Foreign Participation in U.S. Research and Development: Asset or Liability? (1996)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

J Foreign Participation in Privately Funded U.S. R&D Foreign participation in privately funded U.S. R&Di has grown rapidly since the early 1980s. This growth is primarily the result of a surge in direct investment by foreign entities in existing or newly established manufacturing facilities based in the United States. During the 1980s, foreign investors nearly tripled their ownership share from 7.2 to 19.2 percent—of U.S. manufacturing assets (Graham and Krugman, 19951. Also during this period, U.S.-based affiliates of foreign-owned firms increased their share of total private U.S. R&D spending from 6.4 percent (in 1980) to 14.5 percent (in 1992~.2 A second major source of growing foreign involvement in privately funded U.S. R&D has been the proliferation of international alliances. Be- tween 1980 and 1989, U.S. companies entered into over 1,500 technical alli- ances with European and Japanese firms in it&D-intensive industries, such as microelectronics, telecommunications, aerospace, and pharmaceuticals (Hagedoorn and Schakenraad, 19931. The rapid growth of foreign participation in privately funded U.S. R&D has generated four major concerns related to the economic welfare and military secu- rity of U.S. citizens. First, foreign-owned companies and their stakeholders abroad may be ex- tracting more intellectual property and associated economic value from the United States than they contribute to it. Second, U.S.-owned companies may not enjoy reciprocal access to privately funded R&D activities and assets abroad. These asymmetries of access may put U.S.-owned companies at a competitive disadvantage, thereby penalizing their U.S. stakeholders. 39

40 FOREIGN PARTICIPATION IN U.S. RESEARCH AND DEVELOPMENT Third, the rapid increase in foreign direct investment in U.S.-based high- technology companies may leave the nation hostage to a small number of for- eign-controlled suppliers for technically advanced components and subsystems critical to U.S. military security. Furthermore, growing foreign direct investment and international corporate alliances may make it increasingly difficult for the United States to prevent the transfer of militarily sensitive technology to potential enemies. Fourth, it is possible that increased foreign direct investment and trans- national alliances in high-technology industries will foster monopolies. Such monopolies might injure U.S. consumers and compromise U.S.-based compa- nies' access to key components and subsystems they require to make their prod- ucts competitive with those produced abroad. In an effort to assess the validity and significance of these and related issues, the following discussion explores the causes, scope, and nature of growing for- eign participation in privately funded U.S. R&D. THE CAUSES OF GROWING FOREIGN PARTICIPATION Three related trends have fostered the recent surge of foreign participation in privately funded U.S. R&D:3 · The increase in foreign industrial, financial, and technological strength compared to that of the United States; · The internationalization and changing nature of competition and innova- tion in most manufacturing and service industries; and · Recent U.S. trade and technology policies that have raised barriers to for- eign imports of it&D-intensive products and restricted foreign access to U.S. research in certain critical-technology areas. Collectively, these developments have provided powerful economic, techno- logical, and political incentives for foreign-owned firms to access U.S. markets and technological capabilities. The rapid expansion in the number of foreign firms with the requisite techni- cal and financial resources to exploit as well as contribute to U.S.-based R&D activities has been a key aspect of the trend. The growing presence of foreign multinationals in the U.S. economy reflects a general narrowing of the gap in scientific and technological capabilities between the United States and other in- dustrialized countries. Since the mid-1970s, many foreign companies have suc- cessfully entered the U.S. market by utilizing both their unique organizational and technological strengths as well as their newfound financial clout. (See Chap- ter 1, Table 1.1.) Reflecting the growing commercial importance of technology, industrial de-

PRIVATELY FUNDED R&D 41 mend for scientists and engineers has outstripped the supply of these profession- als in some countries, most notably Japan. This, in turn, has provided another impetus for foreign-owned multinational companies to locate advanced techno- logical activities in the United States, where there is an abundance of scientific and engineering talent (Hakanson and Zander, 1988; Serapio, 1994~. Intense global competition, shortening product life cycles, the growing com- plexity of technologies, and spiralling R&D costs have also fostered the interna- tionalization of industrial R&D. Today, firms in most it&D-intensive industries must compete in all major international markets as well as exploit worldwide economies of scale and sources of innovation. In this evolving global competi- tion, firms need the ability to respond quickly to and anticipate customer needs and wants across highly diverse national markets. To achieve this objective, more and more it&D-intensive companies are locating production facilities and a range of advanced technological capabilities, including R&D and design, in the foreign markets they wish to serve.4 The increasingly rapid, interdisciplinary, and costly nature of technological advance in many industries demands that firms seek out and acquire technology developed elsewhere. Yet, the very pace and complexity of technological ad- vance in many industries has made it difficult for firms based in one country to effectively assess and acquire technology developed in another without establish- ing an R&D capability within the nation of interest. The United States has be- come a prime target of foreign multinational activity because of the sheer size, wealth, and sophistication of its domestic market, the strength of its basic re- search enterprise, and its unrivaled capacity for incubating new industries and products (National Academy of Engineering, 1993~. In addition to these economic and technological factors, U.S. trade and tech- nology policies have also provided incentives for foreign firms to expand their manufacturing and R&D presence in the United States. Many analysts believe that existing or threatened nontariff barriers to trade, such as voluntary export restraints, buy-American procurement laws, and domestic-content requirements, have fostered the growth of foreign direct investment in production facilities and subsequently in R&D in some U.S. industries, including steel, automobiles, elec- tronics, and telecommunications equipment. Efforts by the federal government to prevent or regulate foreign access to certain areas of commercially promising U.S. R&D, particularly government-funded R&D, also appear to have encour- aged foreign firms to establish or acquire U.S.-based R&D facilities and enter into marketing and technical alliances with U.S.-owned companies (Chesnais, 1988; Mowery, 1991; Serapio and Dalton, 19941.5 Responding to many of the same factors that have pushed their foreign coun- terparts into U.S. markets, U.S.-owned companies are looking increasingly to foreign firms and markets both as potential customers and as sources of comple- mentary capital, manufacturing capability, and technology.6

42 FOREIGN PARTICIPATION IN U.S. RESEARCH ED DEVELOPMENT FOREIGN DIRECT INVESTMENT Foreign direct investment has been the principal way foreign firms and indi- viduals have become involved in privately funded U.S. R&D.7 Between 1982 and 1992, the cumulative stock of foreign direct investment in the United States grew from $124.7 billion to $430.2 billion. As a share of the total net worth of U.S. nonfinancial corporations, this type of investment increased more than three- fold during the decade, from 3.5 percent to 11.6 percent (Graham and Krugman, 19951. Roughly 80 percent of foreign direct investment in the United States during this period was used to acquire existing U.S.-based businesses; the re- maining 20 percent went to establish new businesses.8 In U.S. manufacturing industries, which account for nearly 90 percent of total U.S. industrial R&D expenditures, the growing importance of foreign direct investment has been even more pronounced. In 1982, U.S. affiliates of foreign- owned companies accounted for 9.8 percent of U.S. manufacturing assets, 6.6 percent of manufacturing employment, and 7.3 percent of manufacturing value- added. By 1991, affiliates claimed 19.2 percent of manufacturing assets, and by 1992, they accounted for 11.6 percent of employment and 15 percent of value- added in U.S. manufacturing industries (Graham and Krugman, 1995; U.S. De- partment of Commerce, 1994b). As foreign direct investment in U.S. industries has increased, so too has the share of industrial R&D funded by U.S. affiliates of foreign-owned firms. From 1982 to 1992, affiliates' share of spending on private-sector U.S. R&D increased from 9.3 percent to 14.5 percent (Figure 3.1~. Over this period, R&D spending by affiliates grew nearly twice as fast as did the domestic R&D expenditures of all U.S.-based companies.9 In 1992, U.S. affiliates of foreign-owned firms per- formed $13.7 billion worth of R&D, employed 104,500 people in U.S.-based R&D activity, and accounted for 12 percent of U.S. jobs in high-technology manu- facturing industries (Florida, 1994; U.S. Department of Commerce, 1993a, l995a).~° As of 1993, foreign nationals owned 1,482, or 6.5 percent, of the nearly 23,000 U.S.-based companies active in high-technology fields (Table 3.1~. Affiliate R&D Expenditures In 1992, U.S. affiliates of European-owned manufacturers accounted for roughly two-thirds of all affiliate expenditures on U.S. R&D. British-owned af- filiates accounted for 15.9 percent of such spending, followed by the Swiss at 15.8 percent, the Germans at 14.4 percent, and the French at 9.3 percent. U.S. affiliates of Canadian companies accounted for 15.7 percent of R&D expendi- tures, and spending by Japanese-owned affiliates represented 12.1 percent of the total. Spending by Japanese-owned affiliates on U.S. R&D has increased much more rapidly than that of other major investing countries since the early 1980s (Table 3.2~.

PRIVaTELY FUNDED R&D 20 15 10 5 o 43 — 9.3 . ~ . ~ 2. ~SS s. i s;~ s .. ~ --s ....s~ ~.s,~.~ , s ,, .s .... , .. i. .... s s i.: ....s s ,,ssjSs, ... ............ ,s ~ . s s. ~. s ~s;. ........ ,., - ,~, ,.:....~,.,s. ~: s:: :::::::::::::::::::::::::::::::::::::: _ ....s.... i..- , , , :, s ................. s sss .. s ~ . ~ ; s s ss .......... s , .s .,.~,..s...;.......... .. .. s s s :::::::: :::::::: .:: ::: -s s ,.,s.,., .. s.? ~ .ss s: . :.~ s ~ ss. ~ .~.^.~.,.( . 9.3 . !.. S~ ~ ~.,~ ~.s. ~ .~ ~ ~ s Ti ~ T ~S .. j j j ~ ~.~.: ? ~ ~.~ i~ -. s ......... j, ... ... S.... s. i........ :: :,s: s: jj: ,:: s:.:j.::: ..... - ~ .... . ~ .................. . ; . ~ s ~. ~ s. ~ .. . j . j s ~ . j ~. . s. .. s ,.,. jj. j j.,....j. j.. ,..,~ , :,s j - . .,,,,,,, j j j, i,, - ~ s. j j., :::::::::::::::::::::::::::::::::::::::: ..... , i. i~ ~ ............... ~,.,.: s . ~ s . j . s . . : s::::::::::::::::::::::::::::::::: ~ :s:.:.:.:,:.:,:,:.: :.:.:.:.:.:~::: .,~....-........................... 2 22:s j2:j: 222:j2.'.j' ::: ::::::::::::::: -::::::::::::: :::2::j:;.:::: .s- : ~,s , ,: :~:: :.,:~ :::,:.:.::s.:.: .;. s..:.,.., sj.j .,. ::::::::::~:: ::,::::::.::~:: :::: :,::s:: :ss:.::: :: i:.: ..... s. - , j ; _ . 9.2 9.7 ......... j .. ~ ~ ~ ~ ~. , j, ,, ,-, ~, i~ ~ ~.~ ~ - S ~ ~.: S.: . ~ . ~ ~. S. ~. .~S . ~ i................. j ~ .S /,. .. i:S, ~j...,,.~.. ~... . - ~.., ... s:,.s..... ............. s :.:.::,:j.:~:,:,:.:,:.:j.:s:~::: j... j sAs s . ' ;"'s22"s::2S:.s: . s . s . ...j j.. ~..j.. .,,, ~., ,, , . , j j .. . s, .. s . . , ' <. . . , ,, ,j-.... . , , .,,,,~,.<s ~ . ... s s. . . s j j . . .j..j - . . .~ . s ...,, . s ., . ... . . - sj . . .~ - : . ... ---s 1 . . ~ ~ ~ . s; ~ ~ j~ , ... s .. ~ ....... ..... ,.~- ~, ~ , .,', s:22..'.2' ....... ~ s i , ..,.,2.;.s;.,.'~ 2 2 .............. ,s ,.,s ,s .. S. -.~ s j ~,,.,.. ~.,j....~.j..... j j j ...... s jjjs s . s ..... ;ss . . s . . s. '.-.2. ?.22.2-2.'?'.'2-22.2. ::::::::::s::::::::::::::::::::::: "2-?2s'2.'.-?j..2,.,?j .... , ?ij, :.:.:..:-.: :.:.:.:~:.:,:-:.:.:. - ."~.v2."2." "'. "." :~ ,,,,~::,:....,.... ~ ,.,,,,j ?i? ?j -:.:::..:.: :, :::: :? s: ?s.: .?j?iT ?.i?j. .,.; s ? ?: y :?. ,, 10.6 ........... ;. ~ ..~ ~....?.~.:,2:..:.. ~:~ :. :2? ?:??.j.~2.j:: ,:~:::.:.:.: :.:-,:.:.:.:.:.:.:.:.:,:~:.: . "',2,'.'.2S..2,2.,2,~., ??,, :?~ .,.,, ?. ~ ,.s ~s, ~ : ?: ::: :.: :.:: :':'~2:?::':2': : s:.:.:.:.:.:,:..: :.:.:.:.:.:.:,:.:. ,2'.'.' ,.:,22'2.:?',2 :::~:::::::::::::::: ::::::..:::::::: '?~ :2:: 5.:?.:.:. .:::?2:.:. .: -;2'. :.. .;':' '.. ? :''-' : :::::::: :::::::::::::::s:::::::: ~ ,.,, ,,, :, 2-2?22.' ::::::::::::::;::::::::::::::i:::::::: :::':~:.:. ,s ,:,:-:,:?,:.s:::.:.s 2. ..?,;' ?:'?..?j"2 :::::~: ::::::.:g::::::::::: , .. ;'?2,22?j..,::,.' 2., ?,?s2,s,2 sss . :::::::~::j:::: ~:s::: ::: .:.2.? ;:.': . .2,..: . ,?..2. ::::::s::::::s:::::::~: ::::::::::. :: :,s:.: :::, s:,~ s:,:.:,:.:: 2 js? si? ...sj: s-. ... "2 . s:'ss ss2i? s "'' s:.:.: :.:::: :. .:.:..:.:.: :.:.: ?..,..s.,..~.,j,...:;.2.2.2...?..2 s :. :.222.?? " ?.: :.::: :,:,: ,: ::.: :. s:.:. : :: ,.,, ,,.,ss '?'"2""""2? """""22? :"2 $::.:.;::.:?s.:::::.:?.:::: ...'. ?2:.2...,'s:...'.:. . ... v::::: ::::::: :~:~::~:::::: :::::::::::: 1 1.8 s......... ~ ? ,,, ~2:~i ?:'.'~.." :.:.:.:.:?:::: :.:~:::::: :.:. :::.:,v:: :?.:~:~:;':':.:s ,,:.,. s ~, ~:~....... :::::::::.:::::.:: ::::::::::::::., ~ "'."'.?~2:".'.'.' ' "" ".'.'.? :.:::::,:,:,:: ::: .:~::?.: :~: :,:.::: :,;: .:,: :: . ,.'.'::.,2..?2~ ::2,.,:?:> :::::s:::::: s:: :j: :::s: .2.~.~''~:..:.' ~-?'''?' :,: i:.:,:,::, :.s:,<:.:~:.: ?....2~... :2.; :2~2? ' :: ,::: ::: ::, ::::::: ::.: .: .: .,. ~,..? ....... :. :j::,:.:~::::,:,:,:,:,:.:.:,:,:, :::.j::::::s:.::::::~::: 2.:~""2.: :, : :j:::::s:::s:::::: :::::: s:S. ..~sj:~.s:...:j?...22 . ??;,.?;i,,,, . '?. ".. '.??2??j?:.'.;.".2.j?j"." "''"2?"?~" ?2 " ?'j2'- 2 :.:: :. :. :.:.:.:.:,:.s:,:.: :.: ,...?i,, ??.?.j..., ..'.:..:. :....22...2. :2.2 ..:'. ....? :....2..i2Jj:j :.2. :2:.:2::: :,:.: .:?ss: :,:: :. , . ,? ?? s:, ... ".2 ?<2? .2 2.' :~::.:-s:::.:,:;:?s ::':.:':.s :':?:: :::?s:.s:s.:::::s l ' ??;,,,.,.i 1 ;, c ,.,.s?.? ?:22s s:: ~:::: ::.:.s::.s:.:.:.::-. :s::::::~:: ::::: ::::::::~:. ,?, 2??.,,,.,?:?j:,,,?i.,~,.??, l ::~:::::s::::::~:: :.S 15.5 ., . A ~ .2 ? ~:,. ~.'.: i:~?: ~-. I ~ ~ :::~: ::::::::::::::::::::s::::: . ~. ~ 's.? s? ?i; ? .2. :'?." 9.3 ~, .?j.??.,,?.,,.,?j;,, ,,,: s l'.:2 .~.2 s. : :.~,, l .s : , ,, .~:, ~ ?: 1 ~ 9 I- ~ ~ 1 3 | ?:i.,.,., ?~ ?~ ~ j ?~ """' ~ V. t ?~: .':''i'.?i l.~:: ::"?5j.22:i2.~ '-l 2.!:?.j?; s ~ S: IS S'' '" -?' i'is-l t=, ., i 1""2~':":"'."' S:.'.'"1 ~ i l :,:., ,? j ',,, I l ~?.:: ~ .i 1,, ,?, ,,.,.s l ..... s s- : 12 9 1 ?i ? fl l ??i ??~l l~ ?s ~ i l s?'. ''1 .:?':i.?..'-~,~ · s::' 22.2.2.25.2 . sl I .?.~, . :? j ,'?,~.l 1"i."~'".:.~''':'~:?'l ~ ,>-, ,:,,-s-,,,:1 ??'?? ~'.?s.:,?~, ~.~ i~,s:?: :~ ?~. "| l? j ?, hi?l |' ???'j??'?"'?i?i"~| i,.,,s;,'-???;,,,' 'i l Us:? j 1 '''?''''?' '?'.'.'? ... 1 i>''. .'? ',' '?;, l I ,j ,? ?,, ,,1 1' j ' ?; ? i ?"?''?:''" ': '"| ... ?, . . .: " s: i~'' ~ ~ '':2:-~il ~ 2 - "" 1 ss?s, , -s-.- 1 1 .- -s , ~ ~ l': '-?':'':'?:?- s l : ':? ,j, ,,~ s -2 -? :' ? . I . -,,, ? .: '-': ?i"'.l :.?i; ? ?? . ?,?.?: I 1' ??'"" ? j ? ? ?? ~ t - i :' ; s i ? ~1 1~;'1 ~ 1 ~ ~ 1 '1 C... 1 ' ~1 1 ' 1 1 1 1 ; ;1 1 1 ., ~ I- - l I ~ l 1 ~ ~l 1~ ~l 1~ 2-~ ~6a~ ~- |? ":j?'"j.2 ?; ?s?j""| | j j s:? j ? j I ;?: 's;?i? ;?i????s: l I j j:?? ? ? j I sjjs s. ~ ' ? j I I ?s?s ? ? I ~ j I 1,? .j? ? ? ? . 1 ????i.,. '?i?i?: ?''.' ?l ? ? :?i. ,?,? ? E ':: ?i s - 1 1 s?: ? :?i "l I """ ~ 1 | ; ??? | | | - . ? ?: ~ ~ ?' ~!, ?j 22,2,.?i:.? 2. l 1 "':?e "'?-'' l ~ ' ' ' ' ? | e, ;j?i, ~? 1 1'' '.'~ ','? ,1 ? ?i . ? ,','--??,,~ E ?ss, . l s.2.2.2.}s: ?~:: ?l | ?, ~ i I - I ' ?': - -: :: :l :,....:??~'',.:? s s ? sl E; js '? 5" '"''l 1"""'"': """'""'''""""l 1 '? s ? ? ~''? | :'':' '',s " :.?' '!' '?~ " :'. .> -? -,.?.~ '~ E''?::2?j:'.:j. ::l I:.:..~.,,., ., l:: . ?s::~:::::s:::? :' .:.::,::~:>,s?: :~:::l 1::: :.? :.:.:.:.:.:~:~:.:...l :.:::::::::::::::::~::. l ~ :: ?ssi?: l ."":i"2".2:'i.e:""":"''l l '?,?.'.''.',? ,,',i l'''.'''e'i:'':'':'i'''3 I .......j.~..i.....3 ' i''" eS ~ :: ~" ~' r.::--:.:.:-::..~:.:.:::.:.:..:' .:s:-ssse:::::.-:s::::::::::s:. '::::::: ::i::::.::::::.::s::::::' '::::s::*..:-s<:::.::::::::' ':::.:: -.:s::::::::::::::::::::, :s::::::s:::::::::.:::s:::::::::::: t'.'e.''.'2-'.2.:<' ' I ' '': "''?' '''" I I - .. ? -e I 1 ? ...? ? 3 1~. '? :.. ,, ~, ~ ~ l ?: .. ?:, ....?'.' ' ' E:::: :::::::::::..::::::s::::::::l · ?* :s :?l l:.:.s::s:.::::?~:~::j:::j::l l:.:.::i:,:.::::: :,::.::::::3 I::~::.:s::2:':':.:.:.:s:T':.:21 :,,,:.i:,::::.i:, .:, ~:,: t,: :s:: :,l l?j:s.~.22:?.s:.:j::'?'?'l I~:~?: ':,? jssl '::::,ss.?j:e::::::':':':sl E:'i:~:::::~:.:.::.:-:.:.:l '::':'::':'::':::':'::::':':'::: E. <: . ?; <?s:?s l Es je '; ''' ': ;| I ~ ? ^e.l 1 - .? e ~ ? 3 1 . ? - -3 "" : e I2 ;' ";' '1 IC..'...;'~.'......'.,...V1 I'''~""'"'"V'G'~1 i''"""'°"''"'s! 1'"" ' - " 1 ... I l t ~ l 1 i: i~21 12-:: 1 1-: ~ ~ r :~:-~-~~ ~ J 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1 993 FIGURE 3.1 R&D spending by U.S. affiliates of foreign-owned firms as a percentage of all privately funded U.S. R&D, 1982-1993. SOURCE: National Science Board (19931; National Science Foundation (1996~; U.S. Department of Commerce (199Sa). Nearly 42 percent of all affiliate spending on U.S. R&D in 1992 was concen- trated in two industries: drugs and medicines (24.8 percent) and industrial chemi- cals (17.0 percent) (Figure 3.24. Roughly another fifth of total expenditures was for R&D in the audio, video, and communications equipment (8.4 percent), com- puters and office equipment (5.5 percent), and instruments and related products (4.4 percent) industries. Spending in each of two other categories, all other manu- facturing industries and all nonmanufacturing industries (the latter comprising wholesale and retail trade, sewices, public utilities, mining, and agriculture), ac- counted for roughly 17 percent of the total. Data on R&D spending and product sales for U.S. affiliates in high-technol- ogy industries underline distinct patterns of specialization among firms of differ- ent national origins (Table 3.3~. For example, Swiss- and British-owned affili- ates together accounted for more than 71 percent of all affiliate R&D spending and over 75 percent of affiliate sales in the U.S. pharmaceutical industry. Like- wise, affiliates of European companies, led by the British, accounted for roughly 90 percent of affiliate R&D spending and sales in the instruments and related products sector. Affiliates of German and Canadian companies performed more than 72 percent of affiliate R&D and claimed more than half of all affiliate sales in the industrial chemicals sector. In the computer and electronic component

