
Joel Waldfogel on Markets, Choice, and the Tyranny of the Market 11/12/2007
What this covers
This is an EconTalk conversation in which Russ Roberts interviews Joel Waldfogel about his book on markets and choice. Waldfogel's starting point is that when fixed costs are large and sellers cannot perfectly price-discriminate, a product is made only if its revenue covers its cost. There is therefore a range of market sizes where a product is worth more than it costs but still goes unmade. He adds that the same logic can run the other way, with too many products, because an entrant ignores the revenue it diverts from rivals. Whether markets get it right then becomes an empirical question. Roberts accepts much of the mechanism but pushes on how common it is, and he draws a line between merely lacking a product, such as a kosher butcher in a small town, and being actively harmed by people unlike you.
The evidence is mostly from media. Waldfogel's central case is newspaper readership by race, where a larger Black population raises readership among Black readers and a larger white population raises it among whites, but each group's presence lowers the other's. He also describes how the New York Times' national rollout drew off cosmopolitan readers and pushed local papers to reposition. The discussion ties newspapers to John Sutton's work on sunk-cost quality competition, and it touches on radio, movies, the internet and cable, and the digital divide. Roberts presses the point that technology and the long tail steadily widen variety, and Waldfogel concedes this for non-local content while holding that local media stay constrained.
The last stretch moves to policy and method. They discuss Essential Air Service, postal rates, and French limits on Hollywood imports, with Waldfogel cautious about prescribing and Roberts raising public-choice objections to efficiency-based subsidies. Roberts also questions how useful efficiency is as a standard, and he argues that models of firms handle quality competition poorly. Waldfogel closes with a forecast that machine-collected purchase data could bring a wave of empirical work.
Joel Waldfogel argues that in industries with large fixed costs, markets reproduce a version of the 'tyranny of the majority': what you can buy depends on how many others share your tastes, and sometimes more people with different tastes actively make you worse off. Russ Roberts largely accepts the mechanism but argues it is not widespread, that technology and innovation steadily erode it, and that political remedies are unlikely to improve on market outcomes.
- With high fixed costs and imperfect price discrimination, socially valuable products can go unprovided, so the presumption that laissez-faire markets are efficient fails in theory.
- Empirical evidence shows negative cross-group effects: newspaper readership among Black readers falls as the white population grows, and the New York Times' national rollout pushed local papers to reposition.
- Roberts counters that low marginal costs and bundling widen variety, that technology such as the internet, cable and the long tail fills niches, and that subsidies such as Essential Air Service are driven by politics rather than efficiency.
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Even in principle, it is hard to know whether such subsidies are justified rather than merely politically attractive. Estimating the relevant surplus requires knowing where the demand curve meets the axis, but the high-value demanders are not observed in price-quantity data. Waldfogel agrees that conventional econometrics struggles here, and notes he once used surveys of valuations to address this kind of problem.
“how would you ever get the information you need to know whether these subsidies are justified as opposed to just politically attractive”
Under fixed costs, markets can also provide too many products. An entrant cares only whether its revenue covers its costs, not whether that revenue is diverted from rivals, so large markets can have excessive product variety. Theoretical presumptions that leaving markets alone yields the right answer therefore fail in both directions under fixed costs, making the outcome an empirical question.
“I enter as a seller if the revenue I get covers my cost I don't mind at all if the revenues were diverted from another firm so there can be a tendency in large markets to actually have too many products”
Newspaper markets stay concentrated as they grow because quality (more reporters and editors) is produced through fixed costs, not marginal costs, and because readers largely agree on what makes a good paper. A firm can therefore raise quality and price to undercut slightly differentiated rivals, since being different brings little benefit. This draws on John Sutton's argument that industries where quality is produced through sunk costs need not fragment as they grow.
“it can make a product quote better but I better I guess I mean more reporters more editors right higher fixed cost but no higher marginal costs”
There is a key distinction between two situations. In the first, people who share my preferences help me get products while people with different preferences don't hurt me; this is a 'who benefits whom' effect that is only remotely like tyranny. In the second, people with different preferences actually reduce what is available to me, which is the true analogue of the tyranny of the majority.
