summary and findings - princeton university

T he end of the Cold War frees the Nation to turn more ofits energies into building a stronger civilian economy.There are hardships in adjusting to a peacetime footingthat demand national attention, but there are opportuni-

ties to grasp as well.l This report concentrates on new opportuni-ties to advance civilian technologies and improve industrialcompetitiveness. Part One asks how government R&D may beput on a new course, shifting from the military goals thatdominated Federal technology efforts for half a century to agreater emphasis on civilian purposes. Part Two considers someoptions for new national initiatives that meet public needs whilefostering the growth of knowledge-intensive, wealth-creatingindustries.

A key issue in Part One is whether the Nation can put to gooduse on the civilian side research talents and institutions that wereformerly devoted to defense. Many diverse R&D institu-tions—in government, universities, and private defense compa-nies—were part of the defense effort, but this report concentrateson three of the Nation’s largest R&D institutions, the U.S.Department of Energy’s (DOE) multiprogram nuclear weaponslaboratories, Lawrence Livermore, Los Alamos, and Sandia.Public concern is fixed on these labs because they are big, they

1

Summaryand

Findings 1

1 This is the second of two reports by the Office of Technology Assessment on theimplications for the civilian economy of the end of the Cold War. The first was Afrer theCold War: Living With Lower Defense Spending, OTA-ITE-524 (Washington+ DC: U.S.Government Printing Office, February 1992). It considered effects of deep, sustainedcutbacks in defense spending on defense workers, defense-dependent cornmunitics, anddefense companies.

2 I Defense Conversion: Redirecting R&D

are publicly funded, and they face a clear need forchange. They still have important nuclear weap-ons responsibilities, including decommissioning,non-proliferation, and environmental cleanup, aswell as modernizing existing weapons; they donondefense energy work as well. But their centraltask, the design of the Nation’s arsenal of nuclearweapons, is much diminished.

A widely asked question is whether the labsshould take up other tasks in place of weaponsdevelopment. Proposals range from radical down-sizing of the labs, with possible closure of at leastone, to using their resources for new nationalinitiatives devoted to peacetime goals. Whatevertheir longer term future--whether they shrink,take on new missions, or do some of both-amore immediate question is whether the labs canwork effectively with industry. This involves twofurther questions: Do the labs possess technologyand abilities that could be of substantial value toindustry? And if so, can these be made availablewithout too much trouble or delay?

Recent evidence strongly indicates that thelabs’ technology, and their ability to develop newtechnologies, are indeed valuable to industry.Despite earlier disappointments in technologytransfer, industry interest in cooperative cost-shared R&D projects is now at an all-time high,and is matched by interest on the labs’ side. Farmore proposals for cooperative R&D are beingmade than can be funded. The answer to thesecond question is less certain. In early 1993,there were still delays and difficulties in signingagreements, partly because of red tape, but alsopartly because DOE, the labs, and their industrialpartners were blazing new trails in government/industry cooperation. It is not yet clear whetherthe way can be smoothed enough to make theprocess work swiftly and easily, or that it can bedone before the new enthusiasm cools. For thenear term, the issue is whether lab/industrypartnerships can yield concrete benefits for indus-try. A few years’ experience should be enough totell whether good results are coming out of themany projects begun in 1992-93, and whether

industry interest in signing new agreements isholding up.

For the longer term future, R&D partnershipswith industry, per se, are not likely to prove asatisfactory central mission for the weapons labs.As public institutions, the labs’ existence is bestjustified if they serve missions that are primarilypublic in nature. The lab technologies that arecurrently exciting high interest from industry aredrawn from the well of public missions of the pasthalf century, especially nuclear defense. As thedefense task fades, other public missions couldreplenish the well. The labs’ traditional missionsare quite broad, encompassing not only militaryand nonmilitary uses of nuclear energy, but alsobasic high energy physics research and appliedresearch into various forms of energy supply anduse, including their environmental implications.There is also a growing interest in expansion ofthe labs’ public missions into newly definedareas, based on expertise they have developed insuch fields as high performance computing, newmaterials, and advanced manufacturing technolo-gies.

Broad expansion of the labs’ missions, byitself, is often interpreted as an effort to ‘‘save thelabs. Another approach would be for the FederalGovernment to set R&D priorities for selectednational initiatives, and then to allocate fundingto whatever performers, public and private, canmake the best contributions. There are few suchcoordinated Federal R&D initiatives; the bestexample is the High Performance Computing andCommunications Program, which is aimed atwell-defined dual-use goals and involves eightgovernment agencies, including DOE and its labs.Up to now, no Federal agency has had both theresponsibility and the authority to coordinatetechnology development efforts in selected areasof national importance.

Selecting areas of national importance that callfor a substantial infusion of public funds for R&Dinvolves political choices at the highest levels ofgovernment. There is no lack of candidates fornew programs. Some of the most attractive are in

1-Summary and Findings 13

the area of sustainable economic growth, thedevelopment of knowledge- and technology-intensive industries that do not burden the envi-ronment. Energy efficiency is almost always acritical element in environmentally benign indus-trial growth.

Part Two of this report opens a discussion ofbroad new initiatives the Nation might adopt toserve peacetime goals. The illustrative case cho-sen for analysis is that of transportation systemsthat offer greater energy efficiency, reducedpollution, and lesser dependence on foreignoil-all public benefits that could justify publicinvestment. The systems include cleaner cars,powered by electric batteries or a combination offuel cells and batteries; intelligent vehicle andhighway systems; and high-speed mass groundtransportation systems, including steel-wheel traincars on rails, such as France’s TGV (Train aGrande Vitesse), and magnetically levitated vehi-cles on guideways.

Without attempting to analyze all the transpor-tation policy issues involved, the discussion herelooks at the systems from a defense conversionperspective. It concentrates on the benefits theseenvironmentally attractive systems might offer inthe way of advancing critical technologies, pro-moting world-class industries, and creating goodjobs—benefits that defense spending often pro-vided in the past—plus their potential for usinghuman talents and institutions formerly devotedto defense. The analysis suggests that nonpol-luting cars, though farther from technologicalsuccess than high-speed ground transportationsystems, hold greater promise for pushing techno-logical frontiers and could, if they succeed, createlarger numbers of well-paid productive jobs inAmerica. There may be other good reasons,however, for government support of the high-speed ground systems.

However desirable they may be, it is not likelythat any of these systems would create nearlyenough jobs at the right time and in the rightplaces to compensate for the hundreds of thou-sands of defense jobs being lost as the Nation

adjusts to post-Cold War military budgets. Someof the initiatives could use the talents of peoplenow working in the defense sector-especiallyresearch scientists and engineers-but the matchwould not be perfect.

This is the second of two Office of TechnologyAssessment (OTA) reports on the implications forthe U.S. civilian economy of the end of the ColdWar. The greatest effects, of course, are relieffrom the threat of global nuclear war and thefreedom to pursue national goals other thanmilitary security. Nevertheless, adjustment todeep sustained cuts in defense spending is notsimple or painless. The first report of thisassessment, After the Cold War: Living WithLower Defense Spending, considered effects ofthe cutbacks on defense workers, men and womenin the armed services, defense-dependent com-munities, and defense companies. It concludedthat there would be hardships-greater perhapsthan the relative size of the cutback suggests,because our economy is burdened with more debtand higher unemployment than in times past, andis under much greater challenge from foreigncompetitors. First aid to affected workers andcommunities, in the form of reemployment,retraining, and redevelopment assistance, canhelp them through the transition. But the bestconversion strategy is a broad one: investment inprograms that train workers well, help businessesperform better, promote technology advance, andinvigorate local and national economic growth.

BACKGROUND

The 1990s are uncharted territory. For the firsttime in half a century, the United States faces nomassive military threat from a superpower foe.Instead, the major challenge is to keep up with theeconomic competition from friendly countries.Some are doing better than we are in industriesthat disproportionately advance knowledge, gen-erate new technologies of wide application, andsupport rising living standards.


