Science, technology and regional economic development

In rectz~t yeass it has become clear that a firm’s competitive position is signi fican tl y in flue, iced by its capacity to innovate. This paper explon?s the hypothesis that if there z;e factors operating at t5e regional level which p’t,lfc then fir!ns Coca&xlin particular reinhibit this capa,! gions may suffer. One such factor might be a lack of ten tralized scientific services. Regions under-represented in terms of a viable scientibic infrastructure may find difficulty in attracting new science-based firms, and firms already in thtb n?gion may find themselves at a comparative dkadvm tage compared with firms located elsewhere. Social forces of tr”;iskind might be playing an important part in enhancitqr the social and economic problems of underdoveioped 1qpbns.

North-Holland

,

regional technology a ewelopnient*

Science t%!icy Research Unit,, Unhersity of Sussex, high ton, England

This paper will attempt to show that there are a prkwi grounds for believing that there are factors operating which inhibit the growth of science-based productiorr at the regional level. For many industries technological advance, even in tkre short run, has become a very important influence on a firm’s competitive performance and consequently decision-m:kkg ;It the plant level is increasingly dictated by technological factors. Among these factors are the scientific iind technological ca.pacities possesl,sedby the scientific infrastructure - i.e. the network of public and semi-public institutions whose interests *focus on F: & D, higher education in science and technoJow, informat:ion gathering and distribution, and technological extension services to1industry and, more specificalIy, th? ease with which these capacities can be used by industry. While a closer association between industry and the scientific infrastructcre may produce gains at the national level in terms of economic growth, this paper is more concerned with its implications for the compeltitive poslition of firms, especially those firms in specific regions. Thus to the extent tha!: a region is disproportionately under-represented in terms of a scientific infrastructure ZK? to the extent that: distance between the scientilk infrastructure and industry has ti kleterious impact upon the effective use of this technology source, then firms in that r:gion will be at a r:ompetitive disadvantage compared witJ1firms in regions not so under-represented. Ilngeneral we should *This paper is drawn largely from Chapter III of a Ph.D. dissertation submitted to the Economics Department of the Uniwczrsityof Edinburgh; ’ ‘The Techszo-Economic l? elationship &tweelz Industry and the Wen;rjic Infrastructure - With Special Reference JO the Nectrc;lnics Industry and Scottis;: Economic Development” July 197 1. My mpervisors were Professor J.N. Wolfe and Professor ?. Vxndome. I am ;21sograteful to Charles Cooper of the Science Policy Rer;earck Unit, 1,Mversity of Sussex, who commenteQ on an ezlrlier k-aft of this paper.

Research P’olicy 1 (1971/1972/ 296419

NortbHo //‘and

29Y

AK. Clark

expect smaller firms t)J sU%‘r a won disadvan aage because of their relative inability to supper t internal R & D capacities. In a region with this characteristic there would be two immediate reslrlts. other thhgs being equlll. (i ) Firms would experience higher costs than those in other regions. (ii) New firms would be discouraged from moving into the region. It is suggested that these impacts would have a negative effect upon economic growth potential in tl-e region, especially if, as seems to be :he case, sciencebased industries are those experiencing fastest growth. Sect2 examines evidence with respect to geographical “clustering” of centres of scientific and technological excellence, and the inJustri,f ;.nplications of this. A certain amount of evidence exists to support th:: hypothesis that scientific institutions and techno/industrial ccbmplexes teml to cor.centrate their activities regionally, although this eviden‘:e is somf-v;vhatlimited. Sect.3 expresses the additionaI costs experienced by firms as a result of their distance from the scientific infrastructure, in more rigorous tertns, qualifying the analysis from the point of view of lack of awareness of new technologies, use of inferior technologies. differential scientific rezptivity of firms, and the diflierenria! experiences of large and small firms. Sect.4 goes on to present a pr zhminary articulation of tlte types of relationships which we should expect to exist between industry and the scientific infrastructure, how important these relationships ape and to what extent they are influenced by locationa! ‘factors. To some extent this discussion draws on conclusions reached as 2 result of empirical work described elsewhere [ 1 1.

2. ‘PHE REGIONAL INCIDENClZOF SCIENT’IFICACTIVITY What evidence exists to substantiate the hypothesis that scientific/R & D activity tends to zluster geographically? Although the evidence on this is rather linrited, su& studies that do exist show that the scientific infrastructure does havt a tendeil!:y LOco]_centrate itself into r’airly well d,eflinedgeographic rl lofzitions. There is also evidence to suggest that the same Ipattern occurs with respect to techno/industrial complexes - i.e. complexes of non-profit r:. search institutions, including universities, te:hnical colleges and governme?rt research and development establishments on the one hand and science-based manufacturing activity on the other - although the quantitative evidence on t%is- is at present somewhat limited. A detailed discussion of the types Bf ic;ter-relationships existing and their signific:inc; !r terms of regional ters is reserved for the subsequent sectilons 01’this paper. The clearest case Is that of the lJnite3 States of America where Cahn a;ld

299

Parthasat Jthi 121 point out that ~hc “human and financial resources involved 1Jna.leffort in K & D - cur! \-ntly ( it 966) involving some 15% of‘the annual hlget - are ~nost unewnly distributed on ;I geographical basis wer the COUL Layand... this distortim has, in consequence produced an imbalance in professional opportunities, which in turn has ;:zoduced a further imbahrnce in the original distribution...“/31 . 965 one-third of US scientific and technological manpower Thfere employed by t te five leading metropolitan complexes which in addition held me R & D contracts awarded by the IDepartment of Defense. hasarathi go on to state that in the SiNTle year “some 45”;6of . budget of $ ,4.5 billion was spent in the single state of California. In t’act over the period 1961-65, California received on the average 38.5% of all federal R & D funds, while the five Midwest States of Illinois, Wisconsin, Ohio, Indiana and Michigan together received only 6%. Or to put it more _;trikingy, the “top” 25 states received in 1963 96.8% of all felderal allocations for R & D, while the remaining 3.2% was divided between the “bottom’- 25. ” [4] . ^.This prczess hasI not, of course. been confined to the scientific infrastructure itself bu t has had far-reaching effects on the locail industrial econom:’ and consequntly upon regional economic growth. In the United States thh: two most stir king examples of this are the grcath of highly specialised “R & D comple~.cs” in the Greater Boston area and in certain area::;of California [5] both o‘ which have benefited substantially from flow:; of governmentsponsored research to these areas. indeed one commentai:or maintains that the scieirtific and engineering manpower pool built up in the Greater Boston area ha:;, “done more to revitalise the Northeast (cf the US) than anyone would have guessed 15- .20 ye rs ago” [6] . These regional benefits applear to have accrued largely as a result of government funding of basic and applied research to institutions both within and without the scientific infrastructure _.- i.e. both to publicly owned research laborarories which may or ma.y not s&con~ ract development contracts out to local industry, as well as direc tly to r:rivate inijustry itself. Shimshoni concluties that “Government expenditure Ii-rashacl :Irn overwhelming impact on new company formation becarl;se of I;;overnment induceo demand and government investment fol technohgy. In I:he Bo,;ton area, about one half of the firms surveyed had more thrn one third of sales, during their first two years (of operation), directly to government cr to prime contractors, For about one third of the companies, government siles amounted to two thirds of their markets*’ [7] . On a lesser sca.le but pAbl:y t:qually important foi*future industrial d;ilvelop*‘The NLtional Aeronautic: and Space Agency.

