so thankful, that they typically succeed.


37 Responses to “Survival of the Fittest”

  1. 1 1 Harold

    Evolution does not produce efficient solutions – a quick look at the peacock should convince. That tail may have been effective so far, but is surely not efficient.

    As far as whether tournaments are worth the cost, does it depend on the purpose of the tournament? If the objective is to find the fastest runner in the world (or USA for example), then a slight lowering of the standard results in utter failure in the objective. Rewards must be very high in order to ensure that the best runner is tempted into the competition.

  2. 2 2 Phil

    I don’t get it. What’s the market failure? The fans are happy. The athletes are happy. Who failed? If you want more cab drivers or sports bloggers, then you can go ahead and pay them more. Don’t blame everyone else.

  3. 3 3 Mike H

    Does this also apply to the market for math blog posts?

  4. 4 4 Maurizio Colucci

    Allow me to challenge this view. Let’s focus on the “arms race” in the sports field (athletics, baseball in your examples). Suppose players did that for fun: they train and compete to such an extent because they enjoy it. In this case, they are now consuming, not producing. So in your framework, unless I am mistaken, the inefficiency would disappear (because your framework does not allow you to say that the way people choose to enjoy their time is inefficient — otherwise you would have to denounce inefficiencies everywhere: “if only you did THIS, instead of playing guitar with your friends, it would be more efficient, so this is a market failure”). And yet this situation is observationally indistinguishable from the one you say is inefficient. In fact it is the exact same situation, only your way of VIEWING it has changed. Does this not raise a problem with your framework, or did I misunderstand something?

  5. 5 5 Ben

    Does this mean that basically all idea/thought-based industries (like entertainment, technology, etc) operate in inefficient markets? If so, would you be an advocate for higher taxes on the entertainment, technology, etc industries?

  6. 6 6 Phil

    The only way an allocation of resources can be inefficient in the sense you mean it (correct me if I am wrong) is if some fan can be made better off without making any athlete worse off.

    Can you identify such a fan and such an athlete?

    If so, then you probably have an interesting business opportunity by splitting the net increases in happiness and taking a middleman cut. You will be making the market even better. Good for you!

    If not, then the market has not failed.

  7. 7 7 Steve Landsburg


    The only way an allocation of resources can be inefficient in the sense you mean it (correct me if I am wrong) is if some fan can be made better off without making any athlete worse off.

    You are wrong because there are people in the world who are neither fans nor athletes.

    Nevertheless, there is likely to be such a fan and such an athlete. Remove half the professional ballplayers from playing baseball and set them to making quilts, while paying them what you’re paying them now. Give the quilts to the fans, who will find that the quality of baseball has declined slightly, but that’s more than compensated for by these nifty new quilts they’ve got.

  8. 8 8 Steve Landsburg


    Does this mean that basically all idea/thought-based industries (like entertainment, technology, etc) operate in inefficient markets?

    The key is the ability of one supplier to supply the whole market. This doesn’t apply to circus clowns, since the audience for a given circus is limited by the size of the tent.

    If so, would you be an advocate for higher taxes on the entertainment, technology, etc industries?

    Just because something is inefficient doesn’t mean that an intervention is likely to make things better. Yes, a well calibrated tax, with the revenue wisely spent, can, in principle, bring about an improvement. But why would we assume that in practice, policymakers would calibrate the tax well, or spend the revenue wisely?

  9. 9 9 Steve Landsburg

    Maurizaio Colucci: You are right in principle about consumption/production, but I expect that a lot of the consumption value for these guys comes from being the best (as opposed to coming from the act of playing) in which case we’re right back to that arms race again.

  10. 10 10 Steve Landsburg


    If the objective is to find the fastest runner in the world (or USA for example), then a slight lowering of the standard results in utter failure in the objective. Rewards must be very high in order to ensure that the best runner is tempted into the competition.

    Agreed in principle, but I do not think that’s ever the objective.

  11. 11 11 S.V.

    I think it is the same for academics. For instance, you can randomly eliminate half of the economists. The consequence would be the elimination of half the noise in economic papers.

    The analysis in not restrict to economists.

  12. 12 12 andy

    It seems to me that this is a natural-monopoly argument. In the late 19th century the monopolist complained about ‘cut-throat competition’; there was lot of ‘waste’ building parallel electric lines, water lines… the state started to regulate these as natural monopolies. Now students are taught that the characteristics of the supply curve (economies-of-scale) cause the industry to become monopolized.