44 o C~ ._ C~ au C~ 4_ V) .= ._ C~ ._ ~o o~ ~S C~ E~ b~ ._ ._ au ._ ¢ .= Ct P" o V o C~ s: O _4 o - , V _ _ ~ C~ _ 4 - ~ o E~ ~ ~ _ o ~ C~ 1 0 ._ — t<: C~ . ~ _ ~ _ E~ ~ C~ ·= ~ o o ~ Q C~ ~i C~ C~ 3 ~o C~ ~ ~ oo ~ ~ C~ ~ ~ o o . . . . . . . . . . . ~ ~ ~ o ~ o o o o o o . . . . . . . . . . . ~ ~ ~ ~ o o o ~ o o o C~ ~ c~ ~ 0 ~ ~o r~ . . . . . . . . . . . ~ ~ ~ C~ ~ o o o o o o Ch ~ ~ ~ ~ ~ ~ ~ ~ . . . . . . . . . ~ ~ o o o o o o o * * ~o C~ CM oo ~ o '~ C .50 o O .o ~ ~ oo ~ ON ~ c-~ c~~ Ir) C~) c~) ¢ ~ ~ ~ ~ ~ ~ ~ o o o * * ~ Ct Z o" 3 oo ~ ~ ~D ~o V =: c~ oo ~ ~i 0 0 o o' o o ~ ~ ~ ~ ~ o m ~ ~ ~ ~ 0 0 0 0 ~ * O 00 0 ~ ~ O. ~ ~n ¢ ~ ~ CM ~ C~ O ~ O O * * ~ ~ v~ ~ ~ 0 u~ ·n ~ ~ ~ ~ . . . . . . . . . . ~ ~ a~ ~ ~ ~ ~ ° ° ° ° ° * t~ ~ L 5 >~ 5 ~ ~ ~ ~ ~ ~ V~ ~ ~ ~ ~ . . . . . . . . . . ~ 00 ~ ~ CN O O O O O * C) C~ o o C) C~ C~ - C~ C~ Ct C~ - Ct C) ._ C) ;- Ct - Ct ._ 0N S ._ C) . c~ ~s e - o c~ u: ~ o c~ ~ 'e ~ · - o ;:: 'e ~ c~ ce - ~ a' o ~ a, ~4 o <4-o-. c~ ~ - o o - o c) oG . - o a' - o o ~ - e . ~ ~ - ~ c) o ~ · l-~ e 'e =: ~ cC ca - 11 * . . E~ o z c=t o m C) ._ C~ V) - Ct o ._ 4_ Ct . . o C~

PRIVATELY FUNDED R&D TABLE 3.2 R&D Spending by U.S. Affiliates of Foreign-Owned Companies, Volume and Percent of Total by Country, 1980, 1985, and 1992 45 1980 Percent $ Millionsa of Total $ Millionsa 1985 Percent 1992 Percent of Total $ Millionsa of Total United Kingdom 312 16.0 748 14.3 2,178 15.9 Gerrnanyb 380 19.5 671 12.8 1,968 14.4 Switzerland 338 17.4 625 11.9 2,159 15.8 France 146 5.4 166 3.2 1,272 9.3 All Europe 1,544 79.3 2,918 55.7 8,956 65.4 Canada 135 6.9 1,550c 29.6 2,151 c 15.7 Japan 88 4.5 267 5.1 1,656 12.1 aCurrent dollars. bGerman data are for the former West Germany only. CData include roughly $1 billion of R&D spending by Du Pont, a U.S. majority-owned company in which the Canadian company Seagrams held roughly 20 percent equity. Seagrams sold its holdings in Du Pont in 1995. SOURCE: National Science Board (1993); U.S. Department of Commerce (1995a). Audio, video, and communications 8.4 Electronic components 2.0 Instruments and - related products 4.4 Wholesale and retail trade 6.7 Services ~ 4.6 / Petroleum 4 ~ Primary and fabricated metals Computer and 2 5 office equipment , I 5.5 k~.~...~..~-~.-.~--~-~ ~~..~ If.-.. . ~ . ~ S?SiS-: S.~:SiS ~ S., ,:. ~ , ~ i, ,~ I, i, i, S ~ ~ So _ :~:22~, ~ ~ -it ~ /' All other non manufacturing 1.1 All other manufacturing 17.0 Pharmaceuticals / 24.8 ,~ ~ Industrial chemicals 17.0 \ \ Food and kindred products 1.8 FIGURE 3.2 U.S. affiliates of foreign-owned firms, percent of total, by industry, 1992. SOURCE: U.S. Department of Commerce (199Sa).

46 ~ US I_ Ct 5 o o En .= o . ~ o o Ct - . V) Cal Cal Ct a: ·_4 Em ._ . ~ c - ._ Cal o Ct em Cal Ct V Cal _ Ct U) C~1 Ct U) U: Ct Cal ~ .o , Ct ~ U) V _ Pa o Ct ~ Cal Cal Cal Cal U: Cal Cal ~ c~ c~ c~ ~ o ~ ~ oo ~ . . ~ . . . . . ~ . . ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~o ~ ~ ~ ~ ~ c~ ~ ~ ~ U~ ~ ~ ~ ~N ~ ~ ~ O O C~ . . ~ . . . . . . . . ~ oo ~ ~ ~ ~ ~ ~ ~ ~ ~ n ~ ~ ~ o . . ~ ~ . . . ~ . . . o ~ Q ~ o o ~ ~ ~ ~ ~ ~ ~ o ~ oo ~ ~ oo ~ . . . . , . ~ . . . °V ~ ~ o ~ ~ o oo . . ~ ~ . . . ~ . . . ~ ~ ~ ~ ~ , ~ _ ~ ~ U) ~ o ~ ~ . ~ ~ ~ . . . ~ ~ ~ Q ~ ~ ~ mo ~ ~ ~ ~ oo ~ ~ ~ ~ ~ U~ . ~-o . . ~ ~ ~ ~ ~ Q ~ ~ ~ ~ ~ _- ~ ~ ~ ~ ~ ~ o. ~ ~ ~ ~ ~ Q ~ ~ oo ~ ~ o ~ ~ ~ ~ oo o ~ ~ ~ ~ oo ~ . . . . . . . . . . . C~ oo ~ V) ~ ~ ~ ~ oo o ~ o o ~ oo ~ ~ ~o o ~ ~ ~ ~ oo CM ~ ~ ~ ~ ~ oo ~o oo ~ oo ON oo oo \~) \~} oo V 0 ~ ~ U~ ~ ~ 0 . . . . . . . . . . . =1 ~ 00 0 t: X A A A ~A U: ._ : - 3 o V - U: ~ C~ V) ~ C~ o ;- _ _ _ _ .= o ;^ C~ 1_ oo oo o ~ oo ~n ~o ~ ~ ~ ~ C~ ~ ~ ~ ~ C~ oo ~ ~ ~ U~ C~ CM ~ ~ C~ C~ C~ ~ oo ~ ~ ~ ~ ~0 ~ O 00 ~ ~ GN C~ ~ ~ ~ =^ ~o e ~ /= ~ ~ s O ~ Ct 5 ~ ~- ~ ~ ~ ~ ~ ~ 2 - o C) C~ c) . - ; - o v - .= ;^ ~ ~ o 'e ~ o ~ c~ ~ - ~ ~ o ·— L~ . ~ . - . ~ o ~ c) o - . ~ . - c~ ~ o ·= · - o o ~ ~ 'e 3 ~ U: ~ o ~ C~ Cq ·— s_, o ~o Ct o ~ _ 5 ·Ct ~ Ct .' t4 C~ ~ ~ ._ C- Ct (~ <¢ 5 ~ C~ o ~o - o ~ ~s o Ct O ._ D ~ U,; - C) ~ ~ ca C~ Ct Ct C~ 11 't Ct `_ ~ Ct U~ ~N ax C) s~ C) C) o Ct s . . V o

PRIVATELY FUNDED R&D Industrial chemicals Pharmaceuticals Computers and office equipment Audio, video, and communications equipment Electronic components Instruments and related products Other transportation equipment Total high technology 47 0 10 20 30 1 987 3 1992 ,,,, 1, 1 1 1 1, 1 40 50 60 FIGURE 3.3 R&D spending by U.S. affiliates of foreign-owned firms in U.S. high- technology industries as a percentage of all privately funded U.S. R&D, 1987, and 1992. SOURCE: National Science Board (1993), National Science Foundation (1996~; U.S. Department of Commerce (199Sa). industries, Japanese-owned affiliates claimed the largest shares of R&D spending and sales. For the most part, data on foreign direct investment and on R&D spending by U.S. affiliates indicate that foreign parent companies have invested in areas in which they have a demonstrated competitive advantage (U.S. Depart- ment of Commerce, 1993a. Sector-by-sector comparisons of R&D spending by affiliates as a share of all privately funded U.S. R&D show considerable variations among industries (Fig- ure 3.3~. Overall in 1992, affiliate spending accounted for 19.3 percent of the total invested by U.S. high-technology companies in R&D. U.S. affiliates of foreign companies accounted for the largest share of total U.S. R&D expendi- tures in industrial chemicals (47.5 percent) and pharmaceuticals (42.7 percent). Affiliate spending on audio, video, and communications equipment R&D (33 percent of the total) also represented a significant share of U.S. R&D expendi- tures in that industry. Affiliates' shares of private-sector R&D spending in elec- tronic components, instrumentation, and in computers and office equipment were 8.1, 8.2 and 7.2 percent, respectively. U.S. affiliates of foreign-owned companies also accounted for a significant proportion of U.S. privately-funded R&D in several other sectors, including the