mjents in the United States is the pirenomenon of t.he “Research Park”, These can best be described as rather specialised industrial estates catering specifically for >cience-based and research-orientated industry. They are very often sited near universiti,:s reflecting the growing confuence of interests between university and industry. The universities can sui?ply a range of :nscialised skihs, knowledge and facilities while industry can supply employment opportun 1ties for gradua es, provide avenues through which universi.ty pcl sonnel can try out ideas commercially and act as a breec;lingground for new spin-off clornpanies. indeed Browne maintains that ‘-it may be that the most significant infiuence upon the success or failure of a research park is the proAmity to, and the interactkqn with, nearby universities.” [8] . [here are now more tlhan 30 university-linked research parks in the US [8] *. A recent study of the scientific infrastructure in Czechoslovakia shows much the same tendency f in’ R $s-D resources to be “located prim: rily in certain regions. Miller and Nejedly [9] examine the regional distribution of R & D llN2tivity in Czechosloyalcia concluding that there :ire 1 .arge regional inequalities 8S measured by data on scientific and technological man.power. 49.2% of all R 4%D university-graduated manpower is concentrated in CLntral Bohemia while a further 27% are employed in two other administrative centres (Brno ratislava). The remainder are dispersed throughout the further eight regions considered in the analysis. These inequalities are only slightly modified when “pure scientists” are excluded**. The: authors go on to perform a simple statistical analysis in which, they test the hypothesis that regional differences are caused by the differing economic characteristics of the regilons. Using data on production, capital assets and employ,ment as regional economic indicators they obtain norm-significa#ntvalues for the correlation coefficients, but using numbers of university faculties and numbers of univerents as indicators of the regional resources committed to education in high and significant values for r. Thus the evidence suggests that R & D centres tend to be sited close to the “traditional focuses” of science and technology -_ the universities, which in Czechoslovakia are located mainly iin the regior s of central administration - but tend not to be attracted towards areas of high economic activity. icipating the criticism that a rough analysil; of this kind - i.e. correlating general economic *ndicators with R & D a.ctivit;l regionally -- ignores differen-

* See ref. f 83. The sme edition of Industtim?Reset rch lists a total of 101 such parks itnthe United Stares and Canada. m That is done by e? clwding scientist:, who work in institu KS mntrolled by the Czech Academy of Sciences and in order to ascertain the distributiw of specifically industm.l R & D cap&y. See ref. [U] , p.8 ei seq.

tial scienliflc receptivity* amongst the regions, Mii;ler and Nejedly exarnirle the relatiol iship between the production levels of specific science-based indu!;tries an4 R & D effort. The analysis is ;igain conducted cross-section&y over elever regions and six diffe.rent science-bas6d industries are studied. The resultant correlation coefficients show a much better fit and in two of the six industries - inorganic chemicals and pharmaceu ticatls - the coefficients are high and slignificanli.Nevertheless iiller and Nejedly conclude that industrial activity and research activity are not linked geographically in any specific way. While tl& analysis does not altogether agree with US experience it is clear that great care should be taken over its interpretation. The relationship li)etween science and production is not one that can easily be picked up by statistical means and a lot depends upon the relevance of particular technoilogies to particular types of producGon. Clearly it is necessary to have information on the technological distribution of these global R & D figures and -the degree to which R St D spending in each technological area has reached an optimum level with respect to the industrial sectors examined, before passing judgement on the influence of industrial activity on regional R & D spending. Thus a situation in which R 8~D investment had reached an optimum level in those technologies relevant to the inorganic chemicals and pharmaceu ti:als industries, but in which there! was under-investment in those technologes relevant to the other four industries analysed, could produce correlation coefficients of the values reached by Miiller and Nejedly which would not necessarily be inconsistent with the hypothesis t at industrial activity is influenced by regional investment in R & D. Moreover the centrally planned nature of the Czechoslovak economy hai an important bearing on the issue since a lot will depend upon the perceived need for industry and research to cooperate closely. To the extent that science-based industry performs little “in-house” R xt D and to the extent that planners do not recognise the advantages to be obt(ained from close contact Jetween factory and laboratory, the results arrived at above may simply rsflect inefficiencil:s in R & D resource aklocation. Shimshoni IIt0] points out that one of the primary factors in the development of an R & D/industrial complex is the tendency for firms to “spin-off” from local laboratories, started by entrepreneurs who were previously laboratory employees and of?en supported for long periods by contracts, technology and know-holw from lthe parent laboratory. One would not expect this tendency to show itself to the same extent in Czechoslovakia. * Le. the industrial structures in csertain regions may be swclt:that there is a relatively science-base and therefore less need for R & D activity.

cot rse, tire analysis is to some extent vitiated by t e nature of the approximirtion to a true measure of data &ich is at bly low in R & D manpower is in fact produc?ag a very ly snaking) ou arch because of a variety of h&kc,~a~i~a~/~ab the nature of the te~hn~~logiesinvolved. s may be carried on in a small number of top-heavy inefficient instituions whilst in the “low”’ regions there may be ahe same number of instituoperating at hi r levels of efficiency because of advantages assost site and inefficiency are necesith smaller size. ou’; thar the value of the R & D sarily highly car lated but merely to effort wtil depend upon :,tva s and baxing correlation analyses 011one particular measure of it ir bound to open the way for bias. Clearly we should require a lot more information before WC: coulcl say that in Czechoslovakia there is no tendency for industry and reslearch to locate themselves in the same geographical areas. What is interestin!: and rather more significant, however, is the regional distribution of scientific activity by itself. sh experience again produces similar patterns. The geographical disThe tribution of government research stations and lsniversities in the UK (1966) shows fairly marked clustering patterns with high densities in the Greater on area, Oxfordshire, Northants, Edinburgh and Aberdeen. Indeed the concentration in the south-east of England is pronounced. Of the 200 or so slations 153, or 75%, are located within 3?4,hoL.rs train journey from London ( 4960 speeds) and 143, or 71%, are located $Iouth-east of a line drawn between the Bristol Channel and the Wash. Of tht! 32 (16%) Scottish stations, 9 are MRC stations *, 12 are ARC stations [ 111 and 5 otners are directly concerned with the fishing industry. The remainde- are t.he Naval Reconstruction estmzh Establishment, the National Engin,:ering Laboratory, the Road Laboratory, the Unit of Grouse and Moorland Ecology, the Royal bsenratory an j the Royal Botanic Gardens. Of these six stations one would c:bxpectonly the first three to have substantial direct relevance to a modern s4ence-based industrial sector in Scotland. Nljr would one! expect the ARC amions to be of much significance in this regard and the same probably holds l.rue with respect to the MRC stations witk the rider that there may be ;idva3t !sges in cios contact between this set of stations and the electronics Industry due to the fact that much of present tlay advance in medical research involves the use of electronics iristrumentaticbn. At the same time jt is ve p that the nature of government R & $1effti;t in Scotland re Rese_wh Council; A C = Ag,ricultural Research CotmciL

