San Jose State Unversty SJSU ScholarWorks Faculty Publcatons Economcs 6-1-1994 Natonal Defense and the Publc-Goods Problem JEFFREY ROGERS HUMMEL San Jose State Unversty, jeff@jrhummel.com Don Lavoe Follow ths and addtonal works at: http://scholarworks.sjsu.edu/econ_pub Part of the Economc Hstory Commons, and the Macroeconomcs Commons Recommended Ctaton JEFFREY ROGERS HUMMEL and Don Lavoe. "Natonal Defense and the Publc-Goods Problem" Journal des Econmstes et des Etudes Humanes: Blngual Journal of nterdscplnary Studes (1994): 363-377. Ths Artcle s brought to you for free and open access by the Economcs at SJSU ScholarWorks. t has been accepted for ncluson n Faculty Publcatons by an authorzed admnstrator of SJSU ScholarWorks. For more nformaton, please contact scholarworks@sjsu.edu.
353 NATONAL DEFENSE AND THE PUBLC-GOODS PROBLEM* l l j :. :. l j ;. : r ~ f ~- Jeffrey Rogers H\_mmel 0 & Don Lavoe 00 1. ntroducton Natonal defense, accordng to the popular deal, s a servce provded by the state to ts ctzens. t entals protecton from aggressors outsde the states jursdcton, usually foregn states. T1e most sophstcated theoretcal justfcaton for government provson of ths servce s the publc-goods argument. Roughly stated, ths argument clams that the ncentve to free -rde nhbts people from provdng enough protecton from fore gn aggresson voluntarly. Thus, t s n peoples best nterests to coerce themselves. Taxa ton s necessary to ensure suffcent mltary expendtures. Many opponents of arms control treat the publc-goods problem as f t alone were suffcent to dscredt any radcal reducton n mltary spendng. We, however, wll challenge ths presumpton. Ths artcle wll not queston the valdty, realsm, or relevance of the publc-goods concept.2 ndeed, we thnk that the core We wsh to acknowledge the nvaluable assstance o f Wllamson M. Evers and joe Fuhrg n workng out the themes of ths paper. Tyler Cowen, Davd Fredman MarshaJ Frtz, M. L. Rantala, Davd Ramsay Steele, Rchard H. Tmberlake, J r., Davd ]. Theroux, and Lawrence H. Whte all gave us h elpful comments upon earler drafts. They do not necessally share our conclusons, however, and we alone are responsble for any remanng errors. Tls artcle s reprnted wth permsson from the book, Anns, Poltcs and the Economy: Hstorcal and Contempormy Perspectves, edted by Robert Hggs. Copyrght 1990, The ndependent nsttute, 134 Nnety-Eght Avenue, Oakland, CA 94603, U.S.A. 0 Professor of Economcs and Hstory, Golden Gate Unversty, San Francsco, CA. 00 Professor of Economcs, George Mason Unversty, Farfax, VA. 1 By "the state" we mean government. We use the two terms nterchangeably, unlke many poltcal scentsts, who use the term the "state" ether for wha t we are callng the "naton".e., the government plus ts subjects, or for some vague ntt: rmedalt: entty whch s less than the entre naton but more than just the government. We recognze tha t the sta te and ts subjects can often be ntrcately nterwoven nto a complex web of mxed nsttutons, but the dstncton s stll fundamental 2 Although we wll not take u p these ssues here, some economsts suggest that the characterstcs that make somethng a publc good are a!jnost never physcally nherent n the good or servce but are rather nearly always a conse quence of choosng one out of many feasble methods for produ cng the good or ser vce. See Cowen-19RS, pp. 53-63; Palmer-1983, pp. l-5, p. 11; Block-1983, pp. 1-34; Roth hard-1981, pp. S32-46; Goldn-1977, pp. 53-71: and Bmbaker- l 97S, pp. 147-61. Volume 5, numero 2/3, }un!septambre 1994, pp 353-3 77.
354 journal des Economstes et des Etudes Humanes servce wthn natonal defense - safety from volence and aggresson - captures the essence of a publc good more fully than economsts have apprecated. But ths essental feature, rather than provdng a sold justfcaton for heavy mltary expendtures, offers one of the most powerful objectons to such a government polcy. We wll frst reexamne the nature of natonal defense n order to clarfy the underlyng goal of mltary spendng. The presumpton that the states mltary establshment automatcally provdes safety from aggresson needs careful scrutny. The taxaton necessary to fuel mltary expanson often generates more publcgoods problems than t crcumvents. Ths leads us to the more general queston of how the free-rder ncentve s ever overcome, despte theoretcal predctons to the contrary. Publc-goods theory seems to msunderstand human nature, by exaggeratng the mportance of narrow self-nterest and confnng attenton to artfcally statc Prsoners Dlemmas. A more socal and dynamc model of human acton s better able to account for the observed fact that free-rder problems are overcome n the real world all the tme. 2. What s a publc good? Economsts have called many thngs publc goods and then endlessly debated whether the label really apples, but natonal defense has remaned the quntessental publc good. Although rarely dscussed n detal, t s unversally nvoked as the classc representatve of the publc-goods category.3 Two characterstcs dstngush a pure publc good from a prvate good, and both are exhbted by the case of natonal defense. The frst s nonrval consumpton. One customers consumpton of a margnal unt of the good or servce does not preclude anothers consumpton of the same unt. For example, n an uncrowded theater, two patrons enjoyment of the same move s nonrval. The second characterstc s nonexcludablty. The good or servce cannot be provded to an ndvdual customer wthout smultaneously provdng t to others. The owner of a dam, for example, cannot provde flood control separately to the ndvdual farmers resdng downstream.4 Although these two characterstcs frequently come n conjuncton wth each other, they do not necessarly have to. Nonexcludablty from the dams flood-control servces s accompaned by nonrval consumpton of the servces 3 Examples of economsts treatng natonal defense as the quntessental publc good nclude Samuelson-1976, p. 159; Buchanan/Flowers-1975, p. 27; and Head/Shoup-1969, p. 567. Among the few attempts of economsts to look n any detal at natonal defense as a publc good are Thompson-1974, pp. 755-82; and Wagner-1975, pp. 199-221. 4 Paul A. Samuelsons two classc artcles, Samuelson-1954, pp. 387-89, and Samuelson-1955, pp. 350-56, are generally credted as beng the frst formal statements of modern publc-goods theory. mportant further developments n publc-goods theory nclude Samuelson-1958, pp. 332-38; Musgrave-1959; Buchanan/Kafogls-1963, pp. 403-14; Demsetz-1964, pp. 11-26; Mnasan-1964, pp. 71-80; Baumol-1965; McKean/Mnasan-1966, pp. 14-23; Davs/Wnston-1967, pp. 360-73; Buchanan-1968; Msnan-1969, pp. 329-48; Head/Shoup/1969; Head-1974, SndaV1979, pp. 532-66. ;. d., d J. ị
/ r j )... r l l.. Hummel & Lavoe. Natonal Defense and the Pu.blc-Goods Problem 355 among the varous farmers, but the owner of a nearly empty theater can stll exclude addtonal patrons. Ye t, accordng to the publc-goods argument, ether characterstc alone causes "market falure" - that s, an allocaton o f resources that s less than Pareto optmal. Thus, ether can be suffcent to justfy state nterventon.5 Even natonal defense s not a pure publc good. Amercans n Alaska and Hawa could very easly be excluded from the U.S. governments defense permeter, and dong so mght enhance the mltary value of at least conventonal U.S. forces to Amercans n the other forty-eght states. But n general, an addtonal CBM n the U.S. arsenal can smultaneously protect everyone wthn the country wthout dmnshng ts servces. n that respect, consumpton of natonal defense s nonrval. Moreover, a technque that defends just a sngle Amercan from the Sovet state wthout necessarly defendng hs or her entre communty and perhaps the entre naton s dffcult to vsualze. That makes natonal defense nonexcludable as well. We are gong to focus, however, only upon nonexcludablty. f consumpton of a servce s nonrval, but busnessmen and entrepreneurs can exclude those who do not pay for t, then they stll have strong ncentves to provde the servce. The most serous "market falure" that s alleged to result s underutlzaton of the servce. Some people wll be prevented from beneftng from the quantty of the servce that has been produced, even though permttng them to do so costs nothng. Furthermore, even ths mperfecton wll dsspate f the market permts dscrmnatory prcng. 6 On the other hand, nonexcludablty creates opportuntes for free rders, who wll pay for the servce only f dong so s absolutely necessary to receve t. From the perspectve of economc self nterest, every potental customer has an ncentve to try to be a free rder. f enough of them act on ths ncentve, the servce wll not be produced at all, or at least not enough of t. Another way to thnk about nonexcludablty s as a postve externalty n ts purest form. Many goods and servces generate addtonal benefts for people other than those who drectly consume and pay for them. There s often no way for the producers of these goods to charge those who receve these external benefts. A nonexcludable good or servce s one where the postve externaltes are not just an ncdental by-product but rather consttute the major beneft of the good or servce.7 5 Much o f the lterature has co nceded that, strctly speakng, very few actual goods or se1vces exhbt ether of the publc-good characterstcs n ts polar form. nstead, n the real world we encounter a range of goods and servces for whch the potental capacty and qualty of noru val consumpton s ncreasng or for whch the costs ofexcluson are ncreasng. 6 We have slghtly understated the supposed "market falure " from nonrval consumpton w th excludablty. The qua ntty of the publc good could also be nonoptmal. There s a vast economc lterature debatng the ntrcaces of nonrval consumpton. Some of the hghlghts nclude Samuclson 1969b, pp. 26-30, Dcmset?.-1970, pp. 293-306; Thompson-1968, pp. 1-12; Ekelund/Hulett-1973, pp. 369 87; Dcmsetz-1973, pp. 389-405; Oakland-1 974, pp. 927-39; Lec-1977, pp. 403-20; Borcherdng-1978, pp. 111-32; and llums/ Walsh-1981, pp. 166-91. 7 On the relatonshp of publc goods and externaltes, see Samuclson-1969a, pp. 98-123; Buchanan 196R p. 75; Head-1974, pp. Hl4-213; and Mshan-1969. ~. l. J..,
