Why do firms do basic research (with their own money ‘Nathan ROSENBERG hes apace, Sued aay, Sd C4 O00, USA “The question to be addressed i: Why do private Sins perform baie tsearch with thee we o> ney? Interet nth question derives from both {naiycal and wtltaran considerations. There is mii! evidence ithe United Stites, which ‘Provides the main eomext fr this paper, support Jeg the view that base research makes a sigii- cant contribution to the productivity growth of the economy [47-1 8 widely beld thar social returns from basic eacarch ate significant and bigher than private ets adie fr this reason nba most such actvids continue 1 be financed by the taxpayer This alo plies that measures simed at increasing sie research by the pevate Sector will be wellare improving I the United States, the federal gorerament inthe years since ‘he Stcond World War ft provided the vart, Inajrty of all funds devoted to basic research [Although the feral sare ha been deslining in recent years and alvough that share is at its lowest ievel in about 20 years, it sil constitutes shout two-thirds of the total [10 the goa sto ensourage the private sector to spend maze money or basi esearch, its neces sary to sat out By asking why they would want {oso inthe frst place: Suppese itis taken as niomati tha private ndusty im basines solely to make money. and therefore that fem are ot [Sins More Mare Clue, Dood Nien, Ea Teas ig Pa prepared to spend money for merely sentimental ‘or humanitarian purposes, Let us assume that they ‘will sped their own money on basic research only ‘when they are reasonably confident that wil Yield rate of return on thi investnent in the {generation of knowiedge that is atleast compare ‘Ble wo therate of return that they would expeet on some other form of investment in more tangibe capital Tis important to think of baie research hy private industry ae a form of investment, and fore wil be sai aboot Uus later But it may be ‘seal to anticipate some of that later discussion here by pointing oat that basic research, from 8 ‘rvate fms pint of wew, i not only aves ‘ent. Is more precisely. aod really By deini- tho, a Tonglerm fnvestment. There piety. ro resonable expectation that these expenditures ‘vl begin wo genevate a eash ow inthe next few yess or so ‘Why, then should private industry be wing to make sich expenditures? The question fe cual (ne for the academe economist a4 well a¢ for Policymakers in both the public and private see- {ors Since the seminal papes by Avtow and Nel- son 2.11] thar been scapes by moa economists tha private enterprise economy fis wo provide sdaquate incentives for investment in knowledge production. ‘There are several reasoas for this serton. Firat, theres inerently a high degre of Saintrable risk and uncersity tha noes 4 ‘we move along the base research end of te esearch spectrum. Secondly ~ and this specie to knowledge asa commodity itis beeved that aowledge one produce isn some meaning sense “onthe shel”, Asa Smsequencs, noclat al/melnstream economics eld that once pro- sled Knowledge was freny avaiable to ali 4723/0 /550 © Wa Eee Sc Pubs BY. ang)eluding tote fms that may hate made no conti- bution whatever to the prducion of the know ge” A clase “tee sider” siuation, la eorain fase, one could counteract ths by creating propery rights in knowledge: bu at all kinds of Enowiedge are patentable in rich a way as (0 preclude a competitor from expoting that knowl bilge, For these resons i hasbeen held th Sint seous problem of apprypiaili Tins Rinaeeing. the resereh save no adequate ig the be ‘he result of insffcent or impstect competion [sithough i ist fer charge apaiast the main (tadion of snodemn mirosconamies that I tends to attsbute almost any problem in resource isle Teeation to inefficient competion. hiss no: ene Of thove caves. It has boo posted out thatthe markt for knowledge is inereally imperfect and thin becuse in order to detrnne the value of information, it would be necessary. a general, 19 enon the information. This o° course creates & feodamentaldifclty because, once the buyer Koows the information, ste tas Bo incentive to ay fori "ae has pone out that society would ins ‘st insfiiem resources in research even under pevetly competitive conditions (2) He als argued {hat « monopolst'spreanvention monopoly Prof its weaken th insentve to invent as compared fo the compaitive situation (2. p. 198-173 Indeed, in one sens contpetion even exacerbates the problem, The lkethood that compeitors wil {Guickly exploit the ureful new tnowledge weakens ‘hen further the incentives on the pat of compet te firme to vest in Howe