Science, tee hnoiogy and Yegionalem nomic development

3u3

importante of the agriculture and fisheries industries. In fact three out of the at .ons lot: ted in Aberdeen, an important fishing centre, are directly concernec: with research into problems connected with fishing. These gecgraphical concentrations do not hold trtle with respect to univl:rsities but id is interesting to note that of the 47 grant-aided industrial research associations only one is located in Sco*tland although two others have substations i-1 Scotland [ 121. Many of these research associations are located in areas where there is a concentration of the relevtant industries (e.g. British Ceramic RA, Stoke-on-Trent; The Cutlery and Allied Trades RA, Sheffi.eld), ng the original aims of the DSIR* which walsto ensure a close connection between research and industrial production by way oC ;etting up .these institutions. Clearly these locational regularities are no accide:,nt. Finally. a further aspect of these general relationships is evidenced with respect to two constituent set s of the scientific infrastructure, the university and the government resear establishment. Not only dc:-esit appear that the formal and informal links between these two sets of institutions are. considerable, there is also evidence that the degree of contac t is substantially influenced by regional factors. In 1965 a Working Party was set up by the Council for Science Policy in the UK to “consider the question of liaison between universities and government scientifi.c research establishments” [ 131. Its remit was to assess the existing situation: and to “examine whether (the facilities possessed by establishmeaits) might be made more accessible for educational and research purposes at universi.ties, and whether the specialised knotirledge of workers in research and development establishments might play a greater part in higher education”. It is clear, however, from an examination of the Working Party’s report that it exlvisagedthe advantages accruing from such contact to be much wider than just this, insofar as substantial economies would in its view, result from closer liaison. These advantages are: (a) Since much of the work performed is cornlmon to both sets of institutions close cooperation avoids duplication of research and wasteful compe tition r scarce manpower, facilities, etc. jn particular the cooperative USA of expensive capital equipment (for example, computers) wtculd economislt3on public money. (b) Boi;h se,ts of institutions are reposij.c\riesojf much tec’hnicalknowledge

* The Dt-!partment of Scientific anti. lndustriar Researc’h. This badly was closed ibwm in 1965 ;;jad most of its functions taken ovelr by the Ministry of Technology r(nwr the Deparltment of Trade and Industry).

&ifthis knowledge would considerably improve the quality c\f s woAd benefit both ts of institutions at a more el. On th 2 univers ty side there woul ution th;lt the ac emit scientist can 1 problems and to in ~e~~t~ststc~~~rds the tykes of p~~ble~~~s ~~~~s~~~r~d~~~~rtant nat~~n~~y and industrial 4tate the rxruitment of sAtably skilled manpower to is!zmt:nts. On the establishment side there would the “gractica! si ifkances of advances in pun scienc? [ 14) made withln the univcrsit-es. The Working Party goes on so point out that university re!;ears,ir~ab~rato~~s are “aften livelier than govc:rnment research laboratorie;s ecause of tne contin s influx af fresh young minds. Each year there is the stimulus that about 3 group is replaced by newcomers, whereas ~o~ernrnent fabo m those going through a rapId expansion, recruit annually a very lirlited number of new young staff... i SC! that)... a wider poiiq of giving fixed-term fellowships, or short appojnt~~~~~~ts~ to a fraction of the younger scientists employed at government la~~~~~~~t~~~~~~ a:ld a more frequent involvement ot’a few of the permanent staff in the t2a::frin and research training of the next generation of scientists :ouid bc very tacncticial in this respect”[ 15). It would seem, therefore, that ttle Workin; I%rt!j believed that government establishments stood to gain just as much frc>mcloser li;;lison as the universities ant it is intcrzsring to note that, 4e evidence ot tained during the enquiry. rhe establishments were ratl.er ive in their welcome for a closer relatonship [ 16) . ‘ritte I altd oral evidence was obtained from all government employers of s rientists, and 41 universities and colleges of advanced technology. This evidcn~e sur~~rn~~rised the existing state of affairs and the views of each institut im a:sto how t e situation could be improved. As t1n il1ustrAon of the type and extent of contact involved it is useful to take the case t)f the national laboratories contrctlled by the former Ministry of ‘echnoLogy, siJr9 their experiences broirdly to parallel those of otner olas;es 01’governn. cnt establishment. ese M:ntech laboratories have ex :ramural re search ancl teaching agreements, consr&ancy arrangements, research f’ellov-shipsand arrangements for the employment of vacation studel-its. Uniaff serve as memb-bl3 of establishment steering committees md also on specialist sub-committees concerned with specific aspects of the eve I ;iIi ramme. Directors of establishments wo+4 &sely with un tments in placing research grants wit5 clnivcrsities and t