48 FOREIGN PARTICIPATION IN U.S. RESEARCH AND DEVELOPMENT primary metal industry (37.2 percent), the fabricated metal products industry (20.2 percent), the petroleum industry (25.4 percent), the nonelectrical machinery in- dustry (10.1 percent), and the food and kindred products industry (17.8 percent). In the automotive industry, affiliates accounted for less than 4 percent of total privately funded U.S. R&D (National Science Foundation, l995b; U.S. Depart- ment of Commerce, 1995a).l2 Organization and Character of Affiliate R&D Activity The R&D activities of U.S. affiliates of foreign-owned companies generally are of three types. First, there is R&D performed by freestanding facilities. Their activities range from basic research and product development to general technical support. These facilities are engaged primarily in R&D, operate under their own budgets, are overseen by their own group of officers, and are located separately from other U.S. facilities of the parent company (Dalton and Serapio, 1993, 1995~. Second, there is R&D predominantly development engineering performed within the business units of U.S.-based manufacturing affiliates. Generally, this R&D is managed by the manufacturing facility or business unit to provide gen- eral technical and organizational support to production. Third, there is R&D performed by third-party contractors to foreign-owned companies. Such contrac- tors may include universities and public- and private-sector laboratories. Few data have been collected on the scope and nature of R&D that supports the U.S. manufacturing affiliates of foreign-owned firms or on the amount of R&D that affiliates contract out to unaffiliated U.S.-based companies.~3 Con- tracting between affiliates and U.S. universities and federal laboratories is exam- ined in detail in Chapter 4. The following discussion focuses on the first type of affiliate R&D, that which is performed in freestanding laboratories. These facili- ties account for the majority of all affiliate expenditures on R&D and have been more extensively inventoried and evaluated (Dalton and Serapio, 1993, 1995; Directory of American Research and Technology, 1994; Florida and Kenney, 19931.14 Freestanding Industrial R&D Facilities Research by Dalton and Serapio (1995) documents that as of 1994, nearly 301 foreign companies had established 645 freestanding R&D facilities in the United States. Of these, 375 were owned by European companies, 225 by Japa- nese companies, 27 by Korean companies, and 8 by Canadian companies (Figure 3.4~. Two industries pharmaceuticals and biotechnology, with 115 facilities, and chemical, rubber and materials, with 110 facilities accounted for more than one-third of all freestanding R&D operations. Another third of all such facilities (238 in total) was in industries in the electronics and information technology fields. The remaining third was dominated by three industry groups automo-

PRIVATELY FUNDED R&D 4 200 49 Europe Japan Korea Canada Other FIGURE 3.4 Number of freestanding R&D facilities in the United States owned by foreign parent companies, 1992. SOURCE: Dalton and Serapio (1995~. tive (53 facilities), instrumentation (43 facilities), and foods, consumer goods, and miscellaneous (55 facilities) (Figure 3.5~. Japanese firms owned 50 percent or more of all freestanding R&D facilities in 6 of the 13 major industrial group- ings defined by Dalton and Serapio. European firms owned the vast majority of facilities in six industry areas (Table 3.4~. In terms of the size of their professional staff, the largest foreign R&D facili- ties are in pharmaceuticals and biotechnology (Table 3.5~. On average, Japanese R&D facilities in the United States are much smaller than those of their European counterparts. Limited 1993 data on freestanding R&D centers (excluding those in the automotive industry) show an average staff size of 160 at European facili-

so FOREIGN PARTICIPATION IN U.S. RESEARCH AND DEVELOPMENT Chemicals, rubber, materials 16.7 \ High definition TV and other electronics Pharmaceuticals and Biotechnology / 17.5 Foods and miscellaneous ~~ 3 consumer goods 8.4 Automobiles 8.1 \ Computers Software 5 9 Instrumentation, 6 5 Controls 6.5 l Metals Opoelectronics 3 Machinery 4.1 telecommunications 4.6 Semiconductory 5.2 FIGURE 3.5 Foreign ownership of freestanding U.S. R&D facilities, percent by indus- try, 1992. SOURCE: Dalton and Serapio (1995~. ties and 45 at Japanese facilities (Dalton and Serapio, 1993~. Eleven of 23 Japa- nese-owned automotive R&D facilities and 4 of 12 Japanese-owned electronics R&D facilities surveyed in 1992 had 20 or fewer employees (Dalton and Serapio, 19931.~6 Data collected by Dibner et al. (1992) indicate that the average number of professional staff (207) working in European-owned pharmaceutical/biotech- nology R&D facilities is roughly 15 times that working in Japanese-owned cen- ters. Most foreign-owned R&D facilities are clustered near major U.S. geographic centers of R&D activity—Silicon Valley/Stanford University (computers, semi- conductors, and computer software), greater Los Angeles (auto design and styl- ing), Detroit (automobiles), Boston/MIT (biotechnology and computers), Prince- ton, New Jersey/Princeton University (software and high-definition television), and Research Triangle Park, North Carolina (biotechnology) (Figure 3.61. The Nature of Affiliate R&D Activity Surveys of affiliate R&D laboratories and their parent companies, case stud- ies of individual companies, and patent data offer some insight into the nature of R&D conducted by U.S. affiliates of foreign firms in different industries. A1- though there are important differences among industrial sectors, most affiliate R&D activity in the United States appears to have two major objectives: to help the local manufacturing affiliate and the parent company meet the demands of

51 A: as · - o v 50 At · - ; - o o At ct I · - c~ · - 4= - · c) ct · - ct $ - o At 7 · ¢ as, Cal _ ~ Ct o Ed C`7 At o 3 I Ct o 0 ~ ~ ~ ~ 0 0 ~ ~ ~ rid ~ ~ t— ~ ~ Cal Cal ~ ~ Us ~ Cal to Cal X ~ ~ ~ ~ ~ CM Us o ~ Ct Z U) ~ ~ ~ ~ ~ Cal , - ~ .-3 ~ o ~ ~ ho ~ ~ ~ U) ~ ~ ~ ~ to Cal I: .= Cal Cal ;^ Cal ._ o Ct o A: ._ ~ ~ c~ cM ~ ~ ~ ~ ~ ~ - x oo ~ ~ ~ ~ ~ —) c~) ~ — ~ vo - cM c~ ~ o ~o ~ cN ~ ~ ~ ~ ~ c~ ~ ~ ~ ~ ~ ~ ~ u: "C ~ 0 2 0 ~ ~, e a ~, C o L-~-o O L, (-= a e ~, ~5 C (004 4— c~ ~_ - ~_ o ~_ o ~D o a~ - ~_ c~ - V o ~_ Q s~ V) ct o - ct o ct . . V o V)

52 FOREIGN PARTICIPATION IN U.S. RESEARCH AND DEVELOPMENT TABLE 3.5 Largest Foreign-Owned Freestanding R&D Facilities in the United States Company (Home Country) Location Staffa 1. Pharmacia(SWE) Upjohn Laboratories: Kalamazoo,MI 1,318 2. Northern Telecom (CAN) Farmington, Ann Arbor, MI; Los Angeles, CA 1,260 3. SmithKline Beecham (UK) King of Prussia, PA 1,198 4. Siemens (GER) Iselin, NJ 1,100 5. Glaxo (UK) Research Triangle Park, NC 1,000 6. Burroughs Wellcome (UK) Research Triangle Park, NC 891 7. Honda (JA) Marysville, OH (2); Torrance, CA; Denver, CO 800 8. Hoechst (GER) Somerville, NJ 716 9. Hoffman-LaRoche (SW) Genentech: San Francisco, CA 672 10. Sony (JA) San Jose, CA 600 11. Bayer (GER) Miles: West Haven, CT 500 12. Glaxo (UK) Sterling Drug: Rensselaer, NY 450 13. Hoechst(GER) Marion MerrillDow,KansasCity,MO 411 14. Nestle (SW) Westreco: New Milford, CT 410 15. Nestle (SW) Alcon Labs: Fort Worth, TX 404 16. Rhone-Poulenc (FR) Fort Washington, PA 400 17. Bayer (GER) Miles: Pittsburgh, PA 389 18. Hoffman-LaRoche (SW) Nutley, NJ 350 19. Toyota (JA) California (4); Ann Arbor, MI 350 20. Rhone-Poulenc (FR) Research Triangle Park, NC 350 21. Unilever (NE) Edgewater, NJ 329 22. Nissan (JA) Farmington, Ann Arbor, MI; Los Angeles, CA 320 23. Northern Telecom (CAN) San Ramon, CA 319 24. Northern Telecom (CAN) Rochester, NY 280 25. PA Consulting (UK) Hightstown, NJ 250 26. Zeneca (UK) Wilmington, DE 245 27. Moore (CAN) Grand Island, NY 235 28. Thomson (FR) Indianapolis, IN 230 29. Mazda (JA) Flat Rock, Ann Arbor, MI; Irvine, CA 213 30. Racal (UK) Sunrise, FL 209 31. Goldstar (KO) United Micro Tech: NJ 200 32. Siemens (GER) Gammasonics: Hoffman Estates, IL 200 33. Siemens (GER) Rolm: BocaRaton, FL 200 34. OSRAM (GER) Sylvania: Danver,MA 200 35. Dainippon Ink & Chemicals (JA) Reichold Chemicals: Research Triangle Park, NC 186 NOTE: CAN = Canada, FR = France, GER = Germany, JA = Japan, KO = Korea, NE = Netherlands, SW = Switzerland, SWE = Sweden, UK = United Kingdom. aProfessional. SOURCE: Adapted from Dalton and Serapio (1995).

53 a, .O C ~ = E m° ~ ~ _ ~ \ ~ ~ ~ ~ ~ A ~ 0 o m ~~m <'I c ce cn / m ~ ~ A< \ ~ ~ , a. . 1 N co ° ~ ma u' _\ o .~ so / u' o

54 FOREIGN PARTICIPATION IN U.S. RESEARCH AND DEVELOPMENT U.S. customers more effectively (e.g., by working both to adapt existing products and processes to U.S. markets, and to design and develop new products); and to facilitate foreign firms' access to the scientific and technical talent in established U.S. centers of technology and innovation (Dalton and Serapio, 1993, 1995; Florida, 1994; Organization for Economic Cooperation and Development, 1994~. In their 1988 survey of the overseas R&D activities of 560 large companies, Pearce and Singh (1992) found that the most prevalent type of R&D activity among U.S. affiliates of foreign firms was applied research to derive new manu- facturing technologies in the industry of the parent company.~7 Less prevalent (in descending order of importance) was applied research intended to: adapt existing products to the local market; derive additional products in new areas of special- ization; derive new products in a current area of specialization; and adapt existing manufacturing technology to the local market. Not surprisingly, the relative importance of different types of applied R&D varied among industrial sectors. For instance, applied research to derive new products was seen as very important by affiliate laboratories in the food, drink, tobacco, and metals industries, and less important in pharmaceuticals, consumer chemicals, motor vehicles, industrial and farm machinery, and electronics and electrical machinery. In contrast, applied research to adapt existing products to local markets was considered particularly important in industrial and agricultural chemicals, motor vehicles, and electronics and electrical appliances, and less sig- nificant in pharmaceuticals and consumer chemicals, and photographic and sci- entific equipment. Research to adapt production processes to local requirements was particularly relevant in food, drink and tobacco, metal manufacturing, and industrial and agricultural chemicals. Such research was least important in phar- maceuticals and consumer chemicals, and industrial and farm machinery (Pearce and Singh, 1992~. Overall, basic research was the least prevalent R&D activity. It was signifi- cantly more common in the pharmaceuticals and consumer chemicals industries, however, than it was in any other manufacturing sector. Survey data also indicate that a significant number of foreign-owned laboratories in the United States, which were established originally to perform applied research, had since added basic research to their portfolios (Pearce and Singh, 1992~. More recent surveys and case studies of affiliate R&D activity document an increase since the late 1980s in the number of research facilities devoted exclu- sively to basic and long-term applied research (Dibner et al., 1992; Florida and Kenney, 1993; Peters, 1991; Voisey, 1992; Westney, 19931. For example, since 1989, NEC, Canon, Philips, Matsushita/Panasonic, and Mitsubishi Electric have all established basic research centers in the United States. Westney (1993) notes that Japanese basic research laboratories appear to cultivate close interac- tions with other U.S.-based basic research institutions. The goal seems to be to support the R&D strategy of the parent company rather than to assist the firm's U.S.-based manufacturing affiliates. Most of these laboratories have relatively

PRIVATELY FUNDED R&D 55 small research staffs and so are believed to be primarily monitoring research and technology developments, not conducting much basic research themselves. (Their small staffs also significantly limit the extent to which these laboratories can reach out and draw upon the work of other U.S.-based researchers.~9) A notable exception is the Princeton-based NEC Research Institute, a facility established in 1988 by the Japanese computer company. As of fiscal 1994, the institute had a permanent staff of approximately 80 scientists and engineers engaged in basic research in the computer and physical sciences.20 There are similarities as well as differences in the objectives and the focus of R&D conducted by U.S. affiliates of foreign-owned firms. For example, Japa- nese affiliates in the automotive, biotechnology, and electronics industries identi- fied acquiring technology and keeping abreast of technological developments and competitors as major objectives for their U.S.-based R&D facilities (Table 3.6~. However, the Japanese biotechnology and electronics industries considered the opportunity to employ U.S. scientists and engineers and to cooperate with other U.S. R&D facilities more important than did the Japanese automotive industry. The automotive and electronics industries assigned far greater importance to as- sisting the parent company to meet customer needs than did the biotechnology industry. Only the biotechnology industry attached significance to the opportuni- ties to take advantage of a favorable research environment and engage in basic TABLE 3.6 Reasons Cited by Technical Executives of Japanese-Owned Firms for R&D Investments in the United States (l=extremely important, 2=important, 3=neutral, 4=unimportant) Automotive Electronics Biotechnology Acquire technology Keep abreast of technological developments Assist parent company in 2 2 meeting U.S. customer needs Employ U.S. scientists and engineers 3 Follow the competition 3 Take advantage of favorable 4 research environment Cooperate with other U.S. R&D 3 laboratories Assist parent company in meeting 1 U.S. environmental regulations Assist parent company's U.S. 2 manufacturing plants in procurement Engage in basic research 1 1 2 1 1 3 2 2 3 4 4 1 2 2 4 4 4 4 2 SOURCE: Adapted from Dalton and Serapio (1995).