production in the fis place The only way 10 seagihen thse inceativs iby offering the fim that conduct the ‘esearch s proprietary contol over the valuable Findings thatthe research generates But such cone {tol ecg in the frm of past Fights ~ puts os nthe other horn of te effiseney lenm. This FS besaue, once knowledge hasbeen produce. it is covey avilable fr odher fms to ule as tell Any restriction of sul uses socially subop: ima) beesse it would depie some firms. of ‘pportenites to aise their producti by making tke of knowledge that is already produced and therefore avaiable to society at no incremental ost To deprive Gems af rly costes opportan= tiesto improve thee efficiency is obviouysubop- ‘al “This is the major conclusion of economic the ony: Market inentives ae insufficiently strong 10 fsenerate the sovally optimal amount of invest ‘ent in research — because of nonappropritbile fier and uncertainties. However, attempts 10 al> levate that problem ~ by allowing firms tap propriate the findings of research ~ create a3, ‘huey serous problem Because they impose ce: ‘ction upon the use of valuable knowlege that Ih aleady toc produced “Tipe economist’ onelosion that nermal market forces do not provide suong incentives forthe performance of research, expecially basic research, Jeauite consistent with obserations of the real ‘word. The obvious empirical acti that the over= ‘wheling majority of private fems do ot finance the performance of any basic research. Basic te- scare fat highly concentrated in two sens {the gest bok of al sch esate is petormed ina very small numberof industrial sectors, and (Gy within these sectors there isa handel of firms, ‘ypialy large firms, that dominate che basic r= search picture ‘Only lnted data are availabe with respect to ‘company spending on basic rereatch, With respect to sectoral concenation, incomplete dati pub- lished bythe Nationa Sience Foundation (10. p 53} indicate tht, in 1984, 61 persantof company financed base research was in four sectors Chemicals ‘$67 milion Electra! equipment 450 Aircraft & Missin 288 Machinery 209 subtotal $1584 /52578~ 61 8! 2 “The question persists: Why do the Firms that do Dosierscerh doi? Alternatively, why do some the ta eae fr af emp nded RAD Sota SOT A058‘ns find it profitable (or expectedly profiable) Economists, 25 we have seen, have stressed the problems asocated wth appropriability as the Iain deerteat to base reteareh ~ the dificliy. Under normal market conditions, of appropriating the enelite generated by the reearch findings But it shoold be sotced, fist ofall thar tus requires an important qualification Even if firm's bse reearch generted many benefits that it ‘ould nor appropriate the mere existence of such ‘onapproprablie is never an adequate explana~ to forthe reluctance to pecform base research, So long asthe perforing firm can eapiae some ofthe Benefits, that might be sufficient for ito 20 Some research, Tes nt necessary to expire al of the benefits ~ indeed it would be wndesable fit id. Research soil deseable precisely be- fase often generates soch widespread andi ‘scriminate benefits All shat is necessary is th ‘markt fores allow te frm to capture enough Of {here benefits to yield a igh at of rtm on its ‘own investment in baie rerereh “Thus, the exsence of spillovers and nonap- propriate that low competitors a tree de is ot 2 decisive case against the performance of R&D (or specially bac tesearchy by private fms IF he production of ew knowledge sen ‘mtes commercial oprortunites to the performer. the relevant caelationsavalves, aot the size of the spillover bot whether the performing. fm in capt enough cf the benefits genecated 10 ‘eld a high rate of rou on is investment Even Inthe exreme case of base research, where there ‘i no prospect of establishing proprictary contre, ‘over the research Tidings commercial benefits ray neverthciews be wey gre “These potential beseits largely tke the Form of what er called “int mover advantage" Tas {s 4 big subject that wil not be addressed ere beyond indicating the eateries into which such advantages may fall. They include a vanety of Tearing expences Firms that move dow Such learning curves fit - whether these curves pet tain 10 ‘cost reductions or performance improve- sents may be able subsequently to exploit the ‘ivantagas conferred asa barier tothe entty of tow firms, irtmoves may be able to acquire Saluable assets - ef. of «geographic location oF 3 rinral deposit whose commercial worth will be ‘vorsbly alfcied by new research findings. To ‘he extent thatthe fading of base research can De eranlatedinta patentable