Why do monopolies exist? Many textbooks point to barriers to entry as a cause of monopolies.
Tyler Cowen and Alex Tabarrok’s textbook says: “In addition to patents, government regulation and economies of scale, monopolies may be created whenever there is a significant barrier to entry, something that raises the cost to new firms of entering the industry.” Greg Mankiw’s textbook goes as far as to say: “The fundamental cause of monopoly is barriers to entry.”
The fundamental cause seems important. We should probably understand that. In addition to being a common topic in economics education, entry barriers frequently pop up in discussions around antitrust and other regulations, so we should be clear about what they are.
So what are barriers to entry? It’s not so simple.
In this post, I will explain a few of the different ways that people have used the term, the problems with those interpretations, and argue for a different approach along the lines of—you guessed it—the UCLA tradition of price theory.
Bain’s Barriers to Entry
The first somewhat-modern industrial organization (IO) formulation of barriers to entry came from Joe Bain, of structure-conduct-performance (SCP) fame. Bain’s definition focuses on the consequences of barriers to entry in terms of pricing behavior and profitability. According to Bain, barriers to entry exist when established firms can set prices above the minimal average costs of production without attracting new entrants. In Bain’s perspective, the key criterion for identifying barriers to entry is whether prices exceed average costs after entry has ceased. Yes, there may be profits today due to some demand shock that firms do not immediately adjust to, but eventually, without barriers to entry, profits will be driven back down. This is a standard Econ 101 argument.
Bain’s perspective highlights the link between barriers to entry and market power, as established firms can enjoy above-normal profits due to their ability to set prices higher than the competitive level.
While Bain’s definition of entry barriers based on the conditions allowing for an elevated long-run price has intuitive appeal for those comfortable with Econ 101 thinking, his analysis falls short in providing a consistent theory to explain the relationship between identified entry barriers and prices. As Dennis Carlton points out, the problems with Bain’s formulation stem from the general problems with the SCP literature.
The problem with Bain’s analysis of entry barriers comes not from his definition but from the standard SCP failure to properly distinguish cause and effect or to distinguish exogenous and endogenous variables. In his framing, entry barriers—such as scale economies, large capital requirements, and product differentiation—exist out in the world (i.e., they are exogenous) and cause price above costs.
But many of the barriers to entry are endogenous. Product differentiation is a choice by firms. Scale economies are partly a choice, possibly based on past investments. Bain’s reliance on the SCP framework overlooks the dynamic nature of competition, the role of endogenous factors in determining profitability, and the potential for innovation to disrupt traditional barriers.
For example, the basic SCP paradigm really can’t account for the role of innovation. In the SCP framing, the set of products is fixed, and the structure of the market for the specific good directly determines the pricing strategies. In reality, entrepreneurs may identify new business models, technologies, or market niches that enable them to enter and compete effectively, even in industries traditionally characterized by high barriers. In reality, the set of products is not fixed. The production functions and cost functions are not fixed. There’s no accounting for this in the baseline SCP model.
The decision to enter an industry is primarily driven by expectations of profitability, taking into account the entrant’s beliefs about all of these factors changing. A potential entrant assesses the profitability of entering an industry by considering not only exogenous factors (such as costs) but also the competitive landscape, the behavior of existing firms, and potential responses to entry. Profitability is a result of the interplay between these factors, including the ability to differentiate products, achieve economies of scale, manage costs, and respond to market conditions.
From an empirical perspective, testing for Bainian barriers to entry requires running regressions across many industries to try to identify above-average profits. The basic idea is that, if investment can be directed to any market, profit should be equalized across all possible investment opportunities that exist in equilibrium. One problem is marginal returns should be equalized, and real data is always about some average return.
Putting aside the marginal-average distinction, running these regressions is problematic, since the form and the intensity of competition can vary significantly across industries. Again, it does not come directly from the number of competitors. Consequently, industries characterized by vigorous competition sometimes exhibit higher levels of concentration compared to those with less intense competition. There is no simple relationship between barriers to entry, the number of competitors, and the level of profits. It is highly misleading to treat the number of firms as a direct consequence of the entry barriers and run your regressions.
Stigler’s Barrier to Entry
In response to some of these issues with Bain’s definition, George Stigler defined barriers to entry “as a cost of producing (at some or every rate of output) which must be borne by a firm which seeks to enter an industry but is not borne by firms already in the industry.” For Stigler, barriers to entry are about a fundamental asymmetry of costs between incumbents and newcomers. These barriers make it difficult for new firms to enter and compete with established firms.
Stigler’s definition emphasizes the costs borne by potential entrants, including both explicit and implicit barriers—such as capital requirements, brand loyalty, economies of scale, and regulatory constraints. Stigler recognizes that these barriers make it difficult for new firms to enter and compete effectively.
As always, there are some grey areas. Suppose Coca-Cola has large brand recognition, which increases the demand for its product relative to a competitor. Is that a barrier to entry for new competitors who don’t have the same recognition? After all, the new brands would have to buy tons of ads to get the same recognition. But it’s unclear whether it is true asymmetry, since one of the reasons Coke has brand recognition is that it spent billions on advertising over decades. If it stopped doing that or if a challenger matched it, the supposed asymmetry would disappear in the long run.
Stigler’s broader perspective recognizes the complex nature of barriers to entry and their potential impact on competition and market dynamics. In contrast, Bain’s definition narrows the focus to the pricing behavior and profitability of established firms, linking barriers to entry to their ability to sustain prices above costs and earn above-normal profits.