aist~eildp jwerect to receive individual research studlents to work, say, to7wards a Ph.D. d :grt c”at no cost to that university. For exar ~plk:, the National PIlysical Laboratory ([NPL) receives re(;ul:srly a number Jf requests to help in teaching specialist courses at universities. Courses we been arranged OIKchemical and engineering standards, mechanir8eeri~g, ship control ;lnJ stability (University of London), on itemperature me:surement (University of Surrey), on mathemati%:s(Brunei University). Alscbthe NPL have arranged a number of joint researclhprojects 7with universities, e.g. the development of a Michelson stellar interferometer (Susserr).,the measurement of spectroscopic line profiles (Oxford), the dl;:velopment of a laser interferometer for seismological purposes (Cambridge). The same geT)eralpattern of relationships seems to exi,st with respect to the gcwernment departmental establishments - particularly the defence departments - with othe bodies, such as -the United Kingdom Atomic Energy Authority (UKAEA) establishments and tith Research Council Stations. The U&WA, for example, had 140 consultants from 30 universities, several members of their staff hnrvejoint posts with universities and there exist a number sf joint research prc jects [ 171. The ‘Research Council sltations, in addition tcl the normal ‘rink:., actually have research units operating “inhouse” withir!, the univers6tjes arld in those cases contacts are particularly close and, it seems, mutual!/y benr ficial. Not only, hottever, are the relationships between universities and government research estatrlishments far-reaching, it becomes quite clear that there is a definite regio131pattern involved. The Working k’a.rtyreceived evidence from 40 universitil::$and colleges of advanced telchnology relating to the nature of the links exisr.ing and to the particular establishments widh which thes’elinks are: maintained. Although little attempt was made to analyse the regional importance c f’ its findings, the detailed evidence presented in Appendix (CI[I 8] has enabled me to perform a rough quantitative anallysis of the position existing a! the time (1967). Out of a total of 38 universities, 26 [nearly 7’0%)h,a;l clos 3 links wi.th establishments in the same geographical area &~ .Iyoughly with &ablishments sited within 50 miles of th,e campus). Of these 26 (only seven (l& of the total:1 had equally important links outside their olNn areas. ‘l’wo (5%) appeared to have contacts which were definiteljr nonregional seven (1857,) alppeared to have little contact of any kind and three (8%) di’d not give sufficient information upon which a judgement could bz made. JI47.u~ despite the rough nature of the analysis it is clear that many rnclre universities had closer relations with neighbrouring establishments &an vice:ve:rsa. Typic al of tl~r:former class sl’ u -liver&y iz;the University of Edinburgh which has the i’ollovrial2 Ii&s with rl:!:searchstations in arId around Edinburgh:

(a) Five Medical Research Council units (MRCj. The honorary directors of three of these are full time members of universe~4.staff while the remainFIg two have staff with honorary senior lecturer status. Each of these units opcxates within the University itself. (b) Seven Agricultural Research Council IJni is (ARC). Ealch of these is closely involved with university departments in research and teaching. (c) There is also contact with the Freshwater Fisheries Laboratory itlochry 1, the Scottish Marine Biological Association’s Oceanographic Laboratory (Edinburgh) and the Roy Observatory (Edinburgh). The only local station with which there is no contact is the Naval Reconstruction Research Establishment (Dunfermline). (d.) The Regional Computer Centre set up “in-house” in the University has pn:,nloted close links with local establiihments. Similar close links are evidenced with respect to the University of Birmingham and the Royal Radar Establishment (RRE’), Malvern, the University of Exeter and the UKAEA station at Winfrith, Dsrset, and th.e University of Strathclyde and the National Engineering Laboratory (NELj, East Milbride. Finally a number of universities mentioned in their evidence that they would like government research stations to be situated nearby. Thus the University of Phewcastle maintained that it “would make much more use of the staff of research establishments if they were s.ituated an ywherc near (the University), but a list of such establishments contains very rew places north of Leeds and still in England” [ 191. Of course, this general legional patterm should not be oversta*.ed. There were a number of cases I.particularly with respec : to UK AEA stations) of links maintained over quite a \-videgeographical distilnce (14berdeen University and Aldermaston, for ex;~mple),but the evidence clear1~4!&ows that. such links were the exception rather than the rule.

3. TI-IEORETIC.4L ANALYSIS ‘The previous section produced a certain amount of evidence to sugges! that scientific activity tends to concentrate itself rcfgionally.There is also a cr:rtain amount of more limited evidence which miglht lead to th’e supposition that ccntres of scientific and technological excelle .acemay exert a location:11pull on thoslz industries whose products are related to the appropriate “‘sciencebase” of these centres, although, for reasons which become clear below we should not expect this phenomenon to reveal itself explicitly. This section d the following one will attempt to provic e a r; “ionale for this g)rQcess,;i rationale already demonstrated tenltatively with respect to the mutual advan-

Science, tehologv

and rqional t~oncunic development

307

tages associated with close linkages between different sectors of the scientific infrastruciure. The focal ooint OPthis, paper is ths importance of science-based production in Y regional develoI,ment) sitlce : 1: is precisely this type of production which exhibits fastest ,economic g:ov:th, While there is some doubt as to whether ve activity cbii ‘be lmked dire&& to growth performance at the vel, then is a lot of evidence to suggest that firms within particular industries do benefit from increased R & D spending within certain broad limits [201. in oonsequencl: if there are factors at the regional level which inhibit the effective tra&aGon of new techniques into production, then it is reasonable to suppose thai these factors would inhibit potential economic development in certain re@ons. ‘W/hatmight these factors be? It is suggested that one imporl:ant variabkAmay be the ready access to new technologies and the important source of potential markets - represented by the scientific infrastruc tUi3. Moreover, The operative phrase in “ready access”, which implies that timI::and distance araypertinent variables. Shimshoni [21] mentions a number of reasons why nearness to research1Iaboratories has a significant influence on the performance of new science-based firms. These are: (a) A reduction in the ccsts o.f acquiring market knowledge. he impl3;rtanceof rapid communications Sor the selling of technologically-intensive products si:rce :t.pplications engineering often requires continuous *technic31interchan;;e between buyer and seller. (C) Serendipity - a ter n denoting the prob’ability of diffusion of Ideas, techniques, etc. through cElanceencounters with knowledgerable!indivicittals. This probabili&. will be greater the larger the number of skills and intczrests represented in a given (sm>ll)lgeographical area. This concept clearly includes the typ:: of informal technical irnterchange whichlcan occur when people with mutually releva3%interests .neet socially. (d) Ready access to spel:ific research results which may be of relevanclefor production*. (e) The availability of a pocl of highly skilled manpower which may, for a variety of reasons (cultural and social), be reluctant to move away from the district and may, therefore, be anxious to take jobs imlocal industry. (f) The avail:rbility of risk: c:Jp..talfrom a findlcial community with tlsllderstanding of the lgroblems and ritjks of scientific en trepreneurslhipC (a) and (b) are peculiarly relate d to the role of the scientific infra+ truct ure as an important market but one wi*[)la key tixtra dimension, i.e. the necessity * Shimshoni finds -.har; this is a relatively unimportanlt factor from his interviwi~s Iwith budncssmen.