This Nation’s success in reaching a peacefulconclusion to 40 years of Cold War will bringsustained cuts in defense spending; that, ironi-cally, threatens to handicap us in rising to newchallenges in the economic realm. Military spend-ing should and will continue to decline. Yetmilitary spending and the military-industrial com-plex are concentrated strongly in things thatincrease our potential for growth-research anddevelopment, technology and knowledge inten-sity. In fact, military spending has sometimesbeen described as America’s de facto technologyand industry policy. If so, it is a blunt instrumentof policy; it is an unfocused and expensive way ofadvancing important commercial technologies.Nevertheless, there is enough commonality inmilitary and commercial applications of somecritical core technologies that defense spendingover the years has strongly supported both. It hasproduced semiconductor chips of various kindsthat find uses in autos and engineering workstations as well as guided missiles; programmablemachine tools that can make parts for fighteraircraft or lawn mowers, tractors, and commercialairliners; computational techniques that modelnuclear explosions or analyze what happens tocars in crashes.

This report focuses on one element of militaryspending that has greatly benefited the U.S.civilian economy—sustained, generous fundingfor research and development. Of course, R&D isnot the only benefit defense spending has be-stowed. Having the Department of Defense (DoD)as a large, reliable first customer for groundbreak-ing new technologies was at least as important; itwas the combination of defense R&D and defensepurchases that launched the semiconductor andcomputer industries. Moreover, R&D is far fromthe whole story in industrial competitiveness.

Figure 1-1—R&D Spending as a Percentage of GDP:United States, Germany, and Japan, 1971-90

Percentage of GDP3

2.7

/

2.4 I

2.1

41.8

IA-————— , , /1971 73 75 77 79 81 83 85 87 89

● United States 1 Germany ● J a p a n

SOURCE: National Science Board, The Competitive Strength of U.S.Industrial Science and Technology: Strategic Issues (Washington, DC:1992), table A-9.

Many other factors are at least as important.Among them are a Nation’s financial environ-ment, whether hospitable or not to long-termprivate investments in technology and productionequipment; training and education of managers,engineers, and shop floor workers; and manage-ment of people, equipment. and the organizationof work to produce well-designed, reliable goodsat reasonable prices.2 Neglect of R&D was not themain reason for one U.S. industry after another tofall behind our best competitors in the 1970s and1980s. Much more important were inattention tothe tasks of improving quality and efficiency,linking design and production, and getting newproducts to market quickly.

Nevertheless, R&D is an essential element inthe mix, and it has been a traditional source ofstrength for the U.S. economy. Today, Americanpreeminence in R&D is fading. By the late 1980s,

2 OTA reports over the past dozen years have analyzed the international competitiveness of U.S. industries, pointed to problems, andsuggested policy options for improving the Nation’s performan W. Recent studies include U.S.-Mexico Trade: Pulling Together or PullingApart (October 1992); Competing Economies: America, Europe, andthe Pac@cRim (October 1991); Worker Training: Competing in the NewInternational Economy (September 1990); Mah”ng Things Better: Competing in Manufacturing (Febnmry 1990); Paying the Bill:Manufacturing and Ametica’s Trade Deficit (June 1988); Commercializing High-Temperature Supercomiuctivity (June 1988); andInternational Competition in Services: Banking, Building, Sojlware, Know-How (July 1987).

1-Summary and Findings 15

Figure 1-2—Nondefense R&D Expenditures: UnitedStates, Germany, and Japan, 1971-90

1992 dollars (billions)120 ~ - -1

I

o~1971 73 75 77 79 81 83 85 87 89

United States ● Germany * Japan


Japan, West Germany, and Sweden all spent ahigher proportion of gross domestic product ontotal R&D than the United States. As for nonde-fense R&D, those nations devoted 2.6 to 3.1percent of Gross Domestic Product (GDP) to thepurpose in 1990, compared with 1.9 percent in theUnited States (figures 1-1 and 1-2). Moreover, theU.S. position is deteriorating. While foreigncountries have stepped up the pace of their R&Dspending in recent years, this Nation’s has stag-nated. In the United States, total and industry-funded R&D hit high points in 1989, haveremained essentially flat in constant dollars since,and have dropped as a percentage of GDP.Government R&D has declined in constantdollars, mostly due to defense cutbacks (figures1-3 and 1-4).

The reasons for the current lackluster R&Drecord in the United States reflect several factors.Declines in military R&D have certainly affectedthe government’s R&D spending and probablyindustry’s as well (figure 1-5). The recession and

Figure 1-3—Nondefense R&D as a Percentageof GDP: United States, Germany, and Japan,

1971-90

Percentage of GDP4 ‘-– ‘ –—- “ —— ‘ —- --r

‘ - 1I

3.5 !

31 9

1 I II 11971 73 75 77 79 81 83 85 87 89

. United States ~ Germany ● J a p a n


sluggish recovery of the early 1990s may havedampened industry’s R&D spending; this hap-pened in the recessions of the 1970s, although notin the turndown of 1981 -82.3 Corporations areburdened with more debt today than in earliertimes when industry’s R&D spending was risingsteadily. Some American companies that weretraditionally the flagship R&D performers ofprivate industry have recently suffered stunning,unprecedented losses. Even innovative compa-nies are now more ready than heretofore toabandon R&D in areas where they see foreigncompetitors ahead of them. Leading corporatelabs that formerly undertook large-scale, long-term R&D projects and produced such innova-tions as the transistor, have been scaled back,broken up, or sold.

Government policy has a variety of options fordirectly encouraging more R&D by private indus-try, but there is also a good case for governmentsharing with industry some of the large risks and

3 possibly, ~5 ~= because defense spend~g was rising so fast during this period that defense COmptitifX were Cofildent R&D ~vestments

would pay off later in large military procurements.


180

160

140

120

100

80

60

40

20

0

Figure 1-4-U.S. R&D Spending by Source of Funding, 1960-92

1992 dollars (millions) — — .

1960 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990

D Federal Government ~ private industry _ Other

SOURCE: National Science Foundation, National Patterns of R&D Resources: 1992 (Washington, DC: 1992), table B-3.

50

40

30

20

Figure 1-5—Federal Budget for Defenseand Nondefense R&D, 1955-93

1992 dollars (billions)

I

10

vIo ! r I I I ~ I I I ~ ~ 1 1 1 1 1 1 1 , 1 1 1 I 1

1955 58 61 64 67 70 73 76 79 82 85 88 91

. Defense R&D t Nondefense R&D

SOURCE: National Science Foundation, National Patterns of R&DResources: 7992 (Washington, DC: 1992), table B-21; NationalScience Foundation, unpublished data.

Figure 1-6-Federal R&D Funds byBudget Function, 1992

8 %

Energy4%

Generalsciencer

4%w“ Space

11%

Health13%

SOURCE: NationaJ Science Board, Sd-andt%grneenng /ndk.akxs-1991 (Washington, DC: U.S. Government Printing Office, 1991), table4-17.

1--Summary and Findings 17

Figure 1-7—R&D for National Defense as aPercentage of Total Federal R&D, 1970-92

30 J I

1970 72 74 76 78 80 82 84 86 88 90 92

SOURCE: National Science Foundation, National Patterns of R&D:1992 (Washington, DC: 1992).

high costs involved in today’s leading edge R&D.Most other advanced Nations do this as a matterof course. There is increasing evidence to showthat, in competition with foreign firms whosegovernments share the costs of developing tech-nologies, American firms are handicapped. Andthe financial environment in the United States hasfor a long time been less friendly than that of ourbest competitors-especially Japan—for long-term private investments in technology develop-ment and equipment.4

The Nation does not inevitably have to lose thebenefits of government supported R&D as de-fense spending declines. The Federal Govern-ment pays for 43 percent of the Nation’s R&Dspending, most of it for defense purposes; somecould be redirected from military to economicgoals. Opportunities to do that are present in

DOE’s nuclear weapons laboratories but they arecertainly not the only candidates. Assuming thatsome former defense R&D spending is rechan-neled to civilian-oriented R&D (instead of beingapplied to many other worthy purposes, fromFederal debt reduction to improved health care),other claimants for public R&D funds includeuniversities, private research laboratories, andcivilian government R&D institutions. The DOEweapons labs have human and physical resourcesthat they are eager to redeploy into dual-use orcivilian efforts, but conversion of defense re-sources is only one consideration in deciding howbest to put public funds into R&D partnershipswith industry.