asoetsfethr dos cam, fisiimovers may be able 10 consolidate {heir market portion through patent protection. Furthermore, buyer switching cove may be sige ifcant and may comutue significant form of protection agaist competes for fis that are First to enter the new produ line ‘Althoush frsemover advantages may thus be sutstatal, there may also be subsantil fst mover disadvantages as well as latomover ade wantager If there are spnieant spillovers of [Knowledge between firms then a ltepover could ‘ain the same knowledge at a lower cost while at the same time, avoiding the major mistakes that the Gstimover made en route Neverthe, for present purposes 1 wish merely to aber that Fratmover advantages may fequently survive the ‘otfeting disadvantages and Seve asa sigafiant incentive for the peformance of basic research 6 One fly obviows, but nonetbelssinpertant statement about base secarch in privat industry ie that mont fem that have engaged int have Bad faidy. strong and. wellentenched postions of ‘market power. Precisely because the potential paj- ff to basi research & 50 long tem. only tras that were reasonably confident of being around in the lang term would be likly to consider the possibilty of king such commitments. Ths the most succesful basic research labs have been in Tems with stong musket postions: Bell Labs (Gspecialybetore divestiture IBM, DuPont, Dos Chemical. Eastman Kodak, etc. GM is reputed 19 have done quite bit af baie research Some ears ‘ag, bot the commitment to base research i aid to have devined along withthe deine in GM's ‘market sire inthe automobile industry (GM has had the largest total RED budget of any povate American fom $3.5 billion i T986. What peo postion of that amount is basic research is aot, public information, But itis probably very smal Another obvious reason why small firms bard ‘vee basie research is tha i research dings fre dificolt co patent and hence she Now of payols cannot be capitalise, then these pavofts bust be appropsated Ma the means mentionedcaer — escatilly through incorporating the Enowiedge in the form of improved goods of proceses, This means of couse, that a larger farket share wil offer the prospect of a higher pavot "A separate but related points that itis nt size sone or market power that ater. Large Fens may be more wiling to undevake basic research then they bave a clvese range of products and rong marketing ae dlsribtion networks chat igereme ther confidence that they wil eventually De able to put the findings of base research 10 Some good commercial se In view of the high Segre of uncerainty tat surounds the outcome patents have, in general. provided more effective Protection to proprietary knowledge inthe phar- Imaceaticl industry than in ms Ube? indus. Ta dhs respect the innovative “output” of sm Biotechnology Firms is likely to be more realy sppropriable than ie the case for stall em i other iadustnes The venture capital industry sems toe tet Ing biotechnology a4 a kind of lotery. The vast "jority of firme wil almost certaily Be losers Svhen the eventual "shake-out™ takes place. but ‘Mane suse innoeaions say eld very Big ‘eumn ~ a5 Genentech bopes wil be the case wih thsve plasminogen activator (TPA) for heart Df hasc research its probably very important to attach vitims {firm #0 have the confidence that i will know how 1 exploit new knowledge at may turn up in tinexpested places and that wl have the come Plementary ses tha will enable ito do so. For fh iluminaing treatment of related issues see ‘ese “The considerable number of small botechnal> cay finns appears to contact the view that Investment in Basic reach regutes a tong ‘marker positon. Iti certainly true that, i the dtug industry proper, the mest active research Dros seen! to be concentrated in the large firms = Merc, Sohason and Johnsen. Lilly, Pie, Upon et, But nthe biotechnology field good all base research coreaty beg petormed by sal companies such a2 Gevente. The relly furious aspect of these small companies that ‘many ofthe have no markeable products at all Sd haraly any have more thar jst fe ‘What seems to be happening here i tha the somal botesoology frm are engaged. basic Fevearch that bebeved to be closet the ome tmeralation stage. AL the same time, i a ‘hiv speculative pame tha is bing financed by ‘venture eaptline. as well a see large ems upd ‘wealthy individuals who are red by the pes Tivo avery high pao ‘What appears to be driving he sal firms that perform basic research in biotechnology is the Frstmover advan that fst mover advantages maybe etal. Mock ‘tthe investment in ths basic esearch took place before