    I would have a different look on this: this actually is a natural monopoly situation, only the consumers didn’t synchronize on the monopolist firm yet. In the moment they do synchronize, the one monopolistic provider could drive the costs down and essentially drive all others out of the business; as long as he keeps the prices down. I woldn’t call a state when we didn’t find the best firm yet a “market failure”.

  13. 13 13 Steve Landsburg


    In the moment they do synchronize, the one monopolistic provider could drive the costs down and essentially drive all others out of the business; as long as he keeps the prices down.

    The problem recurs, though, when new entrants invest vast resources in trying to displace the monopolist with a product that’s only a hair better.

  14. 14 14 Al V.

    Prof. Landburg, You are probably right regarding the inefficiency of sports. We give so many subsidies to sports teams (inexpensive or free stadiums, parking, tax incentives, etc.) that I suspect that many teams would not be able to operate without those subsidies. Major League Baseball operates under an anti-trust exemption that allows it to employ anti-competitive practices.

    In a true free market, I would guess that perhaps 25% of Major League baseball teams would go out of business. The NHL and NBA would shrink by half. Only the NFL would be minimally affected – I suspect no teams would fold, but we would see teams relocate into better markets.

  15. 15 15 Al V.

    Similarly, college sports teams operate as tax exempty entities, and don’t have to pay their players. It’s the best of all possible worlds, in one sense. However, it appears that the end result is that college teams expand their costs to consume their revenues, leading to coaching staffs making enormous amounts of money. In many states the highest paid state employee is the state university’s football or basketball coach.

  16. 16 16 nobody.really

    People generally believed, and some still believe, that the business of running a local telephone exchange is a natural monopoly. One way to deal with this market failure is for government to structure a “tournament” – that is, a competition to win a local franchise.

    But what about the business of running the Yellow Pages?

    Whippersnappers may not recall that, pre-World Wide Web, the Yellow Pages were the main source of information for finding business phone numbers. Businesses paid through the nose for the privilege of placing ads in this publication, because it was the virtually the sole resource for people looking to find all the plumbers or daycare providers in the neighborhood. This was a pretty efficient system for users – and a very costly one for advertisers.

    Other books would compete with the Yellow Pages, selling ads more cheaply in an effort to attract firms away from the phone company’s book. The net result would be consumers receiving, and having to consult, MULTIPLE books. This may have been a better deal for advertisers, but a worse deal for customers.

    What to do? Arguably the optimal policy would be to treat the Yellow Pages as a public utility: grant an exclusive franchise for running the Yellow Pages, but then regulate the advertising rates for the benefit of advertisers (and, ultimately, their customers). The fact that the Yellow Pages was often published by a regulated utility lent some credence to this view, and in fact revenues from the Yellow Pages were often attributed to the publishing phone company, thereby subsidizing local phone service.

    But publishing is a matter of free speech, beyond the power of utility regulation. Thus we ended up with the rival phone books.

    I sense that the Web has largely rendered this debate moot. Yeah, rival phone books still exist, but they don’t represent a big drain on consumers who have access to the Web. Now the books are basically a kind of fat ad circular.

  17. 17 17 Roger Schlafly

    I doubt that you could eliminate half the baseball players, and still be as entertaining. There have been baseball strikes, and open discussion of restarting the teams with strike-breakers and minor league players. The consensus is that the fans would not tolerate the decrease in playing quality. The minor league players are just not good enough.

  18. 18 18 Jonathan Campbell

    Look at Major League Soccer (MLS) in the U.S.: You often hear people say things like “MLS is boring since the players are no good.” This is obviously only true in the context of the implied comparison to the European leagues, rather than as an absolute statement – if it weren’t for our knowledge of the European leagues, we’d be happy to believe that our MLS players are incredibly good.

    If we artificially eliminated a bunch of professional baseball players, people would be happy to go on watching baseball only if they had no benchmark to compare it too – to the extent they could either remember the good old days of 450ft home run shots, or catch wind of a higher level of play in Central America or Japan, they’d be unsatisfied I think.

  19. 19 19 Kenny

    I think that the competition between the daily deal sites only seems inefficient if we only consider its purpose to be provide daily deals. I think the ‘purpose’ of the competition – the tournament as you put it – is to search among the Vast number of possible _ways_ to provide daily deals. When you consider that _that_ function necessarily involves the preferences of suppliers, consumers, and the dynamic networks connecting the two, it’s less clear that there’s a _more_ efficient way to perform that same function.