“there are other people who share my preferences help me but people with different preferences they don't hurt me yeah that's a key distinction”
The digital divide, the lower internet connection rate among African Americans, is smaller in places where Black residents are a smaller fraction of the local population. This suggests the internet serves as liberation, giving access to information and products that is more valuable where offline variety targeting the group is scarce.
“that digital divide is actually smaller in places where blacks are locally a smaller fraction of the population which really can be interpreted as the internet kind of as the liberating opportunity”
Technology has made non-local, non-geographically-specific information abundant for every interest group, through the internet and hundreds of cable channels in many languages. Local information, however, is not resolved this way, so fixed-cost tyranny still operates for local media.
“I think that's particularly true for non local information”
Movies are a mixed case. Technological costs of making and distributing films have fallen sharply, producing fragmentation at the fringes, while the cost of the most expensive movies keeps rising. The blockbuster end still conforms to the fixed-cost logic because it pays to invest heavily in films targeted at mass taste.
“at the same time that we've had the cost of the most expensive movies continuing to go up we do have a bunch of a bunch of kind of fragmentation around the fringes”
Industrial organization offers interesting characterizations of quality competition, notably John Sutton's. But with much consumer heterogeneity and many products, easy welfare statements are unavailable: policies create winners and losers rather than being simply good or bad.
“when you have a lot of heterogeneity of people and lots of different products I'm not sure that easy welfare statements are available”
The evidence for a true tyranny-of-the-market effect is newspaper readership by race. A larger white population raises the share of whites who read and a larger Black population raises the share of Blacks who read, but the cross effects are negative: more people of one group reduces readership among the other. More people of one type actually harms consumers of another type.
“the punchline here is that the cross effects are actually negative the more there are of that group the less it is that this group reads”
Standard theory of the firm takes quality as given and treats quantity as the key decision, which is counterintuitive to business students. For modern firms, quantity is one of the least interesting decisions. The biggest decision is which market to be in: a firm chooses its competitors by choosing quality and differentiating features, and this is the essence of modern competition. Economic models describe it poorly despite work on hedonic markets by Sherwin Rosen, Kelvin Lancaster and Gary Becker, which has not yielded very testable models or welfare conclusions.
“for a firm and a modern economy choosing quality is much more important than choosing quantity”
The textbook perfect competition model is a decent way to capture market forces. Its main flaw is not being too rosy, but being unable to capture how innovation changes markets over time. Complaints about market failure are largely static, because markets themselves often work to correct those failures.
“my biggest complaint about undergraduate perfect competition models isn't that they're too rosy it's that they're not they can never be rich enough to capture that innovative tendency over time”
In markets with large fixed costs, a consumer gets what they want only if many other people share their preferences. Markets therefore do not fully escape the 'tyranny of the majority' that critics attribute to politics, contrary to Milton Friedman's claim that each person can 'vote' for the tie color they want and get it.
“when there are big fixed costs of providing things meaning that a lot of expenditure has to be made by the firm before even the first you know the first widget comes off the line that we can only get what we want if a lot of other people also want what we want”
Because machines now collect vast amounts of detailed data on who buys what, economics may stand on the brink of a renaissance of empirical work that can inform debates about how markets perform.
“a lot of things end up getting collected by machine lots of data exists a lot of you know detailed information about who's buying what I think we we could stand on the brink of a renaissance of empirical work”
Although a newspaper is a bundle of articles that can include content for different groups, some product positioning is not portable. Language is either English or Spanish, and cities with different demographics get papers targeted differently, so a paper cannot appeal to everyone just by adding chunks.
“some aspects of the product though isn't entirely portable I mean to take an extreme example language you know it's either in English or in Spanish”
The standard view that French restrictions on Hollywood imports are crazy, because imports simply add options, is too simple in a high-fixed-cost world. Allowing Hollywood spectacles in may shrink French cinema and change its targeting, leaving some French film lovers worse off. This alone is not an argument for protection, but it shows that trade's effects on differentiated high-fixed-cost products are nuanced.
“as you allow imports into France you might see French cinema shrink and change its targeting which is going to leave some french in a file maybe not me but some folks worse off”
Empirically, as the New York Times rolled out nationally it siphoned off college-educated, cosmopolitan readers, and local papers responded by repositioning toward more locally oriented readers. Importing a high-fixed-cost product therefore does not simply add an option to the existing set; it produces winners and losers through a lumpy, 'wrenching' repositioning.