the resent changes ia ne patent lw tha steed patent protection to Bve organisms. But “is extension ofthe law mast cers steeagien the incentives to do bese reves, Furthermore In understanding why some private firms do Isic reeareh iti ocessary 10 recognize th businewes do not live in ueat. orderly world sere causal relationships are always clealy de- fined and swhere causality always works in One Section only. The business environment is much ‘more Interactive, fll of "Feedbacks" where some “downstream” development reacts back ypon and ere behavior “"upateam’- Perhaps woos im portant, it is fll of unplanned, or seudentl Sevelopments that then tm out 10 have an tm poruant set of consequences oftheir own, Tis essential to emphasize the unexpected and the unplanned, even if — of especally if = renders serious quantification impossible Fact the dificalie in precisely idenying and reas ing the benefits of hase rewarch are hard 10 ‘raggeate. White this might seem to be just an Interesting academic point about the lis of Certain methodologies it hab imporant de- ‘Getonmaking consequences. The post hat been fexpresed suscinctye "Project selection methodol ‘opis of formal, quantiatve aatare reduce the tendency to perozr basic research” [8 ‘art, but only» par. of the problem i that he ‘vipat of asi researc is never some final prod srt which the market place ean atich a Price tap. Rather. the output is some form of new owledge that har no lear dimensionality. The ‘ovipul ia pocular kind of intermediate god that say be used, 04 to produce a final good but {perbaps) to play some farther role i fe invenion of # new final giod. Those connections are, howsurreutrneinariyeificlt tnce wih ay fonfidenc, even expat But even if these eiffiut- ‘es coll be overome, he problems of evaluating the Knowledge. and ef providing an appropiate incentive syrem to veward the knowledge pro ders, would appear to be insuperable, “Thus, it is doubt that busines decision makers often st down and askin an abstract sways Should we do basic eeseaich? How much basic reearch shovld we 40? Obvious. private firs fot no obligation to advance the frombers of basi cence at sch, remamably, they ae alway asking themselves how they can make the mest profitable rate of eeturn on thei investment. Ta this context. my own emphasis on the unes- pected and the unplanned is deliberate, because {he history of base research in Americas industry suggests that every lage part of this researc has been usitentional, Tosti to ay, base research findings of major signieance have emerged a the Unplanned byproduct of the atempt to slve some ‘ery sperifie mdasril problem. The fet is that the dtnctionbetwoen basic research and appli research is highly artical and arbitrary, The ds Uinedon is usualy made to turn upon the motives, for goals, of the persca performing the research thoupht he was doing back around 1870, e would have replied that he was ying to solve some very pracetl problems ceansced with fermentation End puttefaction in te French wine industy. He felved the practical prcbloms ~ but along the ‘way he invented the meder science of bacteril- ‘gy. Simul if tha other great Frenchman, Sack rao had been aed. some ity yen earn, what he thought he was doina, ie answer would fave been that he was tying to improve the cfcency of seam engines." As a byproduct of that pariclar practi interest he created the ‘modem sence Of themmodynamics, ROK Reo ure tne mare ef 1 27 ‘che mes aks eee Buti isnot necesary to go back 1 nineteenth century France Tho two spectacular sientie ‘breakthroughs are cited simply because they were so spectacular. We could, stead, lok at Bell abe in the twentieth century. ack a the end of the 19205, when vansatanie. radiotelephone service wat first exabihed, the service was poor ‘besaute there was lots of ttc. Bll Labe asked a young, man, Karl Jansky, 10 determine the source [Of the noise £0 that it coud be reduced oF iminated. He was given ¢ rotatable antenna ‘work with: Jansky published a paper in 1932 in ‘which he feported thre sourtes of noise: Local Thunderstorms, more dstast thunderstorms nd 3 {Girdsource, which identified as "a steady ies stale, the oxgin of whichis ot known". It was this star noite”. as he labelled it which marke the birth of radio astronomy 12) Sanskys experience (as well asthe eater ex perience of Carnot and Paster) underlines cae Of the reasons why the atlempt to distinguish ‘beeen banc research and applied research it cexremely dificult to cary out in practice. Funda ‘mental breaktvoughs often cceur wile dealing sith very applied or pret problems. Attempt: dng to draw that neon the bats of te mouives of the parton performing the resarch — whether theresa concern with acquiring useful knowledge (applied) as opposed to a purely disinterested search for new knowledge (basic) ~ is in my ‘opinion, 2 hopeless quest: Whatever the ex ante Intention in undertaking reearch, the hind. of Knowledge actually acquired i highly wapredic- able. Historizally, some ofthe most fundamental Seintie breakthroughs have come from people Ike Carnot, Paster and Jansky, who hough they were doing very applied research, and who would Undoubtedly have sad 40 they had been asked atthe time. [But the distinction breaks down in ancther way as well We have to ditngush between the mo- tives ofthe individual cenit and the matives of the fim that employs them. Many scent in Deva industry could honesty say that they are Etiempting to advance te trots of ase ses ‘ie Knowledge, without any interest in possible pplication At the same time the motivation of the research managers who decide fo finance Te earch in seme base fell of cince, may be ‘Mrongly motivated by expectations of eventually ‘etal tindngs Ths, Bell Lape decied toport base research in astrophysics Because of its lationship to the whole fle of problems and possiblities in micowave trantmason, and espe lly the use of eommanication stelites for soch purposes It tered out tht tery hgh fequen en ain and other atmosphere conditions be ‘ame major soarees of intrleace in transmis: Sion. This sour of signal lst was a continuing oncom inthe development of stelite eomman= ‘htons I was out of such patel concerns that Bell Labs desided 10 employ Amo Penzas and Rober Win, Pencar and Wien would ae ‘oubtely have been ignant if anyone had su fested that they were doing ayihing cher than hse reeurch, They fir observed the cosmic Ickground radiation, which sow taken a cone Firmation ofthe "big bang” theory of the forma thom ofthe uivese, while they were attempting © Idenaty and measure the vargas sours of alse inher anteana and inthe atmosphere. Although Ponzias and Wikon didnot now atthe Sie the character of the background radiation that they discovered war jest what ha been postlaed fart by couologits favoring the "big bang™ theory, Pena and Wilson appropeately shared & Nobel Prize for this fining. The finding was Shout as bate as bate sience can get, and isin ‘no wey diminished by observing thatthe fez that ployed them did s0 because they hoped to improve the quality of satelite tansmision (12) ‘The pales between the fundamental ds: coves of Jans and Pensis and Wilson i ery have deliberately examine thse instances of bse resentch emerging out 9 praca and op plied concerns because they provide valuable nr fnto the question of how busi essarch pets { be card aut in private indus. Te is often ‘Grd out unintentionally. Its, moreover dif Fie to understand if one insists on drawing ‘sharp distinctions between base and applied re Sarch om the bess of the motivations of those fevocing te esearch, ff, T would go much ferters When bavi research fp industry is io Inte from the ext of the firm. whether organiza tionally or geographical, it key to bevome rie and unproductive, The story of base re search in industry suggests that iis hel to be Inoststective when Hi highly interac with the work. or the concerns. of applied sien td engineer. This i baause the high technology industries are continually throwing up problems, Gifiulies and anomalous observations that te ‘most unlikely o aseur outside of high technol ‘gy content. High techacogy industries provide & Sight vantage point for the conduct of basic ‘een, but in order for ssendsts to expt she potential of the industrial environment ts neces intracion with other components of the ine Austad word. Bel Labs Before divestiture s probably the best example of place where the Frsituional environment was most hospitable for ‘asi esearch “The emphasis on interactions and feedbacks sogpists 4 way of thinking about basic research that T boot. is potentially fruitful. That i, the performance of basic esearch may be thought of sa ticket of admission to an ifoeration net- trork Ths network includes a variety of ions tion flows with no. parcular attempt 10 di tinguish or eluify nto asic or applied cate tones, Thote iss high degre of interact, even fmbracing work that goes oo within the reat of Development as well at Research tis worth obuerving thatthe attempt cas: sity coeur into asie and applied categories i paricly hard to take seriously in some areas nd disciplines. eg. in the realms of heal ‘medicine and aprcuire. A strict application of the tat common