N.G. Clark

308

for close user/supplier contacts when selling a c;ophisticatE;d capital gorjd. In ai:idit:ion there are two reasons which Shimshoni does not mention but which are clearly important. These are: (g) Ready ACCESS to centralised faciliities st!ch as computing and library services which it may be beyond the cost of individual entrepreneurs to d on a sufficient scale. (h) Possibilities for consulting ke:y individvlals from the scientific infrastructure with respect to particular technolol;ies. ~4 special case is that of seoandment for a certain length of time. It is apparent that centralised inputs of this kind wi!Il play an important role in the production of science-based commoclities. Indeed what has been &qlonstrated by Shimshoni for the new smal’i firm will, bsy extension, hold true for the larger, more established firm although we s:hrnuld expect the impact to be P.ess strong. Goins back, then, to the locational pull of the scientific infrastructure and discussing it in ecsonomic terms, it seems reasonable to hypothesise that the important variable is the cost of acquiring relevant technologies which increases in direct prclportion 2s the time required to obtain the technologies. Thus, for example, where close technical ir terchange is necessary between buyer and seller both of which are located at some Distance from each other, periodic visits will be necessary and the cost inYolved will be equal to the direct cost of travisl plus the foregone prc)dLI:tion of the engineer(s) in question. Evidently the lirm located close to the market will be at a comparative advantage and twl;, firms operating in &fferent regidns, one close to a number of laboratories and the ofher some dist;ince away, but producing the same product under lhe same economic environment would exhibit different cost functions. More rigorously, let the relevant production function be 0

....X.~!

=f(x1,q,

XJ

I

where x

are the conventional factors of production (labour, ,apital, etc.), -xt represents technological input, 0 representts output. Then the appropriate cost function is represented by 1,

x2,

._.,

x,

c= PlXl + P2x’:2 + -‘* + P,,Xn + p*x, ?

/

wl-rerr: pl, p2, . . . . pn are conventional factor prices, pt is the price of technological input and C is the total costs. ,hssuming that firms attempt to m;lximise prOfits - i.e. minimise cost5 for g:ven Ieve& of output - and that prices and quantities of all other factors remain the same, then for a given level of output but differing prices of techncjlogf the total cost differential is represented by

Science, itechnobgy Lnd regiorzaleconomic devehpmer,t.t

309

where ,rk”) p’ repro sen t the prices of technology to the firm located close to and the firm “oca~ed at some distance from the scientific infrastructure respectively. In the :lbove anal!/si:il technology has been treated as a single factor of production but it is possible to generalise to tak, into account the different “‘elements” of technolog? ’ required for prolduction. In this case the COSI differential is given by

where there are jr elements ‘of technological input. Evidently for operatilonal. purposes it may be necessary 10 separate out these elements since pric:e dzf,, ferential due to regional factors will be different in each case. The overaI!l hypothesis is, thesefore, that rteE’erisparibus any advantage which one firm has over another- in terms of ready access to technological inputs, can be expressed in terms of the above cost differential. There are, howevea., a nlurn.. ber of points which require clanification. (1) To begin with technology is conventionally regarded as a factor which shifts production possibility curves rather than producing movements, along any particular curve. This is related to the fact that whereas conventional inputs’ can be regarded as contributing directly to production at a @ven moment in time, ahe impact (Oftechnoldgy is seen., correctly, as inf!luerlcing production over a wider time perspective and is therefore included. as an exogenous variab!e in most. production furlctilons. Salter /122] for e:c.arnple, views technology as being embodied in cllpital stock and consequently as being beyond the control of’ thle firm in the sholrt run. New technology will be introduced via new capital equipment only wlhen the discounted stream of future benefits exceeds costs (including normal profits) by an amount greater than foregone net benefits using the old equipment. Where techrmlogy isi regarded as being disemboidied it is still viewed as the result of cumullative investments in innovation during past periods (Minasian [23] , foli er;ample., achieves his most significan t results on technical changl: in the chemical in., dustrli. using a distributed l;ag ynodel of K 6~ID activ:lty.) Nevertheless, it is still possible conceptuJly to regard each firm - in the abovrz example -- ;ISusing essentially the same techlnol ogy, whether embodied or disembc>diczd, but having different Zorrg YIAUIcosts because of unequal acc,ess to external tech. nolo;gzr. Of course, enormous problems would present themselves were empirical measurements o!f such cost differentials to be attempted*. * NoI-mally, of tours :, a distincticsn is made betwee, the “long run” a.nd the “very ioiig run”. the distinction being that in the long run the entrepreneur varies his Axed capital while in the very lo 11;run ractica i cc anges in technology add a felrther dimensi~~nto the entrepreneur’s procLuction possl.bilities. While this may be true in the case of ver~r radical innovations, like the jet (.ngine, I would argue that for science-based industry in

at%~d if there arc net di~~ren~es due to ac~l expressron becomes

payments for techno

A@= p;x;’ - p;.x; + x;‘(g” -g’) , where xt”, xtr are respectively the optimal and suboptimal technoiogies. An afivc scht?me might be to regard each firrl as having a different proba&+y of acquiring the correct technology. This takes into account the fact tllat not all firms will be aware of the optimum set of technological inputs. because of differential locations with respect :o technologl sources. In this case t’neappropriate production function is

where &xt) = the probability of thy optimum technology being utilised. The cost differential is now given by AC = p;p”( : ‘It j .- p;lqA$

i.lsing the same nomencl;lture as before. This i; clearly analogous to the preGous Gase since flrrns with low p(xt) will tk:nd on the average to utilise

general technological chan,ge takes place over much shorter tiime horizons, either through new capital stock (Salter) or through new disembodied technologies using the same - and/or new -- capital sroc‘l~ The important point is that technological c2lange involves costs and that, from the Ijoint of view of an analysis which tries to separate out one particular element of these costs in a gencxal sense, the ldistincticns between “short run”, “long run”, and “‘veryloLAg:run” are academic. * i.e. assuming that the entrepreneur will at the rjarg, 7 trade off inferior quality of product against the relatively higher cost of obtaining external technologies. This dces not seem to be an unreasonable assumption.