Ị. 356 journal des Economstes et des Etudes Humanes Clearly, the justfcaton for the states provson of natonal defense does not stem from any major concern that n ts absence protecton servces would be produced but underutlzed. Rather, t stems from the assumpton that, unless taxaton or some other coercve levy forces people to contrbute, natonal defense wll be nadequately funded and therefore ts core servce of safety from aggresson wll be underproduced. t s ths wdely held but rarely examned asswnpton tha t we wsh to queston.. 1 3. What s natonal defense? :.., Before we ca n explore the free-rde r dynamcs of the states mltary establshment, we must clarfy the meanng of the term "natonal defense." The publc-goods justfcaton for mltary expendtures rests upon a fundamental equvocaton over exactly what servce natonal defense entals. When economsts dscuss nato nal defense, the core servce they usually have n mnd, explctly or mplctly, s protecton of peoples lves, property, and lberty from fore gn aggressors. Ths also appears to be what people have n mnd when they fear foregn conquest, partcularly n the case of the Amercan fear of Sovet conquest. People throughout the world beleve that ther own government, no matter how dsagreeable, defends them from foregn governments, whch they thnk would be even more oppressve. Ths defense of the people s not synonymous wth another servce that goes under the same "natonal defense" label: protecton of the state tself and ts terrtoral ntegrty. Hstorcally, the state has often embarked on mltary adventures unrelated to the defense of ts subjects. f ths were not the case, people would requre no protecton from foregn states n the frst p lace. Many Amercans serously doubt that the U.S. bombng of North Vetnam and Camboda had very much to do wth protectng ther lberty. One defense-budget analyst, Earl Ravena}, contends that nearly two-thrds of U.S. mltary expendtures goes toward the defense of wealthy alled natons n Europe and Asa and has lttle value for the defense ofamercans. s The dstncton between the two meanngs of natonal defense does not apply only when the state engages n foregn nterventon or conquest. Even durng unambguously defensve wars, the state often systematcally sacrfces the defense of ts subjects to the defense of tself Such unversal war measures as conscrpton, confscatory taxaton, rgd economc regulaton, and suppresson of dssent aggress aganst the very ctzens whom the state s presumably protectng. People beleve the state defends ther lberty; n fact, many end up surrenderng much of ther lberty to defend the state. People of course may consder some trade-off worth t. They may accept the costs and rsks of the states protecton n order to reduce the rsks and costs of foregn conquest. But n most dscussons of natonal defense, the aggressve acts. ~ 8 Ravenal-1984. See also Russett-1970, pp. 91-126.
l r f \. Hummel & Lavoe. Natonal Defense and the Publc-Goods Problem 357 taken by the government aganst ts own subjects are arbtrarly excluded from the dscusson. t s ths frequently overlooked cost whch s suggested n Randolph Bournes famous observaton: "War s the health of the State. "9 n other words, the natonal nterest and the publc good do not automatcally concde. We do not deny the possblty of an ncdental relatonshp between the defense of the state and the defense of the people. But n the next secton, we wll present general reasons why we thnk ths relatonshp s not as strong as usually supposed. Before we can do that, we must fully expose the conceptual gulf between the two meanngs of natonal defense. The pervasve doctrne of natonalsm obscures ths fundamental dstncton. Natonalsm treats natons as collectve enttes, applyng prncples drawn from the analyss of ndvdual nteracton to the nternatonal level. n a war between two natons, the natonalst model focuses on essentally two partes: naton A and naton B. As n fghts between ndvduals, one of these two natons s the aggressor, whereas the other s the defender. As a result, the model axomatcally equates protectng the state wth protectng ts subjects. The basc flaw n the natonalst model s ts collectvst premse. Although the model nforms many of the formal economc analyses of nternatonal relatons, t represents a glarng example of the fallacy of composton. The state smply s not the same thng as ts subjects. Democraces are sometmes referred to as "governments of the people," but ths s, at best, rhetorcal sloppness. The state and the people nterpenetrate one another and n complex ways, but they clearly do not have exactly the same purposes or nterests. Consequently, any conflct between two natons nvolves not just two partes, but at least four the state governng naton A, the state governng naton B, the people wth the (ms)fortune to lve under state A, and the people wth the (ms)fortune to lve under state B. Whatever the merts of a dspute between states A and B, the dspute need not dvde a sgnfcant porton of people A from people B.lo Abandonng ths collectvst dentfcaton of the State wth ts subjects exposes the crtcal nsght about the natonal-defense servce. f one s truly concerned about defense of peoples lves, property, and lberty, then the transfer of ther captal cty from one locaton to another s not ntrnscally sgnfcant. n some cases t mght even be thought an mprovement. Many Amercans are convnced that the terrtory consttutng Russa s n a very real sense already conquered - by the Sovet government. Some even beleve that the Sovet people would fare better wth Washngton, D.C., as ther captal cty. What ultmately matters s whether transferrng the captal cty brngs the ctzens a net loss or gan... <!!t.~,:-... : 9 Bourne-1964, pp. 65-104. A general substantaton (or refutaton) of Bournes observaton has so far.).. not attracted the professonal energes of any hst01an, perhaps because they feel no need to belabor..! the obvous. There are lots of stud es showng the growth of state power n partcular countres durng pattcular wars, but vety few that even treat a sngle country durng more than one war, or more than a sngle country durng one war. A few exceptons that have come to our attenton nclude, Rosster-1948; Ekrch-1956; Hggs-1987; Tlly-1975; Tlly-1 985; and Hale-1985. 10 We cte examples of economc models exhbtng the natonalstc fallacy ofcomposton below. One of the very few wrtten challenges to the natonalstc model s Rothbard-1974, pp. 70-80. We have profted greatly from ths pathhreakng essay. l:
358 journal des Economstes et des Etudes Humanes l,.! The danger therefore s not foregn conquest per se, but the amount of power the conquerng government can successfully weld. n the fnal analyss, protecton from foregn states s not a unque servce. t s a subset of a more general servce: protecton from aggresson by anyone- or any state. Whether we formally label an oppressve state "foregn" or "domestc" becomes a secondary consderaton. Pe ople admttedly may hghly value ther own states preservaton and glorfcaton, n and of tself. Ther government s m ltary establshment may drectly enter ther utlty functons, the same way ther favorte baseball or football team does. But natonalsm s not just a subjectve preference. t s also a postve socal theory, as legtmately subject to crtcsm for ts polcy recommendatons as any other. The mltarys coercve fundng unfortunately prevents people from revealng ther true preferences about natonal defense d rectly and unambguously. Some ctzens may stll want a huge and expensve mltary establshment even f they dscover that t gves them less protecton than they thought. But meanwhle, an examnaton of whether mltary expanson truly does defend p eoples lves, property, and lberty s stll n order.ll... :1. ;.J " j... ~, ~.s,! ~,J,) q. >,l : :lj.~:~. : 4. The Free-Rder Dynamcs of Government nterventon When Paul Samuelson frst formalzed publc-goods theory, many economsts unreflectvely subscrbed to what Harold Demsetz has called the nrvana approach to publc polcy. Demonstratng some "market fa lure" wth respect to an abstract op tmum was consdered suffcent to justfy state acton. Economsts assumed that the costless, all-knowng, and ben evolent sta te could smply and easly correct any falure. Snce then, e conomsts have become far more realstc. Publc-goods theory has advanced to the pont where t s now an exercse n comparatve nsttutons. Demonstratng "ma rket fa lure" s no longer suffcent. One must compare the market wth the state, not as one wshes the state would behave n some deal realm, but as t must behave n the real world. To justfy state acton, one must show the agents of government have the capacty and the ncentve to do a better job than p artcpants n the market. Can the state provde the publc good wthout costs that exceed the benefts? And s there some ncentve structure that would concevably ensure that t do so?l2 Economsts wthn the feld of publc choce h ave done some of the most mportant work on the comparatve capabltes of the state -by applyng publcgoods nsghts to poltcal acton tself. They h ave come to the realzaton that the free-rder ncentve does not only arse for market enterprses. As Mancur Olson ~. 11 For a purely formal approach to peoples utlty functons wth regard to natonal defense, see Wagner-1975. 12 Demsetz makes the comparson berween the "nrvana" and "comparatve nsttutons" approaches n Demsetz-1969, pp. 1-3. See also Coase-1960, pp. 1-44; Buchanan-1962, pp. 17-28; and Turvey-1963, pp. 309-13.