criterion for basi research — research that undertaken withost» concern for rata applications = eocld easily lead 10 the onclsion thatthe National Institutes of Health [re not deeply involved im base research, or tha current university and industrial research in the Feil of biotechnology contin ao basic rsarch ‘which i beard To conducting ie resource surveys the NSF defines basic rveach ay esearch tha has 35 Ghjecve "a fuller knowledge or understanding of (he subject understudy. eater than a practical pplication thereof” By contrast. applied research iS research drested toward pining “knowledge oF findestanding necessary for determining he meane by which a recognized and specif need nay be met 9}. These definitions appeat to mean {hats ifthe Nationa Institutes of Heals dzeted 3ranjr research thrust into cellar biology to pro- ‘de the enonladgs necesay forthe development ‘ofa vaccine again AIDS, ora cre for 8 pectic form of cance, that none of the resulting esearch could be classed at asi, Ht dificult to se Snhy the doerminaticn to eal witha paricalar dscase cannot give the to rsecch that provides fuller knowledge ct understanding of the sub- jet understudy" even hen there i praca! pplication” in mind. Here again the inteodwcton ‘of motives, or goal, ees than useful athe NSF Js ford to acknowledge with respect specfcally ‘to research in private dustry Thus the NSF ads {o is definition of Easie research the following ‘qualification "To sake into account industrial goals, NSF ‘modifies this definition for the indy sector Wo indicate that Be research advances sein tie knowledge not having specific commercial objectives. although such investigations may be In fields of present or potenti interest to the ‘eponing company” (9, ‘A furher point that needs tobe eaphasized is that there afe a munber of actives that ae essential tothe suces of business firms in high technology industes that depend heavily vpon 8 bse research capability, even i that cup oes not play a direc fle in solving industrial problems. For one thag, fies often need to do Basic research in oeda to understand better how and where to conduct resaren of a more applied [Link] that rust be # major reauon for ‘he performance of basic teseaech in private in dusty. Many firs need to have a basi research ‘apablity Decne that capability i eseatial 10 making efecve dcisons about their applied re Search settles, For another thing. a basic re Search capability is essential for evaluating the ‘outcome of much applied reverch and for perce: Ing it pose implestions, Tn providing a depe level of understanding of natural phenomens, asi research can provide ‘alunble guidance 1 the deco which there is. high probably of payoifs to more applied research, [a this sent, William Shoekle's educe thon in old state physes during the 1980+ may Ihave been exteal to be dexslon at Bell Labs 10 "ook for subatitute for the vacuum tube in the realm of sesicoaductr materials ~a search that Tea avecoy to the invention of the ttnsisor. fp this eapct a basic research capability is extential product line of the fim an the kinds of process technologies that ought to be adopted. I ean aso be thought of, therefore as providing some defen: sive eapbily ~ ofeing protection agaist the possibilty of = mex, competitive prod intro: ‘ced fom an unexpected direction Tm an even more geneal sense, asi research capably i offen indispensable in order to moai= {or and 10 evaluate Teseach being conducted ‘leewhere, Mont baie research in the Unite Sates 's conducted within the wiversty community. bt in order to“ plugin” to there esearch center and to exploit th knowledge that generated there, 8 fim must ave some in-house capability. A fim is rmoch less likely to bene fom universiy te Search unless it so performs sme basi rerarch, "Ths point is Important aio in identifying & serious limitation in the way economiss fison bout scone knowledge and research in gene fal As T suggested eater, such knowledge is ‘garded by economists at beng “on the hell” nd costly avalale to all comers once it has been produced. But this models seriously flawed because it feguenily requires a evbetaial ce search capably to understand, interpret and to pprane knowledge that has Been placed upon the shelf - whether baie at sppied. The cot of maintaining this expabiity is high, beaute It is "ely to require a enre of in howte sees who fan do these things. And, in cede to maintain {ach cadres the lm mus be willing to let them perform basic research. The mont effective way #0 femain effectively plugged into the scenic net. work to bea participant in the research