56 FOREIGN PARTICIPATION IN U.S. RESEARCH AND DEVELOPMENT research. At the same time, the automotive industry placed much more emphasis on assisting its manufacturing affiliates with local procurement, and assisting the parent company in meeting U.S. environmental regulations than did the other two industries. The U.S.-based R&D facilities of European automotive and electronics com- panies have not received as much scholarly attention as those of the Japanese, therefore it is difficult to draw cross-national comparisons in these sectors. Data gathered by Dalton and Serapio (1995) indicate that European-owned automotive R&D facilities are engaged in a narrower range of activities (mostly automotive design and styling) than their Japanese-owned counterparts. In the electronics industry, U.S. affiliates of European-owned companies tend to concentrate their R&D activities in fewer but larger R&D laboratories than is true for the Japanese. This is also the case in the biotechnology, pharmaceuticals, and chemicals sec- tors, industries in which European-owned firms are world technological leaders (Dalton and Serapio, 1995; Dibner et al., 1992; Peters, 1991, 1992~. In the electronics, pharmaceuticals, and biotechnology industries, European- owned freestanding R&D facilities appear to be less specialized (focusing on more than one product or technology area) and enjoy greater autonomy than those owned by the Japanese.22 Compared to their Japanese counterparts, European- owned U.S. research laboratories in both sectors appear to be more heavily ori- ented toward supporting the technical needs of their U.S.-based manufacturing affiliates and less focused on drawing on U.S. research and technical talent to support the technology strategies of their parent companies (Dalton and Serapio, 1995; Kummerle, 1993a,b; Peters, 1991; Pisano et al., 1988; Westney, 1993~.23 These observations are consistent with the reliance of Japanese-owned firms in most industries on licensing technology developed overseas.24 Finally, although a lack of data makes it difficult to draw comparisons between European and Japa- nese affiliates in this regard, several of the newly established Japanese electronics R&D facilities in the United States appear to be focused on technology or re- search that is deemed critical to the long-term technology strategy of the parent company but is in an area in which the parent company does not yet possess significant capabilities (Kummerle, 1993a,b; Voisey, 1992~.25 INTERNATIONAL CORPORATE ALLIANCES A second major vehicle for foreign participation in U.S. industrial R&D has been corporate technical alliances. Corporate alliances involve sustained col- laboration between independent firms in R&D, product development, produc- tion, or marketing. They demand ongoing contributions of technology, capital, or other assets by the participants. International corporate alliances have a long history in many U.S. indus- tries from oil and chemicals to power generation and automobiles. Indeed, since 1945, such alliances have accounted for a large share of foreign investment by

PRIVATELY FUNDED R&D 57 U.S. manufacturing companies (Hladik,1985~. Since the mid-1970s, the number of domestic and international corporate alliances involving U.S. companies has grown rapidly, particularly in it&D-intensive industries. This trend is a result of companies' efforts to respond to increasing international competition, the rising costs and risks of applied research and product development, and the increasingly interdisciplinary nature of innovation in many industrial sectors.26 Since the early 1980s, alliances between U.S.- and foreign-owned companies in the United States has also been encouraged by changes in U.S. antitrust law.27 During the l950s and 1960s, corporate alliances were based nearly exclu- sively on joint marketing ventures. Today, the emphasis is on more technology- intensive activities, such as joint R&D, development, and production. Corporate technical alliances can assume many forms, including technology cross-licensing agreements, joint technology development, technology-acquisition (equity) in- vestments, second-sourcing agreements, servicing contracts, and outright joint ventures. Between 1976 and 1987, the number of international R&D joint ventures entered into by U.S. companies grew on average by more than 17 percent per year (Hladik and Linden, 1989~. As of 1987, nearly half of such joint ventures tracked by Hladik and Linden were in four it&D-intensive industries: electronics, com- puters, semiconductors, and instrumentation. More recent surveys (Peters, 1992) indicate that licensing is the predominant form of technical alliance in the phar- maceuticals industry, joint ventures are preferred in telecommunications, and the chemical industry relies on both approaches equally. In another analysis, Hagedoorn and Schakenraad (1993) documented a surge during the 1980s in the number of newly established international technical alli- ances (both equity and nonequity arrangements) involving U.S. firms. Most of these alliances were concentrated in three areas: information technology, bio- technology, and new materials (Figure 3.7~.28 However, other U.S. industnes, such as aerospace, automobiles, chemicals, and steel, also experienced a prolif- eration of transnational technical alliances in the 1980s (Mowery, 1988a; Peters, 1991; Vonortas, 1989~. Technical alliances between U.S. and European companies have been con- centrated in the areas of information technology, instrumentation, and medical technology. Most U.S.-Japanese corporate technical alliances have been in the areas of automotive technology, instrumentation, and medical technology.29 Technical alliances between U.S. and European firms have generally focused more on R&D and less on enhancing market access than those between U.S. and Japanese firms (Hagedoorn and Schakenraad, 1993~. A 1990 Department of Commerce survey (Dalton and Genther, 1991) of U.S.-Japanese corporate link- ages in six industrial sectors30 found that most of the alliances were concerned with the production or development of new products rather than with research. As the preceding discussion indicates, corporate technical alliances (whether national or international in scope) may yield significant benefits to the companies

58 500 - 300 - 200 - 100 - 0~ FOREIGN PARTICIPATION IN U.S. RESEARCH AND DEVELOPMENT Europe-Japan U.S.-Japan U.S.-Europe 1 985-89 1 980-84 Biotechnology New materials Chemicals Automobiles Information technology FIGURE 3.7 Number of new transnational corporate technology alliances in the United States, by industry and partnership nations or region, 1980-1984 and 1985-1989. SOURCE: National Science Board (19931. involved and society in general by fostering innovation, technology transfer, pro- ductivity gains, new products and processes, and economic growth more broadly. At the same time, increased alliance activity in high-technology industries that are already highly concentrated at the national and international level also carries with it a greater potential for collusive behavior by the firms involved. There has been little effort to assess or document the anticompetitive effects of the most recent wave of corporate alliances. However, the history of cartelization in many industries during the first half of the twentieth century illustrates the potential costs to society of domestic or international alliances that result in monopoly abuse (Hexner, 1945; Stocking and Watkins, 19461. OPPORTUNITIES AND RISKS The committee finds little value in debating whether foreign participation in U.S.-based privately funded R&D, either through foreign direct investment or corporate alliances, is generally good or bad for U.S. economic and national secu-

PRIVATELY FUNDED R&D 59 rity interests. Clearly, growing foreign involvement in the United States' indus- trial R&D base has costs and risks as well as benefits and opportunities. While the positive and negative consequences of this foreign involvement can be readily described, it is virtually impossible to quantify the associated benefits and costs. The committee nonetheless believes it worthwhile to explore whether specific costs and risks can be isolated. If so, these may then be assessed and either reduced, eliminated, or at least better managed through policy actions in the pub- lic or private sector. Debate over the costs and risks of growing foreign participation in U.S. in- dustrial R&D activity has focused on four major issues: . the failure of foreign participants to provide an adequate economic or technological quid pro quo for the benefits they receive; . the lack of reciprocal U.S. access to foreign-based privately funded R&D assets and activities; · the dangers of such participation to U.S. military security; and · the dangers of such participation to U.S. economic security (i.e., its effect on the ability of U.S.-owned companies to access technologies critical to their competitiveness in world markets). The following discussion assesses each of these four concerns, and it identi- fies the benefits and opportunities of foreign investment in privately funded U.S. R&D that policymakers may need to consider as they shape responses to the concerns. Do Foreign Participants Offer an Adequate Quid Pro Quo? The growth and nature of foreign-controlled industrial R&D have led a num- ber of observers to challenge its net contribution to the U.S. economy and tech- nology base. Since more than 80 percent of foreign direct investment in the United States during the 1980s went to acquire existing U.S.-based businesses, it is estimated that a majority of the recent increase in affiliate R&D spending in the United States has come from the acquisition of existing U.S. R&D operations rather than the establishment of new R&D operations by foreign-owned compa- nies.3i This has led some to question whether foreign participation in U.S. indus- trial R&D has been truly additive or has merely displaced would-be U.S. owners of R&D assets, technology, knowledge, and their associated revenue streams. Limited evidence suggests that several recently established U.S. R&D facili- ties owned by Japanese firms in the electronics and pharmaceutical/biotechnol- ogy industries are drawing on areas of U.S. research strength that are relatively new to the Japanese parent company yet are viewed as critical to the parent's long-term technology strategy (Kummerle, 1993a,b; Voisey, 1992; Westney, 19931. For the most part, however, data on foreign direct investment suggest that foreign parent companies have invested in U.S.-based high-technology assets in

60 FOREIGN PARTICIPATION IN U.S. RESEARCH AND DEVELOPMENT areas in which the parent company has a strong export position or demonstrated competitive advantage. In other words, foreign companies appear to be trying to exploit their company-specific competitive advantages in the U.S. market rather than trying to buy their way in to areas of U.S. competitive advantage (U.S. De- partment of Commerce, 1993a). Nevertheless, two issues remain troubling. First, are the U.S. affiliates of for- eign-owned firms performing enough high-yield R&D in the United States given the volume of their sales in the U.S. market? Second, are they removing more intellectual property and associated economic value than they are contributing? R&D Intensity of Affiliates Some argue that foreign-owned firms with U.S.-based manufacturing facili- ties are not performing their fair share of R&D in the United States. Indeed, the average it&D-to-sales ratio (one measure of R&D intensity) of affiliates was lower than that for all U.S.-based firms in five of seven major high-technology manufacturing industries and in manufacturing industries as a whole in 1992 (National Science Foundation, 1994; U.S. Department of Commerce, 1993a, 1 995a). This has led several analysts to suggest that the foreign parent companies of U.S.-based affiliates are deliberately and unfairly retaining in the home coun- try certain R&D projects and other high-value-adding activities (and their associ- ated jobs) that might otherwise be conducted in the United States. This argument has also been used to support policies that would place performance requirements or other restrictions on foreign firms seeking to invest in U.S. high-technology industries (Gaster, 1992; U.S. Congress, Office of Technology Assessment, 1994~. In 1992, foreign-owned manufacturing affiliates spent less on R&D relative to their total sales than did U.S.-based manufacturing companies overall (Figure 3.81. Nevertheless, the average intensity of all affiliate R&D doubled during the past decade, from 0.5 percent in 1980 to 1.1 percent in 1992.32 There is signifi- cant variation among industries, however. For example, in the pharmaceuticals, industrial chemicals, and primary metals sectors, the R&D intensity of U.S. affili- ates was slightly higher than or equal to the average for all U.S.-based firms in these industries. Affiliate R&D intensities were only marginally lower than those of all U.S.-based companies in the audio, visual, and communications equipment, electronic components, and instruments and related products sectors. However, the R&D intensities of affiliates that manufacture computers and office equip- ment, motor vehicles and equipment, and other (nonautomotive) transportation equipment were far below the U.S. industry average in these sectors. There are also significant international differences in the R&D intensity of U.S. affiliates of foreign-owned firms (Table 3.7~. To a large extent, these differ- ences reflect international variations in the composition of foreign direct invest- ment and affiliate sales in the United States. For example, Swiss and German

PRIVATELY FUNDED R&D Industrial chemicals Pharmaceuticals Computers and office equipment Electronic components Audio, video, and communication equipment Instruments and related products Primary metals Motor vehicles and equipment Other transportation equipment* All manufacturing 61 _ All U.S.-based companies Affiliates 1 1 1 1 1 1 1 1 1 o * Includes aircraft and missles. 5 10 15 20 FIGURE 3.8 Ratios of R&D to sales in percent for all U.S.-based companies and U.S. affiliates of foreign firms, by industry, 1992. SOURCE: National Science Foundation ( 1 9961; U. S . Department of Commerce ( 1 99S a,b). direct investment in U.S. industry is concentrated in pharmaceuticals and indus- trial chemicals industries that account for nearly half of all U.S. affiliate R&D spending. These same two industries were responsible for 34.4 percent and 35.7 percent, respectively, of Swiss- and German-owned affiliates' total U.S. manu- facturing sales in 1992. By contrast, much direct investment in the United States by British and Japanese manufacturing firms as well as sales by their U.S. affili- ates are in significantly less R&D intensive industries (U.S. Department of Com- merce, 1995a). Even within industries that are heavily dependent on research and develop-

62 au o Ct Cal ; - ·_4 o _ sit o Lo o Cq Ct .~ ¢ I o C) ~ Cal au 'a Cal o ~ o o o ·_d Cal o o V C _ ~ t_ ~ ~ lo, CN ~ o ~ CM ~ Cal ~ ~ ~ C) o ·3 Cal ~ ~ ~ ~ ~ ~ ~ ~ sac it ,C C Via ~ ~ ~ ~ ~ ~ ~ ~ ~ C ~ -° ~i ~ ~ ~ ~ ~ ~ ~ ~ c c .-o -o O 00 ~ ~ <) ~ ~ ~ ~ Ch O . _ ~ -_~0_ e ~ ~ ~ ~ ~ ~ ~ ~ O ~ _ .= O _ _~0~0~0= =0 ~ C ~ ~ _ ~ ~ ~ O ~ ~ ~ ~ _ v° ~ ~ ~ ~ ~ ~ ~ O ~ <= ~ - ^ ~ D (t U ~ ~ O ~ O ~ . e' ~ ~ ~ ~ ~ , ~ g _ ~ '

PRIVATELY FUNDED R&D 63 meet, there are major differences in R&D intensity among U.S.-based affiliates of different national origin. For instance, the it&D-to-sales ratios of Japanese- owned affiliates in industries that manufacture computers and office equipment, and instruments and related products in the United States are slightly higher than the average for all foreign-owned affiliates. The Japanese ratios are higher than the average for all affiliates that manufacture computers and office equipment, and other (non-automotive) transportation equipment and are significantly lower than the affiliate average for the industrial chemicals, pharmaceuticals, and au- dio, video, and communications equipment industries (Table 3.7~. The fact that the average R&D intensity of affiliates is currently lower than that of all U.S.-based companies and that international variations exist in affiliate R&D intensities does not, in the committee's view, shed much light on either the motives of foreign companies, "equity" issues in general, or, most important, the economic consequences of foreign direct investment (including investments in R&D) for U.S. citizens. The committee does not place much stock in the argument that lower affiliate R&D intensity is the result of "conspiracy" by foreign firms (or by foreign indus- try-government collaborations) to retain high-value-added activity within their home markets. To begin with, the vast majority of industrial R&D in all major industrialized countries is conducted within the home markets of the R&D-per- forming firms (Patel, 1995; Roberts, 1995a). As noted above, an individual company's decisions about undertaking R&D in a given market, the level of that investment, and how this activity should be managed are shaped by many differ- ent factors. These include, among other things, the relative size, projected growth, and special regulatory or other requirements of the target market, as well as the availability of trained scientists, engineers, and technicians, and other elements of a strong technological infrastructure in that market. The potential for economies of scale and scope may also figure in, as may the firm's sense of how important it is to conduct R&D proximate to foreign customers, suppliers, or competitors. Numerous case studies make clear the significant variation in both R&D intensity and spatial organization of R&D activity among firms—even firms of the same national origin within virtually every industry. This is not to say that corporate nationality is irrelevant to the scope, organization, and character of affiliate R&D activity. As noted, leading Japanese electronics companies as a group appear to retain more of their R&D activity within their home market and exercise greater centralized control over R&D activities abroad than do their U.S. and European counterparts (Kummerle, 1993a,b; Westney, 1993~. Nevertheless, the motives behind any single firm's decision to conduct R&D in the United States are multivariate and highly specific to the needs of that firm. Indeed, the realities of corporate R&D practice are much at odds with the rather simplistic notion that a single "appropriate" or"fair" it&D-investment-to-sales ratio exists for companies within a given industry in a given geographic area.33 Fairness or equity arguments regarding affiliate R&D intensities are not en-

64 FOREIGN PARTICIPATION IN U.S. RESEARCH AND DEVELOPMENT tirely compelling either. After all, the R&D intensities of U.S.-owned overseas affiliates are significantly lower for all manufacturing industries as a whole as well as for most high-technology industries individually than are those of for- eign-owned affiliates in the United States (U.S. Congress, Office of Technology Assessment, 19944. Finally, although affiliate R&D activity can yield multiple direct and indirect economic benefits to the United States (see Chapter 2), the R&D intensity of an individual affiliate in and of itself says little about the overall contribution of that affiliate to the U.S. economy or technology base. Indeed, the correlation between R&D spending levels and commercial performance is at best weak at the level of the firm within any given industry (Roberts, l995a,b). For these reasons, as well as because of the complexity and variability of the highly firm-specific calculus involved in decisions regarding R&D investments, the committee believes strongly that the U.S. government should avoid imposing R&D performance standards or taking other policy actions aimed at compelling private companies to increase the R&D intensity of their activities in the United States. The Value of Affiliate R&D Limited data suggest that most of the R&D performed in the United States by foreign-owned firms, whether in freestanding laboratories or in production facili- ties, appears to be oriented toward meeting the immediate technical needs of the firm's U.S.-based production efforts design work, incremental, process-ori- ented, applied R&D, and applied R&D related to U.S. technical standards or regulation. No industrywide or sector-specific data exist on the level of basic or long- term applied research performed by U.S. affiliates of foreign-owned companies. Surveys and case studies suggest such research is rather limited in scope in all but the pharmaceuticals and biotechnology industries. These studies also indicate that a growing number of foreign firms are developing a resident capacity for basic research. Such capacity appears designed primarily to draw more effec- tively on the research competence and creativity of U.S.-trained scientists and engineers as well as on the basic research activities of U.S. companies, universi- ties, and federal laboratories (Brooks, 1994; Pearce and Singh, 1992; Voisey, 1992; Westney, 1993~. U.S. affiliates of foreign-owned firm have increased their patenting activity in recent years. Nevertheless, the proportion of foreign-owned patents awarded to laboratories of U.S. affiliates remains relatively small overall (Table 3.8~. Af- filiates of European-owned firms account for a significantly larger share of their parent companies' patenting activity than do affiliates of Japanese companies (Patel, 1995; Patel and Pavitt, 1991~.34 This is consistent with the findings of Roberts (1995a), which show that major Japanese companies spend less than

