What is a Barrier to Entry

7/28/2019 What is a Barrier to Entry

1/6


2/6

WHEN ARE SUNK COSTS BARRIERS TO ENTRY?ENTRY BARRIERS N ECONOMIC ND ANTITRUST NALYSISt

What Is a Barrier o Entry?

By R. PRESTONMCAFEE, HUGO M. MIALON, AND MICHAEL . WILLIAMS*

In the Papers and Proceedings of the Forty-eighth Meeting of the American Economic As-sociation, Donald H. Wallace (1936 p. 83)proposed a research program hat proved vision-ary: "The nature and extent of barriers o free

entry needs thorough study." Fifteen years later,Joe S. Bain published a book that was the firstthorough study of entry barriers.

In this book, Bain (1956) defined an entrybarrier as anything that allows incumbents toearn above-normal profits without inducing en-try. He believed that economies of scale andcapital requirements meet his definition becausethey seem to be positively correlated with highprofits. George J. Stigler (1968) later defined anentry barrier as a cost advantage of incumbentsover entrants. With

equalaccess to

technology,scale economies are not an entry barrier accord-ing to this definition, and neither are capitalrequirements, unless incumbents never paidthem.

With respect to scale economies and capitalcosts, the definitions of Bain and Stigler are atvariance, which has resulted in controversyamong economists and antitrust awyers, bothover the definition of an entry barrier, and thequestion of whether scale economies and capitalcosts each constitute one. The present article isan attempt o resolve the controversies concern-

t Discussants: Steve Berry, Yale University; Michael H.Riordan, Columbia University; Michael Whinston, North-western University.

* McAfee: Humanities and Social Sciences, CaliforniaInstitute of Technology, Pasadena, CA 91125 (e-mail:[email protected]); Mialon: Department of Economics,Emory University, Atlanta, GA 30322-2240 (e-mail:[email protected]); Williams, ERS Group, 2000 Pow-ell Street, Suite 500, Emeryville, CA 94608 (e-mail:[email protected]). We thank Sue Mialon for valu-able comments.

ing the concept of barriers o entry. We begin bycontrasting the definitions of an entry barrierproposed in the economics literature. We thenintroduce a classification system to clear up theexisting confusion, and we employ it to assess

the nature of the barriers posed by scale econ-omies and sunk costs.

I. History of the Concept

In chronological order, the seven principaldefinitions of an entry barrier proposed in theeconomics literature are as follows.

Definition 1 (Bain, 1956 p. 3): A barrier toentry is an advantage of established sellers in an

industryover

potentialentrant sellers, which is

reflected n the extent to which established sell-ers can persistently raise their prices abovecompetitive levels without attracting new firmsto enter the industry.

Definition 2 (Stigler, 1968 p. 67): A barrier oentry is a cost of producing (at some or everyrate of output) that must be borne by firmsseeking to enter an industry but is not borne byfirms already in the industry.

Definition 3 (James M. Ferguson, 1974 p. 10):A barrier o entry is a factor that makes entryunprofitable while permitting established firmsto set prices above marginal cost, and to persis-tently earn monopoly return.

Definition 4 (Franklin M. Fisher, 1979 p. 23):A barrier o entry s anything hat prevents entrywhen entry is socially beneficial.

Definition 5 (C. C. von Weizsacker, 1980 p.

400): A barrier o entry is a cost of producingthat must be borne by a firm seeking to enter an461


3/6

AEA PAPERS AND PROCEEDINGS

industry but is not borne by firms already n theindustry, and that implies a distortion in theallocation of resources from the social point ofview.

Definition 6 (R. Gilbert, 1989 p. 478): An entrybarrier is a rent that is derived fromincumbency.

Definition 7 (Dennis Carlton and Jeffrey Per-loff, 1994 p. 110): A barrier o entry is anythingthat prevents an entrepreneur from instanta-neously creating a new firm in a market. Along-run barrier to entry is a cost necessarilyincurred by a new entrant that incumbents donot (or have not had to) bear.

Bain's definition s flawed in that it builds theconsequences of the definition into the defini-tion itself. Moreover, one can imagine an indus-try with competitive pricing due to the presenceof many incumbents but with no possibility ofentry (e.g. by government iat); such an industryhas no entry barriers according to Bain'sdefinition.

Stigler's definition avoids tautology by identi-fying an entry barrier n terms of its fundamental

characteristics, mphasizinghe differential osts

between incumbents and entrants. However, thepresent tense "is" in the definition is confusing.Literally, the definition implies that a cost thatonly entrants (not incumbents) have to beartoday is an entry barrier, ven if incumbents hadto bear it in the past (when they entered themarket).