process "These assertions require some qulieston ard shading Moch san be acoomplihed in monitoring tnd evaluating many Mind of rvearch etnies Conducted clewhere by inhouse persoanel who fare strongly motivated and who place a high valve ‘upon such activities, The Japanese have tfestivsly ‘demonstrated these possibile in the las 390 years for so. Neverteless, the sues of tht monitoring apabiity wil often be determned by the sophie Ueation of the in-hoase staff in eating the signtcance of baie reearsh findings. Moteoves, Japan's monitoring achievements Were cated out Drimaaly with respect to technological rome While fsa was stil in “eaten” mode, rather than with respect to research that wa at oF nearthe sions rontirs. Foran iesghful eatmeat of elated ives see Abramov (I A final factor tha influences the wlingnes of private firms to finance basi euch isthe role EF the federal procurement proces, particularly tniiary procurement. The eusence of this ence fous market sbviouly iftoenes the REED dei ‘Sons of private firs that want to improve tir “bit and ther eligi for goverment ale tary procorement contracts An obvious way 10.40 tha b to signal ons capabilites by peforming RED of the relevant sore This practice ie drelly encouraged hy government sponsorship of design Sd echnical competion in whic potential con- tenetorsparipate, a lest parily at eis Fn frpense. According t© ove se. of estimates in {OBA about 30 pereen of RED expenditures by private industry was stimulated by the prospect of Securing. government. proeutement "conracts (primarily éefence) [5] There are 2 number of froblems wih thete eximates. Furthermore. for Sur preset purpoves, they do not disagrepate {otal R&D ito separate comporents such as sie researc, Nevenhels. they suggest that Tage Shae of private RAD may ot 3e directed toward poral comnercial markets where they might ‘outnbue ie to productivity growth and in proved compettivenes in domestic or intern Tonal markets ater, hey mar be shaped by the tems toe federal goverameat. Tp adtion. the reporting of RED expenditure inevitably cone tains certain amount of subjesive judgment, and the numbers reported wil undoubtedly be in- ‘acne bythe dese to redex bites. The harp increases ip reporied RAD expenditures ‘that have ovnsonall fllowe quickly pon the Introduction of new tax incanives for R&D in ‘several counties may have bana more apparent than real “The growing role of defense R&D inthe federal budget suggest that these concerns are of neess- ‘ne sigiieance, During the 198s there hasbeen 8 Inajorinereas in the relative importance of m= tary R&D. In 1980 defense RAD consited 30 percent of all Tederal RAD expendicres “Thereafter petoent in 1986 (9. p. 226) Although its not Enlil clas what the impact of this growth has rose steadily 19 an (estimated) 72 fof use sesareh spending withia the feseal Sedge. This is because defense R&D expend ‘res are very highly development-intensive com ‘ited to nondefense R&D expentures, Weapons Stems involve immense developmen costs ahd Skew the federal R&D budeet heavily in that ‘selon as the following date ivote (3 Dries eps Benton Be 3 Thus, the growing role of military procurement plus the ising share of defense in the federal RED budget may be exercising major indiret cifects pon the performance ofthe evi eson- ‘ny, It is obviously esoaial to examine mare Carefully the effete of lage mullay purchases Upon the composition of all R&D activities ine ‘lading those in what are regarded a the civilian Stor It ie alko important fo erame the range of seis within the huge defense R&D budget For example, even though only 32 percent of defense RAD is cased as basi, that is 32 peroet ofa extremely large suber. and i Constitutes a sgaficant fraction of all basic re- earch hats fnaneed bythe federal government. iow can these base research actives be char- setrzed? On what categorie of problems are thay concentrated? What connections do they have ‘wih company-finded R&D? Where are. they Complements and where are they sabsuttes? What rete prospects thatthe outpt of miltary R&D tray bene technologies of vale tothe iin 38 irall at to the military sector? Tt seems apparent that these questions do not admit of categorical oF general answers Thee ae reasons to beleve that the valve of spillovers from the mlitry 10 the ciuitian seater has changed substantially over the pst 40 years or 30. These spillovers may very Considerably. depending upon the specific come Potion of research proects inthe defense R&D portfolio. It is especially important 19 krow whataxe the prospects for he emergence of genuinely wal