PRIVATELY FUNDED R&D 65 their leading U.S. and European competitors on overseas R&D as a percentage of their total R&D investment. There is little reason to believe that the privately funded R&D conducted by U.S. affiliates of foreign-owned firms differs significantly from that of U.S.- owned affiliates abroad. Indeed, what limited comparative evidence there is sug- gests that the importance of different types of R&D varies significantly among industries and among firms within the same industry both for U.S.- and foreign- owned multinational companies. Several studies have noted differences in R&D focus between European- and Japanese-owned affiliates in the U.S. biotechnol- ogy and electronics industnes. However, for the most part, there is relatively little variation in R&D type among companies of different nationality within the same industry (National Science Foundation, 1991; Pearce and Singh, 1992; Serapio, 1994~. There ale few data to support the notion that U.S.-owned firms in a given industry do significantly more basic and long-term applied research overseas than do their foreign-owned competitors in the United States. In fact, patenting data suggest that as a group, U.S.-owned firms do a notably smaller share of their total patent-yielding R&D work in foreign markets than do European-owned firms in the United States (Patel, 1995) (Table 3.8~. Nevertheless, there is one aspect of international R&D activity in which Japanese companies on average appear to TABLE 3.8 Geographic Location of Large R&D Firms' Patenting Activities, By Home Country Location of Patenting Activity Region or Nation in Which "Abroad" Patenting Activity Took Place Firm's Home Country United (number of firms) Home Abroad States Europe Japan Other Japan (139) 99.0 1.0 0.8 0.2 - 0.0 United States (243) 92.2 7.8 6.0 0.5 1.3 Italy (7) 88.2 11.8 5.3 6.2 0.0 0.3 France (25) 85.7 14.3 4.8 8.7 0.3 0.6 Germany (42) 85.1 14.9 10.4 3~9 0.2 0.4 Finland (7) 82.0 18.0 1.6 11.5 0.0 4.9 Norway (3) 67.9 32.1 12.7 19.4 0.0 0.0 Canada (16) 67.0 33.0 24.9 7.3 0.3 0.5 Sweden (13) 60.8 39.2 12.6 25.6 0.2 0.8 United Kingdom (54) 57.9 42.1 31.9 7.1 0.2 3.0 Switzerland (8) 53.3 46.7 19.6 26.0 0.6 0.5 Netherlands (8) 42.2 57.8 26.1 30.6 0.5 0.6 Belgium (4) 37.2 62.8 22.2 39.9 0.0 0.6 All Firms (569) 89.1 10.9 4.1 5.6 0.3 0.8 SOURCE: Patel (1995).

66 FOREIGN PARTICIPATION IN U.S. RESEARCH AND DEVELOPMENT invest greater resources, if not always perform more effectively, than do their U.S. or European counterparts: monitoring and drawing upon R&D activity and technology beyond their corporate and national boundaries (Mansfield, 1988a,b; Roberts, 1995a).35 The Balance of Technology Flows The United States exports significantly more technology than it imports from abroad, and U.S.-owned multinationals are the principal gateways through which technology enters and leaves the country. Nearly 80 percent of the flow of U.S. technology exports, as measured by the receipt of royalties and licensing fees, is between U.S.-owned parent companies and their foreign-based affiliates.36 Unaffiliated foreign-owned companies paid U.S. organizations and individu- als $3 billion for the use of proprietary technology and know-how in 1991, roughly $2 billion more than unaffiliated foreign firms received from U.S. citi- zens for the use of foreign technology. In unaffiliated technology trade, the United States runs a net deficit with Europe and a large surplus with Japan (Figure 3.9~. In contrast, royalty and licensing fee data show that the U.S. affiliates of foreign-owned firms import significantly more technology from their foreign parent companies than they export to them or to other unaffiliated firms abroad. Net payments by U.S. affiliates to their foreign parents increased from $378 mil- lion in 1980 to $2.1 billion in 1991 (U.S. Department of Commerce, 1993a). That year, payments of royalties and licensing fees to parent firms were largest for affiliates of British-owned companies (Figure 3.101. Japanese, Swiss, and German parent firms received fewer payments from their affiliates. In addition to imports of patented or copyrighted technology, which can be readily quantified by data on royalties and licensing fees, U.S. affiliates of for- eign companies, particularly Japanese companies, have also imported into the United States advanced production technologies and methodologies. These have included technologies embodied in advanced manufacturing equipment, unpat- ented production technology and know-how, and organizational innovations, such as concurrent engineering and just-in-time and total quality management tech- niques. Collectively, these less quantifiable flows of technology and know-how are believed to have contributed significantly to increased productivity in many U.S.-based industries (Florida, 1994; Jaikumar, 1989; Kenney and Florida, 1993a; National Academy of Engineering, 1993; Westney, 1993~. Contributions to the technological strength of the United States by U.S. af- filiates of foreign-owned firms appear to vary from industry to industry. For example, a U.S. General Accounting Office study (199Ob) concluded that foreign direct investment in the U.S. chemical industry, which represented more than 30 percent of the industry's assets in 1990, would bring a flow of new technology to the United States. Similarly, sectoral studies of U.S.-Japan technological link- ages conducted during the late 1980s and early 1990s suggest that the transfer to

PRIVATELY FUNDED R&D All countries European Community Asia and the Pacific South and Central America Japan Germany United Kingdom South Korea 67 Receipts 1 1 1 1,,, 0 500 1,000 Payments , 1 1 1 1 1 1, 1 1 1 1 1 1 1 1 1 1,500 2,000 2,500 3,000 Millions of dollars FIGURE 3.9 U.S. royalty and licensing fee receipts and payments resulting from tech- nology trade between unaffiliated U.S. and foreign companies, 1991. SOURCE: National Science Board (19934. NOTE: U.S. payments to both South and Central America and to South Korea were less than $500,000. U.S. businesses of organizational and managerial innovations by the U.S. affili- ates of Japanese companies has been significant in the automotive and steel in- dustries but relatively inconsequential in the consumer electronics industry (Cusumano and Takeishi, 1991; Florida and Kenney, 1992; Kenney and Florida, 1 993a; National Research Council, 1 992a). Similarly, case studies of international corporate alliances indicate that the United States is a net technology importer in some industries and a net exporter in others. Most technology transfer within international corporate alliances in the aircraft and biotechnology industries, for example, has consisted of exports of U.S. technology to other countries (Mowery, 1988b; National Research Council,

3,000 - 2,500 2,000 1 ,000 500 F0~# ~~\ ~ .. RE3~^ ~ ~~r 7-~ ~ ~ Others . . Germany 1 980 1 985 1 991 Payment hno~gy Ida) Switzerland Japan United Kingdom 1 980 1 985 1 991 Receipts (Technology oudiows) FIGURE 3.1D AeI1~[e myth Ed bcensing Sac payments Ed mceipl~ by county of Ukim~e BeneOci~ Owner 1980, 1983, Ed 1991. SOURCE: U S. Dep~enl of Com- meme U993~

PRIVATELY FUNDED R&D 69 1992c, 1994b; Pisano et al., 1988~. However, case studies of alliances in the automotive, steel, integrated circuits, and robotics industries reveal that U.S. firms which collaborate with foreign companies gain access to not only financial re- sources, but also technology particularly production technology and other assets not available from other U.S. firms (Lynn, 1988; National Research Coun- cil,1992b; Steinmuller, 1988; Womack, 1988; Womack et al., l990~. Christelow (1989) has observed that most U.S.-Japanese joint ventures have occurred in in- dustries in which Japanese companies appear to have a demonstrated competitive advantage. An exclusive focus on technology imports and exports ignores the technol- ogy and know-how that foreign-owned companies underwrite, develop, apply, and diffuse throughout the U.S. economy. In U.S. industries that are net export- ers of technology as well as in those that are net importers, foreign-owned firms have complemented existing U.S. assets with valuable assets of their own. These include capital, intellectual property, managerial and organizational know-how, advanced manufacturing equipment, and knowledge of and access to foreign mar- kets. In many industries, these assets have enhanced the productivity of existing U.S.-based assets, including indigenous R&D capabilities. This has occurred both directly, as those assets are used in foreign-owned facilities, and indirectly, as affiliates of foreign-owned firms increase the competitive pressure on indig- enous U.S. producers in many industries (McKinley Global Institute, 1993; Organization for Economic Cooperation and Development, 1994~. Without the complementary R&D assets provided by affiliates of foreign-parent companies, there likely would have been significant delays in the commercialization and dif- fusion of many important product and process innovations developed within the United States. The recent experience of the U.S. biotechnology industry is illustrative. Ana- lysts generally agree that foreign direct investment in and technical alliances with foreign firms have resulted in a net export of technology by U.S. companies. Industry analysts also find that foreign direct investment in the biotechnology sector has had a net beneficial effect on the industry itself (Dibner et al., 1992; National Research Council, 1992c; Pisano et al., 1988~. As a recent Commerce Department study concluded, foreign direct investment "allowed [U.S.-based] companies to survive, retained jobs, [and] increased investment in plant and equipment and R&D to develop new products that might have been dropped due to lack of funding" (U.S. Department of Commerce, 1991~. "Lost Opportunities'' and Technology Stripping There has been considerable discussion in recent years about the so-called lost opportunities resulting from foreign direct investment in the United States, particularly from foreign acquisitions of U.S. high-technology start-up com- panies (Koprowski, 1991; Schrage, 1990~. In the opinion of the committee,

70 FOREIGN PARTICIPATION IN U.S. RESEARCH AND DEVELOPMENT however, this is a highly speculative concern based largely on faulty assump- tions. For example, many of those who decry such supposed lost opportunities seem to assume that an American purchaser of the now-foreign-owned high-tech company would have managed that firm more effectively than the foreign pur- chaser. At the same time, while virtually impossible to verify empirically, foreign acquisitions of U.S. high-technology companies probably have in some cases resulted in the transfer of profits and economic activity abroad that might other- wise have remained in the United States. However, the committee believes it is equally if not more likely that in the absence of a foreign investor, the U.S. R&D assets or intellectual property in question might have remained idle, moved abroad, or been less effectively managed, thus yielding smaller economic returns to U.S. citizens. Indeed, many of the most celebrated instances of foreign acqui- sitions of U.S.-owned companies attest to the fact that there were no U.S. buyers waiting in the wings.37 Some observers also assert that foreign firms deliberately "strip" U.S. tech- nologies by buying small U.S. high-technology companies. However, there are few data to either confirm or refute this claim. Clearly, nothing prevents foreign firms from transferring codified or otherwise readily transferrable technology from the U.S. companies they have acquired to production or R&D sites over- seas. Neither is there anything stopping U.S.-owned companies from licensing, selling, or otherwise transferring proprietary technology to affiliated and unaffili- ated firms abroad. Yet, the principal R&D assets of these small high-technology firms are the individual and collective capabilities of the highly motivated entre- preneurial scientists and engineers who work in them, not the companies' patent portfolios. And, as any manager of U.S. industrial R&D will attest, these highly specialized human assets are very difficult to move from one region of the United States to another, let alone across national borders (Mowery and Teece, 1993J. Ultimately, the committee believes it would be ill-advised for the federal government to attempt to second-guess or otherwise shelter U.S. corporate secu- rity markets in the hope of preempting lost opportunities or technology stripping caused by foreign acquisitions of U.S. companies. This opinion is based on the high degree of risk and uncertainty associated with the development and com- mercialization of any new technology, and on the highly complex process firms undertake when deciding to invest in a particular technology or set of R&D assets. Asymmetries of Access The debate over the costs and benefits of foreign participation in privately funded U.S. R&D has focused primarily on the lack of reciprocal access, its costs to U.S. citizens, and the need to level the playing field internationally. Whether the result of discriminatory public policies, collusive private practices, or long-

PRIVATELY FUNDED R&D 71 standing differences among national systems of corporate finance and governance, barriers that deny U.S.-owned companies access to privately owned R&D capa- bilities in other nations are seen to impose costs both on those firms and on their U.S.-based stakeholders. These costs take the form of foregone exports, market share, profits, and economies of scale and scope. There are also costs associated with more limited or less timely access to leading-edge technological capabilities of foreign suppliers, competitors, scientists, engineers, and customers. The United States has long been the leading proponent of liberal treatment of foreign direct investment worldwide. In international forums, as well as in its bilateral economic relations with other countries, the United States has consis- tently advocated liberalization of international trade and investment, and has sup- ported the national, or nondiscriminatory, treatment of the affiliates of multina- tional companies by their host governments. Furthermore, the United States' market-driven, shareholder-based system of corporate finance and governance has greatly facilitated foreign investment. Until relatively recently, most of America's trading partners in Europe, Asia, and Latin America have to varying degrees regulated and restricted foreign direct investment as well as discriminated against foreign-owned firms operating within their borders. During the past decade, however, public policies governing for- eign direct investment in most industrialized and industrializing countries have been significantly liberalized. Explicit restrictions on foreign investment have been eased or lifted in most industrial sectors. Moreover, in many countries, these changes have been accompanied by privatization of state-owned industries and a general opening up and development of domestic financial markets, includ- ing those for corporate securities (Organization for Economic Cooperation and Development, 1992a, 1993~. At the international level, there has also been sig- nificant progress toward liberalizing the treatment of foreign direct investment. Most notable in this regard are the recent efforts of the Organization for Eco- nomic Cooperation and Development (OECD) aimed at liberalization of interna- tional capital movements, and the market-opening investment provisions of the U.S.-Canada Free Trade Agreement, the North American Free Trade Agreement (NAFTA), and the Uruguay round of multilateral trade negotiations under the General Agreement on Tariffs and Trade (GATT).38 Despite these trends, however, significant impediments to foreign direct in- vestment remain in a number of major economies. For example, in Germany and Japan, complex cross-shareholding and bank-holding arrangements continue to impede foreign acquisitions of indigenous companies (Organization for Economic Cooperation and Development, 1992a; U.S. Congress, Office of Technology As- sessment, 1994~.39 The impact of these impediments on the flow of foreign direct investment appears very pronounced in the case of Japan but is difficult to discern in the case of Germany. Comparisons of foreign direct investment in the major industrial- ized countries with direct investments they make in other nations' economic sys-

72 FOREIGN PARTICIPATION IN U.S. RESEARCH kD DEVELOPMENT terns reveal relatively moderate investment asymmetries for Germany, Great Bnt- ain, and the United States (Figure 3.1 l). Japan, in contrast, invests more than 20 times as much in other countries as it permits in foreign direct investment within its own borders. The causes of Japan's anomalous position are the subject of intense debate.40 At the same time, the success of some U.S. firms in gaining access to the Japa- nese market suggests that the "access problem" may be at least in part the result of U.S. corporate practices in some industnes. Inwardly focused technology de- velopment strategies, the "not-invented-here" syndrome, and a general lack of 25 20 15 10 5 o 20.5 1.3 2.5 1.2 Japan United States United Kingdom Germany FIGURE 3.11 Ratio of direct investment in other nations' economies to amount of for- eign direct investment in home nation, selected countries, 1990. SOURCE: U.S. Con- gress, Office of Technology Assessment (1993~.