Stigler's definition is narrower than Bain's:some costs are barriers according to Bain andnot according to Stigler; but all Stiglerian bar-riers meet Bain's definition.

Ferguson's definition follows Bain's, butwith the additional requirement hat incumbentsearn monopoly profits. Pricing above marginalcost is not sufficient for incumbents to persis-tently earn above-normal profits. Incumbentsonly earn above-normal profits if prices exceedaverage cost. Prices may not exceed averagecost even though they exceed marginal costbecause of price or quality competition amongincumbents.

Fisher's definition follows those of Bain and

Ferguson, but it is normative rather than posi-tive. For Fisher, an entry barrier is socially

harmful only if potential entrants make a calcu-lation that is different from the one that societywould want them to make in deciding whetherto enter an industry hat possesses the barrier nquestion.

Von Weizsacker's definition is also norma-tive, but it follows Stigler's definition. He ar-gues that a cost differential is an entry barrieronly if it reduces welfare. His point is that thenumber of firms in a Cournot industry can begreater han the socially optimal number. n thiscase, entry barriers serve a socially desirablepurpose.

Gilbert's definition ocuses on the advantagesof incumbents rather han the disadvantages ofentrants. According to him, an entry barrier s

the additional profit that firms can earn as a soleconsequence of being established in the indus-try. This definition has an immediate problem nthat a profit is not a barrier.

Carlton and Perloff offer two definitions.They argue that the first is not practical becauseit implies that any capital requirement is anentry barrier, and that any industry in whichentry takes time has an entry barrier. They notethat the term "barrier o entry" s often used torefer to both costs of entering and the time

requiredto enter.

However,to our

knowledge,they are the first to propose a definition thatexplicitly includes a time dimension.

Unfortunately, hey avoid the timing issue byconsidering only entry barriers n the long run.They argue that a firm can only earn profits inthe long run if it has an advantage over potentialentrants, which leads them to adopt a modernversion of Stigler's definition. Notice that theirversion clears up the confusion about thepresent tense "is" in Stigler's definition.

Scale economies and capital requirements reentry barriers according to Bain's definition be-cause they seem to be positively correlated withprofits. Scale economies are not an entry barrieraccording to Stigler's definition provided en-trants and incumbents have equal access to tech-nology. Capital requirements are not Stiglerianentry barriers ither, unless the incumbent neverpaid them.

Fisher also claims that capital requirementsare not entry barriers according o his definition.Consider, as Fisher does, an industry hat firms

can only enter if they make a large capitalexpenditure. A firm will not enter if the profits

462 MAY 2004


4/6

WHEN ARE SUNK COSTS BARRIERS TO ENTRY?

that it anticipates in the long run will not besufficient to justify the initial capital cost. But,argues Fisher, this is exactly the calculation hatsociety would want the potential entrant tomake. The capital expenditure would be so-cially wasteful if it did not guarantee a rate ofreturn that exceeded that which it could earnelsewhere. Fisher therefore concludes that, ac-cording to his definition, capital requirements,no matter how large, are not entry barriers.

Note, however, that Fisher's argument ig-nores consumer surplus, which enters into soci-ety's calculation, but not into the potentialentrant's calculation. The addition of anotherfirm to the industry could increase competition,and hence consumer surplus, enough to com-

pensate for the entrant's profit loss in society'scalculation. Governments have operated firmson this theory, to create price competition wherethere would otherwise be a natural monopoly.

Capital requirements r scale economies maynot constitute entry barriers according to vonWeizsacker's definition. To prove this, vonWeizsacker models an industry with scale econ-omies and shows that the number of active firmsin the Cournot equilibrium with free entry ex-ceeds the number of active firms that would

maximize social surplus. The cost savings thatarise from firms taking greater advantage ofscale economies more than compensate for thereduction in total output from having fewerfirms. In such an industry, additional entry bar-riers enhance welfare.

Capital requirements can be entry barriersaccording to Gilbert's definition, especially if asignificant proportion of them are sunk. Sunkcosts generate earnings that would be lost if afirm exits the market; n this sense, sunk costsare exit barriers. Exit barriers can affect entryby influencing the incentives of incumbents. Ifincumbents cannot exit without considerablelosses, then their threats of aggressive post-entry behavior are more credible, which detersentry and earns them higher profit. Thus, exitbarriers or incumbents create entry barriers.

Moreover, sunk costs increase an entrant'slosses in the event that entry fails, which makesthe incumbent's threats of aggressive post-entrybehavior more frightening. Thus, exit barriersfor entrants create entry barriers. n these ways,sunk costs provide rents to incumbents and,hence, are entry barriers according to Gilbert's

definition, while they are not according toStiglerian definitions, since all entrants mustbear them equally. Similarly, Carlton and Per-loff exclude sunk costs as entry barriers undertheir second definition) since there are no sunkcosts in the long run.