sar” technologies from delanse R&D spending Tn the past there have beon specific eviian technological systems that have relied sobtan- tal benefits trom miliary REED. At onetime of nother inthe past fony years, military and space R&D have made major contnbutons to commer- ial jet aieraft Gnsting airframes, jet engines ‘snd avons) to computers to semiconductors. t0 emmunication atellies, aad 10 nuclear power, ‘Although tis relatively eaey wo identity specific technologies where mutary R&D has generat important evan benefits, measuring the size of ‘eve benefits i far moe diffe. Even so, hare fare szong resons to believe thatthe aatre of ‘hess silver is changing ove ie and ha, i ‘many ofthe industries mentioned the connection between mibtary and commercial research pro fas has grown move tenuous. Ar the reqe ‘ments of miliary RAD have become concerned sith an increasingly arcane set of seeds of modern ‘weapons system, they have moved further apart from he roquirements of evilian markets. fo some rity technologies seems thatthe song em Dass om product perlormance and improvement nd he neglect of cost considerations. has erated 2 gap that has drastically reduced the possibltes of significant splloves to the civilian economy. For a further decusicn of spillovers, se (3) Atthough this pape isnot primarily concerned with questions of poley, it may nevertheless be Spproprate to make a final observation of direct, felevance to policy. Tat i that base research i in very ral sens, 5 long-term investment and needs to be thought aboot in such tenns. Basi ‘esearch represents a commitment of resoures to Certain present ses tat may eventually have & fGeancial payoll, but thee isan unusually High degre of uncertainty stached to this possibly. ‘Wha is clea is that the payoff it comes. is zy unlikely to come inthe ner term. ‘Although thinking about basic research in the same way that an cccaomist thinks about long {erm investment in tangible goods does not ex aust what can usefuly be suid about basic Ye- imteteeuat exercise. The point is that ve have heen considering that portion of basic research that it financed by private industry. where des Sons are expected to be made in terms of cael tions of present cose 4nd prospective benefits, ‘Thos, anything that strengthens the prosper for contua!Tnancil retues is Ukely 19 stengthen ‘he wilingnesst0 perform basic research All the sual forces that Would strenghen the willigness to commit nancial resources to longterm pro- jects become direc relevant to dessins com (coming basi research Ia fact T would suggest that the limas tein thinking about how 12 influence base research deiions theough gover- rent pois ist ask: Doss this action improve the prospect for deriving financil return (veo tually) fom any useful products that may de generated by the hase research? From this polat Sf view, all goverment macroeconomic polices fat improved’ the economic environment for long-term basnes investment would alo ineease ‘he willingness of busines to spend more on base research Basie research, [n eder1 be success quires the making of stable. longterm commit ‘ents. Put negatively. itis Wkly tobe discouraged by enti and unexpected changes in the busioss environment. It i favored by the reduction of ‘inceraimes, by inereasng business confidence land by the sense of sable foture prospect, includ ing the confidence tha government polices them selves wil not be subject to fequent change, Not least important, iis favored by Tow interest cates and reductions inthe cost of capital. Is iet= enily Vue of al long-term investirens Finally, a grester confidence in the strength of one's downstream commerciabzation capabilites Should increase the willingness to peform babe ressarsh, by irengthening the prospect that the fim will capture a larger share of the potential downstream benefits that may be generated by suc research 1 Mos Arann, Cane Up Fron Ate, sd ating Behn, Yun of tora Ho hoe 8) 1 Kenta, Pomona Wore a Alcano Resour iv The Rat ed Dereon of 1 Cepia Bain, ral Sepa AD1M 2 nites Prodan, RAD até Bae Rech ‘oe Pm tae ne Re arama 15) Pk cher, Poe ei RAD Spa ‘iin Prem Canc ers opt 1 eta Br non, Sat (mn Siin Mast’ Bese Reser an run fn ci Aor Eos Re Be (0 Brat Non, ncn een Be md Apt ins Roce ees Romp 1 ew ene Fm, Soe Ins ms 1 Nat Sen Faden, Nano Pts See are Rac Wat (01 Ret Neon he Sale coor ae ene [euch Jaret ate enone ane 939, (na Nae Renee. Jee the ik Bor (Cape Urey Pe Cine 9 ch tnx Naan eine Spe rm nary 2 D Spreng the US Ege ne Nata Wer Sof Tat and Randy ito es) Sonne Bi etn tne OE St dC 114 Bow Tene ratng fm Tahoe anonate, ibe tity ah Po yn) 8-8