THE STRUCTURE OF FEDERAL R&DThe U.S. Government is a major force in the

Nation’s research and development, and defensedominates the government’s share. In 1992, theFederal Government spent $68.2 billion overallfor R&D out of a national total of $157.4 billion;$41.5 billion of the Federal share was defense-related. 5 Health is a distant second to defense inFederal R&D, followed by civilian space andaeronautics, energy, and scientific research(figure 1-6). At times in the past, defense has beeneven more dominant, reaching a recent peak of69 percent of Federal R&D in the mid-1980s(figure 1-7).

The leading performers of federally fundedR&D are private companies, which account for 45percent of the total.6 Eighty percent of their workis for DoD, and the National Aeronautics andSpace Administration (NASA) occupies most ofthe rest. Universities and colleges, which receive

d For discussion of the reasons and principles for government-industry collaboration in developing technologies with commercial promise,see U.S. Congress, Office of Technology Assessment Mak”ng Things Better, O’IA-ITE-443 (Washington DC: U.S. Government PrintingOffice, February 1993) ch. 2; and Competing Economies, OTE-ITE-498 (l%shingto% DC: U.S. Government Printing Office, October 1991),ch. 2; also, John Alic, Lewis Branscomb, Harvey Brooks, Ashford Carter, and Gemld Epstein, Beyond Spinofl: Mi/itary and CommercialTechnologies in a Changing World (BostoG MA: Harvard Business School Press, 1992), ch. 12.

5 National Science Foundation, National Patterns ofRiW Resources: 1992, by J.E. Jankowski, Jr., NSF 92-330 (Washington DC: 1992),tables B-3 and B-21.

b National Scienee Foundation, Selected Data on Federal Funds for Research and Development: Fiscal Years 1990,1991, and 1992, NSF92-319 (Washington, DC: July 1992), table 9.


Table l-l—R&D by Selected Government Agencies and Laboratories, FY 1992 (millions of dollars)

Department/Agency Total R&D Total Lab Intramural FFRDCs

Department of Defense . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Department of Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .National Aeronautics and Space Administration . . . . . . . . .Department of Health and Human Services . . . . . . . . . . . .

National Institutes of Health . . . . . . . . . . . . . . . . . . . . . . .Department of Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . .Department of Commerce . . . . . . . . . . . . . . . . . . . . . . . . . . .

National Institute of Standards and Technology. . . . . . .National Oceanic and Atmospheric Administration . . . . .

Department of the Interior . . . . . . . . . . . . . . . . . . . . . . . . . . .National Science Foundation . . . . . . . . . . . . . . . . . . . . . . . .

$38,7706,4998,5439,7818,2531,256

539186337562

2,102

$11,5964,6983,4992,0391,559

826431144272482211

$9,890449

2,6131,9661,486

826431144272479

89

$1,7074,249

8867473

●

o003

123

“Indicates amount less than $50,000.KEY: Federally Funded Research and Development Centers.

SOURCE: National Science Foundation, Federal Funds for Research and Development: Fiscal Years 1990, 1991, 1992, Volume XL, NSF92i322(Washington, DC: 1992), table C-O..

15 percent of the government’s R&D budget, areless defense-dependent. They are the biggestperformers in the areas of health and generalscience, with a substantial presence as well indefense, energy, and agriculture.

Laboratories owned or principally funded bythe Federal Government receive 35 percent ofFederal R&D funds. Their growth and strengthare largely a phenomenon of post-World-War-IIyears, and their work reflects the Nation’s priori-ties during that period. About half the $25 billionthey received in 1992 went for defense, withaerospace, energy, health, and agriculture sharingmuch of the rest (table l-l).

In considering how to redirect R&D resourcesfrom military purposes to strengthening the civil-ian economy, this report concentrates on thegovernment’s own research institutions. Althoughtwo-thirds of defense R&D dollars are spent inprivate industry, public policy has a stronger andmore direct influence on the conduct of govern-ment R&D than on how private firms managetheir laboratories and research teams. (Box 1-Abriefly describes some public policies related totechnology conversion by defense companies).The report therefore focuses on governmentlaboratories that, up to now, have put most of theireffort toward military goals.

I Federal LaboratoriesThe often-quoted figure of “more than 700”

Federal laboratories summons up a rather mis-leading picture of a national network of largewell-equipped research centers. In fact, the Fed-eral research, development, testing, and evalua-tion (RDT&E) system includes everything fromsingle offices staffed by a handful of people tosprawling weapons testing centers like the FlightTest Center at Edwards Air Force Base inCalifornia, or large campuses with thousands ofresearchers, such as the National Institutes ofHealth (NIH) in Bethesda, Maryland. SomeFederal labs are owned by the government andmanaged and staffed by Civil Service employees(government-owned, government-operated, orGOGOs), like the labs of the National Institute ofStandards and Technology (NIST) and most DoDlabs. Some, including many of the biggest, aregovernment-owned but operated by universities,companies, or non-profit institutions acting ascontractors (GOCOs); these include all nine of theDOE multiprogram labs and NASA’s biggest lab,the Jet Propulsion Laboratory. Some are ownedby other institutions but do virtually all of theirwork for the government (Federally FundedResearch and Development Centers, or FFRDCs)like the Lincoln Laboratory at the Massachusetts


Box l-A—Conversion of Military Technologies by Defense Companies

Among private defense companies there is no lack of military technologies that might be adapted for use incommercial products. Some major companies, in fact, have taken steps to reorient a portion of their R&D towardcivilian applications. For example, Westinghouse Electronics Systems, TRW, Martin Marietta, and LockheedElectronics are using information, data processing, and remote sensing technologies of military origin for suchcivilian uses as air and highway traffic control systems, drug interdiction, and office security systems. l Althoughmost of the customers so far have been civilian government agencies, and sales are small compared to defensecontracts of the recent past, opportunities for converting technologies are certainly there and could be sizable.Nevertheless, there are serious barriers to technology conversion by private firms. The barriers are not so muchat the technical or engineering level, but rather at the broader level of how the company operates.

Many studies and reports have called attention to the gulf in company culture and management practicebetween defense and commercial firms.2 During 45 years of Cold War, most large defense companies and defensedivisions of diversified corporations withdrew from commercial markets into what has been termed the “defenseghetto.” The reasons are several. Defense contractors that make complex weapons systems or major subsystemsare geared to producing at low volume while meeting very exacting demands for technical performance. Bycontrast, the emphasis in the commercial world is on high-volume production that combines product reliability withaffordable cost. And while some U.S. commercial industries have fallen behind their best foreign competitors ingetting new generations of products to market quickly, they are years ahead of defense industries. The time fromdesign to production for military systems is often 15 to 20 years, compared to 3 to 5 years for many commercialitems. Furthermore, major defense companies typically have little acquaintance with commercial marketing anddistribution. DoD prime contractors have very few buyers to deal with and no need for a distribution network.

Department of Defense (DoD) requirements are another major source of division between commercial anddefense companies. DoD often imposes rigid, detailed specifications and standards, not only for the product itselfbut also for the process of manufacture. These “roil specs” and “roil standards” have blocked technologicalprogress for defense applications in fast-moving fields such as electronics, and have locked into defense contracts

technologies that commercial companies no longer produce. Even more important are the government’s specialauditing, review, and reporting requirements for defense contractors, which are intended to guard against wasteand fraud but which also impose heavy extra costs. A leading reason why many companies keep defense andcommercial work separate is to avoid burdening the commercial business with overhead from the defense side.

It is therefore hard for defense contractors to combine their defense business with commercial production,

or to change from one to the other. Technology conversion, per se, might not be such a formidable challenge. Butif defense companies are to adapt their military-generated technologies to civilian use, they must make themselves

into civilian, or at least dual-use, companies. This is no small task.