    I’m also puzzled why you think this “Darwinian competition” isn’t market competition. Isn’t it, if anything, a specific sub-category of market competition? What are the necessary and sufficient conditions for ‘market competition’?

  20. 20 20 Jimbino

    Justification for market competition does not require that it perfectly efficient, just more efficient that all the alternatives, given the external constraints.

  21. 21 21 Steve Landsburg

    Jimbino: agreed, 100%.

  22. 22 22 Seth

    “But if we could randomly eliminate half the players, we’d lower the quality of competition just a little, while freeing up a lot of ex-ballplayers to start businesses, drive cabs, or write sports blogs. The market fails to find that outcome.”

    I read this and the other two post and it’s not clear to me why the market isn’t finding this outcome, or maybe why it would be a better outcome. Who would these superstars play? The Washington Generals?

    Or maybe I don’t understand why level of competition is the only social value of a marginal sports player worth considering.

  23. 23 23 mobile

    Surely the market could be served just as well with 10 firms instead of 350. Or maybe it’s 20 firms? Or just 3 firms? How to decide how many firms? And which ones? Darwinian competition can decide, and who’s to say that that isn’t the best way to do it?

    To use another baseball analogy, we don’t just hand the World Series trophy to the team with the highest payroll, or the team with the most All-Stars, or the team selected by a due diligence sabermetrics committee in some annex of the Department of Commerce. We play the games and see who wins.

  24. 24 24 Al V.

    Bill James once pointed out that athletic skills are distributed in a standard (normal) distribution across the entire population. Thus, taking baseball as an example, Albert Pujols is at one extreme of the distribution, and an infant or my 90 year-old neighbor is at the other extreme, while I am probably somewhere in the middle. No as talented as Pujols, or the local high school players, but more talented than Mrs. Boone, my wife, or my daughter.

    So if we look at the distribution of talent of players actually in the major leagues, there are very few highly talented players, and very many with just the minimum skills to make it to the majors. Of course, some people with the talent to play in the major leagues may not be in the majors – they may be playing NBA basketball, soccer in Greece, or on Wall Street. But most of the most talented baseball players will be in the major leagues.

    And because of the distribution of talent, reducing the number of major league spots would remove a very few of the most talented players, and largely remove spots for the least talented. Thus, I would expect that the reducing the number of major league players would increase the talent level by a small amount.

  25. 25 25 Al V.

    @Seth, I would argue that the market doesn’t find the outcome because it’s not a truly free market. Sports teams are subsidized in many ways, including discounted access to facilities, tax breaks, and baseball’s anti-trust exemption. If we eliminate the subsidies I expect that many teams would go out of business. As stated previously, I wouldn’t be surprised if the NBA would shrink to 12 to 16 teams.

    One of my personal peeves is the interstate trucking industry. Freight in the U.S. is carried mostly in one of four ways: air express, river barges, semi-trucks, and railroads. Railroads operate at a disadvantage because all of the other methods receive an implicit subsidy via discounted access to their “roads”. Trucks pay fuel and freight taxes, but it doesn’t nearly cover their portion of highway costs; air express companies pay little of the cost of the air transit system; and barges pay virtually nothing for free use of our rivers. Only railroads pay all of their own costs.

    This is a little bit of exaggeration, as there are costs for non-railroad transporters, and hidden subsidies for railroads, but in the largest sense I’m correct here.

  26. 26 26 Silas Barta

    I thought Steve_Landsburg’s post was satire…

  27. 27 27 Harold

    The incomes of “superstar” traders and bankers has rocketed over the last decades – CEO’s also. Top executives earned about 40 times average wages in the 1980′s now its hundreds of times (haven’t actually checked the figures, but the principle is correct.) I think there is a huge loss of value, because these people could be adding to our knowledge in fields of physics, chemistry and mathematics that could allow us to produce new and valuable things. Instead they are being drawn into a profession where they use their talents to skim off a few extra bucks on complex cycles of money circulation. Is this a version of superstar inflation? The companies need to be a little bit better than the competion, so have to pay ever higher and higher.

    There are two points. It may be that the value added by these people in finance is greater than it would be pursuing science. I don’t see it myself. Also, it may be that this is different from “superstar” distortion. which I think is quite possible.

    Personally, I think this is the biggest cost of our emphasis on banking, hedge funds and related “wealth generation”. It is taking away talent from areas of real wealth and life enhancing activities.