II. Economic Analysis

As we have shown, the concept of an entrybarrier has a rich and confused heritage in eco-nomics. To clear up the confusion, we offer thefollowing new classification of entry barriers.

Definition 8: An economic barrier o entry is acost that must be incurred by a new entrant and

that incumbents do not or have not had to incur.

Definition 9: An antitrust barrier o entry is acost that delays entry and thereby reduces socialwelfare relative to immediate but equally costlyentry.

All economic entry barriers are antitrust bar-riers. However, many antitrust barriers are noteconomic barriers. Antitrust s a larger categorythan economic.

When free entryleads to the efficient number

of firms, if a market has no antitrust entry bar-riers, then it is efficient. If it has no economicentry barriers, hen it is eventually efficient. Anantitrust entry barrier n a market that is other-wise efficient reduces welfare relative to what itwould have been in the absence of that barrier.

The presence of an antitrust ntry barrier doesnot necessarily mean that a merger should bedisallowed. The net change in welfare resultingfrom the merger could still be positive. Rather,the presence of the antitrust barrier means thatwelfare would be higher if that barrier did notexist.

In our analysis, we also find it useful todistinguish between direct and reinforcing bar-riers, as follows:

Definition 10: A primary barrier to entry is acost that constitutes a barrier o entry on its own.

Definition 11: An ancillary barrier o entry is acost that does not constitute a barrier o entry by

itself, but reinforces other barriers to entry ifthey are present.

VOL. 94 NO. 2 463


5/6

AEA PAPERS AND PROCEEDINGS

A group of small primary barriers may con-stitute a significant entry barrier. A group ofsmall ancillary barriers does not commonlyconstitute a significant entry barrier unless otherprimary barriers are also present. However, insome cases, large ancillary barriers an combineand reinforce each other to form a large primaryentry barrier.

A particular ancillary barrier may produce aprimary entry barrier only when combined witha restricted class of other ancillary barriers, orreinforce only a restricted lass of other primaryentry barriers. If a market possesses no entrybarrier rom either class, the ancillary barrier nquestion does not deter entry.

III. Scale Economies

Bain argued hat scale economies are an entrybarrier. Incumbents may have already builtplants of efficient scale. If the added output ofthe entrant's efficient plant is large relative toindustry demand and existing output, pricecould fall below the entrant's per-unit cost, andentry could be unprofitable.

This argument assumes that the entrant ex-pects the incumbent to maintain its pre-entry

outputevel even after

entryhas occurred. Once

the new firm has entered, the incumbent maywant to reduce its output to prevent its profitsfrom falling to zero. But then the entrant's prof-its might also be prevented rom falling to zero,and entry might be ex ante profitable.

However, the incumbent's output reductionwould prevent the potential entrant's post-entryprofits from falling to zero only if it causedsome fraction of customers to switch from theincumbent to the entrant. Customers may beloyal to an existing brand because continuing tobuy it involves less risk than trying a new one.Therefore, scale economies deter entry only ifcustomers are sufficiently loyal to the incum-bent's brand. Hence, scale economies are ancil-lary entry barriers hat reinforce primary entrybarriers such as brand oyalty.

We now argue that scale economies are an-titrust, not economic, entry barriers. Bain's no-tion, that entry may be deterred because theindustry only has room for a fraction of theefficiently scaled plant, is generally a short-run

phenomenon that does not persist as existingplants are replaced. Moreover, in an industry

experiencing growth, the fractional issue isclearly temporary. Thus, scale economies arenot economic, but may be antitrust, entry bar-riers if they delay entry and thereby reducesocial welfare.

In a technical appendix o this paper availablefrom the authors upon request), we have con-structed model in which new firms ace Coumotcompetition with incumbents f they choose toenter, and in which (i) scale economies do notdelay entry on their own, (ii) brand loyaltydelays entry on its own, and (iii) brand oyaltydelays entry even longer in the presence of scaleeconomies. Thus, according to the model, scaleeconomies are ancillary barriers hat exacerbatethe entry delay caused by brand oyalty.

Does the additional delay caused by scaleeconomies necessarily reduce social welfare?For an important class of demand functions(including linear demand), social welfare underCourot competition is higher than social wel-fare under monopoly, because the profit lossincurred by the incumbent s not large enough tooffset the price reduction that benefits consum-ers. In these cases, scale economies are ancillaryantitrust ntry barriers, ince they delay entry byreinforcing the entry-deterrent ffects of brand

loyalty,and

therebyreduce social welfare.