J 1 1 Hummel & Lavoe. Natonal Defense and the Publc-Goods Problem 359... ~~.. has demonstrated, the free-rder ncentve can arse for any group, especally poltcal groups wantng to nfluence state polcy. Ths mparts an nherent publcgoods character to all poltcal decsons.l3 Assume that one of us wshes to change some.state polcy that we personally fnd partcularly onerous - for nstance, to repeal a tax. We are members of a farly large group that wll beneft f the tax s repealed. f enough of us contrbute money, tme, or other resources to brngng about the taxs repeal, we wh succeed and all be better off. The money we save n taxes wll more than remburse us for our effort. Once the tax s repealed, however, even those who dd not jon our campagn wll no longer have to pay t. We cannot exclude them from the benefts of the taxs repeal. They wll be free-rders on our poltcal efforts. Just as n the case of a nonexcludable good n the market, every potental benefcary of the tax repeal has an ncentve, from the perspectve of economc self-nterest, to try to be a free-rder. f enough of them act accordng to ths ncentve, the tax wll never be repealed. Publc choce economsts call ths result "government falure," completely analogous to the "market falu re" caused by nonexcludablty. Of course, ths example grossly oversmplfes the problem. Under a democratc state, people do not drectly purchase changes n state polcy; they vote for them. Or more precsely, some of them vote for representatves who then vote on and bargan over state polcy. f the tax repeal example were completely accurate, nearly every ntentonal beneft provded by the state would be a pure prvate good, smlar to the current salares of poltcans and bureaucrats. Wth votng, poltcal entrepreneurs and vote maxmzng frms (whch are called poltcal partes) have some ncentve to provde us wth our tax repeal, even f we do not poltcally organze, n order to entce us to vote for them. 1 4 Ths ncentve, however, s not very great. Frst of all, votng tself gves rse to a publc good. An ndvdual must expend tme and other resources to vote, but he or she can avod these expendtures by free-rdng on the votng of others. Only n the very remote case where the voter antcpates that, a sngle vote wll decde the electons outcome does ths ncentve to free -rde dsappear. Consequently, the poltcal entrepreneur must have some reason to expect that we wll vote at all. And f we do n fact vote, he must n addton have some reason to expect that the tax repeal, among all the other competng ssues, wll affect how we vote. Our formng a poltcal organzaton to repeal the tax gves hm reason to beleve both these thngs.15 n short, unorganzed groups have some nfluence upon the polces of a democratc state. But other thngs beng equal, groups that organze and campagn for polces have a sgnfcant advantage. That s presumably why they organze. l ~ /, 13 Downs-1957; Buchanan/rullock-1962; Olson-1971; Nskanen-1971 ; Tullock-1967; Breton-1974 ; and ~,fcker-1983, pp. 372-80. Wagner-1966, pp. 161-70; and Frohlch-1971, stress the poltcal-entrepreneur thess. Olson responds brefly n the second e dton of The Logc ofcollectve Acton, pp. 174-75. Bany-1978, pp. 37-40, and ~rdn-1982, pp. 35-37, go nto the weakness of ths thess n greater detal. Extended dscussons of the outcome of votng as a publc good ndude Barze/ Slberberg-1973, pp. 51-58; Meehl -1977, pp. 11-30; Bucbanan/Brennan-1984, pp. 185-201; and Barry-1978, pp. 13 19.
360 journal des Economstes et des Etudes Humanes! and campagn. t strans credulty to suppose that all the people who pour vast sums of money nto poltcal lobbyng are utterly mstaken n the belef that they thereby gan some leverage on polcy. The common observaton that specal nterests have nordnate nfluence upon a democratc state s wthout doubt emprcally well founded. Two varables affect the lkelhood that a group wll overcome the freerder problem and successfully organze. These varables operate whether the group s tryng to attan nonexcludable benefts on the market or from the state. The frst s the sze of the group. The smaller the group, ceters parbus, the more lkely the members are to organze successfully. The larger the group, the more dffcult t s to nvolve enough of them to secure the publc good. The second varable s the dfference between the value of the publc good to the members of the group and the cost to them. The greater ths dfference, ceters parbus, the more lkely they are to organze successfully. ndeed, f ths dfference s great enough, one sngle member mght beneft enough to be wllng to pay the entre cost and let all the other members of the group free-rde. The smaller ths dfference, on the other hand, the more essental becomes the contrbuton of each potental member.16 The democratc state therefore makes t much easer to enact polces that funnel g reat benefts to small groups than to enact polces that shower small benefts on large groups. Because of ths free-rder-nduced "government falure" the state has the same problem n provdng non excludable goods and servces as the. market wth one crucal dfference. When a group successfully provdes tself a publc good through the market, the resources t expends pay drectly for the good. n contrast, when a group successfully provdes tself a publc good through the state, the resources t expends pay only the overhead cost of nfluencng state polcy. The state then fnances the publc good through taxa ton or some coercve substtute. Moreover; the group that campagned for the state-provded publc good wll not n all lkelhood bear very much of the coerced cost of the good. Otherwse, they would have had no ncentve to go through the state, because dong so then costs more n total than smply provdng themselves the good voluntarly. nstead, the costs wll be wdely dstrbuted among the poorly organzed large group, who may not beneft at all from the publc good. Ths makes t possble for organzed groups to get the state to provde bogus publc goods, goods and servces whch n fact cost much more than the... ṫ : r._ -. t : 1 6 One of the clearest expostons of these factors appears n Davd Fredmans neglected Fredman 1973, pp. 185-88. See also Fredman-1986, pp. 440-47. Olsons taxonomy of groups - prvleged (small); ntermedate, and latent (large) n 7be Logc of Collectve Acton treats the two factors - group sze and relatve cost of the publc good - smultaneously and thereby slghtly confuses the ssue. Hardn-1982, pp. 38-42, clarfes Olsons taxonomy, correctly pontng out that a prvleged group (one n whch a sngle member values the publc good enough to pay ts entre cost) could theoretcally be qute large. Admttedly, there s some ambguty about whch cetera reman pares when group sze s vared. Some scholars have consequently challenged the clam that larger groups have greater dffculty overcomng the free-rder ncentve. See Frohlch/Oppenhemer-1970, pp. 104-20; Chamberln-1974, pp. 707..J.6; and McGure-1974, pp. 107-26. The best resoluton of these questons s Hardn-1982, pp. 42-49, pp. 125-37.