  28. 28 28 Seth

    Thanks Al V. I agree with your comments about the market distortions introduced by subsidies, but I don’t think that was Steve’s point.

    I think his point was that the possibility of the big payoff attracts too many folks to the sports-likes markets, even sans subsidies, and some of those marginally good folks could being doing something more productive.

    I guess I can’t visualize how we can know this. It seems the implicit assumption is competitiveness is the only social value worth considering in sports. But there are others. Like having someone for the superstars to play against and to play with and having more cities for those superstars to play in, which is made possible by attracting the marginally competitive ball player into the market.

  29. 29 29 Gabriel

    Isn’t this true of all hot new businesses? Many initially compete, the best survive in the long run, consumers benefit. Often the first firm makes a splash and temporarily yields high profit, then is chased by a large number of competitors trying to cash in and the shift in “profit” temporarily skews to consumers, and then balanced is reached with the better performers providing a high level of service at reasonable margin. The result is often a few large national brands and a number of smaller regional niche players. This applies just as well to pizza as to electronic stores as to clothing stores as to, soon, daily deal sites.

  30. 30 30 Steve Landsburg

    Gabriel: The difference between pizza stores and daily deal sites is that each pizza store can serve only a limited number of customers, whereas a single daily deal site can serve the entire market. If you carefully work through the logic that reveals that market outcomes are efficient (which I am not supplying here, but which is in many textbooks), you’ll find that, for that reason, it applies to the pizza stores but not to the daily deals.

  31. 31 31 Gabriel

    I’m not sure I agree that one deal site can serve the entire market. I’m VP of eCommerce for a national retailer and we’ve run different deals on different sites for different reasons, not the least of which are significant demographic differences in their subscriber bases. This is not that different from online coupon sites, which have been around longer and invariably cater to different niches, from broad to very specific. Or, if you go back further, to mailing list houses for direct mail. Online email lists may have a different technology, but it’s just a modern application of the old direct mail list model, and anyone who’s familiar with either knows that there’s no one-size-fits-all provider.

  32. 32 32 Robert Wiblin

    I don’t doubt your conclusion Steve but can you give a diagram or set of equations demonstrating that agents in a tournament put an inefficiently large effort into improving their products?

    Comparing the basketball case with Groupon seems to confuse things. Beyond some point basketball spectators arguably just want to see the best basketball players (relative to other humans), but it doesn’t matter how good they are in absolute terms. So improving anyone of everyone’s skill doesn’t change our enjoyment – it’s zero sum.

    But the same might not be true of the winner takes all discount site. People really would benefit from a better designed site. And if that site is serving everyone, it’s important that it be well designed! A lot of effort will go in to winning, but with millions of customers the benefits are also large.

  33. 33 33 Doc Merlin

    But the darwinian evolution is the only thing that actually matters. You can’t measure value outside of this context.

    While human rationality allows us to escape /some/ of the traps of evolutionary markets, ultimately the main mode of entreperneuralism is evolutionary. It is throwing metaphorical spagetti up against the wall to see what sticks.

    Furthermore, there is no way of knowing ahead of time, if the results will be good ones or not. We have to wait and find out. Think of it this way. /We/ are the product of such an arms race. The internet itself is the product of such an arms race. The arms races are what allows us to find local maxima in utility, and also to escape those local minima when they aren’t productive optimal enough.

  34. 34 34 Doc Merlin

    5 am is a bad time to respond. My post had a number of typos. The last minima should be maxima.

  35. 35 35 Max

    An example of an arms race which has been getting some attention lately is high frequency trading. Trading profits depend on latency, but shaving microseconds from trading latencies doesn’t really benefit the end users. So the trading firms are spending a ton of money on “improvements” of little, if any, value.

    That’s an extreme case, but it seems to me that races are a common situation. It’s a basic fact of capitalism that relative skill is rewarded, not just absolute skill. Or to put it in a more provocative way, the most skilled people are overpaid.

  36. 36 36 Doc Merlin


    “Or to put it in a more provocative way, the most skilled people are overpaid.”
    How do you measure being overpaid? What does it mean to be overpaid? Are you arguing that their marginal cost – marginal benefit to their employer is negative? I find that unlikely.

  37. 37 37 Steve Landsburg

    Doc Merlin: I think it’s pretty clear that “overpaid” in this context means that *social* marginal cost exceeds *social* marginal benefit.

Comments are currently closed.