Despite the difficulties, some defense companies are making the attempt. Besides the major defense

contractors who are dipping a toe into t he water of commercial markets, there are many smaller companies who

1 Forfljrtherdiscussion of theoutfookforand experience of conversion by defense COmpaniOS, see U.S. Con9ress,Office of Technology Assessment, Afler the Cold War: Living WWJ Lowr Defense Spending, OTA4TE-524 (Washington,DC: US. Government Printing Office, February 1992), ch. 7.

2 See, for ~xample, u-s, ~ngress, ~fi~ of Te~no@y Assessment, ~ol~ln~ the Edge: hfahta.hh~ the fMfer)Se

Tbchno/ogy Base, OTA-ISG420 (Washington, DC: US. Government Printing Office, April 1989); Integrating CornmercMand Military Technologies for Nationa/ Strength: An Agenda for Change, report of the CSIS Steering Committee onSecurity and Technology (Washington, DC: The Center for Strategic & International Studies, 1991); John A. Alic, et al.,Beyond Spinoff: Military and Commercial %chnologies in a Changing 14hld (Boston: Harvard Business School Press,1992).


Box 1-A--Continued

see their only salvation in the civilian economy. Some are getting help from State programs. For example,Connecticut the State that tops the list in economic dependence on the private defense industry, providesconverting firms with various forms of financial aid, including both conventional low-interest loans andsuccess-dependent investments in new product development, to be repaid in royalties. Even wit h help, these firmsface years of effort and uncertain prospects.3

The Federal Government has very broad interests, both military and civilian, in encouraging defense firmsto convert to more civilian production and to integrate the military and civilian sides of their business. Most of theFederal programs are framed to promote the development of dual-use technologies and integrated companies.Efforts to raise the share of DoD purchases off the shelf from commercial vendors are at least 20 years old,4 butthe incentive to do so today is far stronger today as defense budgets tighten. The same motive is pushing Federalpolicymakers toward removing some of the burdens of military accounting requirements? Moreover, new laws andpolicies already allow defense companies to recover more of their own R&D expenses--for dual-use as well asstrictly military technologies-as allowable overhead on government contracts.6

These changes should help to breach the walls of the defense ghetto and support a more effective, efficientdefense industrial base. However, defense contractors still face the need to find more commercial business or elseshrink, or possibly perish. At least one Federal program is explicitly directed at helping defense-dependentcompanies enter the commercial marketplace with dual-use products. The $1.7 billion defense conversionpackage that Congress passed in 1992 includes a $97 million Defense Dual-Use Assistance Extension Program.It provides cost-shared grants to centers sponsored by Federal, State, or local governments that offer defensefirms technical assistance in developing, producing, and marketing dual use products; it also provides forgovernment-guaranteed loans to small defense businesses. For the most part, however, Congress took a broaderview of defense conversion and threw open to all firms-whether or not they are defense-dependent--new orenlarged technology development and diffusion programs. Two of the new programs, each funded at $97 millionfor fiscal year 1993, are a manufacturing extension program supporting State and local agencies that help smallfirms adopt best practice technologies, and a regional technology alliance program, which concentrates onapplications-oriented R&D for locally clustered industries. In addition, several hundred millions of dollars wereprovided for government-industry R&D partnerships to develop critical dual use technologies

In sum, the issue of technology conversion by defense companies quickly turns into broader policy areas.From the standpoint of military interests and requirements, civil-military integration is highly desirable; but it is notclear whether that can be achieved better by trying to turn defense firms into dual-use companies, or by formingR&D partnerships with commercial companies for defense needs (as ARPA does, see ch. 5 of the full report) andby changing DoD’s acquisition policies to allow more purchases from companies whose essential nature iscommercial. From the standpoint of the nation’s economic performance, a very broad definition of conversionseems most desirable. This implies a policy approach that offers transition assistance to defense companiesstruggling to survive in the commercial world while opening technology diffusion and development opportunitiesto all companies equally.

3 ~ev~ pmbw, “timpanies struggle to Adjust As U.S. Cuts Military Budget,” The N8w Yofk mm-, Feb. 10,1993.

4 U.S. bngre~, offb of Technology Assessment, &JMlhg Future Sscurfty: &rateg/es bf R6SfIUCfUd17~ fheDefense Tbchno/ogyand/n&stria/ Base, OTA-ISC-530 (Washington, DC: U.S. Government Prlntlng Office, June 1992),pp. 99-103.

5 In January 1w3 the Advisory Panel on Streamlining and Codifying Acquisition 1.awSSubmitt8d to DoD ib reporton reforming the body of aoquisitkm law; at this writing the Department had not yet responded.

6‘rhig is in~petint research and development, or IR&D, an important source of funding for defense ~mpanies’development of teohnologles with no speolfic weapons application. IR&D Is destined to beoome less important asprocurement deofines, since it is reoovered from government oontracts.

1--Summary and Findings I 11

Institute of Technology, sponsored by the AirForce.

It is also sometimes mistakenly assumed thatall the Federal labs have an untapped potential forcontributing to the Nation’s economic perform-ance, but that is an exaggeration. Some alreadyhave longstanding close relations with industry.Examples are NIST’s labs, which have a centralmission of serving industry’s needs; the NASAaeronautics labs, with their history and explicitmission of R&D support for the aircraft industry,civil as well as military; and the NIH labs, withsubstantial research that is of immediate interestto the pharmaceutical, medical devices, andbiotechnology industries. No doubt some of theselaboratories could improve their links with indus-try, but they are not starting from zero.

DoD has the biggest budget of any Federalagency for its laboratories--$11.6 billion in1992; this includes not only R&D laboratories perse but also testing and evaluation (T&E) centers,such as the Air Force’s Arnold EngineeringCenter in Tennessee and the Navy’s WeaponsCenter at China Lake, California. Less than halfof DoD’s total budget for the labs is spentin-house; the rest is passed through to outsideperformers, mostly defense contractors.7 Withfew exceptions (e.g., the science-oriented multi-program Naval Research Laboratory), the De-fense Department’s R&D labs pass through wellover half of their budgets while the T&E centersspend more than half in-house.8 overall, more

than $5 billion was available for in-house RDT&Ein DoD facilities in 1992.

The next biggest spender was DOE, with $4.7billion. 9 In contrast with the DoD labs, most of thefunding DOE provides its labs is spent in-house,and in fact is supplemented by about $1 billionfrom other Federal agencies, mostly DoD. DOElabs also differ from most DoD labs (and mostother Federal labs as well) in that most areGOCOs.

For this report, with its focus on redirectinggovernment R&D resources from military tocommercial or dual-use applications, DOE nu-clear weapons labs and DoD labs are mostrelevant. The former are of prime interest, forseveral reasons. The term ‘‘weapons labs” usu-ally refers to Los Alamos and Lawrence Liver-more, which design nuclear warheads, and San-dia, which develops field-ready weapons usingthe warheads. These labs are in a class bythemselves. Their collective budgets were over$3.4 billion in 1993, and together they had over24,000 employees.

10 Nuclear weapons-related

activities accounted for 51 to 60 percent of theiroperating budgets (least for Lawrence Livermore,most for Los Alamos); if the labs’ work for theDoD is added in, funding for military-relatedactivities ranged from 67 percent at LawrenceLivermore to 78 percent at Sandia. However, agrowing share of activities funded by the nuclearweapons accounts is not, properly speaking,military. Nonetheless, funding for the labs from

7 Department of Defense In-House RDT&E Activities for Fiscal Year 1990, prepared for the OtXce of the Secretary of Defense, Ofilce ofthe Deputy Director of Defense, Research and Engineering/Science and Technology (Rkshingtonj DC: The Pentagoq n.d.). This documentreports spending for total and in-house RDT&E activities in 91 Army, Navy, and Air Force facilities, employing about 100,000 civilian andmilitary personnel. Spending for the total RDT&E program was $8.4 billioq with $3.9 billion (46 percent) spent in-house in f~cal year 1990.These figures are not exactly comparable with R&D data collected by the National Science Foundation. They are mostly limited to RDT&Eactivities where funding for in-house RDT&E is at least 25 percent of the in-house portion of the facility’s budgec they do not include spendingin FFRDCS. See also Michael E. Davey, ‘‘Defense Laboratories: Proposals for Closure and ConsolidatiorL’ Congressional Research Service,The Library of Congress, Jan. 24, 1991, p. CRS-6.