j Hu mmel & Lavoe. Natonal Defense and the Publc-Goods Problem 361 : J,j J benefcares would be wllng to pay even f excluson w ere possble and they could not free-rde. n ths manner, the state generates externaltes, and ones that are negatve. Rather than overcomng the free -rder problem, the state benefts free -loaders, who receve bogus publc goods at the expense of the taxpayers. Provson of these goods and servces moves the economy away from, not toward, Pareto optmalty. When the bogusness of such publc goods s obvous enough, economsts call them transfers. What s the upshot of ths "government falure" for natonal defense? n the case of defendng the state tself, we are dealng qute clearly wth a servce that the state has enormous ncentves to provde. f ths s a nonexcludable good or servce at all, then t s a publc good that benefts small groups very hghly. But n the case of defendng the people, we are talkng about, n the words of Davd Fredman, "a publc good... wth a very large publc." The benefts, although potentally great, are dspersed very broadly.17 Thus, to the extent that the free-rder obstacle nhbts market protecton of lberty, t rases an even more dffcult obstacle to the states ever undertakng that vtal servce. The state has strong ncentves to provde natonal defense that protects tself and ts prerogatves, but t has very weak ncentves to provde natonal defense that protects ts subjects lves, property, and lberty. Ths explans the common hstorcal dvergence between defendng the state and defendng the people. Furthermore, there s a perverse nverse relatonshp between the peoples belef that the state defends them and the realty. To the extent that they accept ths natonalstc concluson, ther poltcal resstance aganst the domestc states aggresson, however weak because of the publc-goods problem to begn wth, decreases further. Ths s most notceable durng perods of actual warfare. The belef of the states subjects that t provdes protecton actually reduces the amount of protecton they enjoy, at least aganst the domestc state. Natonalsm thus results n an ronc paradox. t vews the state as a protecton agency, but ths very vew contrbutes to the possblty that the state wll take on the lteral role of a protecton racket. Those who declne to p ay for the states protecton become ts vctms. Ths n tum gves the state an added ncentve to fnd foregn enemes. For wthout a foregn threat, the justfcaton for the states protecton becomes far less persuasve. S Our remarks have thus far been confned to the democratc state. They apply, however, even more strkngly to the undemocratc state, nsofar as there s aq.y sgnfcant dfference between the poltcal dynarrucs of the two types. We beleve that many economsts have overemphaszed the operatve sgnfcance of 17 Fredman-1973. p. 189. Lee-1985, p. 46, makes the same obsejvaton abou t the poltcal producton of natonal defense, but because he does not recognze the dltncton between defendng the state and defendng the people, he arrves at a much dfferent concluson: vz., democratc states w ll u nderprod uce mltary defense relatve to undemocratc states. 18 A smlar pont s mad e by Bouldng-1963, pp. 3-27. He refers to the worlds competng mltary organzatons as "mlorgs" and nssts that, n contrast to any other socal enterprse (ncludng polce.protecton), mltary organzatons generate ther own demand. "The only justfcaton for the exstence of a mlorg s the exstence of another mlorg n some other place... A polce force s not justfed by the ~exstence ofa polce force n anmher town, d1at s, by ~mother nsttuton of the same knd" (p. 10).
362 journal des Economstes et des Etudes Humanes formal votng. Both types of states are subject to the nfluence of groups that marshal resources n order to affect polcy. Formal votng only makes t possble for some changes to manfest themselves faster and less panfully. Our argument does not rule out the possblty that the state mght actually defend ts subjects. Whereas the dfference between the poltcal dynamcs of democratc and undemocratc states s overdrawn, states do dffer markedly n the amount of aggresson they commt aganst ther own subjects. f we automatcally assume that a conquerng government can weld as much power over foregn populatons as t does over ts domestc subjects, then a relatvely less oppressve government wll, n the process of defendng tself, provde some protecton for ts subjects. But ths s often only an unntended postve externalty. Moreover, a mltary polcy desgned prmarly to defend the states prerogatves wll generally dffer from what would be suffcent for the protecton of ts subjects. Ths dfference may unnecessarly nvolve the people n dangerous mltary commtments and adventures. Ther lves, lberty, and property, beyond beng sacrfced to the nterests of the domestc state, wll then be at greater rsk from foregn governments as well. Even when counterng oppressve governments, natonal defense therefore generates negatve externaltes that may more than offset the possble postve externalty. Above all, the value of ths defense hnges entrely upon the assumpton that conquerng governments can oppress a foregn populaton more fully and easly than can that populatons domestc government. But ths assumpton s hghly smplstc. t treats the power of the state as exogenously determned. Yet, f our concern s for the protecton of peoples lves, property, and lberty from any state, then a states oppressveness becomes the most crtcal varable of all. One states mltary polcy mght not only drectly affect the lberty of ts own subjects, but t mght also ndrectly nfluence the power of opposng states. Only a more sophstcated under-standng of oppressons fundamental determnants can tell us how best to ward off foregn aggresson. 6. The free-rder dynamcs ofsocal consensus To ths pont, our conclusons have been somewhat pessmstc, justfyng Earl Brubakers observaton that the free-rder assumpton makes economcs a dsmal scence.19 Based on that assumpton, nether the market nor the state has much ncentve to provde any drect protecton of peoples lves, property, and lberty. To the extent that hstorcal accdent has resulted n marked dfferences n the power ofvarous states over tqer own subjects, some such protecton mght be produced as an unntended externalty of the states effort to protect ts own terrtoral ntegrty. But that very effort at self-protecton wll also have a sgnfcant countervalng negatve mpact on the degree to whch the state aggresses aganst ts own subjects. 19 Brubaker-1975, p. 153. l
Hummel &Lavoe. Natonal Defense and the Publc-Goods Pmblem 363 Attrbutng a dfference to hstorcal accdent, however, s smply another way of sayng that the dfference s unexplaned. Not untl we explan the marked dfferences n domestc power of the worlds states wll we fully comprehend the relatonshp between protectng the state and protectng the people. One nave explanaton common among economsts s the publc-goods theory of the state. Ths theory often rests upon a sharp dchotomy between two types of states, usually democratc and undemocratc. Undemocratc states accordng to ths theory are lttle better than crmnal gangs, run by sngle despots or small groups of olgarchs essentally for ther own personal ends. The subjects of these states suffer under ther rulers but can do very lttle about ther plght. Any effort on ther part to change the stuaton, whether through volent revoluton or other means, produces an outcome that s a publc good; agan, we are caught n the free-rder trap.20 Democratc states, n contrast, are the result of socal contracts. Accordng to the publc-goods theory of the state, people create democratc states to solve the free-rder problem. At some obscure moment n the past, they drew up consttutonal rules n whch they agreed to be coerced n order to provde publc goods for themselves. Over tme, because the freerder problem genefates "government falure," democratc states have a tendency to fall under the nfluence of specal nterests. Perhaps better consttutonal decson rules could allevate ths decay. Nonetheless, democratc states always retan vestges of ther publc-goods orgn. That s why they aggress aganst ther own subjects far less than do undemocratc states. 21 We do not have to turn to the readly accessble hstorcal evdence to refute ths nave theory about the orgn of democratc states. The theorys proponents qute often do not lterally beleve t. nstead, tl1ey vew the theory as merely explanng the conceptual nature rather than the concrete orgn of the democratc state. Ether way, however, the theory has an nner contradcton. Creatng a democratc state of ths nature s a publc good tself. A very large group must n some manner have produced t. Because of the free-rder problem, they have no more ncentve to do that than to revolt aganst an undemocratc state or to provde tl1emselves an y other nonexcludable beneft.22 A more realstc alternatve to the publc-goods theory of the state s what we can call the socalconsensus theory of the state. All states are legtmzed monopoles on coercon. The crucal word s "legtmzed." Ths legtmzaton s what dffe rentates states from mere crmnal gangs. Any socety n, whch people refran from regularly kllng each other enjoys some knd of socal consensus. No government rules through brute force alone, no matter how undemocratc. Enough. 2 For the argument that revoluton s a publc good, see Tullock-1971, pp. 89 99, whch became wth mnor alteratons one of the chapters of hs book, Tullock-1974. Tullock dstngushes between what be calls "explotatve" and "cooperatve governments, rather than democratc and undemocratc, but the two classfcatons are almost dentcal.. 21 The publc-goods theory of the democratc state s stll stated best n Baumo-J 965, p. 57. 22 Kalt-1981, pp. 565-84, pnponts the contradcton n the publc-goods theory of the state. 111e stll devastatng, classc, pont-by-pont refutaton of the socal contract remans Spooncr-1966. See a lso Eve rs-1977, pp. lfls-94, whch traces the lteral noton of a socal contratt all the way back to Socra tes.