“as the New York Times rolls out and becomes available in more and more markets what happens is that the local papers reposition so the New York Times siphons off these college-educated kind of cosmopolitan readers”
Daily newspapers are the extreme case. A typical US city has essentially one metropolitan daily, and that single product must choose whom to target. The more people who like one thing, the more it targets them and the less appealing it becomes to others, which is genuinely analogous to political tyranny of the majority.
“a typical US city has one daily newspaper”
Markets offer a product when the seller's revenue exceeds its cost. Only with perfect price discrimination does revenue equal total consumer valuation. Since perfect price discrimination is essentially impossible, there is a range of market sizes where social value exceeds cost but revenue does not, so products that ought to exist don't get made. This is a standard result from 1970s–80s industrial organization theory.
“it selects the product where the revenue available to the seller exceeds the cost of providing them”
Estimating whether a product's total value (the area under the demand curve) exceeds its fixed costs requires knowing the valuations of high-value demanders. These are not observed in price–quantity data, so conventional econometrics struggles to locate where demand hits the axis. Waldfogel agrees and notes he once used valuation surveys to address this.
“it's hard using conventional econometrics techniques to learn something about how that of where the demand curve hits the axis”
Economic efficiency is a utilitarian concept: add everyone's benefits, subtract costs, and treat people as interchangeable. Roberts, a Chicago-trained economist taught efficiency and deadweight loss as central, has become less enamored of it with age and now spends almost no time on it in his micro class.
“it's a very utilitarian concept and you know I'm a Chicago grad I was taught efficiency at the heart of all of my graduate training and I've gotten less enamored of it as I've gotten older”
Roberts raises two public-choice objections to efficiency-based subsidies. First, they may reflect political power, such as influential senators or grandfathering after deregulation. Second, once instituted, the political process ignores efficiency and focuses on redistribution. Beyond that, policymakers cannot obtain the information needed to know whether a subsidy is justified rather than merely politically attractive.
“anytime you Institute these kind of subsidies the political process is going to ignore efficiency I would argue and focus mainly on the opportunity to redistribute”
Minorities in non-diverse places will find no Hispanic radio station, Afghan restaurant or kosher butcher until a critical mass exists. This is not a tyranny of the market, because adding people unlike you does not make your life harder.
“you have to get to a critical mass to support the Hispanic radio station or the Afghan restaurant or the kosher butcher”
Waldfogel is wary of proposing new policy based on theory, even with empirical support. Instead he points out that existing government subsidies already look designed to address fixed-cost underprovision in small markets: air service subsidies, some public broadcasting subsidies (though many don't fit), and telecommunications subsidies where demand density can't cover fixed costs. The policy question is whether to abolish these on laissez-faire grounds or recognize that they may enhance efficiency.
“I always get nervous when people you know propose new new policy initiatives based on theoretical ideas even with empirical support”
Subsidies like Essential Air Service and the postal service's uniform rates are cross-subsidization, which is nice for the subsidized and not for everyone else. An alternative explanation is political power: Mason City's representatives and senators may be powerful, though here the subsidy reflects historical grandfathering after deregulation. Once subsidies exist, the political process tends to ignore efficiency and focus on redistribution.
“a lot of cross subsidization is the technical term right where a fixed rate a flat fee is charged for everyone even though they don't have the same costs”
Economists know too little about how real markets work because the profession theorizes about a narrow spectrum of choices that the math can handle, rather than richer dynamic pictures. Industrial organization in the 1950s and 60s studied how outcomes vary by market size and number of firms in a primitive and not particularly effective way, then moved away from it without much progress. Waldfogel's book is valuable for actually looking at what happens.
“we spend a lot of time theorizing about what might happen in a very narrow spectrum of choices because that's what the math works on”
Under the Essential Air Service program, small inconvenient cities such as Mason City, Iowa (reportedly about $1 million per year to Mesaba Airlines) receive subsidies for air service. Eligibility was grandfathered to cities served under airline regulation that lacked commercial entry after deregulation, eligibility ends once commercial entry occurs on its own, and roughly 50 cities participate.