IV. Sunk Costs

Many firms are capable of paying large cap-ital costs if entry is worthwhile. If capital mar-kets are efficient, raising capital is no moredifficult for profitable arge-scale projects thanfor profitable mall-scale ones. And even if cap-ital markets are inefficient, perhaps because ofasymmetric nformation bout ndustry prospects,they do not necessarily fail in larger measurewith large-scale projects than with small-scaleones.

Capital-market mperfections avor wealthierand more experienced firms over entrepreneurswithout track records, but the former are notnecessarily the incumbents. Some entrants arelarge diversified firms that build new plants inan industry. Microsoft entering the internetbrowser business is an instance where the en-trant was larger than the largest incumbent. Inindustries where the primary potential entrants

are large diversified irms, arge capital costs arenot entry barriers.

464 MAY 2004


6/6

WHEN ARE SUNK COSTS BARRIERS TO ENTRY?

Nevertheless, capital costs can indirectly dis-courage entry. Instead of being entry barriers ntheir own right, they often reinforce other entrybarriers, by magnifying the risks. Thus, when asolid reputation s necessary to enter an indus-try, large costs make it difficult or impossible totest the market; nstead, the entrant must com-mit large resources to enter. If large sunk costsare associated with entry and entry s unsuccess-ful, the entrant's losses are large. In such asetting, the threat of aggressive behavior by theincumbent may deter entry. The greater he po-tential loss, the more frightening s the threat ofaggressive behavior. By magnifying risks, cap-ital requirements einforce other entry barriers.Therefore, capital requirements are ancillarybarriers, specially if a significant proportion ofthem are sunk.

Capital requirements are not economic entrybarriers, since incumbents had to bear capitalcosts in the past similar in size to those thatentrants have to bear today. However, they maynevertheless be antitrust entry barriers. Sunkcosts cause firms to delay entry because of theiroption value. The option of entering s lost oncethe firm enters. With uncertainty about marketconditions, this option has value. Thus, dynamic

entry is delayed relative to a static world.In the technical appendix o this paper (avail-able from the authors upon request), we argueformally that sunk costs, like scale economies,are ancillary antitrust ntry barriers. We presenta model in which (i) sunk costs do not delayentry in the absence of uncertainty and (ii) un-certainty does not delay entry in the absence ofsunk costs, but (iii) uncertainty and sunk costscombine to delay entry until the realization ofuncertainty. For an important class of demandfunctions, efficient entry is in advance of the re-alization of uncertainty. n these cases, sunk costsand uncertainty re ancillary antitrust ntry barri-ers that combine and reinforce each other to pro-duce a primary antitrust ntry barrier.

V. Conclusion

The presence of entry barriers s the centralsubject of contention in numerous antitrust

lawsuits. Usually, a merger in a particularindustry cannot permanently reduce competi-tion if new firms can easily enter the industry.Therefore, to prove that mergers are sociallyharmful, antitrust authorities must usually, atthe very least, demonstrate the presence ofentry barriers. To do so, they must rely on adefinition, and for this, they have often turnedto economists.

Unfortunately, economists have not yet beenable to reach broad consensus over the defini-tion of an entry barrier, and this has probablyhindered the development of efficient antitrustpolicy. In the hope of facilitating consensus, wehave highlighted two aspects of entry: (i) theeffect of alleged entry barriers on the timing of

entry and (ii) the effect on the timing of entry ofthe interaction between different alleged entrybarriers.

REFERENCES

Bain, Joe S. Barriers to new competition. Cam-bridge, MA: Harvard University Press, 1956.

Carlton, Dennis and Perloff, Jeffrey. Modern in-dustrial organization. New York: HarperCol-lins College Publishers, 1994.

Ferguson, James M. Advertisingand

competi-tion: Theory, measurement, act. Cambridge,MA: Ballinger, 1974.

Fisher, Franklin M. "Diagnosing Monopoly."Quarterly Review of Economics and Busi-ness, Summer 1979, 19(2), pp. 7-33.

Gilbert, R. "Mobility Barriers and the Value ofIncumbency," n R. Schmalensee and R. D.Willig, eds., Handbook of industrial organi-zation. Amsterdam: North-Holland, 1989, pp.475-535.

Stigler, George J. The organization of industry.Chicago, IL: University of Chicago Press,1968.

von Weizsacker, C. C. "A Welfare Analysis ofBarriers o Entry." Bell Journal of Econom-ics, Autumn 1980, 11(2), pp. 399-420.

Wallace, Donald H. "Monopolistic Competitionand Public Policy." American Economic Re-view, March 1936 (Papers and Proceedings),26(1), pp. 77-87.

VOL. 94 NO. 2 465

What is a Barrier to Entry

Documents

Transcript of What is a Barrier to Entry