364 joumat des Economstes et des Etudes lfumanes of ts subjects must accept t as necessary or desrable for ts ru le to be wdely enforced and observed. But the very consensus whch legtmzes the state also hnds t.23 The socal consensus bears lttle resemblance to the mythcal socal contract of publc-goods theory. Whereas the socal contract s generally conceved of as an ntentonal poltcal agreement, agreed upon explctly at some specfc moment, the socal consensus s an unntended socetal nsttuton, lke language, evolvng mplctly over tme. Sometmes, the evoluton of the socal consensus can be very volent. Often, partcular ndvduals or even farly large groups wll strongly dsagree wth certan features of ther socetys consensus. But at all tmes, members of socety are socalzed nto the consensus n ways that they only dmly grasp, f at all. 2 4 Consder a classroom flled wth average Amercan ctzens. Ask for a show of hands on the followng queston: how many would pay ther taxes n full f no p enaltes resulted from nonpayment? Very few would rase ther hands. Ths shows that taxaton s nvoluntary. Then ask the group a second queston: how many thnk taxes are necessary or just? Ths tme, nearly every hand would go up. Ths shows that taxaton s legtmzed. 2 5 Of course, one of the reasons Amercans generally vew taxaton as legtmate s because they thnk t s necessary n order to provde publc goods. All ths proves, however, s that, although the publc-goods theory of the state s utterly worthless as an objectve descrpton of the states orgn or nature, t s very valuable as an deologcal ratonalzaton for the states legtmzaton. t performs a functon analogous to that performed by the dvne rght of kngs under monarchcal states or by Marxst dogma under Communst states. The socal-consensus theory of the state suggests that fyou conducted the same survey about taxaton upon a group of average Russ ans lvng wthn the Sovet Unon, or a group of average ranans lvng under the Ayatollah (and you could guarantee them complete mmunty regardless of how they answered), you would get smlar results. These foregn and "evl" undemocratc states are not exogenous a nd alen nsttutons mposed on ther subjects by sheer terror. They ;! l.., j f t... :. t 23 Snce the defnton of the stare (or government) s somethng poltcal scentsts cannot even agree upon, ours wll obvously be controversal. By "legtmzed" (a postve adjectve), we of course do not ean "legtmate" (a normatv e adjectve). Most economsrs should have no dffculty concevng of the state as a monopolstc coercve nsttuton, but noneconomsrs mght balk. Members of the general publc appear to have a bfurcated defnton of the state, dependng on whether t s domestc or foregn. They vew hostle foregn states as smply monopoles on coercon, just lke c1mnal gangs, whch s why they fear foregn conquest. They overlook the legtmzaton of these states. On the other hand, that s the only element they seem to recognze about the domestc state, overlookng or at least deemphaszng the coercve element. Ths dchotomy s only a cmder verson of the dstncton made by publc-goods theory between democratc and undemocratc states. For an extended defense of the mplcatons ofour unversal defnton, see Rorhbt~rd-1974, pp. 34-53.. 24 One of the earlest observatons that a socal consensus alujays legtmzes the state s a Boete-1975. Other wrters who have snce put forward a socal-consensus theory of the state nclude Davd Hume, Hume-1963, pp. 29-34; Mses-1966, pp. 177-90; and Sharp-1973. 25 We are confdent about the emprcal results, havng conducted the test ourselves many tmes. m f L..
) j ( l j l.. ) ~...1 ; : l. l ~?> A. 7 Hummel & Lavoe. Natortal Dejense and the Publc-Goods Problem. 365 are complex products of the culture, atttudes, preferences, and deas, whether explct or mplct, t~at preval wthn ther socetes.26 The vast deologcal and cultural dfferences among the peoples of the world are what explan the marked dfferences n the domestc power of ther states. The consensual constrants upon states dffer n content, but all states face them. The Sovet leaders fully realze ths, whch s why they devote so many resources to domestc and foregn propaganda. The shftng socal consensus also explans the many changes n the form and power of the state over tme. Although professonal economsts tend to gnore the deologcal and cultural components of socal dynamcs, professonal hstorans gve these factors the bulk of ther attenton. Hstory records that n the not-so-dstant past the world was entrely n the grp of undemocratc states, whch permtted ther subjects very lttle lberty. Democratc states evolved from undemocratc states. States that now must tolerate a large degree of lberty emerged from states that dd not have to do so. Publcgoods theory s n the awkward poston of theoretcally denyng that ths could have happened. t rases an across-the board theoretcal obstacle to every concevable reducton n state power that benefts more than a small group of ndvduals. 2.7 The socal-consensus theory, n contrast, attrbutes ths slow progress, sometmes punctuated wth volent revolutons and wars, to deologcal changes wthn the socal consensus. Thus, hstory s lttered wth drastc changes n state power and polcy that resulted from successful deologcal surmountngs of the free-rder obstacle. The Mnutemen volunteers who fought at.concord Brdge could not even come close to chargng all the benefcares of ther acton. They produced tremendous externaltes from whch Amercans are stll beneftng today. The Aboltonst movement produced such a cascade of postve externaltes that chattel slavery a labor system that was one of the worlds manstays no Jess than two hundred years ago, and had been so for mllenna - has been rooted out everywhere across the entre globe. We could multply the examples endlessly. 28 ndeed, the exstence of any voluntary ethcal behavor at all faces a freerder obstacle. Socety s much more prosperous f we all cease to steal and cheat, but the sngle ndvdual s better off stll f everyone else behaves ethcally whle he or she steals and cheats whenever able to get away wth t. Thus, everyone has a powerful personal ncentve to free-rde on other peoples ethcal behavor. f we all succumbed to that ncentve, socety would not be possble at all..,...,.! 26 Zaslavsky-198.3, has actually conducted far!} relable surveys among Sovet subjects, whch ndcate qute unambguously that the Sovet state s legtmzed. Good sngle-volume hstotes that mpart an apprecaton for the domestc sources of the Sovet state are Danels-1985, and Hoskng-1985. An ntroducton to the varous nterpretatons of Sovet hst01y by Amercan scholars, wrtten from a revsonst slant, s Cohen-1985. 27 Ths awkward poston s clearest n Tullocks Socal Dlemma. The new Socety for nterpretve Economcs, codrected by Lavoe/Klamer (Economcs Department, George Mason Unversty) s a welcome excepton to the geneml neglect among economsts of cul tural and deologcal dynamcs. We also cte some specfc exceptons below. 28 The premer work on the role of deas n the Amercan Revoluton s Balyn-1967, whle a work that explores the nternatonal repercussons of the revoluton s Palmer-1959-64. A magsteral survey of the nternatonal hstory of chattel slavery s Davs-1984. On the emergence of the nternatonal aboltonst movement, see hs Davs-1975.
366 journal des Economstes et des Etudes Humanes We must avod the mstaken mpresson that the governments polce forces and courts are what prevents most stealng and cheatng. To begn wth, the ntal creaton of such a polce and court system (at least under government auspces) s another publc good. But far more mportant, the polce and courts are only capable of handlng the recalctrant mnorty, who refuse voluntarly to obey socetys norms. A cursory glance at varyng crme rates, over tme and across locatons, clearly ndcates that the total stealng and cheatng n socety s far from solely a fun cton of the resources devoted to the polce and the courts. Certan neghborhoods are less safe, makng an equal unt of polce protecton less effectve, because they contan more asprng ethcal free-rders. f all members of socety or even a substantal fracton became ethcal free-rders, always stealng and cheatng whenever they thought they could get away wth t, the polce and court system would collapse under the load. 2 9 n short, every humantaran crusade, every broad-based deologcal movement, every wdely practced ethcal system, relgous and non-relgous, s a defant challenge hurled at the neoclasscal economsts justfcaton for state provson of publc goods. The steady advance of the human race over the centures s a seres of successful surmountngs of the free-rder obstacle. Cvlzaton tself would be totally mpossble unless people had somehow crcumvented the publc-goods problem.30 7. Beyond the free-rder ncentve f what we have been sayng so far s even partly correct, there must be a serous flaw n publc-goods theory. Howard Margols ponts out that "no socety we know could functon" f a ll ts members actually behaved as the free-rder assumpton predcts they wll. He calls ths theoretcal falure freerder "overkll".31 Despte ths flaw, publc-goods theory explans a great deal, whch s why t remans so popular among economsts. t explans why so many elgble voters do not waste ther tme gong to the polls. But t fals to explan why so many of them stll do go. (We thnk an nterestng emprcal study would be to determne what percentage of economsts who accept publc-goods theory volate ther theoretcal assumptons about human behavor by votng.) t explans why the progress of cvlzaton has been so panfully slow. But t fals to explan why we observe any progress at all.. 29 Among the economsts who recognze the publc-goods nature of ethcal behavor are Buchanan 1965, pp. 1-13, and Buchanan-1975, pp. 123-29; McKenze-1977, pp. 208-21; and Notth-1981, pp. 11-12, 18-19, 45-46. See also Parft-1984. 30 Rothbard-1981, p. 545, makes a smlar observaton: "Thus the free-rder argument proves hr too much. After all, cvllzaton tself s a process of all of us free-rdng on the achevements of others. We all free-rde, every day, on the achevements of Edson, Beethoven, or Vermeer." 31 Margols-1982, p. 6. See also McMllan-1979, pp. 95-107, Smth -1980, pp. 584-99 and Schneder/Pommerehne-1981, pp. 689-704.. f r