“Mason City gets a subsidy of a million dollars a year so that Mesaba Airways or Airlines finds it profitable to serve them”
Roberts argues that because advertising revenue matters greatly relative to subscription revenue, and an extra reader costs little, newspapers have strong incentives to broaden their audience. Diversity then shows up within the paper rather than across many providers: bigger papers, wider coverage, more reviews, and neighborhood editions, such as the Washington Post's area-specific sections.
“because the advertising revenue is such an important part of the total revenue relative to the subscription rate... their incentives to widen the audience very very widely”
Blockbuster movies and books carry high fixed costs, largely marketing and star salaries, with reverse causation: a costly movie must attract a large audience. Yet a vast, varied niche market coexists. About 300,000 books were published in America last year, many aimed at specialized audiences, and falling movie production costs make reaching small audiences financially viable. Roberts is therefore skeptical that fixed costs matter much in the French film story.
“America last year 300,000 books were published”
Because advertising revenue is a large share of newspaper revenue relative to subscriptions, and extra readers are cheap, papers have strong incentives to widen their audience. Diversity then appears within the paper rather than across papers: bigger papers, wider coverage, more reviews, and neighborhood-specific editions such as the Washington Post's regional sections.
“because the advertising revenue is such an important part of the total revenue relative to the subscription rate or the individual purchase revenue their incentives to widen the audience very very widely”
The real reason newspapers and radio don't fragment with market size is not high fixed costs per se. It is enormous economies of scale from the very low marginal cost of serving additional consumers, such as printing an extra copy.
“isn't the real difference the fact that additional consumers are relatively cheap it's not so much that there are high fixed costs but that there's enormous economies of scale”
The political process often imposes a single uniform level where nothing inherent requires it. Examples are the uniform national postal price and public schools, though schooling can be supplemented with private lessons or private schools. In such cases the tyranny of the majority is a political choice rather than an inevitability.
“the price is uniform across all areas and that's a political decision it need not be that way”
The two economists mostly agree that firms seeing profit opportunities are how products reach people and are most of the solution to market failure. Their disagreement is about how much the stuff markets don't provide could be improved upon, and whether the political process would take the right steps to do so.
“or that if it could be improved on whether the political process is likely to make the right the right steps in the right direction”
If preferences over newspapers were highly heterogeneous, for example if New York had 20 languages each spoken by 5% of the population, the newspaper market would fragment into slivers, each resembling a small-town market.
“if you know in New York there were 20 different languages spoken by each by 55 percent of the population in New York would presumably fragment”
There has never been a better time to be curious about the world. Market performance should be measured by the information and variety available to, for example, a Hispanic consumer, not by the number of newspapers, and by that measure the variety markets produce is extraordinary.
“I don't think there's ever been a better time to be curious about the world than today”
The perfectly competitive model, which takes about half the semester in many courses, is a neat model of homogeneous products and a good intellectual benchmark and first approximation. Besides dynamics, it misses product differentiation, so relying on it obscures important features of real markets.
“it's a neat model but it is a model of you know homogeneous products”
Essential Air Service appears to target exactly the markets theory predicts the market would not serve. In principle, with price-quantity data one could estimate demand curves and compare the area under them with capturable revenue and fixed costs to evaluate such subsidies.
“interestingly it looks like it's targeting the markets that are that they're you know described by what the theory would suggest markets wouldn't serve”
Radio is a moderate fixed-cost example. A typical US city has about 20 stations; many cities have no stations targeting certain listener types, while larger cities have enough 'preference compatriots' to support programming for many listener types.
“a typical US city has you know maybe 20 stations on the dial for some types of listeners you know many cities have no stations targeting that type of listener”
If big fixed costs relative to market size cause problems, the natural solutions are reducing costs or enlarging the market. Trade is the main way to enlarge markets, and falling transportation costs for physical goods, plus the internet for virtual goods, have clearly expanded consumer choice.
“the natural solution to that is either reducing the cost or increasing the market size and trade is this the first”
No economist would deny that fixed costs reduce choice; for example, you can't buy a car without a back seat because tailoring every car is costly. Through innovation and competition, consumers nonetheless get far more of what they want than 100 years ago. The real question is whether anything should be done about it.