8 Ibid. ~ 1990, tie R&D labs spent $24 bfllion of heir tot~ $5.8 bfiion RDT&E budget in-house (41 percent); the T&E CtXlkXS Sj)ellt $1.6

billion of $2.7 billion (59 percent) in-house.

g Note Mt ~ese figures are only for R&D performed in government-owned, - operated, or -funded labs. DoD’s total 1992 budget authorityfor R&D, excluding expenditures for R&D plant and e@pmen4 was about $38.8 billion. DOE’s was $6.5 billion.

10 ~s ~u~ o~y re~a employees. On-site contract employees amount to many more. IrI 1993, Sandia’s 8,450 regular employees Waesupplemented by 2,000 on-site contract employees; Los Alamos, with about 7,600 regular employees, had some 4,000 on-site contractors.


the nuclear weapons accounts rose in FYs 1992and 1993 (in constant dollars, taking inflation intoaccount), but this growth was largely due to bigincreases for a massive environmental cleanupjob, plus rising amounts for non-proliferationwork, decommissioning existing weapons, andsafety and security of the remaining nuclearstockpile, all of which are funded by the nuclearweapons accounts.

The fact is that the nuclear weapons labs arelooking at a future that is very different from theirpast. Their mission of nuclear weapons design isfading; in 1993, no new nuclear weapons werebeing designed. Among Federal R&D institu-tions, the nuclear weapons labs face the clearestneed to change with the end of the Cold War.

1 The DOE Laboratory ComplexDOE’s laboratory complex consists of the nine

multiprogram laboratories (including the weaponlabs) that are usually called the national labs, pluseight single-program energy labs.1l They arefunded by six program areas: Defense Programsand related nuclear weapons offices, which in-cludes work in all aspects of nuclear weaponsdesign, safekeeping, non-proliferation, and envi-ronmental restoration of the damage from 50years of weapons work; Energy Research, whichsupports fundamental scientific research; theNuclear Energy, Fossil Energy, and Conservationand Renewable Energy Programs, which concen-trate on applied energy R&D; and the Environ-

mental Restoration and Waste Management Pro-gram.

In 1992, the weapons labs got over one-half ofthe funding for all the labs in the DOE complex.The biggest part of their funding comes fromDOE’s atomic energy defense weapons account(including Defense Programs and related nuclearweapons offices); DoD contributes an additional,though declining, share (figures 1-8, 1-9, and1-10). These labs have fluctuated in size over thelast two decades. In the early 1970s as theVietnam War wound down, their budgets werecut substantially (in constant dollars). With thenew emphasis on energy supply and conservationprograms in the Carter years, the weapons labsdiversified into more nondefense work; both theirenergy and defense funding rose. Then in themilitary buildup of the 1980s, nuclear and nonnu-clear defense work grew rapidly,12 pushing theweapons labs’ budgets up 58 percent from 1979to 1992 (in constant dollars), while the energylabs’ funding rose 15 percent (figure 1-11).13 Thebudgets for the three labs combined continued toclimb through 1993, when their funding wasalmost two and one-half times what it was at thelow point in 1974 (figure 1-12). Only LawrenceLivermore took a substantial cut in 1993; fundingfor Sandia and Los Alamos continued to rise.

Although details of the FY 1994 budget werenot yet available as this report was completed,cutbacks were probably in store for the weaponslabs as well as the rest of the defense establish-

] 1 me nu~r of DOE labs differs as counted by various sources. If small, specialized labs are included, the number cm be m @I@ m 29.The figure of 17 comes from Secretary of Energy Advisory Board, A Report to the Secretary on the Department of Energy NationalLaboratories (mimeo), July 1992. The other national labs are the six energy multiprogram laboratories: Argome National Laboratory,Brookhaven National Laboratory, Idaho National Engineering Laboratory, Lawrence Berkeley Laboratory, Oak Ridge National Laboratory,and the Pacific Northwest Laboratory. DOE’s eight single-program laboratories include: Ames Laboratory, Continuous Electron BeamAccelerator Facility, Fermi National Accelerator Laboratory, National Renewable Energy Laboratory (formerly the Solar Energy ResearchInstitute), Princeton Plasma Physics Laboratory, Stanford Linear Accelerator Center, Stanford Synchrotrons Radiation Laboratory, and theSuperconducting Super Collider Laboratory.

12 Much of tie non-nucl~ defense work was for the Strategic Defense ~~ve.

13 U.S. Dep~ment of Enm~, Unp”bfished dam from tie ~titutio~ P1- Da@~e, US DOE ST-31 1, These CdCUkitiOIIS include the

Idaho National Engineering Laboratory (INEL) among the energy labs. INEL is sometimes categorized separately as a “nuclear energy”laboratory because its work is concentrated largely in producing nuclear materials (mostly for weapons) and handling nuclear wastes. Argonne,Brookhaveu Lawrence Berkeley, Oak Ridge, and Pacific Northwest National Laboratories are considered “energy research” laboratories.Excluding INEL, the total funding for the energy research labs rose about 10 percent horn 1979 to 1992.


Figure 1-8-Nuclear Weapons and DoD Funding for Sandia National Laboratories

—-

1978 1980

I 1 1

1982 1984 1986 1988 1992

~ Nuclear weapons B DoD funding = Other

NOTES: Operating budget only. DoD funding data not available prior to 1977.

SOURCE: Sandia National Laboratories.

Figure 1-9—NucIear Weapons and DoD Funding for Lawrence Livermore National Laboratory

1993 dollars (billions). .

1972 1974 1976

1 I

1978

~- ~ Nuclear weapons

1980 1982 1984 1986

n DoD funding = Other

1 I 1 1 I 1 r i I 1 IJ

1988 1990 1992


SOURCE: Lawrence Livermore National Laboratory.


Figure 1-10—Nuclear Weapons and DoD Funding for Los Alamos National Laboratory

1993 dollars (billions)—

1.2

1

0.8

0.6

0.4

0.2

0 b 1 1

1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992

m Nuclear weapons ~ DoD funding _ Other


SOURCE: Los Alamos National Laboratory.

Figure 1-1 l—Funding for DOE Multiprogram Laboratories in 1979 and 1992 (in millions of 1992 dollars)

623

623

851

1979

Los Alamos

LawrenceLivermore

Sandia

472 IdahoNational

214PacificNorthwest. . . — — .

213 LawrenceBerkeley

228 Brookhaven——

435 I Argonne

—.———.

538

1

Oakridge

1011

1023

1277

1992

Los Alamos

LawrenceLivermore

Sandia

614

363

214

273

391

551

IdahoNational

PacificNorthwest

LawrenceBerkeley

Brookhaven

Argonne

Oakridge

Weapons labs

NOTE: Operating budget only.

Energy labs Weapons labs Energy labs

SOURCE: U.S. Department of Energy, DOEMu/t@rograrn Laboratories: 1979 to 1988A Decade of Change (Washington, DC: Sedember, 19921:U.S. Department of Energy; Los Alamos National Laboratory, Lawrence Livermore National Laboratory, Sandia National Laboratories. “ “


Figure 1-1 2-Combined Funding for LawrenceLivermore, Los Alamos, and Sandia

National Laboratories, 1970-93

1993 dollars (billions)4

I II I

3

2

“ = _ - = ”

1

1970 72 74 76 78 80 82 84 86 88 90 92

NOTE: Operating budget only.