Hummel & Lavoe. Natonal Defense and the Publc-Goods Problem 367 Before we work out the mplcatons of ths theoretcal flaw for the ssue of natonal defense, let us dgress brefly and try to dentfy t. t must nvolve some weakness n the theorys assumpton about human behavor. We make no pretensons, however, to beng able fully to resolve the.weakness. Because ths very ssue sts at the conjuncton of publc-goods theory and game theory, t has become one of the most fertle areas of nqury wthn economcs and poltcal scence over the last decade. All we can do s modestly offer some tentatve thoughts about the sources of the weakness. Two possbltes suggest themselves. Ether people do not consstently pursue the ends that the free-rder assumpton predcts they wll pursue, or they pursue those ends but usng means nconsstent wth the assumpton. We wll take up both of these possbltes n order: 1. Do people consstently pursue ther self-nterest, as the free-rder assumpton defnes self-nterest? Publc-goods theorsts have offered not one but two motves that should cause a person to behave n accordance wth the freerder assumpton. The most obvous s narrow economc self-nterest. Ths end does provde a suffcent reason to free-rde, but vsualzng someone choosng a dfferent end s qute easy. Smple altrusm s not the only alternatve that wll volate ths narrow assumpton. People may desre socal mprovements - lbetty, justce, peace, etc. - not smply for ther materal benefts, but as ends n and of themselves, ndependently present wthn ther utlty functons. Patrck Henry may have been engagng n poltcal hyperbole when he exclamed "Gve me lberty or gve me death!", but he was stll expressng a wllngness to pay more for attanng lberty than ts narrow economc returns would cover. Perhaps ths wllngness should be called deologcal; no matter what we call t, t appears to be qute common n human hstory.32 Mancur Olson s the most promnent publc-goods theorst to argue that a second motve beyond narrow economc self-nterest justfes the freerder assumpton. Ths second motve apples even to the ndvdual wth deologcal ends - fthe group s large enough. He contends that only ratonalty n the plrsut of whatever end the ndvdual chooses s strctly necessary. The ndvdual wll stll choose to free-rde, because for a publc good requrng a large group hs meager contrbuton wll have no perceptble effect on attanng the end.33 We could object that an ndvdual s contrbuton to a cause s often not contngent n any way upon the causes overall success. Consequently, how much the ndvdual thnks hs acton wll affect the probablty of success s often. \ l 1! 32 Several scholars are movng n ths drecton. For nstance, Dawes-1980, pp. 169-93; Brubaker-1984, P~ - 536-53; Bany-1978; Hggs-1987, chapter 3; and North-1981, chapter 5. Even Mancur Olson suggests thts approach n Olson-1968, pp. 96-118, whch contrasts economcs, the study of ratonal acton, wth socology, the study of socalzaton. But the most ambtous effort along these lnes s Margols-1981, pp. 265-79. We ~hould note that we attach the adj ectve "narrow" to the term "self-nterest" to ndcate the ~tsage that tnvolves seekng partcular, usually selfsh, goals. Ths s to dstngush t from the broader usage of the term, whch can encompass any goal, ncludng altrusm. Whether ndvduals do n fact pursue ther ~:rro_~ self-nterest s a queston subject to emprcal ver fcaton or falsfcaton, bllt ndvduals by fnt!ton always pursue ther broad self-merest. 33 Olson-1971, pp. 64-6S.
368 journal des Economstes et des Etudes Humanes,,. rrelevant. Some people refuse to ltter, for nstance, fully aware that ther refusal wll have no perceptble mpact on the quantty of ltter. Such ndvduals gan satsfacton from dong what they beleve s proper, regardless of ts macro-mpact. n addton to a sense of rghteousness, deologcal movements can offer ther partcpants a sense of soldarty, of companonshp n a cause, that keeps many loyal no matter how hopeless the cause.34 But ths objecton concedes far too much to Olson. As phlosopher Rchard Tuck has cogently ponted out, Olsons noton of "ratonalty" f consstently obeyed precludes some everyday actvtes. t does not just apply to an ndvduals contrbuton to the effort of a large group; t apples just as forcefully to the cumulatve actons of a sngle person on a large ndvdual project. Olsons "ratonalty" s smply a modern varant of the ancent phlosophcal paradox of the Sortes. n one verson, the paradox argues that there can never be a heap of stones. One stone does not consttute a heap, nor does the addton of one stone to somethng that s not already a heap. Therefore, no matter how many stones are added, they wll never consttute a heap. (Of course, n the other drecton, ths paradox argues that there can never be anythng but a heap of stones.) One more dollar wll not make a perceptble dfference n a persons lfe savngs. One days exercse wll not make a perceptble dfference n a persons health. f the fact that the ndvduals mperceptble contrbuton goes toward a group rather than an ndvdual effort s what s decsve, then we are smply back agan at the motve of narrow self-nterest. No doubt, ths type of "ratonalty" does nfluence some people not to undertake some actons under some crcumstances. But just how compellng people fnd t s demonstrated by the mllons who vote n presdental electons, despte the near certanty that the outcome wll never be decded by one persons vote.35 2. Do people pursue ther self-nterest but n a manner nconsstent wth the free-rder assumpton? Olson, agan, has suggested one way that ndvduals mght effectvely organze despte the free-rder obstacle. Groups can lnk ther efforts at achevng nonexcludable benefts wth excludable by-products. Such byproducts nclude low group-rate nsurance and professonal journals. The ncentve provded by these by-products h elps counteract the ncentve to be a free rder.36 34 Hggs-1987, chapter 3, heavly emphaszes the role of deologcal soldarty. Coleman-1966, pp. 49-62, postulates an ndvduals psychc nvestment n collectve enttes. Buchanan and Brennan, "Voter Choce: Evaluatng Poltcal Alternatves," thnk that ths symbolc dentfcaton s the major motvaton behnd votng. 35 Tuck-1979, pp. 147-56. We can salvage Olsonan "ratonalty" under two strct condtons. When(!) a threshold level of resources s necessary before any of the publc good becomes avalable whatsoever, and (2) people end up payng wha tever resources they contrbute, rrespectve of whether they reach the threshold or not, t becomes ratonal not to contrbute f a person predcts that the threshold wll not be reached. n that specal case, he or she would smply be throwng away resources for nothng. Notce that these two condtons apply more frequently to obtanng publc goods through poltcs whch s often a wn-or-lose, al-or-nothng, stuaton - t11an to obtanng publc goods on the market. n partcular, t apples to votng. Hardn-1982 pp. 55-61, analyzes the frst of these condtons, for whch he employs the term "step goods." 36 Olson-1971, pp. 132-68. Olson also refers to excludable "by-products" as "selectve ncentves." Looked at another way. the byproduct theory converts a full publc good nto a postve externalty of a prvate good..... +J! j J..l ( ~
3 70 journal des Economstes et des Etudes Humanes Prsoners Dlemmas, wth two or more people, cooperaton frequently becomes ratonal for even the most narrowly self-nterested ndvdual. What ths work mples s that n many real-world dynamc contexts, deologcal a ltrusm or some smlar motve beyond narrow self-nterest may not be necessary at all to counterbalance the freerder ncentve.40 8. Concluson. ~. :. We have seen that puttng domestc lmtatons upon the power of the state s a publc-goods problem, but nonetheless one that n many hstorcal nstances for whatever reason has been solved. We have also seen that natonal defense, n the sense of protectng the people from a foregn state, s a subset of the general problem of protectng them from any state, domestc or foregn. Because of "government falure," the domestc mltary establshment tself can become the greatest threat to the lves, property, and lberty of the states subjects. The danger from mltary expanson, moreover, s not confned to ts domestc mpact. By threatenng the opposng naton, t cannot even unambguously guarantee greater nternatonal safety. The same threat that deters can also provoke the opposng sdes mltary expanson. Perhaps the factors that already provde protecton from the domestc state are the very factors to whch we should turn for protecton from foregn states. The same socal consensus that has voluntarly overcome the free-rder obstacle to make the Unted States one of the freest, f not the freest, naton may be able to overcome the free,.rder obstacle to protect Amercan freedom from foregn states. Nearly all ofus desre a world n whch all states have been dsarmed. Of course, most of the formal economc models of nternatonal relatons are not very sangune about ths eventualty. Yet our analyss ponts to two possble shortcon.ngs n such models and suggests at least a glmmer of hope. Frst, they are generally bult upon a statc formulaton of the Prsoners Dlemma, whereas dynamc formulatons are more realstc and more lkely to yeld cooperatve outcomes. Second, they generally commt the natonalstc fallacy of composton, gnorng the nteractons of the state wth ts own and foregn populatons. Lke the publc-goods theory they emulate, these models are very good at explanng the cases where dsarmament fals. They do not do so well at explanng the cases where t succeeds- as for nstance, along the U.S.-Canada border snce 1871.41. f. r 40 Hardn-1971, pp. 472-81; Taylor-1976; Hardn-1982; and Axelrod-1984. Axelrod corfnes hmself to two-person dynamc Prsoners Dle mmas, whle both Taylor and Hardn consder n-person terated games. For a good revew of the growng lterature on n-person games, see Dawes, "Soda) Dlemmas." 41 Brtan and the Unted States demltarzed the Great Lakes n the RushBagot Treaty of 1817. The process of dsarmng the entre border was not complete untl 1871, however. Both Noel-Baker-1951:1, and Bouldng-1963 apprecate the sgnfcance of ths example. Economc studes of nternatonal relatons that share these weaknesses nclude l.ee-1985 and Tullock-1974. Most of the economc work n these areas has focused upon allances. See for nstance Olson-1966, pp. 266-79; Olson/Zeckhauser 1967, pp. 25-63, Sandler-1 975, pp. 223-39; and Sandler-_1978. [!