PRIVATELY FUNDED R&D tools with which to learn from posed barriers for many U.S. firms. 73 joint ventures and alliances have been self-im- Regardless of its origins, for many Americans Japan's impenetrability to for- eign direct investment raises serious questions about the merits of continuing to provide foreign investors free entry into the United States. Moreover, in some quarters, the access problem contributes to growing skepticism regarding the ef- fectiveness of international economic negotiations in this area. The costs associated with asymmetries of access are relatively easy to de- scribe, but they are difficult to quantify and allocate among the various U.S. and foreign stakeholders (i.e., customers, company shareholders, and workers). At a time when foreign direct investment is an increasingly important engine of growth in world trade,4i sheltered foreign markets impose costs on most U.S. stakehold- ers as well as on many foreign consumers. In large foreign markets that are particularly rich in technological resources, such as Japan, one would expect there to be substantial costs to the United States imposed by barriers to foreign invest- ment. Indeed, in some cases, asymmetries in access to markets have weakened the bargaining position of U.S.-owned companies in the negotiation of alliances with foreign-owned firms.42 Moreover, barriers to foreign direct investment may undercut the competitiveness of U.S.-owned firms, both in the protected market and in global markets generally, by denying them access to important material, financial, and technological resources. The continuing existence of barriers to foreign direct investment has caused some observers to challenge the effectiveness of U.S. government efforts to ne- gotiate reductions in these barriers within various bilateral and multilateral fo- rums, such as the GATT, the OECD, the Asian-Pacific Economic Cooperation (APEC) Forum, the NAFTA, the U.S.-Japan Structural Impediments Initiative (SII), and bilateral investment treaties. In order to force the pace of progress in this area, particularly with the Japanese, the federal government has resorted to more aggressive unilateral actions, such as sanctions imposed under Section 301 of the 1988 Trade Act, quasi-extraterritorial application of U.S. antitrust law, and the use of legal standards based on principles of reciprocity43 and conditional national treatment (Bayard and Elliott, 1994; Beltz, 1995; Coalition for Open Trade, 1994: Graham and Krugman, 1995; Tyson, 1992~. Some advocate even more forceful changes in U.S. policy, such as imposing more extensive disclo- sure requirements on foreign-owned affiliates, increasing the screening of foreign acquisitions in the name of broadly defined national economic interests, or im- posing extensive economic performance requirements on foreign-owned firms (Gaster, 1992; Gaster and Prestowitz, 1994; Tolchin, 1993; U.S. Congress, Of- fice of Technology Assessment, 19941. While some of these unilateral actions may appeal to a collective sense of fairness, it is not at all clear to the committee that they will advance the short- or long-term economic interests of U.S. citizens. Such measures run a serious risk of discouraging more "good" foreign direct investment (with its associated ben-

74 FOREIGN PARTICIPATION IN U.S. RESEARCH AND DEVELOPMENT edits to the U.S. economy) than "bad." In many respects, these unilateral mea- sures undercut the efforts of the United States to champion open markets and the free flow of goods, services, investments, and technology in bilateral and multi- lateral negotiations. There is also a risk that such measures will invite retaliation by America's major trading partners at significant cost to U.S.-owned multina- tional companies and their U.S.-based stakeholders. Given the magnitude of U.S. direct investment abroad, the dominant role of U.S.-owned multinational compa- nies in U.S. exports, and the growing importance of overseas markets and sources of technology to virtually all U.S. industries, the potential costs to U.S. citizens of retaliatory actions are not trivial. Ultimately, the committee believes that the benefits to U.S. citizens of for- eign direct investment and of U.S. direct investment overseas are substantial and, on balance, outweigh the associated costs. Moreover, the committee believes that more restrictive unilateral actions designed to force U.S. access to closed markets abroad actions beyond those already provided for in existing U.S. trade law are more likely to delay rather than advance progress toward more liberal treatment of foreign direct investment and trade worldwide. Implications for U.S. Military Security The U.S. military is perceived to be heavily dependent on the industrial and technological capabilities of resident and nonresident foreign-owned companies.44 The scope and nature of this relationship are poorly documented, however. Pro- curement regulations that place particularly stringent security requirements on foreign prime contractors generally discourage these firms from working directly with the Department of Defense (DOD). Nevertheless, with DOD's tacit bless- ing, U.S.-owned prime contractors routinely subcontract with foreign-owned sup- pliers for critical components and subsystems. The U.S. military's reliance on foreign-owned companies is certain to increase with time, given the growing importance to the military of technologies that have both civilian and military applications and the current and growing strength of foreign-owned firms in many of these dual-use technologies (Alic et al., 1992; National Research Council, 1995). The growth of foreign participation in privately funded U.S. R&D may have a positive effect on U.S. national security. Given the global leadership position occupied by many foreign-owned firms in many areas of dual-use technology, foreign direct investment and international corporate alliances can improve DOD access to innovative technological capabilities important to national defense. For example, Sony Corporation's 1989 acquisition of the U.S. semiconductor equip- ment manufacturer Materials Research Corporation (MRC) prevented the U.S.- owned company from going bankrupt, thereby maintaining a "domestic location and relatively assured fU.S.] access to 60 percent of the world's production capa- bility for sputtering materials" (Defense Science Board, 19901. According to the

PRIVATELY FUNDED R&D 75 Defense Science Board, had MRC gone bankrupt, the United States might have had assured access to only 2 percent of world production capability. Similarly, both the Defense Science Board (1990) and the National Research Council (1992c) have concluded that joint ventures and other types of technical alliances between U.S. and Japanese companies in the semiconductor industry have helped strengthen resident U.S. capabilities in important areas of dual-use technology.45 On the other hand, America's national security may be jeopardized in several ways by increasing foreign involvement in the U.S. dual-use technology base. First, when foreign firms acquire or establish U.S.-based companies in high-tech- nology industries that serve both civilian and military markets or enter into tech- nical alliances with U.S.-owned companies in these sectors, militarily sensitive technology may be more easily transferred (intentionally or not) to current or potential future enemies of the United States. Second, despite the possible benefits to the United States, foreign acquisi- tions of privately held U.S. technological capabilities may reduce the timeliness or increase the expense associated with gaining access to leading-edge or emerg- ing technologies controlled by these companies. Foreign-owned companies may withhold or delay access to dual-use technologies either because they are com- pelled to do so by their government or are seeking a competitive advantage. Third, foreign participation in the nation's dual-use industrial base may pose a more long-term national security risk. As a result of their expanding presence in U.S. commercial markets, foreign-owned firms may be better positioned to take away domestic market share from U.S.-owned competitors in critical dual- use industrial sectors. This, in turn, would enable them to shift sourcing for advanced technological components away from U.S.-based suppliers to suppliers located overseas or beyond the reach of U.S. national security laws and regula- t~ons. The federal government has several options for reducing the risks associated with foreign involvement in the nation's defense technology base. The risk of intentional or unintentional transfer of militarily sensitive U.S. technology by foreign-owned companies appears to be fairly well contained by current laws and procedures.46 All companies operating within the United States are subject to U.S. export control laws (Export Administration Act of 1979, P.L. 96-72~. The Committee on Foreign Investment in the United States (CFIUS), established by executive order in 1975 and formalized by Congress in the 1988 Omnibus Trade and Competitiveness Act, is charged with reviewing foreign acquisitions that might constitute a threat to national security. Under the same 1988 law, the president was given the authority to investigate and block foreign investments that threaten national security. (See box.) Current rules governing DOD contracting impose stringent security require- ments on foreign firms and their U.S. subsidiaries. Under the DOD's program on foreign investment, control, and influence (FOCI), the Defense Investigative Ser- vice can require foreign owners of U.S.-based affiliates that want to work on

76 FOREIGN PARTICIPATION IN U.S. RESEARCH AND DEVELOPMENT The Committee on Foreign Investment in the United States In May 1975, in response to concerns over a surge in foreign petrodollar investment in the United States by members of the Organization of Petroleum Exporting Countries (OPEC), President Ford established the interagency Committee on Foreign Investment in the United States (CFIUS). CFIUS was intended to serve as a central point for gathering and analyzing information on national security-related foreign investments in U.S. firms. CFIUS is run out of the Treasury Department and relies on data collected by other agen- cies, including the Commerce and Defense Departments, the Securities and Exchange Commission, and the Census Bureau, as well as on the voluntary filings by companies involved in a planned foreign acquisition of potentially militarily sensitive U.S. industrial assets. In 1988, Congress formalized the CFIUS as part the Exon-Florio amend- ment to the Omnibus Trade and Competitiveness Act, which expanded the authority of the president to investigate and stop investments that threaten U.S. national security. Exon-Florio established two key requirements for blocking proposed foreign investment: 1) there must be a finding that the foreign entity might take action that could impair U.S. security; and 2) there must be a finding that provisions of law other than the International Emer- gency Economic Powers Act do not provide adequate authority to protect national security. CFIUS does not have the authority or resources to assess broad concerns about foreign involvement in the U.S. defense industrial base or technologi- cally strategic industries. In practice, this means that CFIUS first considers whether a proposed foreign investment is linked to national security. (How- ever, Exon-Florio does not define the term "national security." The legislation also does not provide a precise definition of what firms and technologies are considered critical to U.S. national security.) If CFIUS concludes that a foreign investment would hurt the national inter- est, the situation may be reviewed by more senior government officials. Over the years, CFIUS has attempted, apparently with some degree of success although without legislative authority, to modify objectionable aspects of par- ticular foreign investments, such as the unauthorized access to classified information or technology, rather than preventing the investment altogether. CFIUS also must assess whether other U.S. laws are inadequate to pro- tect national security. It can use provisions in other laws, such as the Export Administration Act, the Defense Production Act, and antitrust laws, to block foreign investment, when necessary. The CFIUS process has been criticized as reactive and case specific. Exon-Florio does not require foreign investors to notify CFIUS of proposed investments, although fear of forced divestiture at a later date may motivate foreign investors to notify CFIUS in advance of any transaction. The Exon- Florio amendment has also been criticized because it does not require for- eign investments in nonpublicly traded companies to undergo CFIUS review (Graham and Krugman, 1995; U.S. General Accounting Office, 1 990a, 1 994b).

PRIVATELY FUNDED R&D 77 classified contracts to either relinquish managerial control of their defense busi- nesses to U.S. citizens or to meet other demanding performance or control crite- ria. The FOCI program appears generally effective in protecting classified infor- mation (Defense Science Board, 1990~. However, the program is a disincentive to foreign acquisitions of U.S. companies involved in defense work and discour- ages affiliates of foreign firms from contracting directly with the Department of Defense (Defense Science Board, 1990; Graham and Krugman, 1991~. Recent trends raise doubts about the ability of U.S. policies to assure access to militarily critical technologies over both the short and long term. Under the federal government's emergency powers authority (Defense Production Act), the DOD may requisition materiel or services from any domestic firm in times of national crisis. However, as in other situations, the United States seems to have little recourse should foreign-owned firms withhold their most advanced technol- ogy from their U.S.-based affiliates and customers. Although there are a number of economic and political disincentives for do- ing so, at least one or two times in recent decades foreign-owned firms have either been compelled by their home governments or opted on their own to with- hold or delay the transfer to the United States of technologies deemed important to U.S. defense. In 1983, the Japanese government reportedly pressured the lead- ing Japanese producer of ceramic materials, Kyocera, to stop supplying through its U.S. subsidiary ceramic nose cones to the U.S. Tomahawk Missile program (Graham and Krugman, 1991~. More recently, Japanese semiconductor equip- ment and materials manufacturers may have withheld their most advanced tech- nologies (many of them considered dual use) from some of their U.S. customers, including the semiconductor, semiconductor manufacturing equipment, and com- puter industries (U.S. General Accounting Office, l991b). Similarly, questions have been raised about the ability of U.S. antitrust law as currently enforced to anticipate and prevent the emergence of monopolies in niche defense markets highly specialized markets served by a small number of suppliers. Accordingly, mergers and acquisitions, whether by U.S.-owned or foreign-owned firms, pose a risk of monopoly if they further diminish the number of competitors within a given segment of the defense supplier base. Because of the limited data available, it is difficult to draw general conclu- sions about the risks to national security posed by denied, delayed, or monopoly- priced access to private technological capabilities, all of which can result from foreign direct investment or mergers and acquisitions. Several observers believe that in the niche defense markets, current monitoring efforts and enforcement of U.S. antitrust law may be inadequate to address the monopoly risks posed by mergers and acquisitions, whether instigated by foreign- or U.S.-owned compa- nies (Graham and Krugman, 1991, 1995~. The task of identifying vulnerable niche sectors is hampered by the federal government's lack of clearly defined, agreed-upon criteria or procedures for determining whether the technological capabilities of a particular company are, in fact, militarily critical.47

78 FOREIGN PARTICIPATION IN U.S. RESEARCH AND DEVELOPMENT Even more difficult to assess is the impact of foreign direct investment on the long-term health of domestic suppliers of dual-use technologies. Although U.S.- owned firms have been displaced in world markets by foreign-owned companies in a number of high-technology product areas, one can only speculate whether the influx of direct foreign investments has accelerated or helped slow this trend. In some instances, foreign parent companies have probably diminished U.S. techno- logical capabilities in certain dual-use areas, either by shifting the R&D activities of acquired U.S.-based companies abroad or by shifting sourcing of advanced technological components from U.S.- to overseas-based suppliers. There is no reason to assume, however, that U.S.-owned firms or resident capabilities more generally would not have experienced a similar contraction when exposed to ex- port competition from abroad. In other instances, such as the Sony takeover of Materials Research Corpora- tion or the U.S-Japanese joint ventures in semiconductor manufacturing men- tioned above, foreign direct investment and the associated pressure of new com- petition have probably helped strengthen important domestic capabilities in the area of dual-use technologies. Despite the difficulty of assessing the long-term consequences of foreign direct investment for particular industry niches, the committee believes that U.S. regulations that discourage foreign-owned firms from contracting directly with the DOD or investing in existing U.S.-based defense contractors may pose greater costs and risks to U.S. national security than they prevent. The committee also believes that the federal government must develop more sophisticated capabilities for assessing and addressing the risks and capitalizing on the opportunities presented by the growth of foreign involvement in the nation's dual-use technology base. Greater scrutiny of mergers, acquisitions, and corporate alliances in dual-use industries, particularly in niche technology areas, might reduce the risk of anticompetitive behavior by both foreign- and U.S.- owned firms (Graham and Krugman, 1991, 19951. More important, to serve U.S. national security interests more effectively, DOD needs to define clearly the cri- teria and procedures for identifying militarily critical technological assets and broaden its portfolio of strategies for managing inevitable U.S. dependence on foreign technological capabilities. Implications for U.S. Economic Security Closely related to the above concerns is the broader question of whether foreign direct investment may help foreign-owned firms acquire monopoly con- trol of established and emerging commercial technological capabilities critical to U.S. long-term economic growth and development. Again, however, there are few data that validate this concern. As noted, some evidence suggests that Japanese firms have withheld their most advanced technologies from some U.S.-owned companies in the semicon-

PRIVATELY FUNDED R&D 79 ductor, semiconductor manufacturing equipment, and computer industries (U.S. General Accounting Office, l991b). Without timely access to these technolo- gies, most U.S. companies in these industry sectors could not sewe as effectively as they might the needs of their worldwide customers. The effect on the Japanese firms involved is less clear. Their growing involvement in U.S.-based R&D ac- tivity in these industries, through direct investment or joint ventures, may have either enhanced or weakened their ability to manipulate the market to gain com- mercial advantage (National Research Council, 1992c). Furthermore, broader concerns have been raised about the long-term effects of foreign acquisitions of U.S. niche technology companies or high-technology start-ups on the ability of U.S.-owned companies to access emerging critical ci- vilian technologies. Much more than any other industrialized country, the United States has relied since World War II on technology niche companies for a dispro- portionate share of major product and process innovation (Mowery and Rosenberg, 1993~. Since the mid-1980s, foreign acquisitions of these firms ap- pear to have increased significantly. Data gathered by Spencer (1991) and the U.S. Department of Commerce (1993a) underline the particularly large appetite of Japanese investors for U.S. high-technology start-up companies during the late 1980s and early 1990s (Table 3.9~. These data also suggest that foreign investors have targeted U.S. companies in a select number of technology areas deemed critical or emerging by the U.S. government (Council on Competitiveness, 1991; National Critical Technologies Panel, 1993; U.S. Department of Commerce, 1990~. Still, inventories of foreign acquisitions of U.S. high-technology firms offer little if any insight into the relative importance of the technological capabilities acquired, let alone the cumulative effect of these purchases on the nation's capa- bilities in a given area of technology. For example, such data do not provide information on sales, assets, or employment for either the firms acquired or their niche industrial sector. Nor do these data shed any light on the performance of the acquired firms subsequent to their takeover. Without such information, it is impossible to assess the extent and significance of foreign control in any high- technology niche sector. Many high-technology industries are already highly concentrated at the na- tional and global levels. Hence, it is likely that at least some of the many recent mergers and acquisitions have significantly reduced competition in particular civilian industry sectors. For this reason, the committee believes that the federal government should intensify its scrutiny and regulation of all U.S.-based merg- ers and acquisitions. However, there is no evidence to suggest that foreign- owned firms are any more likely than their U.S.-owned counterparts to engage in anticompetitive activity in the United States. Indeed, in terms of the sheer number of acquisitions and mergers concluded in U.S. technology-intensive in- dustries each year, foreign acquisitions are dwarfed by those involving U.S.- owned firms.