“if you don't want to backseat with your car you don't have that choice”
Restaurant variety grows dramatically with city size (Fergus Falls to Minneapolis to New York City), with more cuisines, styles and price ranges. Newspapers do not show the same explosion: New York has only a few more papers than Minneapolis.
“New York City does have a few more papers in Minneapolis but nothing like the explosion of restaurants”
Cross-subsidization, such as a flat postal rate charged to all despite differing service costs, benefits the subsidized at the expense of everyone else. It is acceptable only insofar as we care about the beneficiaries.
“a lot of cross subsidization is the technical term right where a fixed rate a flat fee is charged for everyone even though they don't have the same costs”
The tyranny-of-the-market argument applies only to markets where fixed costs are an important part of production. For goods like neckties, production is roughly constant returns to scale, so consumers get abundant choice.
“this is not about all markets about markets where fixed costs are an important part of the process of production”
Some people being unhappy with market offerings is no cause for action and can be entirely efficient. If it costs $1 million to create something worth $10 in value, it shouldn't be created, and the person willing to pay $10 ought to be unhappy.
“if it costs a million dollars to create something that has $10 in value you know it shouldn't be created”
If a few people value an Afghan restaurant enormously and everyone else values it little, a single price may not cover the restaurant's fixed costs, even though total benefits exceed costs. Fixed costs combined with the inability to price discriminate cause a forgone net gain.
“a single price might not be able to cover the costs of that the fixed cost of the restaurant even though some of these people would get enormous value out of it existing”
Some goods are inherently provided at a single level for everyone, such as missile defense or the size of the US Army. For these, whether provision is public or private (private provision being hard because of free riding), the tyranny of the majority is unavoidable.
“it's very difficult to get two levels of national defense two levels of an army right there's just there's one US Army we all are going to get that same level”
An alternative explanation for French restrictions on American movies is rent-seeking to enrich French filmmakers. Waldfogel agrees this is often the case, and says the fixed-cost argument only complicates the picture for differentiated, high-fixed-cost products.
“the alternative explanation for why French the French government restricts imports of American movies is to enrich French filmmakers”
Someone wanting a sophisticated blockbuster will likely be disappointed because blockbusters cater to the common denominator. Outside it, however, there is an explosion of choice: as Chris Anderson's 'long tail' describes, technology has lowered inventory costs at Amazon, iTunes and other web retailers, so even the smallest niches get filled.
“what I find incredible is the explosion of choices outside the common denominator”
Technology has lowered the cost of stocking inventory at Amazon, iTunes and other online sellers. As Chris Anderson's 'long tail' argument holds, even the smallest niches get served in a market economy, while sophisticated blockbuster lovers are left to the mass common denominator.
“technology has lowered the costs of stocking inventory at Amazon and iTunes and other places on the web so in many ways the smallest niches are still filled”
In many cities, many weekly papers target specific audiences, but the daily market is different: there are very few targeted dailies, and Spanish-language dailies exist in only a few cities.
“in many cities there there are a lot of weekly papers targeting specific audience the daily market tends to be a little different”
Roberts accepts the negative cross-effect mechanism. Low marginal cost of adding readers pushes a white-oriented paper to add more white-oriented content as the white population grows, making it less interesting to Black readers; fixed and marginal costs interact here. He doubts the phenomenon is widespread beyond newspapers.
“that low marginal cost of adding readers is going to push a white oriented newspaper to add more white stuff”
Waldfogel describes his contribution as modest. Much is known about government allocation failures, but little empirical evidence exists about how markets produce tyranny-like interdependencies among consumers. He aims to add data to the market side of the ledger, so people can decide whether they are happy with how markets work.
“less has been said about based on empirical evidence about how markets entail similar kinds of features”
Even a critic of market optimism should recognize that markets provide the lion's share of the solution to fixed-cost choice problems.
“I still think markets are the lion's share of the solution to these kinds of problems”
Roberts questions how many undergraduates are actually taught a purely rosy view of markets as opposed to the market-failure view. He describes himself as a big proponent of markets and private initiative who does not claim markets are perfect.
“I wonder how many undergraduate economics majors are taught the purely rosy view that you suggest as opposed to the alternative view which is the market failure view”
Introductory remark that the guests hold very different views and a lively discussion is expected.
“you and I hold very different views on the subject so I hope we'll have a lively discussion”