3t1

the enalys~ in pr~..hakility terms is in highlighting nge is occurring so rapidly, iv.3 must be the exception rather rd it is irnplietl that firms loca red close to centres of scientific excellence will be at a cmrqxmtti~*eadvmtuge. No&e also that there are sca!e factors present. Thus the relative advantages to a firm located near one particular laboratory m,ay in general be small, sine 3 it is unlikely that this laboratory by itself will be :apable of supplying mor: than 3 small number of the various Wemen ts” of technology required for producti~~n. It is suggested that there c!xists for Imost industries api essential1 disconrirmity between technolo;g and production such that an appropriate: “technology package” for a given product All contain inputs from a variety of scientific and engineering dMplines. Fo - example, the production of electronic instruments requires inpu fs from metallurgy, semi-conductor tech,nology, paint technology, spring technolo~ and a variety ofother disciplines anrl sub-disciplines. Clearly the a&hntages fill bz greatest for the firm situated close toi a number of laboratories, const&.ting ;a complex of relevant disciplines, and we should, there Fore, expect Ip(x,) to be directly related to the number of laboratories withiri the geographical environment of the firm. Thcl relative success of university-ceiltred research parks in the United ,Stcite:;provides some justification for this Brypothesis. (3) ‘We should expect the size of firm to have a significant inflwerice on differemial cost structures. Economies of scale have not been taliceariinto account in the specification of the production function but it is evident that the larger firm may, for a variety of rczsons, be able to offset disadvantages associated wit distance from t echnolofy source:‘s.bi:aving aside questions of econom:es of scale and market mperfecitions, one important advantage tlrrhlich the l;_u-g,e firm possesses is the alility to rely very ~n?uch more updn “in-hOou,se” technological activity since t:hcfoverhead costs of an R & D departmerrt can be covered more easily. This w 11 be particularly important where new \ f:ch* Another result may be that lahck:)f awareness of new techlnologicaf opportunities will reduce the scope for product a’Pw~,w@tltiunand this could1be true even whe:re no new investment is required. For exannIe,* a plain? producing polyethylene can very simply be converted to the producticln 1If ethylene oxide and in a situation of under-utihsation cf capacity the knowledge tYldf this caSnbe done (and. also, of course, o’f hc w it is done) would be extremely value 11cfor the firm. 1 am grateful to Charles Cos:lw:r of the. Science Policy Researclh Unlit for pointing this out to me.

312

N. G. Clark

!?rPlogie$*are developed in secret which give ihe firm a technical lead over r:ompeWors, but it may also have significance insofar tij it obviates the necessity to go outside the firm for certain technclogical elements, such as computer services. &her things being equal,, therefore, it would appear that the srnal!er firm stands to gain more [or lose les:;) from close proximity to the scientific infrastructure although again this ‘till vary from industry to in(dustry [24] since for some products a relatilrely small firm may be a yort a viable R & D department. 14) Similarly, it would appear that the economic signif?cance of such technology costs will vary from industry to industry. For example, industries with a high scientific receptivity would be affectted morz than industries at the “mature” end of the product cycle. Also industries relying on a small number of specific well-defined technologies would be affected less than industries \.%ich depend upon a wide spectrum of technical disciplines. (5) Finally, abstractin, from industrial differences of the types mentioned above, it is plot possible to state how important technological factors are in relation to other factors known to be impojrtant in industrial location. For exam+e, nearness to supplies of a key raw matc:rial or closeness1 to large ma&e ts uf ;r non-technical nature might outweigh the disadvantages of distarIce from sources of technical knowledge. Thus, although we have seen that ‘there is a certain amount of evidence which suggests the significance of the technological factor it is still to be shown that its relative significance is sufficiently important to justify, for example, policy action. This analysis is not intended as a rigorous treatment of the costs of technology. Quite clearly it could be refined extensively, as we have seen, to include market imperfections, inter-industry differences, scale ef’fects, etc.*; and it would have to be extended in thi.s fashion if a viable statistical analysis of differential cost structures due to technologil:al factors were to be attempted. However, for present purposes it is sufficjent to state the nature of the problem. This has been done mathematically for clarity of exposition.

4. SOME FURTHER ELABORATKIN In sect.3 it was tentatively hypothesised th:it firms located at some distance from concentrations of scientific and technological activity would experienze * One point, for example, which has n :t be’en discllssed explicitly is that there may ve;y well exist economies of scale in cc~~mmunicati,)n, such that the effort involved .n azquiring a particular technology ma)- prodllce knefi s, not only with respect to th,\t technology, but ~Iso in terms of new technological opportunities not hitherto recokMed.

Science, r ~chnology

~;ofdregion,21economic develop went

313

greater costs of acquiring appropriate technologies Ihan firms situateid close to such zozrcentrations; and a ldelineati~n Wal!; made of the types of technology input that I ,ay be pertinent. Hn this section s”iidI tiitizt:!ztF &se technologic;il factors in more d&ail and at the same time attempt to specify what sorts of advantages accrue to the firm as a result of close direct links with the scientific infrastructure. Tlhis discussiorr will not be systematic: in the sense that it will consist largely of an exploration, in (G! piori terms, of the nature and value of links which we slaI:)uld expect to encounter. Although the scientific infrastructure is clearly much mc re br;goadl:ybased, 1 shall confine my remarks predominantly to academic bodies - i.e. university and technical colleges - and to government research laboratories, since these seem to be the most important generators of new technologies. 4.1. Actidemic bodies An important regional impact of‘ a new scientific institution such as a university, is the spin-off of specialised R & D and/or production companies from the parent body*. These are essentially started odffthrough trained manpo~wer perceiving a demand for particular technology and/or products based on their own work, leaving the university and setting up on their own. The markets for these products may be the parent university itself, other universities, other scientific institutions, SCience-based indu.stry., or any combination of these. Such companies often tend to set u;r in the immediate environment of the parent body, even where there is no “researc;lCIpark” of the kind discussed in sect.,& and it is interesting tc peculate on the realsons for this. One important factor might very well be the natural reluctance of skill4 personnel to leave an area where they have built up strong social ties both1 personally and as regards their families (schools, housing, etc.). Inertia of this kind may be more powerful in the case of personnel who have had their initial training in the corresponding institutiorr. Secundly the new entrepreneur may wish 20 remaj :I in the: same area bec,ause of easy access to former colleagues who have specialisec: insights into particular aspects of thle relevant technologies. In fact it is very likely that such colleagues will act as c:orrsultants to the new firms. Tlrira’l~ . closeness to the parent university may be desirable insofar a:; there ar: a number of specialised facilities available, uniquely related to thl.: technologies of the new firm. ln&ed we should expect this to be t.he ~asc since the new product has, almost by definition, originated from R & D conducted within the parent laboratory. Let us suppose, for example, that a anew type of * See xfs. [ 2,591. Both studice.; pro Arce iwpressave evidence testifying of thi 3 phenomenon.