80 o oo ~o C~ C) o V _ Cq C~ C~ ;^ s~ Cq U) G) ._ s: C~ o o o E~ .= . U) . o C~ o ._ ._ Cq ._ 5 ¢ ._ s~ o E~ ~ Ct E~ o o ._ m C~ 1 o £ ~ o ~ E~ £ C~ o 3 o C) ._ V) o C) o £ ~ ~, ·£ ~ 5 ~Q ~ C~ ._ o C) C) £ o v C~ Ct C~ ._ v C) C~ o C) U. Ct ._ C) C~ ~ ~ ~ ~ C~ X ~ ~ ~ ~ ~ oo ~ ~ vo ~ o ~ o o o ~ X U~ CM ~ o ~ o ~ o o U~ C~ ~ ~ u~ ~ C~ CM ~ ~ O O ~D ~) U~ CN ~ ~ ~ O C~ O C~ 0 M0 C~ _ ~ ~ O ~ O ~ O O ~ ~ ~ ~ ~ O O 00 O ~ O ~ ~ ~ ~ ~ ~ 00 ~ ~ CM ~ CM ~ ~ ~ ~ ~ O O O ~ M0 oo d" V~ ~ ~ O O O O O C~ O ~ ~ ~ ~ O .~ ~ t- C) Ct ~ ~ C) Ct . V V ~ 3 ~ ~ U) ~ ~ C~ U, Ct ~ _ Ct o ·3 ~o C~ C~ U, Ct V £ ~o C~ rO . . o

PRIVATELY FUNDED R&D 8 SUMMARY During the past decade, foreign participation in privately funded U.S. R&D has grown significantly, both through direct investment and intercorporate tech- nical alliances. In 1990, U.S. affiliates of foreign-owned firms accounted for a sizeable fraction (20 percent or more) of U.S.-based R&D activity in several large manufacturing industries. These included industrial chemicals, pharmaceu- ticals, audio, video, and communications equipment, and primary and fabricated metals. Roughly two-thirds of total affiliate R&D spending in the United States is accounted for by companies based in Canada, the United Kingdom, Germany, Switzerland, and Japan. Since 1980, Japanese-owned affiliates have increased their share of total affiliate R&D activity faster than has any other major investing coun- try. For the most part, foreign parent companies have invested in areas in which they have a strong export position or demonstrated competitive advantage. Comparative surveys of U.S.- and foreign-owned multinational companies across a range of industries suggest that the motives for engaging in R&D in a foreign market, as well as the type of R&D activity, vary in importance primarily according to the industry sector, not the company's nationality. Overall, the two most frequently cited motives are to help the local manufacturing affiliate and the parent company meet the demands of U.S. customers, and to improve access to U.S. scientific and technical talent. Most major foreign-owned R&D facilities are clustered near major U.S. cen- ters of R&D activity, and most affiliate R&D performed in the United States appears oriented toward meeting the more immediate technical needs of the af- filiates' U.S.-based production facilities. Thus, affiliates concentrate largely on design work, incremental process-oriented applied R&D, and applied R&D re- lated to U.S. technical standards or domestic regulation. Growing foreign involvement in the nation's industrial R&D base brings with it costs and risks as well as benefits and opportunities. One key question, however, cannot be answered definitively: Do foreign nationals take away more technology and associated economic value than they return to the United States through their participation in U.S. industrial R&D? The evidence suggests that in general, the technological contributions of foreign-owned firms through affiliates or technical alliances vary from industry to industry. Thus, in some industries, foreign firms are net exporters of technology; in others, they are net importers. The few quantitative measures that exist confirm that overall, U.S.-based affili- ates of foreign-owned firms import significantly more codified technology from their parent companies than they export to them or to unaffiliated firms abroad. Case studies show that foreign-owned companies, and Japanese companies in particular, have imported significant amounts of advanced production technology and methodologies into the United States in several industries. Foreign involvement in U.S.-based industrial R&D has in some cases re- sulted in lost opportunities for U.S.-owned firms, as well as in foregone wealth

82 FOREIGN PARTICIPATION IN U.S. RESEARCH AND DEVELOPMENT for their U.S.-based stakeholders. In other cases, however, foreign firms have created opportunities and wealth for U.S.-owned firms and Americans generally as they transfer technology, know-how, capital, and other complementary assets to the United States. On balance, the committee considers the growth of foreign direct investment in the United States and the proliferation of transnational corporate alliances to be generally positive trends that enhance the productivity and wealth of the United States and its trading and investing partners overseas. Furthermore, the commit- tee believes that for the purpose of assessing its consequences for the U.S. economy, foreign participation in privately funded U.S. R&D cannot be sepa- rated meaningfully from the larger trends that carry it. Asymmetries of access to the economies and innovation systems of the United States and other industrialized nations have affected profoundly public perceptions and federal policies on foreign involvement in U.S.-based R&D. During the past decade, America's trading partners have liberalized their policies on foreign direct investment- the most important avenue of access to privately funded R&D activities abroad—making them more similar to those of the United States. Nevertheless, significant impediments to open access the product of structural barriers, public policies, or collusive or discriminatory corporate prac- tices remain in some major economies. These impediments have led some to call for aggressive unilateral action by the U.S. government. The committee, however, considers many such proposals ill advised and, in their stead, urges the government to use existing policies to hold itself and its trading partners account- able to international agreements. The government should also redouble its efforts to negotiate solutions to these asymmetries of access in bilateral and multilateral forums. The public debate about national security has focused almost exclusively on the difficult-to-assess risks that accompany growing foreign involvement in par- ticular it&D-intensive industries. Current national security regulations and pro- cedures appear to minimize the risk that militarily sensitive U.S. technology will be transferred to foreign-owned companies. However, the utility of these rules and procedures is of questionable value for addressing the medium- to long-term risks of delayed or denied access to militarily critical technological capabilities posed by foreign direct investment or mergers and acquisitions. Two sets of issues inform the debate on whether existing measures intended to protect national security are sufficient. On the one hand, current monitoring efforts and methodologies associated with the enforcement of U.S. antitrust laws may be neither extensive nor strong enough to address the monopoly risks that mergers, acquisitions, and corporate alliances pose in niche defense markets, whether instigated by foreign- or U.S.-owned companies. Moreover, the task of identifying vulnerable niche sectors is made difficult by the federal government's lack of clearly defined, agreed-upon criteria or procedures to determine whether a particular company's technological capabilities are critical to the military.

PRIVATELY FUNDED R&D 83 On the other hand, little consideration has been given to the costs and risks to national security posed by existing procurement regulations, which discourage foreign-owned firms from contracting directly with DOD or investing in existing U.S.-based defense contractors. The committee believes such concerns merit greater attention in the national security debate. This is especially true given the current relatively high level of U.S. dependence on foreign sources of component technology, the growing importance of technologies that have both civilian and military applications to the nation's military needs, and the growing strength of foreign-owned firms in many of these dual-use technologies. There is little evidence that increasing foreign involvement in U.S. indus- trial R&D through direct investment and alliances has damaged U.S. economic security. The ability of U.S.-based companies to access the technologies, compo- nents, and subsystems required to make their major products competitive with foreign producers does not appear to be impaired. Many high-technology indus- tries are already very concentrated at the national and global levels. Hence, it is likely that at least some of the many recent mergers, acquisitions, and alliances have fostered monopolies in particular civilian high-technology industries. How- ever, the actions of foreign-owned firms within the U.S. economy do not suggest that they are any more likely to engage in anticompetitive activity than their U.S.- owned counterparts. The committee anticipates that the fundamental trends that have fueled the experience of foreign involvement in privately funded U.S. R&D will continue to spur its growth into the next century. Carried by expanding international trade and foreign direct investment, global technical and economic capabilities will be distributed more evenly among an expanding population of industrialized coun- tries, and competition and innovation in most manufacturing and service indus- tries will become increasingly internationalized. NOTES 1. Privately owned U.S. companies financed 59 percent of all R&D performed in the United States in 1994 and an estimated 90 to 95 percent of all privately funded R&D (National Science Foundation, 1995b). 2. Data on R&D spending by U.S. affiliates of foreign-owned firms come from the Annual Sur- vey of Foreign Direct Investment in the United States, conducted by the Department of Commerce's Bureau of Economic Analysis. The Survey of Industrial Research and Development, conducted by the Bureau of the Census for the National Science Foundation, provides data on the total amount of privately funded U.S. R&D. Changes introduced recently in the methodology and scope of the Survey of Industrial Re- search and Development have resulted in significant upward revisions in estimates of R&D funded by U.S. companies. Beginning in 1992, survey statistics are based on annual sampling. Previously, samples were selected every 5 to 7 years. In addition, "For 1992, the sample size was increased from approximately 14,000 to approximately 23,000 firms to better account for births of new R&D-per- forming establishments in the survey universe, to survey more fully and accurately R&D activity in the service sector, and to gather more current information about potential R&D performers" (National Science Foundation, 1993b). With the upward revision in the volume of U.S. company-funded R&D,

84 FOREIGN PARTICIPATION IN U.S. RESEARCH ED DEVELOPMENT there has been a corresponding downward revision of the share of R&D conducted by U.S. affiliates of foreign-owned firms. For instance, the proportion of R&D conducted by affiliates in 1992 fell from 17.1 percent in the old data series to 14.5 percent in the new data series (National Science Foundation, l995b, 1996; U.S. Department of Commerce, 1995a). 3. For a more extensive overview of the many factors that have contributed to the international- ization of industrial R&D, or more specifically to the decisions of many multinational firms to estab- lish R&D subsidiaries abroad, and an extensive bibliography on the subject, see the review article by Cheng and Bolon (1993). 4. Illustrating the expectations that some foreign firms have of their U.S.-based R&D facilities, Serapio and Dalton (1994) quote the following observation of an executive of a large Japanese auto- motive company: "Prior to establishing an R&D center in [the United States], our engineers in [the U.S. plant] had to work with the R&D and technical centers in Japan. We were sending faxes to each other all the time and our engineers took many trips between Japan and the United States. The process of developing and producing a car for the U.S. market or correcting an engineering problem was very time consuming. We have eliminated this lengthy process by establishing a U.S. technical center [in close proximity] to our U.S. plant, sales office and suppliers in the United States. We expect to shorten the time needed for concurrent design and development, concurrent development and engi- neering, and working on design and engineering issues for our vehicles in or near production." (p. 29) 5. Recent surveys of Japanese multinationals in a number of industries suggest that the fear of diminished access to U.S. technology has played an important role in the decision of many Japanese companies to establish an R&D presence in the United States (Ministry of International Trade and Industry, 1992; Serapio, 1994). As discussed in greater detail in Chapter 4, the U.S. Congress has placed restrictions on foreign corporate involvement in publicly funded R&D in American universi- ties, federal laboratories, and other U.S.-based institutions performing government-funded research. For example, foreign participation in the Department of Commerce Advanced Technology Program, Advanced Research Projects Agency-coordinated Technology Reinvestment Project, and cooperative research and development agreements with federal laboratories is conditioned on reciprocal access to comparable government-funded R&D initiatives abroad. Several U.S. industry-led consortia, includ- ing the Semiconductor Manufacturing Technology Research Corporation (SEMATECH), the Na- tional Center for Manufacturing Sciences, the U.S. Display Consortium all partially funded with public monies exclude foreign participants. In 1987, a U.S. government-sponsored symposium on high-temperature superconductivity (HTS) excluded foreign participants. Subsequent Reagan administration proposals for additional research funding for HTS included provisions designed to prohibit or restrict foreign access to the results of publicly funded research in this area (Mowery, 1991). 6. For documentation of the growing interest and involvement of U.S.-owned multinational com- panies in the advanced technological capabilities of foreign firms and foreign countries see, for ex- ample, Dalton and Serapio (1995), Mansfield et al. (1979), Mowery (1991), National Science Foun- dation (199Oc, 1991), and Peters (1992, 1993b). 7. The U.S. Department of Commerce defines a foreign investment as direct when a foreign investor acquires a stake of 10 percent or more in a U.S. firm. The 10-percent criterion, although arbitrary, is meant to reflect the idea that a large stockholder will generally have a strong say in the operations of a company, even if that stockholder does not have a majority stake. Data from 1988 show that on average, the foreign parent controlled 78.8 percent of its U.S. affiliate's equity. Calcu- lations by analysts at Commerce's Bureau of Economic Analysis indicate that raising the definitional cutoff for direct investment to 20 percent or even 50 percent would increase only slightly estimates of the amount of foreign direct investment in the United States (Graham and Krugman, 1991, pp. 9-11). Hence, most experts consider the potential for understating or overstating the level of foreign control with the 10 percent criterion to be small. There is one notable case of relevance to this study in which a foreign firm with a moderate equity share in a U.S. company, according to the Department of Commerce criterion, "controls" a very large R&D portfolio. Until spring of 1995, the Canadian

PRIVATELY FUNDED R&D 85 Bronfman (Seagram) family owned a 23 percent stake in DuPont, which invests over $1 billion annu- ally in R&D worldwide. On April 7, 1995, the Bronfman family sold its stake in DuPont, thereby removing DuPont and its R&D spending from the ledger of foreign-owned affiliates. 8. See U.S. Department of Commerce, Survey of Current Business, May issue, various years, for data regarding the share of foreign direct investment accounted for by foreign acquisitions of existing U.S.-based firms and that accounted for by the establishment of new U.S.-based companies by for- eigners. 9. "Special tabulations were prepared by [the Commerce Department's Bureau of Economic Analysis] to reveal R&D expenditures in the United States of those firms in which there is majority foreign ownership—i.e., 50 percent or more. For 1990, the 10-percent foreign ownership threshold results in an estimated $11.3 billion foreign R&D investment total. R&D expenditures of majority owned U.S. affiliates of foreign companies were $8.4 billion. "Funding trends of these two groupings are quite similar. From 1980 to 1990, inflation- adjusted R&D spending of majority-owned foreign firms was up 350 percent, whereas that of firms with 10 percent or more foreign ownership (including majority-owned firms) rose slightly more, 370 percent" (National Science Board, 1993, p. 125, footnote 77). The particularly rapid growth of affiliate R&D spending between 1987 and 1990 included several multimillion dollar acquisitions by foreign firms of U.S. pharmaceutical companies with large R&D budgets, such as Glaxo, SmithKline Beecham, and Genentech (Dalton and Serapio, 1993). 10. Majority-owned affiliates of foreign companies performed approximately $10.7 billion of R&D in 1992, roughly $750 million more than was spent by U.S. companies and their foreign subsid- iaries on overseas R&D that year (National Science Foundation, 1994, Table SD-5; U.S. Department of Commerce, 1994a, Table N-1). 11. Florida and Kenney (1993) go so far as to conclude that "foreign R&D investment in the U.S. is largely the province of corporations which are global technology leaders and that it is primarily used to consolidate that position of technological advantage." (p. 30) 12. The volume of affiliate R&D spending in the U.S. motor vehicle and equipment industry is arrived at by adding affiliate R&D expenditures that are classified in the Bureau of Economic Analy- sis' survey of affiliate R&D as "manufacturing R&D" with those that are classified as "wholesale trade R&D." Department of Commerce analysts acknowledge that much of the R&D spending clas- sified as "wholesale trade R&D" by affiliates in the motor vehicles and equipment industry is, in fact, R&D performed by U.S. manufacturing establishments of foreign-owned firms. 13. In 1992, U.S. affiliates of foreign-owned firms performed $170 million of R&D for the U.S. federal government and $689 million of R&D for other unaffiliated U.S.-based organizations. Ninety- four percent of the R&D performed by affiliates in 1992 was for the affiliates themselves (U.S. De- partment of Commerce, 1995a). For one of the most extensive, though by no means comprehensive, inventory of U.S.-based organizations that perform industrial research, development, and design, see the Directory of Ameri- can Research: Organizations Active In Product Development for Business, published annually by R. R. Bowker, New Providence, New Jersey. 14. Freestanding R&D facilities appear to play the greatest role in two industries: pharmaceuti- cals and chemicals, which together account for nearly half of all affiliate R&D expenditures. Whether freestanding facilities account for a majority of affiliate R&D expenditures in discrete manufacturing industries (i.e., automobiles or consumer electronics) is more of an open question (Peters, 1992). 15. Dalton and Serapio (1995) define a foreign R&D facility in the United States "as a free- standing R&D company (i.e., a company engaged mainly in R&D) of which 50 percent or more is owned by a foreign parent company." (pp. 16-17) A 1993 Department of Commerce report (1993a) noted that data collected earlier by Dalton and Serapio (1993) "include numbers of design studios, which are not considered research facilities by the National Science Foundation, and thus, possibly represent a significant overstatement of re- search facilities in the United States as much as 80 percent above actual R&D facilities, according