Hummel & Lavoe. Natonal Defense and the Publc-Goods Problem 3 71 The domestc producton of dsarmament s tself a publc good, confrontng the same free-rder obstacle that confronts every nonexcludable good and servce. Should a majorty n any one naton come to endorse ths polcy, the narrow - or not so narrow - specal nterests who beneft from an armed state would undoubtedly be wllng to commt vast resources to keepng a huge mltary establshment. Thus, lke all sgnfcant gans n the hstory of cvlzaton, the dsarmng of the state could only be accomplshed by a massve deologcal surge that surmounts the free-rder obstacle.. 1. f!f. ~ J f ~. r,... ~1...... ;
3 72 journal des Economstes et des Etudes Hunu-xnes References Axelrod, R. (1984), 7be Evoluton ofcoopera ton, New York, Basc Books. Balyn, B. (1967), deologcal Orgns of the Amercan Revoluton Cambrdge, Harvard Unversty Press. Barry, B. (1978), Socologsts, Economsts and Democracy, Chcago, Unversty of Chcago Press. Baumol, WJ. (1965), "Welfare Economcs and the Theory of the State", 2nd ed. Cambrdge, Mass., Harvard Unversty Press. Becker, G. (1983), "A The01y of Competton among Pressure Groups for Poltcal nfluence", Quarterly journal ofeconomcs, n 98, August. Block, W. (1983), "Publc Goods and Externaltes: The Case of Roads", journal oflbertaran Studes, n 7, Sprng. Boete, E. de la (1975), The Poltcs of Obedence: The Dscourse of Voluntary Serutude (1574); reprnted., New York, Free Lfe Edtons. Borcherdng, Th. (1978), "Competton, Excl uson, and the Optmal Supply of Publc Goods", journal oflaw and Economcs, n 2l, Aprl. Bouldng, K.E. (1963), "The World War ndustry as an Economc Problem", n the collecton he coedted wth Emle Benot, Dsarmament and the Economy, New York,.Harper & Row. Bourne, R. (1964), "The State", n War and the ntellectuals: Essays by Randolph Bourne, 1915-1919, ed. Carl Resek, New York, Harper & Row. Breton, A. (1974), The Economc Theory ofrepresentatve Government, Chcago, Aldne. Brubaker, E.R. (1975), "Free Rde, Free Revelaton, or Golden Rule", journal oflaw andeconomcs, n 18, Aprl. Brubaker, E.R. (1984), "Demand D sclosures and Condtons on Excluson", Economcjournal, n 94, September. Buchanan, J.M. (1962), "Poltcs, Polcy, and the Pgovan Margns", Economca, 2nd ser., n 29, February. Buchanan, J.M. (1965), "Ethcal Rules, Expected Valu es, and Large Numbers", Ethcs, n 76, October. Buchanan, J.M. (1968), The Demand and Supply of Publc Goods, Chcago, Rand McNally. Buchanan, J.M. (1975), The Lmts of Lberty: Between Anarchy and Levathan, Chcago, Unversty of Chcago Press. Buchanan, J.M. & Brennan, G. (1984), "Voter Choce: E: ~luatng Poltcal Alternatves", Amercan Behavoral Scentst, n 29, November/ December. Buchanan, J.M. & Flowers, M.R. (1975), The Publc Fnances: An ntroductory Textbook, 4th ed. Homewood, ll., Rchard D. rwn. Buchanan, J.M. & Kafogls, M.Z. (1963), "A Note on Publc Good Supply," Amercan Economc Revew, n 53, January. Buchanan, J.M. & Tullock, G. (1962), The Calculus of consent Logcal Foundatons of Consttutonal Democracy, Ann Arbor, Unversty of Mchgan Press.,.,.
Hummel & Lavoe. Natonal Defense and the Publc-Goods Problem 3 73 ~ -. f l : :. Burns, M.E. & Walsh, C. (1981), "Marke t Provson of Prce-Excludable Publc Goods: A General Analyss",journal ofpoltcal Economy, n 89, February. Chamberln, J. (1974), "Provson of Publc Go~ods as a Functon of Group Sze", Amercan Poltcal Scence Revew, n 68, June. Coase, R. (1960), "The Problem of Socal Cost", journal oflaw and Economcs, n 3, October. Cohen, S.E. (1985), Rethnkng the Sovet Experence: Poltcs and Hstory snce 1917, New York, Oxford Unversty Press. Coleman, J.S. (1966), "ndvdual nterests and Collectve Acton", Papers on Non-Market Decson Makng. Cowen, T. (1985), "Publc Goods Defntons and Ther nsttutonal Context: A Crtque of Publc Goods Theory", Revew ofsocal Economy, n 43, Aprl. Danels, R.U. (1985), Russa: The Roots of Confrontaton, Cambrdge, Harvard Unversty Press. Davs, D.B. (1975), The Problem of Slavery n the Age of Revoluton, 1770-1823, thaca, Cornell Unversty Press. Davs, D.B. (1984), Slavery and Human Pmgress, New York, Oxford Unversty Press. Davs, 0. & Wnston, A. (1967), "On the Dstncton Between Publc and Prvate Goods", Amercan Economc Revew, n 57, May. Dawes, R.M. (1980), "Socal Dlemmas ", Annual Revew of Psychology, Demsetz, H. (1964), "The Exchange and Enforcement of Property Rghts", journal oflaw and Economcs, n 7, October. Demsetz, H. (1969), "nformaton and Effcency: Another Vewpont", journ.al oflaw and Economcs, n 12, Aprl. Demsetz, H. (1970), "The Prvate Producton of Publc Goods", journal oflaw and Economcs, n 13, October. Demsetz, H. (1973), "Jont Supply and Prce Dscrmnaton", journal of Law and Economcs, n 16, October. Downs, A. (1957), An Economc Theory ofdemocracy, New York, Harper &Row. Ekelund, R.B. & Hulett, J.R. (1973), "Jont Supply, the Taussg-Pgou Controversy, and the Compettve Provson of Publc Goods", journal oflaw and Economcs, n 16, October. Ekrch, A.A. Jr. (1956), The Cvlan and the Mltary: A Hstory of the Ame1can Antmltarst Tradton, New York, Oxford Unversty Press. Evers, W.M. (1977), "Socal Contract: A Crtque", journal oflbertaran Studes, Vol., Summer. Fredman, D. (1973), The Machne1y of Freedom: Gude to a Radcal Captalsm, New York, Harper & Row. Fredman, D. (1986), Prce The01y: An ntermedate Text, Cncnnat, South-Western. Frohlch, N. & Oppenhemer, J.A. (1970), " Get By wth a Lttle Help from My Frends", World Poltcs, n 23, October. Frohlch, N., Oppenhemer, J.A. & Young, O.R. (1971), Poltcal Leadershp and Collectve Goods, Prnceton, Prnceton Unversty Press...,.!