to the importance

314

IV.C. Clark

semi-conductor has been developed in a universyty research laboratory and thal: a number of engineers, previously laboratory employees, have set up a nlew company to produce this component commercially because they em i&m age that it has a viable market potential. It ma:! very well be the caze that clerttain testing and standardisation facilities are required which are directly relevant to the new product - and may even bc specific to that product but. are housed within the laboratory and are too expensive for the new small ’ company to invest in. Clearly there will be consiclierablelocational advantages in siting the new factory close to these facilities In addition it will.normally Ike the case that moving from a laboratory proto type to a commercial model will involve changes in design and/or raw materials and alt this stage the “de-bugging” of the product may require the use of expensive apparatus and/or technical expertise which is ody availaIde in the parent laboratory. IEvidently extensive discussion may be necessary bet ween the parent and the firm and there will be, in consequence, a.dvantagesin geograplhicalproximity. RxuM”, the research in the university laboratory, upon which the new product has been based, will normally be part of an on-going programme lwhiclzmay very well provide advances in technollogy relevant to that product. In this guise the parent laboratory can be seen as a continuous source of new tel,:hnclogy giving the spin-off company a conrpetitive edge over its rivals. Ther z may also be advantages to the university. For example, the “problems” experienced by the new company may be useful vehicles for the training of students who may be required to carry throui;h and solve simple research problems as part of their degree work. F’#!r&, more general facilities (e.g. computing and library Tfacilities)will be availalble free or at low cost to the new company. Finalry, the role of the parent as a market or, possibly more importantly, as a testing ground for a wider market, should not be overlooked. What has been discussed above relates to the spin-off company which has by definition close links with the parent laboratorr. Evidently, however, there is amother class of “non-spin-off” company for ,which the advantages of close raladonships with a university department rna;lrbe just as beneficial. Such a company may have commenced manufacture in close physical proxirnation to a university for a variety of reasons. It may have been started up by a man who had received his initial ,training in the area, who had gone to work with a larger corporation elsewhere but had subsequently returnea to the original :rrea for personal reascns, or because he anticip,ated some sort of benefi.t from eing carriecl on in the university laboratory. This company would 0e essentially en repreneurial but could not 04: described as “spin-off””in the e of c.~mplany might be a sales :lffiliT;t
Science, technology

and regional economic development

315

elsewhere which has decided to go itlto manufacturing (W partial manufacturing) of the Televant products. A special cat;e is that of the sales affiliate to an overseas company. v Flit may commence production. for a variety of reasons. Thus differzntial tariff policies on ‘in al as opposed to intermediate goods., I(jr the necessity to adapt a product i;o render it suitable for the home market may create an incentive in favour of setting up manufacturing (and possibly R & D operations) within the host econ~omy.One factor influencing locatiolnal choice might be the existence of a university laboratory possessing skills relevant to the company’s activities. A posr;ibly more iml3ortant incentive might be t!lle existence of regional devlelopment policies der;igned to encourage new firms into an area - e.g. tax holidays, grants, low rlfnt fat tory space, etc. Now althsug’l these non-spin-off companie!; may not initially experience such close technological inks with the labocatory - excepl:ing 4fose compainies which have mo~etl into an area to take advantage of specific technolog3za.l capacities extant in that area - it is to be expected t’hAtthey will become aware of the technologies and facilities available, an~dit would be surprising if some of the:;e were not exploited. In this way closer links will be forged of the kinds mention& above such that ideas, techniques and insSightsdevelolped in the laboratory w’,zlbecome integrated into the production activities 0%the firms. Furthlzmore, the firms may see new commerrzial opportunities in partilouse” developments alld may shift their product structures into cular activities which arc more in line wit!1 these techniques, often by sfferinlg facilities for university personnel within the firm. Shimshoni places stress on this mechanjisrn,m;tintaining that technology “is o:fitentransferred most effectively b!y thle movement of a man into an existing company rather than starting his own. New technical companies have been particularly effective “hosts” for :;uch movements” [ 25,26] . 4.2. Gowmrnerlt rvsearch laboratwies We should expiect the relationships here to follow similar lines to those with academic bodies *. Thus the government laboratory may spin off new companies and rot as a I :chnol ~gy source to other companies sited in its ww area. Library %,ndinkxmation services, computing facilities, more speciakd facilities su& as test:, calibratior I and precision engineering facilities, skilled manpowr, e lx., will each act as technology “gene:ratws” or “diffusers”’ and could play a significant role in twhno1ogical developments in the local R & D economy. * This is lrq#:itly recognised by i he SutkleJand Committee’s Report insofar as it emph2;i;eS the groqrving confkence of resfxir,-‘1-rink-W., **bet ween universities i3rlClgWernnrenl rrrsearch laboratories. See 1.9. [ 11 J .

316

N. G. Cimk

)_~owever,it is convenient to mention three possible differenc:es. FirstlJJ, to t’he extent that work related to national securit;r is being performed in government latloratories the technology potential for the local economy will be reducs;d. A prioriwe should expect this tc, he thi: case with respect to laboratories administered by departmentmsof dcfence. Secortdly, we migh the, government laboratory to play a more pos.tive role in the fostering of local innovative capacities, since whi.!ethe univeisity’s predominant task is to educate and train students, the government laboratory is predominantly orientated towards problems of national (including industrial) importance and this has clear implications for private induastry. Thus, for example, the ‘National Engjine:ring Laboratory (YEL) has rece-1tly undertaken to introduce specific engineering design facilities using compuier technolou with a view to (acting as a centre which can help small local firms to improve the engineering design of their products [ 24:] *. This pcint is also exemplified dy the third important difference between government laboratories and acadentic bodies, viz their differential importance with regard to the marketing of technclogically intensive products. Besides their importance to industry as actu;tl markets for sophisticated equipment, government laboratories 3cf as intermediaries between buyer and seller insofar as they are liriked with -- and often eontrolled by - government departmen as. These departments are in effect i large and significant public sector mar’kei and to the extent that related laboratories are performing riate to this irarket, it may be very necessary for firms within this fi&j : TVmai,: +lin close links with the lab,)r itory?. Thus a naval research establishm~tnt performing research into intp wvetl forms of navigational equipmenu may demonstrate the need for, par titular types of navigational aids (suclt as, for example, specialised radb devices), or a transport laboratory pltrforming research into improved road safety may come up with a scheme which requiires t!*~commercial production of special instruments (such as, for e,xample, fog warning apparatus)., 11-leach of these cases, while it rnaj occasic nally be the practlcle of the I;r,boratory to manufacture a batch cf the relevant products itself, in general Slilch activity would take it outside it5 eecognised sphere of competence as primarily a research organisation Moreover, it will not normally possess the facilities for large-scale production. !n such cases it is the: practice for these laboratories to place contracts with industry bJt to ensure that the necessary sltandards and specifications are adhered to, atld it would ,appear that there ‘;See ref. [ 271. This is only one of a ntim ber of exalnplcs of t%le NEL taking an acti-Je ro:e in srimulating the local industrial en~iironment. Tht: la oratory views this as one of it5 basic tasks. *i*See ref. [ 11 which provides some evidence on this p Iin’,