SOURCE: Lawrence Livermore National Laboratory, Los AlamosNational Laboratory, and Sandia National Laboratories.

ment. In any case, further changes in directionappeared certain. Announcing a new technologyinitiative in February 1993, President Clinton andVice-President Gore committed the Administra-tion to altering the mix of government R&Dsupport; the share for civilian technologies wouldbe lifted from 41 percent in 1993 to over 50percent by 1998, they said.14 While emphasizingthe part to be played by a strengthened Depart-ment of Commerce, they also promised a reviewof all laboratories managed by DOE, NASA, andDoD ‘‘that can make a productive contribution tothe civilian economy,’ with the aim of devotingat least 10 to 20 percent of their budgets to R&Dpartnerships with industry.

DISPOSITION OF THE DOEWEAPONS LABORATORIES

The end of the Cold War has raised persistentquestions about the future of the weapons labora-tories. First, what if anything do the labs have tooffer beyond their traditional work in nuclear and

nonnuclear defense-in particular, what do theyhave to offer that is truly valuable to civilianindustry and national competitiveness? Second,assuming the labs have outstanding capacities intechnologies of importance to industry, howreadily available are these capacities? Can thelabs work in partnership with private companieswithout crippling delays or red tape? Finally,assuming private industry can get reasonableaccess to valuable capacities in the labs, how dothese partnerships fit into a national technologystrategy? What place does cooperative government/industry R&D in large expensive national labora-tories have in a broader scheme for technologydevelopment and diffusion that will help U.S.industries keep up with the world’s ablest compe-titors? Answers to these questions are not easy,and some can come only as the fruit of severalyears’ experience.

1 Opportunities for Technology TransferThe human talents and physical equipment in

the three weapons labs are often described asamong the Nation’s freest. A central question iswhether these resources fit with the needs ofindustry. Some skeptics have doubted that tech-nologies dedicated to the exotic demands ofnuclear warhead and weapon design could be ofany use to civilian industry, but this view is toonarrow. It is not in the final weapons system itselfthat synergies with commercial needs are mostlikely to occur, but rather in core competencies,technologies and production processes. Box 1-Bsummarizes the core competencies claimed byeach of the three weapons labs (see ch. 4 for moredetail).

In a report on industry relations with theFederal labs (mainly DOE labs), the private sectorCouncil on Competitiveness concluded that thereis clearly “extensive overlap between industryneeds and laboratory capabilities. ” Citing aninformal poll of several of its member companies,

1.$ ~e~ident W~~J4 Clhton and vic~~~ident ~~fl @re, Jr., Techno[ogyforAmerica ‘S ECOrWmI”C Growth, A New Direction to Build

Economic Strength, Feb. 22, 1993.


Box l-B-Core Competencies of DOE’sNuclear Weapons Labs

Lawrence Livermore National LaboratoryMeasurements and diagnosticsComputational science and engineeringLasers, optics, electro-opticsManufacturing engineeringElectronic systemsEngineered materials Applied physics and chemistryAtmospheric and geosciencesDefense sciencesBioscience

Alamos National LaboratoryNuclear technologiesHigh-performance computing and modelingDynamic experimentation and diagnosticsSystems engineering and rapid prototypingAdvanced materials and processingBeam technologiesTheory & complex systems

Sandia National LaboratoriesEngineered materials and processesComputational simulations and high-performance

computingMicroelectronics and photonicsPhysical simulation and engineering sciencesPulsed power

SOURCE: Lawrence Livermore, Los Alamos, and Sandia NationalLaboratories, 1993.

the Council said that industry rated severaltechnologies as major technical areas in whichthey need assistance.

15 The technologies included

advanced materials and processing, advancedcomputing, environmental technologies, and man-ufacturing processes, testing, and equipment. Thelabs specified these same areas as ones in whichthey have unique capabilities that could helpindustry. Three out of four of these areas have

contributed to andweapons program

been supported by the nuclearfor decades, and the fourth,

environmental technologies, is now a prominentpart of the program.

Examples of synergies are numerous, espe-cially in computer modeling and simulation. Allthree weapons labs have demonstrated mastery inhigh performance computing. They were the firstcustomers of early supercomputers and wereclose collaborators in developing both hardwareand software (the relation between Los Alamosand Cray Research was especially close). Theyare still leaders today as early purchasers andcontributors to the design of massively parallelmachines and software. Applications of comput-ing power developed in the labs for weaponspurposes have already found many civilian usesand have the potential for many more. Forexample, computer codes developed to model theeffects of nuclear explosions have been adapted tomodel crash dynamics and are widely used in theauto industry.

In addition, each of the labs has distinctiveassets. One of Lawrence Livermore’s particularstrengths is in laser technology. Sandia, with itsexperience in engineering weapons that containnuclear warheads, has special facilities and expe-rience in advanced manufacturing technologies,in particular for semiconductors. Sandia’s Com-bustion Research Facility at Livermore, Califor-nia, is a magnet for university, industry, and otherweapons lab researchers in a variety of fields,including ‘‘lean-burn’ combustion of hydrocar-bons in auto engines. Los Alamos has tradition-ally concentrated on basic scientific research; itsmeson physics laboratory attracts university andother laboratory researchers, and it is a center forthe development of complexity theory in mathe-

15 council on competitivene55, Zndwtv as a Customer of the Federal Laboratories (Washingto% Dc: co~cil on Competitiveness\

September 1992), p. 10.16 DoE$s enaw ~seuch ~bs ~50 ~ve some di5fictive fac~hties and asse~ of interest to industry. For example, IBM hM ~d

Brookhaven’s synchrotrons storage ring as a source of x-rays for advanced lithography technology for semiconductors, and several companiesuse Oak Ridge’s High Temperature Materials Facility for development of advanced ceramics.


matics. All three labs are leaders in developingadvanced materials.16

Behind the specific technologies in which thelaboratories excel are their human resources andtheir experience with state-of-the-art equipment.Leaders at the labs claim unique capacities to takeon large-scale projects where science makes adifference, engineering is also required, andteamwork is essential; the multidisciplinary ap-proach is ingrained in the labs, they say, Recog-nizing the contribution of universities, especiallyin scientific research and in training new genera-tions of researchers, they see the labs as havingthe additional capacity to marshal the people andspend the time required for tackling big, long-term problems. And they believe their ability toconcentrate on the long term is a distinctiveaddition to privately funded industrial R&D,which generally has a shorter term focus—especially since some of the Nation’s leadingcorporate labs have been scaled back or dis-banded. The DOE labs’ role can be seen asintermediary between the universities, the sourceof most basic research, and industry, which turnsnew technologies into commercial products andprocesses. Their best contribution may be theability to carry scientific concepts into large-scaledemonstration projects. (Figure 1-13 schemati-cally represents the roles of universities, industry,and the DOE labs in various aspects of R& D.)

Assuming that the labs do have technologicalresources of potential value to industry, thereremains the question of whether they can worksuccessfully with industrial partners to transfertechnology to the commercial realm. Until the1990s, most of the evidence suggested that theanswer was no. A few Federal agencies and theirlabs have long worked effectively with the privatesector, but most—including DOE--+oncentratedon their public missions and gave relatively littleattention to technology transfer. Despite urging

Figure l-13-Capabilities in SemiconductorTechnology

SOURCE: Los Alamos National Laboratory.

from various commissions and internal evalua-tions, despite several laws in the 1980s pushingtechnology transfer, there was not a great deal toshow for it.

Since 1989, the picture has changed, withseveral significant developments. First, the Na-tional Competitiveness Technology Transfer Act(NCTTA) of 1989 allowed the contractor-operated DOE labs, for the first time, to signcooperative research and development agree-ments (CRADAs) with industry.17 Although itwas possible for the labs to undertake cooperativeprojects before, and some had done so, CRADAshave some significant advantages, including clear-cut legislative authority and the ability to protectintellectual property generated in the projects foras long as 5 years. Cooperative projects with thelabs often have a good deal more appeal forindustry than simply licensing existing technol-ogy, because so much of what the labs have tooffer needs extensive development before it isuseful to commercial firms.