86 FOREIGN PARTICIPATION IN U.S. RESEARCH AND DEVELOPMENT to one NSF analyst." Research by Peters (1991, 1992) also suggests that data gathered by Dalton and Serapio considerably overstate the number of "true" R&D facilities owned by foreign companies in the United States. 16. The relatively small size of many Japanese-owned freestanding R&D facilities in the United States has led some observers to suggest that these facilities may serve mainly as "listening posts" rather than research units (Fusfeld, 1994; Peters, 1991). Serapio (1994), however, notes that the smaller size of Japanese R&D facilities in the United States could be explained by their more special- ized orientation and relative newness. This may be true in the case of many recently established Japanese-owned R&D facilities in the U.S. biotechnology and electronics industries, which conduct mainly basic research. For further discussion, see pp. 54-55. 17. The overall response rate to the Pearce and Singh survey was 28.9 percent. Parent companies supplied 163 of 296 usable responses; subsidiaries supplied the remaining 133 responses. The re- sponse rate from 181 U.S.-based affiliates that were sent the survey was roughly 30 percent. The authors note that response rates were poor for French-owned companies and that there were too few Japanese companies in their sample to say much about them. 18. NEC Research Institute was established in Princeton, New Jersey, in 1989. It focuses on software development, artificial intelligence, and machine learning (Noguchi, 1989). Matsushita established its Information Technology Laboratory in Princeton the following year to conduct basic research in computer graphics, document processing, and systems software. In 1991, Hitachi set up a high-definition television research laboratory in Princeton. That same year, Mitsubishi Electric set up a basic research laboratory in Cambridge, Massachusetts, that focused on next-generation parallel processing and supercomputers. Canon established the Canon Research Center in Palo Alto, Califor- nia, in 1990 to do research on data compression, optical recognition, and network architecture. Florida and Kenney (1993) note that these facilities tend to focus on areas of research "where Japanese indus- try lags the United States, such as parallel computing, software development, and artificial intelli- gence." This may explain their proximity to leading U.S. centers of R&D in these fields. Florida and Kenney also note that these laboratories "are designed to generate and harness new sources of knowl- edge, which leverages existing corporate technological capabilities and enhances long-range corpo- rate development efforts." 19. Dibner et al. (1992) note that because of their large staffs, European-owned biotechnology R&D facilities have many more interactions with U.S. university- and industry-based researchers than Japanese-owned facilities with smaller R&D staffs. R&D managers generally agree that there is a minimum efficient size for R&D laboratories that varies from industry to industry. Below this minimum, laboratories are unlikely to be successful either at performing in-house research or drawing upon extramural research. 20. From comments by NEC Research Institute Director William Gear before the NAE study committee at a November 9, 1993, workshop. See, also, Chapter 4, p. 110. 21. Of the 12 Japanese-owned electronics R&D facilities surveyed by Dalton and Serapio (1993), six focused on applied research (new applications for existing technology, design customization), two performed basic research exclusively, and four conducted both basic and applied research, although basic research was limited. Ten R&D facilities performed prototype testing, evaluation, and produc- tion, designed new products or modified existing product designs; seven did parts evaluation; five designed parts; seven conducted joint R&D with other research organizations; and seven employed university research professors as consultants. In contrast, six Japanese automotive affiliates performed vehicle testing and evaluation, emis- sions certification, scanning of regulatory trends, technology scanning, and advanced concept design; four of the six firms surveyed did parts and materials evaluation and design; two performed joint research with a U.S. partner and produced prototypes of near-production vehicles; and three compa- nies were involved in the development of local parts suppliers. 22. See, for example, Dibner et al. ( 1992) and National Research Council ( 1 992c) for a discussion of affiliate-conducted R&D in the pharmaceuticals and biotechnology industries. For affiliate R&D

PRIVaTELY FUNDED R&D 87 in electronics, see Florida and Kenney (1993), Kummerle (1993a,b), Peters (1991, 1992), Voisey (1992), and Westney (1993). 23. According to Serapio (1994), Japanese companies are under additional pressure to internation- alize their R&D activities as a result of a growing shortage of highly skilled scientific and engineering manpower in Japan. 24. See, for example, National Research Council (1992b), U.S. Congress, Office of Technology Assessment (1993, 1994), Organization for Economic Cooperation and Development (1994). See, also, Figure 3.9. 25. Summarizing the results of a 1992 Mitsubishi Research Institute survey of Japanese compa- nies with overseas R&D facilities, Kummerle (1993b) notes the following: Out of 28 labs which the 7 leading Japanese electronics companies have established abroad during the last 5 years, l5 were built with the intention to create new knowledge in targeted areas, 7 facili- tate local adaptation of products, 3 support complex local production facilities, and 3 were built because of local political pressure. These figures mean that more than half of the new labs serve as facilitators for learning and creating new knowledge abroad. In the pharmaceutical sector, numbers are even more striking: 7 out of 9 labs abroad were established for purposes of local learning and local creation of new knowledge. Moreover, the majority of electronics and pharmaceutical compa- nies expressed an intent to establish at least one more laboratory abroad during the next 5 years.... The research focus of the new 'learning-creating laboratories' is generally on areas with strategic importance for the firm but which are still mastered insufficiently by the company. Out of 5 labora- tor~es that Canon has founded in Europe and the United States since 1988, none is concerned with research in optics: the focus is on either computer languages, image processing software, or tele- communications. This is in line with Canon's intent to shift from an 'Optical Technology Company' to a 'Total Image and Information Processing Company.' 26. Hagedoorn and Schakenraad (1993) attribute the recent growth of corporate interest in techni- cal alliances to three primary motives. First, firms view alliances as a way to strengthen their research capabilities in the face of rapid technological change and the need to monitor and exploit external sources of science and technology. Second, by leveraging the know-how of alliance partners, compa- nies can expand their ability to develop and apply new technologies. Third, alliances enhance firms' access to foreign markets and help them to seek out new business opportunities abroad. 27. Under the National Cooperative Production Amendments of 1993 [P.L. 103-42], which amended the National Cooperative Research Act of 1984 [P.L. 98-462], R&D and production joint ventures located within the United States and registered with the U.S. Department of Justice are exempted from the treble damages liability of U.S. antitrust laws. If any of the firms involved in a joint venture are foreign-owned, the home country of that firm must accord national treatment to U.S. firms with respect to antitrust treatment of similar joint ventures in the country. 28. The Department of Commerce (1991) found that the number of alliances involving biotech- nology firms in the United States increased from 30 in 1981 to 400 in 1988. In 1981, 30 percent of such alliances included a foreign partner and in 1988, 45 percent had one. Peters (1992) notes that during the 1980s, the number of international strategic alliances grew eightfold in the telecommunica- tions industry, around sevenfold in the pharmaceuticals industry, and sixfold in the biotechnology industry. 29. Peters (1993a) cites the example of NEC to show the complementary relationship between the growth of Japanese R&D facilities in the United States and the growth of technical alliances in the computer industry. In 1991, 2 years after establishing its basic research facility in Princeton, New Jersey, "NEC signed a contract with AT&T for a comprehensive package in semiconductor develop- ment for the next 5 years. NEC also entered a joint development agreement with Hewlett Packard to develop tools for microprocessors and microcomputers. Other companies having joint technology agreements with NEC include Grumman (supercomputer), American Microsystems (microprocessors), 3M (optical memory system), Summit Micro Circuit (a venture company to develop high-speed static random access memory chips), Hughes (weather satellite), Adobe Systems (desktop publishing soft- ware), General Electric (international PC network), Tektronix (gate array design software)." (pp. 6-7)

88 FOREIGN PARTICIPATION IN U.S. RESEARCH AND DEVELOPMENT 30. The six industrial sectors were aerospace, computers and peripherals, computer software, semi- conductors, semiconductor manufacturing equipment, and biotechnology. 31. See note 8. Graham (1992) "guesstimates" that roughly two-thirds of the increase in Japa- nese-controlled R&D activity in the United States during the late 1980s resulted from Japanese acqui- sitions of existing U.S. operations. 32. Between 1980 and 1992 the average it&D/sales ratio for manufacturing affiliates increased from 1.6 percent to 2.6 percent (U.S. Department of Commerce, 1993a, 1995a). 33. For further discussion, see Graham and Krugman (1995), Organization for Economic Coop- eration and Development (1994), Ozawa (1991), and U.S. Congress, Office of Technology Assess- ment (1994). 34. A number of factors may explain this variation between European and Japanese firms. Differ- ences in the industrial composition of European and Japanese foreign direct investment may be one determining factor. The fact that direct investment by the Japanese is more recent in most U.S. industries than is that by European firms may also play a role. (Many U.S. affiliates of European- owned companies have added R&D capabilities after having had a U.S. manufacturing presence for decades.) Westney (1993) also points out that Japanese organization and management of affiliate R&D in the electronics industry may have a negative effect on Japanese patenting activity in the United States. Specifically, Japanese electronics firms have chosen to establish large numbers of small highly specialized facilities in the United States, each with a narrow technology mandate yet networked to the parent companies to provide input into technologies developed at the parent firms' facilities in Japan. 35. It should be noted that there is nothing unfair or unethical about foreign companies establish- ing technology scanning capabilities in the United States, or U.S. companies doing the same abroad. Indeed, the rapid pace of technological innovation in many sectors and the high costs of R&D have made it imperative that companies become much more effective at scanning for and acquiring tech- nology developed beyond their institutional borders. 36. Royalties and license fees are payments for the sale and use of intangible property rights, such as patents, copyrights, franchises, trademarks, industrial processes, know-how, and other intellectual property rights. Some observers argue that multinational corporations, U.S. and foreign, frequently use these fees as a form of "transfer pricing," that is, shifting costs from subsidiaries in low-tax countries to a high-tax country in order to minimize tax obligations. (U.S. Department of Commerce, 1993a, p. 73). 37. See, for example, the Council on Competitiveness (1993) case study of the Committee on Foreign Investment in the United States (CFIUS) review of the sale of Semi-Gas System Inc. to the Japanese company, Nippon Sanso KK (pp. 137-155). Dalton and Genther's (1991) survey of foreign acquisition of U.S. electronics companies notes that most acquired companies were small, in need of a capital infusion, and were having difficulty obtaining credit or raising equity. See also the reflec- tions of Materials Research Corporation's CEO Sheldon Weinig (1990) on his company's purchase by Sony. Graham and Krugman (1995) also review several CFIUS cases. 38. The Organization for Economic Cooperation and Development (OECD) has adopted two codes directed at international investment, the Code of Liberalization of Capital Movements and the Code of Liberalization of Current Invisible Operations, which are, in principle, binding on all OECD mem- ber states. In addition to these codes, all member states currently adhere to a "National Treatment Instrument," which obligates them to grant national treatment, with some exceptions, to companies controlled by investors from other OECD countries. The OECD is presently considering a "Wider Investment Instrument" that would, among other things, bolster national treatment provisions. Both the U.S.-Canada Free Trade Agreement (FTA) and the North American Free Trade Agreement (NAFTA) commit (with some exceptions) their respective signatory states to national treatment of enterprises owned by nationals of other signatory states, limit the screening of acquisi- lions by nationals from other signatory states, call for free repatriation of capital and earrungs, restrict recourse to performance requirements on investments that affect trade between signatories, and pro-

PRIVATELY FUNDED R&D 89 vice for dispute settlement mechanisms (the option of binding arbitration). Finally, the Uruguay Round agreement signed in mid-1994 obliges signatories, in principle, to refrain from imposing local- content requirements and export performance requirements on foreign-owned firms operating within their borders. For a detailed assessment of these and related initiatives, see Graham and Krugman (1995). 39 For a more comprehensive review of foreign barriers to international trade and investment, see Office of the U.S. Trade Representative (1995). 40. To get a feel for this debate, see Japan Economic Institute (1991), Krugman (1991), Lawrence (199la,b), and Saxonhouse (1991). 41. It is estimated that intraf~ trade (it l ) trade between the affiliates and parent multinational companies now accounts for more than 40 percent of total world trade. Between 1983 and 1992, U.S.-Europe lL l accounted for 43 percent of all U.S.-European merchandise trade, with U.S.-owned multinationals claiming 43 percent of total U.S.-Europe IFT and European-owned multinationals 57 percent. During the same period U.S.-Japan IFT accounted for 71 percent of total U.S.-Japan mer- chandise trade. However, reflecting a large asymmetry in the volume of U.S. foreign direct invest- ment in Japan compared to Japanese foreign direct investment in the United States, Japanese multina- tionals accounted for 92 percent of total U.S.-Japan lE-l (U.S. Congress, Office of Technology Assessment, 1994). 42. Many U.S. companies enter into alliances with foreign firms in order to achieve greater access to their foreign partner's home market. In some cases, policy-related or structural barriers to access provide the foreign company with additional leverage in its negotiations with its U.S. partner. On the whole, however, transnational corporate alliances appear to have eased the entry of U.S.-owned and U.S.-based companies into foreign markets and national innovation systems. 43. The recent focus within segments of the U.S. trade community on "specific" or "tit-for-tat" reciprocity (with its emphasis on reciprocity in absolute levels of protection) marks a significant departure from the pursuit of "general" reciprocity (i.e., reciprocal changes in the level of protection with a commitment to unconditional Most-Favored-Nation treatment) that has characterized the nego- tiation of multilateral trade agreements during the post-World War II era. The former focuses on outcomes, such as comparable market access, while the latter focuses on the process of liberalization, such as market opening. For further discussion, see Bayard and Elliott (1994), and Cline (1982). 44. See Defense Science Board (1990), Institute for Defense Analysis (1990), Moran (1993), National Defense University (1987), The Analytic Sciences Corporation (1990), and U.S. General Accounting Of lice (199lb, 1994b). 45. In May 1993, the only remaining U.S.-owned supplier of mainstream semiconductor lithogra- phy equipment, SVG Lithography Systems, Inc. (SVGL), announced its intention to enter into a 10- year contract with the Japanese firm Canon. The contract would have given Canon access to all of SVGL's current and future Micrascan scan-and-step technology in return for an infusion of capital and Canon's assistance with manufacturing lithographic steppers. Negotiations between the two Grins became protracted, and ultimately the deal fell through when the SEMATECH consortium members put together an alternative financing package that SVGL accepted . . . Nevertheless, many observers agree that the SEMATECH alternative would not have materialized had Canon not given SVGL its seal of approval by entering into negotiations with the small U.S. company in the first place. Ultimately, it was Canon's reputation and the prospect of Canon backing up the struggling U.S. company with its deep pockets, manufacturing technology, distribution networks, etc., that turned the tide in customer perceptions of the company's viability. See Randazzese (1994) for a well-docu- mented assessment of the SVGL-Canon deal. 46. Aside from the oft-cited sale of sensitive U.S. submarine technology to the Soviets by Toshiba Corporation in the early 1980s, there is only limited anecdotal evidence to suggest that foreign-owned films have leaked sensitive U.S. technologies to potential adversaries (Graham and Krugman, 1991). 47. See, for example, the criteria and procedures put forward in Graham and Krugman (1991, 1995), Moran (1990), and U.S. General Accounting Office (1994b).