31 3

are obvious advantages in close c )cp:ration between the laboratory and industly. at XI early stage. From tht r Glt of view of Ihe laboratory, firms with relevarlt 14~nologicA expertise 111 a~ be enormously helpful in attacking particular dcA~pmentA problems oc isa pointing out that if a certain type of product is eventuallvD requireId, then initial development should follow certain specific pa tbs. For example, the firm may recommend the use of a certain type of raw material bec;;use it is in plentiful r;upply or because it possc~s technical characteri:)tics which give the fin311product greater colmmercial viability. In this way il can guide R & D into areas such that the raw material in question can be effectively u;ed. From the point of view of the firm, close contact of this kind fill keep it cognizant O[ %e latest developments in particular areas, and may ljrery wt:ll lead to futute production contracts sin:ce the firm’s engineers will have tin closely inv Yved at all stages of the R & D process, and will doubtless t:i :it an advantage when tenders are put out to offer. Even closer links are :;ometimes established where firms actually perform R 8~D themselves on a I;levelopment contract basis. More gene:rally such contact will fulfil the role of j:eeping the firm in touch with particular market trends. 4.3. Some furlhez aspects Two final points should be mzd.2. Firstly, re-emp:hasising the discussion in sect.2 on the discontinuity bervbeen technology and production, it is clear that the advantages associated Gth close link:- between science and industry will be very much greater uherc 1 Crm can locate itself clbse to a conrzplex of scientific institutions such that ,t will have ~:--~s ‘0 a wider spectrum of technological inputs. Also to thta e:ctent to thie area and produced backward 1inkq;es in terms of new firms set up to produce food processing plant. It also benefited 104 agricultural production.

318

could, in turn, prod\rce further backward link,rges in tc3rms of new intermeMe product firn is, linked on the one hand to the technology source represented by the scientific infrastructure anc on the other to the firms requiring the existenx of ;pecialised componf n t rranufat t urers. Tire re is evidence that dynamic growth of this nd has men place with respec electronics industry in the south-east England 1291. Firmally, returning to the question of firm size, it should be pointed ou is often the small technologically-intensive firm which is able tts compete with the giants in certain areas. The advantages whicn the largx fSrm possesses in terms of marketing facilitie ‘3,access to sources of finance, risk-taking, etc. are of‘ten those which accr ere to the firm at the stage when th, products and their assoc~ ited teAnologies are already established. On the other hand American experience has shown that smaher firms, relying on close rnterchaarge with scientific labort\tories and concentrating solely on ;J particular procluct, can make substantial inroads into a well-established market [30) *. It is possible that measures clesigned to encourage the growth of such firms could play a significant role in ftxtering, the industrial health. of an underdeveloped region.

KEFI” RENCES N.G.Clark. “Thas Technc~logicul Relutior /ship tletween Firms in t!?e E~ecttonics Industry end Outside RtTsearch Bodies” in J.N. Wolfe c,ed.) ‘Regional P:qects 01 Industrial Pofic_v” @dint urgh University Press, 1972). AM. Cahn and A. Parthasarathi “The impact q’a Government-Sponsored I/rtiversity Resemch Laboratotys on the Lccal R d D ~‘conomy”, M.I.T. Occasional Paper (12 January 1967). Ibid., pp. 1 .and 2. Cahn and Parthasarathi. op.cit., p.2. “Aspects of’ Scientific Entrepreneurship”, Ph.D. thesis, Harvard University, Cambridge, Mass. (May Y966) see p 6. M.M. Wyatt, “Where to Build ;~fRxearch Lab”, fndustriul Rwearch (March 1962) p.27. D. Shimshoni, opxit., p.6. T.B. Browne, ‘“The Changing Research Parks”, Jndusmal Research (May L966) p.41. K. Miiller and R. Nejedly, “Regiorlal Distribu<‘ion of Research and Developweltt”, Research Policy 1(1971/1972) 32(. D. Shimshoni, op. cit., chapter I. G. Sutherland (Chairman), “Rzpxt of the Working Party, on Liaison Between Universities and Government Rcscarch Establishments” (Lofldon, HX.S.O., 1967) Cmnd.3222. See Appendix F, p. 1%.

:@Shnmshoni mentions this in connection with inlovatian; in the proc:e%s control and physical analytic sectors 1:f the scientific instrumc nts industry.

Science, xe&no/ogy anti regional economic development

319

MinIstry o:f Technolou, “Technical Services for lndustr+v“, (London, 967). This vokme gives details of all outside research bodies and the services they rovidt: for industry. C. Sutherland, ~p_cit.. p. 1. ‘111c following discussi,on is largely based upon this document. fbi& p.3,4. I&LB.,y.Xi. G. Sutherland

(Chairman), op.dr., p.9.

Ibid, pp.96- 186.

lbk#., p.151. See, for example, E. Mansfield, ‘industrial Research and Technological Iprrtovutit& (Norton, 1968) and “The Economics of Technical Change” (Longmans, 19ri9); J.R. Minasian, “The Economics of Research and Development” in “The Rade and Direction of Invenrive Activity” (Priricetaln University Press, 1962); also N.G. Clark, “The Techno-Ei c>nonticRelationship Between Industry and the Scientifrl: Infrastnrcture - With Special Reference to the Electronics IndustrTl and Scottish Economic Development ’ ‘, Chapter II, op.&:. above, note on first page.

D, Shimshoni, opcit., pp.6,7. W.E.G. Salter, “Productivity and Technical Change” (Cambridge University Press, 1966). J.R, Minasian, “The Economics of Research and Development”, in “The Rare and Direction of Inventive Activity” (N.B.E.R., Prince&n University Press, 1962. See C. Freeman, “‘Research and Development in Electronic Capital Goods ‘, .N.I.E.R. (1965), No.34, who &uzes a lot of empleasis upon inter-industry differences in the minimum necessary IR& D capacity. WI D. !Shimshoni, opcit., p.3. 1261 There is also similar evidence relating to British experience. See, for example, B.E,. Launder and G.A. Webster, ‘YIniversity Re:iearcir and the Considzations .4lpcting its Commercial Exploilation” (London, T.D.C., January 1969). 1271 J. Fowler, Research in Scorl.and -1, Scotland (Ckober 1968) pp.49-53. 1281 AX. O’Sullivan, Technology, Ilrelarnd ( 197 I ); Development o,T science-basea’ industry in the Cork region.

Toothill (Chairman), “Inquir) Into The Scottish Economy - Report of Q by the Scottish Council (Development and industryl “‘, Edinburgh ( 196 1). WI D. Shimshoni, op.&., pp.3-5.

WI

!‘. N.

Co,mmittee Appointed