17 G- labS ~d b~~ @ven tie ~u~onty t. sign c~As ~ 1986, in me Feder~ Tw~oIogy Tr~sfer At of 1986, and ~ecutive Order

12591, issued by President Ronald Reagan in April, directed Federal agencies to delegate to GOGO lab directors authority to negotiate termsof CRADAS.


Second, by 1992, top officials of the Adminis-tration as well as Congress were actively pushingtechnology transfer from Federal R&D programsand labs. DOE claimed technology transfer as a“formal, integrated mission” of all its labs, withthe primary goal of ‘assisting U.S. based compa-nies in the global race for competitive technolo-gies. ’18 In February 1992, President GeorgeBush launched a National Technology Initiative,with 15 conferences around the country at which10 Federal agencies19 invited industry to makecommercial use of government-sponsored re-search.

Interest on the part of industry has been un-precedented—a third major factor. No doubt thiswas partly because the power and prestige of thePresident and his cabinet officers were nowbehind the program. At the same time, many inU.S. industry had come to recognize that theyneeded the government as a partner in R&D,especially for high-risk, long-term, expensiveprojects.

Fourth, there is a new pot of money forcooperative R&D projects—at least for the DOEweapons labs and for Defense Programs (DP) inthe energy labs. NCTTA and subsequent legisla-tion encouraged the labs to build cooperativeprojects with industry into their R&D programs tothe maximurn extent practicable,20 and to set agoal of devoting 10 percent of their DP funds tocooperative agreements.

21 But to give the CRADA

process a jump-start, Congress also directed that$20 million of Defense Programs’ R&D funds infiscal year 1991 be explicitly set aside for

cooperative projects with industry; the sum wasraised to $50 million in 1992 and at least $141million in 1993.22

Finally, the labs themselves now have a power-ful motive for making technology transfer acentral mission. During the 1980s, while Con-gress was urging this mission on the labs, it wasat the same time providing steep rises in fundingfor both nuclear and nonnuclear defense work.Little wonder that the weapons labs, which sawtheir nuclear weapons and DoD funding swell bymore than half in the 1980s, should redouble theirconcentration on their historic defense missionand that a new mission of working with industryon commercially promising technologies shouldbe relatively neglected. The end of the Cold Warand the dissolution of the Soviet Union hasupended these priorities. Although some in thelabs still believe they will get the biggest part ofa shrinking defense pie, many of the labs’managers and researchers know their defenseresponsibilities must decline.

This combination of factors means that now,for the first time, there is broad, significantinterest in lab/industry partnerships. Evidencecan be seen in the fact that in July 1992 there were1,175 CRADAs joining private companies andFederal laboratories, compared with 33 in 1987.By November 1992, DOE’s CRADAs numbered292.23 It is noteworthy too that for every CRADAsigned with DOE weapons labs there are manymore that did not make the cut. The competitionfor getting CRADAs approved and funded is nowkeen.

18 U.S. Department of Energy, “The U.S. Department of Energy and Technology Transfer, ” mimeo, n.d.

19 p~cipa~g agencies incl~d~ the r)ep~ments of Commerce, Energy, Transpo~tioG Defense, hterior, Agriculture, and Health and

Human Services as well as NASA, the Environmental protection Agency, and the White House ~lce of Science and Technology Policy,20 me D~e~e Autho~tiOn Act for Fiscal y-s 1992 and 1993, sec. 3136 (enacted in 1991).

21 U.S. Semtej co~ttm on tied Semices, National D#en~e Authorization ~t for F&al year 1993: Report, report 102-352, to

accompany S. 3114.22 rbid. A@ me c~ton A&s&ationpropo=d in ~ch 1992 @ @ aside anadditio~ $47 million from DPR&D fid.!i fOr COO~rative

projects; a set-aside of $47 million from other DOE programs was also proposed.n ~s fiWe ~cludes w DOE labs, not the weapons labs alone. Data provided to O’E4 by the U.S. Depa@ent of Enmgy.

-—

1 Roadblocks to Technology TransferDespite the unprecedented interest in coopera-

tive lab/industry projects, the process of gettingagreements actually signed got off to a very slowstart. In some cases, lags were due to unfamiliarity—on industry’s side as well as DOE’s—and somewas due to bureaucratic foot dragging at DOEheadquarters. It took well over a year for DOE toput in place some of the basic procedures forsigning CRADAs. From 1989, when DOE’snational labs gained authority to sign CRADAs,to early 1991, only 15 CRADAs were signed.Since then the pace has picked up, with close to300 agreements signed by 1993 and the time fornegotiations becoming shorter. Even so, some ofthe many companies keenly interested in the labs’technological offerings were still expressing im-patience with the time and expense involved.Possibly, the windows to cooperative R&D thatwere opened so recently might close if thedifficulties are not soon solved.

REASONS FOR DELAYIn early 1993, it still took 6 to 8 months or more

to nail down most individual CRADAs--startingwith the submission of a proposal, which itselfmay have taken many months to develop in talksbetween lab and industry researchers. Much of thedelay is laid at the door of DOE headquarterscontrol, though some also occurs at the labs andat DOE field offices; company legal counsels arealso named as sources of delay. The progress ofCRADAs at DOE labs is often compared unfavor-ably (but not altogether fairly) with the process atother Federal labs—in particular NIST labs,whose parent agency, the Department of Com-merce, has delegated most of the authority to sign


CRADAs to lab directors. NIST agreements areoften out the door in a few weeks. Some in theprivate sector have strongly advocated givingboth authority and money for CRADAs to the labdirectors, with DOE exercising control throughevaluations of the labs’ performance and budgetsfor subsequent years.24

This solution is possible and might well speedup the process, but it is not as simple as it mayseem. First, the legal authority for CRADAs inGOCO labs (e.g., the DOE labs) is quite differentfrom that in GOGOs (e.g., NIST labs and mostDoD and NASA labs25). NCTTA requires that theparent agency must approve every joint workstatement (the first step in preparing a CRADA)from GOCOs as well as the CRADA itself; underthe Federal Technology Transfer Act of 1986,GOGO labs may go ahead with a CRADA so longas the parent agency does not disapprove within30 days. This difference in the laws reflects afairly common attitude in Congress that someGOCO contractors, laboratory directors, and re-searchers are less reliably committed to publicpurposes than the government employees whostaff GOGOS.26 Congressional oversight cover-ing details of lab operations is seen as partlyresponsible for DOE’s close management ofmany of the labs’ doings, including CRADAs.

Other factors-probably still more important—are size and visibility. DOE’s national labs,especially the weapons labs, are far larger thanmost other labs in the Federal system, theirCRADAs involve much more money, and theyget much more scrutiny. DOE feels obliged to beabove reproach on issues such as fairness ofopportunity for companies wishing to work withlabs and requirements that jobs resulting from

M see, for emple, Comcil on competitiveness, IrtAS~ as u customer of the Federa/ Luborutoties (w-tom N: September 1992).25 one ~jor NASA lab, me Jet ~p~sion ~boratory at tie c~~omia ~stitute of Technology, is a GOCO, but in my case NASA labs

do not use CRADAS. They have their own legal authority to make cooperative agreements with industry under the 1958 Space Act, and havelong done so.

26 Those holding MS view do make some distinctions among GOCO contractors ad the ~bs ~ey manage; some are seen as more responsiveto public purposes than others. One contractor that has received little criticism is Sandia Corporation, a subsidiary of AT&Tj which has managedSandia National Laboratories for $1 per year since 1949. However, AT&T announced in 1992 that it would not renew the Sandia Corporationcontract the following year. AT&T’s long stewardship of Sandia comes to an end in September 1993.


lab/industry R&D partnerships stay in the UnitedStates.

Finally, some delay is inherent in the systemDefense Programs at DOE headquarters hasdevised to exercise guidance over a cooperativeR&D program that has grown to substantial size.By far the largest sum of money available forDOE CRADAs is in Defense Programs, in theset-aside from the atomic weapons RDT&Eaccount for cooperative agreements and technol-ogy transfer. The set-aside was $141 million infiscal year 1993 and was planned to rise to $250million by 1995. Most of the projects DP fundscome from the three weapons labs, since they arethe leading performers of atomic weapons R&D,but several of the energy labs also have some DPfunding.