374 journal des Economstes et des Etudes Humanes Goldn, K.D. (1977), "Equal Access vs. Selectve Access: A Crtque of Publc Goods Theory", Publc Choce, n 29, Sprng. Hale, J.R. (1985), War and Socety n Renassance Europe,.1450-1620, New York, St. Martns Press. Hardn, R. (1971), "Collectve Acton as an Agreeable n-prsoners Dlemma", Behavoral Scence, n 16, September. Hardn, R. (1982), Collectve Acton, Baltmore, Johns Hopkns Unversty Press. Head, J.G. (1974), Publc Goods and Publc Welfare, Durham, N.C., Duke Unversty Press. Head, J.G. & Shoup, C.S. (1969), "Publc Goods, Prvate Goods, and Ambguous Goods", Economc journal, n 79, September. Hggs, R. (1987), Crss and Levathan: Crtcal Epsodes n the Growth of Amercan Government, New York, Oxford Unversty Press. Hoskng, G. (1985), The Frst Socalst Socety: A Hstory of the Sovet Unon from Wthn, Cambrdge, Harvard Unversty Press. Hume, D. (1963), "Of the Frst Prncples of Government", n Essays, Moral, Poltcal, and Lterary(l74l-42); reprnted. London, Oxford Unversty Press. Kalt, J.P. (1981), "Publc Goods and the Theory of Government", Cato journal,, Fall. Lee, D.R. (1977), "Dscrmnaton and Effcency n the Prcng of Publc Goods",journal oflaw and Economcs, n 20, October. Lee, D.R. (1985), "The Sovet Economy and the Arms Control Deluson", journal ofcontemporary Studes, n 8, Wnter/ Sprng. Machan, T.R. (1982), "Dssolvng the Problem of Publc Goods", n The Lbertaran Reader, :r"otowa, N.J., Rowman and Lttlefeld. Margols, H. (~981), "A New Model of Ratonal Choce", Ethcs, n 9l, january. Margols, H. (1982), Selfshness, Altrusm, and Ratonalty: A Theory of Socal Choce, Cambrdge, Cambrdge Unversty Press. McGure, M.C. (1974), "Group Sze, Homogenety, and the Aggregate Provson of a Pure Publc Good under Cournot Behavor", Publc Choce, n 18, Summer. McKean, R.N. & Mnasan, J.R. (1966), "On Achevng Pareto Optmalty -Regardless of Cost", Western Economcjournal, n 5, December. McKenze, R.B. (1977), "The Economc Dmensons of Ethcal Behavor", Ethcs, n 87, Aprl. McMllan, J. (1979), "The Free Rder Problem: A Survey", The Economc Record, n 55, June. Meehl, P.E. (1977), "The Selfsh Voter Paradox and the Throw-Away Vote Argument", Amercan Poltcal Scence Revew, n 71, March. Mnasan, J.R. (1964), "Televson Prcng and the Theory of Publc Goods",journal oflaw and Economcs, n 7, October. Mses, L. v. (1966), Human Acton: A Treatse on Economcs, 3rd rev. ed. Chcago, Henry Regnery. Mshan, E.J. (1969), "The Relatonshp Between jont Products, Collectve Goods, and External Effects", journal ofpoltcal Economy, n 77, May-june.. ; :~,..,!.!.
Hummel & Lavoe. Natonal Defense and the Publc-Goods Problem 3 75! f l ~.. r ;. (,._,,!-..... : Morgenstern, N. & Morgenstern, 0. (1953), The Theory of Games and Economc Behavor, 3rd ed. Prnceton, Prnceton Unversty Press. Musgrave, R.A. (1959), The 1beo1y ofpublc Fnance: A Study n Publc J:.conomy, New York, McGraw-Hll. Nskanen, W.A. Jr., ( 1971), Bureaucracy and Representatve Government, Chcago, Aldne-Atherton. Noel-Baker, Ph. (1958), The Arms Race: A Programme for World Dsarmamen t, London, Atlantc Books..North, D.C. (1981), St1ucture and Change n Economc Hstory, New York, W. W. Norton. Oakland, W.H. (1974), "Publc Goods, Perfect Competton and Underproducton",]ournal ofpoltcal Economy, n 82, September/October. Olson M. (1968), "Economcs, Socology, and the Best of All Possble Worlds", Publc nterest, n 12, Summer. Olson, M. (1971), The Logc of Collectve Acton: Publc Goods and the Theory ofgroups, 2nd ed. Cambrdge, Mass., Harvard Unversty Press. Olson, M. Jr., & Zeckhauser, R. (1966), "An Economc Theory of Allances", Revew ofeconomcs and Statstcs, August. Olson, M. Jr., & Zeckhauser, R. (1967), "Collectve Goods, Comparatve Advantage, and Allance Effcency", n ssues n Defense Economcs ed. Roland N. McKean, New York, Natonal Bureau of Economc Research. Palmer, R.R. (1959-1964), The Age ofdemocratc Revoluton: A Poltta/ Hstory ofeurope and Amerca, 1760-1899, 2 vols Prnceton, Prnceton Unversty Press. Palmer, T.G. ( 1983), "nfrastructure: Publc or Prvate?", Polcy Report, n 5, May. Parft, D. (1984), Reasons and Persons, Oxford, Clarendon Press. Rand, A. (1964), "Government Fnancng n a Free Socety", n The Vrtue ofselfshness: A New Concept ofegosm, New York, New Amercan Lbrary. Rapoport, A. (1974), "Prsoners Dlem ma - Recollectons and Observatons", n Game Theory as a Theory of Conflct Resoluton, Dordrecht, D. Redel. Ravenal, E.C. (1984), Defnng Defense: The 1985 Mltary Budget, Washngton, Cato nsttute. Rosster, C. (1948), Consttutonal Dctatorshp: Crss Government n the Modern Democraces, Prnceton, Prnceton Unversty Press.. Rothbard, M. N. ( 1974), "War, Peace and the State», n Rothbard, Egaltaransm as a Revolt Aganst Nature: And Other Essays, Washngton, Lbertaran Revew Press. Rothbard, M.N. ( 1981), "The Myth of Neutral Taxaton", Cato journal, n, Fall.. Russett, B.M. (1970), What Prce Vglance? The Burdens of Natonal Defense, New Haven, Yale Unversty Press. Samuelson, P.A. ( 1954), "The Pure Theory of Publc Expendture", Revew ofeconomcs and Statstcs, n 36, November. Samuelson, P.A. (1955), "Dagrammatc Exposton of a Theory of Publc Expendture", Revew ofeconomcs and Statstcs, n 37, November. ;.! :: j. l
3 76 Journal des Economstes et des Etudes Humanes Samuelson, P.A. (1958), "Aspects of Publc Expendture Theores" ReVew ofeconomcs and Statstcs, n 40, November. Samuelson, P.A. (1969a), "Pure Theory of Publc Expendture and Taxaton", n Publc Economcs: An Analyss ofpublc Producton and Consumpton and ther Relatons to the Prvate Sectors, ed. ]. Margols & H. Gutton, London Macmllan. Samuelson, P.A. (1969b), "Contrast Between We lfare Condtons for Jont Supply and for Publc Goods", Revew ofeconomcs and Statstcs, n 5.J l February. Samuelson, P.A. (1976), Economcs, loth ed. (wth Peter Temf}), New York, McGraw-Hll. Sandler, T. (1975), "The Economc Theory of Allances: Realgned", n Comparatve Publc Polcy, ssues, Theores, and Methods, eds. Lske, C., Loehr, w. & McCamant, ]., New York, John Wley & Sons. Sandler, T.M., Loehr, W. & Cauley, J.T. (1978), The Poltcal Economy of Publc Goods and nternatonal Cooperaton, Denver, Unversty of Denver Press. Schneder, F. & Pommerehne, W.W. (1981), "Free Rdng and Collectve Acton: An Experment n Publc Mcroeconomcs", Quarterly journal of Economcs, n 9l, November. Sharp, G. (1973), The Poltcs ofnonvolent Acton, Boston, Porter Sargent. Slberberg, E. (1973), "s the Act of Votng Ratonal?" Publc Choce, n 16, Fall. Smth, V.L. (1980), "Experments wth a Decentralzed Mechansm for Publc Good Decsons", Amercan Economc Revew, n 70, September. Sndal, D. (1979), "Publc Goods, Propery Rghts, and Poltcal Organzatons" nternatonal Studes Quarterly, 0 23, December. Spooner, L. (1966), No Treason: The Consttuton ofno Authorty(1870); reprnt ed., Larkspur, Colo., Pne Tree Press. Tannehll, M. & Tannehll, L (1970), The Market for Lberty, Lansng, Tannehll. Taylor, M. (1976), Anarchy and Cooperaton, London, John Wley & Sons. Thompson,.E.A. (1968), "The Perfectly Compettve Producton of Collectve Goods", Revew ofeconomcs and Sta tstcs, n 50, February. Thompson, E.A. (1974), "Taxaton and Natonal Defense", journal of Poltcal Economy, n 82, July, August. Tlly, Ch. ed. (1975), The Formaton of Natonal States n Western Eu rope, Prnceton, Prnceton Unversty Press. Tlly, Ch. (1985), "War Makng and State Makng as Organzed Crme", n Brngng the State Back n, ed. Peter B. Evans, Detrch Rueschemeyer, and Theda Skocpol, Cambrdge, Cambrdge Unversty Press. Tuck, R. (1979), "s There a Free-Rder Problem, and f so, What s t?" n Ratonal Acton: Studes n Phlosophy and Socal Scence, ed. Ross Harrson, Cambrdge, Cambrdge Unversty Press. Tullock, G. (1967), Toward a Mathematcs of Poltcs, Ann Arbor, Unversty of Mchgan Press. Tullock, G. (1971), "The Paradox of Revoluton", Publc Choce, n 9, Fall.,. r