In some ways, Stigler’s definition is a weaker version of Bain’s, simply implying there are fewer barriers to entry in Stigler’s world. For example, scale economies may be a barrier to entry for Bain, but not for Stigler, if another firm has access to the same technology.
Are We All Wasting Our Time on Barriers to Entry?
There are lots of problems with both Bain’s and Stigler’s definitions. Often, thinking in terms of barriers to entry pushes you into thinking in terms of the short run and long run. In the short run, no entry is possible: the types of goods and the number of sellers are fixed. On the opposite side is the long run, where all adjustments have already taken place. The long run has the cost and benefit of brushing aside any dynamics of how markets move toward that new state.
Unfortunately, the long run is unhelpful for many policy question people would have, like “What will happen to prices in two years if this merger goes through?” Instead, we need to think about how costs differ over different time horizons, stretching from the immediate to the long run and everything in between. Entry is more costly tomorrow than it will be in the next 10 years, but it may not be impossible tomorrow. We don’t need the short run and long run, as Armen Alchian pointed out (although not in the context of barriers to entry).
In addition to Alchian’s insight, his UCLA colleague Harold Demsetz was the one to point out the fatal flaws of the SCP paradigm, including barriers to entry.
Demsetz gives the example of a government that issues 100 taxi medallions, which can be sold in an open market. Stigler’s definition would not classify this entry restriction as a “barrier to entry” since outside and inside firms both can access medallions at a market-clearing price. There is no cost differential. Refusing to label it as a “barrier to entry” distorts the use of language. Demsetz argues that using language in an unnatural way can invite confusion in antitrust and regulatory proceedings, where precision and clarity are crucial considerations. And Bain’s definition doesn’t help us either, since the price of the medallion would dissipate profits, so the price is not above the minimum of average costs.
On the other side of Stigler’s definition, are all cost differences considered barriers to entry? Suppose the incumbent has the best owner and CEO. He can’t be hired away like the taxi medallions can. Should we call that a barrier to entry? It’s a cost differential, even though it won’t show up on any accounting book, so Stigler should call it a barrier to entry. To me, it would be highly misleading (at least, for policy purposes) to lump something like that together with government-granted monopolies as barriers to entry.
The reason I’m uncomfortable with calling a better CEO a barrier to entry is that, as Harold Demsetz pointed out, calling something a barrier to entry bakes in the normative conclusion. The implication is that the barrier is bad in some sense. The flip side is labeling something as competition with the implication that it is good. The label is a way to pre-judge something as bad without explicitly saying so.
In an ideal world, none of us would assume a barrier to entry is bad (or competition is good). But we do. For example, in Cowen and Tabarrok’s definition (perfectly fine for Econ 101), they say a barrier to entry is “something that raises the cost to new firms of entering the industry.” That’s perfectly fine for a 101 reader, but I’m sure many students will conclude barriers are bad. After all, high costs are bad! That’s why they are called costs. If they were good, they would be benefits. But raised costs compared to what? It’s not clear in general. In the medallion example, we can imagine a world without medallions, so that’s probably the implicit comparison.
But other cases are not so easy, as Demsetz points out in a 1982 article. Are patents barriers to entry since they raise costs? Probably. The law says you can’t copy my invention, which raises the cost of entering the new industry relative to a world where you could copy my invention.
But are laws against theft barriers to entry? If I’m not the strongest person around, there is probably someone out there for whom, absent the laws against theft, physically taking my invention from me would be the lowest cost way of entering the market. Is the law a barrier to entry? Surely not.
If all property is a barrier to entry, it is hard to see what the term buys you. Property rights are everywhere. No one labels property rights as barriers to entry, but that’s exactly what they are in Demsetz’s formulation. Everything is just property rights.
Instead of thinking about barriers to entry, the more general formulation is to think in terms of property rights. Some firms have some property rights. Other firms have different property rights. It depends on the institutional arrangements, the policies in place, and lots of other factors. But if we are trying to understand the market, we need to place these barriers to entry in the context of many property rights. Any profit that a firm makes is just a return to those property rights, as I explained in the case of markups.
I see Demsetz as proposing an alternative and an abandonment of the old barriers to entry ideas. I agree.
On the policy side, luckily, modern antitrust has moved beyond the SCP paradigm in many ways. Barriers to entry are not invoked to “explain” or condemn profits as unearned. The focus is not on whether entry barriers exist but what are the forces that push toward quicker or slower entry? Can policy change the timing and magnitude of entry?
For example, the past few iterations of the Horizontal Merger Guidelines issued by the U.S. Justice Department (DOJ) and Federal Trade Commission (FTC) effectively highlight the significance of entry in merger analysis, without assuming it will fix everything without entry barriers. That’s the key. Entry is not a panacea, but it is a key feature of markets! According to past guidelines, entry becomes relevant when it is timely, occurring within a specific timeframe (such as two years), and of sufficient magnitude to prevent price escalation beyond current levels. The DOJ and FTC are saying we should focus on the speed at which entry erodes any price increases resulting from a merger, rather than solely considering whether it eventually achieves this erosion because of the lack of entry barriers.
Soon, we will have new guidelines written under FTC Chair Lina Khan and DOJ Assistant Attorney General Jonathan Kanter. Hopefully, they continue to highlight the significance of entry in merger analysis. Given Khan’s desire to turn back the clock on antitrust, that’s no guarantee. We will see.