DP managers believe that strategic direction isessential in a program of this size, and that itshould be a coherent part of multilab initiatives todevelop dual-use technologies. As of 1992, DPmanagers planned to fund initiatives in semicon-ductor lithography, flat panel displays, a broadarray of automotive and transport technologies,and advanced materials and ceramics. Severaltimes a year, DP issues a call for proposals fromthe labs and potential industry partners for R&Din these areas.

27 Dp then reviews the proposals in

two steps (see ch. 4 for details); the purpose of thereview process is to minimize unnecessary dupli-cation and encourage complementarily.

All of this precedes the preparation of a jointwork statement and CRADA that, by law, DOEmust review. The agency has formally delegatedto DOE field offices responsibility for these twofinal reviews, which can take up to 120 days, but

in practice has shrunk to less than 90 in mostcases.28 DP aims to keep the time from the formalsubmission of a lab/industry proposal to approvalof the work statement and CRADA to no morethan 6 months, and eventually reduce it to 4months.29 This goal had not been met by early1993.

The time for negotiating CRADAs will proba-bly decrease as everyone becomes more familiarwith the exercise; it was already somewhatshorter in 1992-93 than a year or two earlier.There were still delays at several points in thesystem, however; and there is some inherentdelay in a system that aims for strategic direction,coherence, and selection on merit among compet-ing proposals.

FUNDING BOTTLENECKSUp to now, the DP set-aside has been the source

of nearly 70 percent of DOE’s funds for CRA-DAs. Another option is to use program funds,rather than tapping into a special set-aside. IndeedCongress has urged DOE to use this route, writinginto law that the labs are to use all their weaponsR&D funding to the “maximum extent practica-ble” for cooperative agreements and other formsof technology transfer, and using committeereport language to suggest that at least 10 percentbe devoted to the purpose.30 At present, this ishard to do. At the beginning of each budget cycle,DOE and the labs establish how they will spendtheir program funds and allocate lab budgets toindividual projects. Afterwards, it may be diffi-cult for lab project leaders to adjust the focus orscope of project work to accommodate theinterests of a potential industrial partner. Aproject that has been significantly redefined needs

27 ~ere may be only one cd for proposals in f~cid yerU 1993.

28 ~cor~g t. tic ~w, ~~ review of tie jo~t work statement must be Completed in 90 days, and review of the CRADA in 30 &JY3.

Although the labs have proposed submitting both documents at once, and keeping the time to 90 days, some of the field offkes have taken theposition that the review periods should be sequential. However, in practice, nearly all the reviews have been completed within the 90 days.

29 AS noted, ~S Wholeprwess com~ on top of tie time tit tie ~b ~d company r~~chem take to define the work they want to do tOgethtX.

The same is true of other Federal labs, such as NIST; the CRADA approval process starts after much prelhinary work has been done by theresearchers involved.

~ r)epartment of Defe~ Authorization Act for Fiscal Years 1992 ~d 1993, Public hw 102-190, sm. 3136.


the approval of lab managers and DOE headquar-ters.

In DOE programs outside DP, funding forCRADAs has been meager. For example, GeneralMotors held a “garage show” at its technicalcenter in Warren, Michigan, in January 1992 toacquaint hundreds of company engineers andscientists with technologies available at DOElabs. The meeting was a success, with enthusiasmon both sides. The upshot was that GM research-ers identified over 200 interesting cooperativeprospects, afterwards winnowed to 25 formalproposals. About half of these proposed to use DPfacilities, and were eligible for funding from theDP set-aside. The other half were submitted tovarious energy programs; only 2 received fund-ing, compared with 14 submitted to DP. Accord-ing to GM, this was because money outside DPwas lacking.

The DP set-aside is not a bottomless well. In itsJune 1992 call for proposals, DP received 398first-round submissions, requesting $170 millionin first year funding from DOE; these were laterwinnowed to 184, requesting $79 million.31

Eventually, 61 were funded with first-year fund-ing of $40 million (matched by an equal amountfrom the industry participants). In November1992, a call for proposals for a still smaller pot ofDOE money—about $25 million-drew hun-dreds of proposals. Even if the DP set-aside wereraised to $250 million a year, many proposalswould fail to make the cut.

LEGAL BOTTLENECKS

Just as there is a genuine tension between thegoal of fast action on CRADAs and that ofcoherent, strategic direction of cooperative tech-nology development, so there are some realconflicts regarding legal agreements between thelabs and industry. One source of disagreement isprotection of intellectual property.

The public interest in allowing private compa-nies rights to intellectual property developed inpart at taxpayer expense has been recognized in aseries of laws, starting with the Stevenson-Wydler Act of 1980. Companies that put theirown money into cooperative R&D with govern-ment labs are interested in exclusive rights toresulting inventions.32 If they can’t get thoserights, at least for some period, they are not likelyto find much appeal in the project. On the otherhand, government also has an interest in broaddiffusion of new technologies, especially thosepartly funded by public funds.33

NCTTA allows the labs to protect intellectualproperty generated in a CRADA for up to 5 years,and further exempts from the Freedom of Infor-mation Act any intellectual property companiesbring to the CRADA (thus protecting againstdiscovery by competitors). Although industrywelcomed the changes under NCTAA, somepotential industry partners still consider theprotection of intellectual property insufficient,especially for software. However, some in gov-ernment foresee trouble down the road if thebalance tips too far, and intellectual propertydeveloped in part at the expense of the taxpayeris held too tightly by CRADA partners. DOE does

s 1 Full m~tiyew tifig requested was $778 million for all the CRADA proposals submitted, and $3%? million for the winnowed list. ~ese

numbers represent DOE’s share, to be matched by industxy.

32 Subjat, tit is, t. the government’s royalty-free use of the invention for is own pvses.

33 me U.S. patent system, which protects intellectual property and rewards inventors with exclusive rights for a nar of years, also hassome positive technology diffusion effect.s in its requirement for disclosure of the technical workings of the patented device or process. Althoughothers camot freely copy the patented device, they may be able to invent around i~ i.e., devise another version with help from the disclosure.NC’ITA not only provides patent rights to CRADA partners, but also protection for another form of intellectual property, or proprietaryinformation that is not patented. Data that is generated by industrial partners in CRADAS may be kept free from disclosure under the Freedomof Information Act for up to 5 years. In some industries (e.g., computer software) protection of data is more important than patent rights.


not take a direct hand in negotiations overintellectual property in CRADAs or other cooper-ative agreements; it assigns the rights from labactivities to the contractors who operate the labs,and the terms are largely up to the labs and theirindustrial partners, within the general limits set bythe law. Nevertheless, DOE can if it wishesexercise some oversight over the labs’ handling ofintellectual property rights, and the issue remainsa live one for public policy.

An attempt to compromise and settle the prob-lem for a whole industry was part of the umbrellaCRADA for manufacturing process technologiessigned between four DOE labs (the three weaponslabs and Oak Ridge) and the National Center forManufacturing Sciences (NCMS) on behalf ofitself and member companies. The CRADA gaveNCMS exclusive rights to license commercialapplications in fields covered by the project’s taskstatement for 30 months after project completion.The terms are similar to those used by NIH andare somewhat more generous to industry thanthose of NIST, two agencies generally consideredsuccessful in transferring technology from gov-ernment lab to industry. However, the agreementis coming unraveled. Some NCMS membercompanies are dissatisfied with the terms; inparticular, they want to widen the field of use(breadth of application) to which their intellectualproperty rights apply. In another industry, com-puter systems companies are insistent on protect-ing the source code for software developed in labpartnerships; without this protection, they argue,their investment in the software will gain themnothing.

There is no simple or obvious solution to theproblem of balance in disposing of intellectualproperty rights. It is not just in DOE labs thatthese rights can become a thorny issue. They areoften a sticking point in negotiations with otherlabs as well, including NASA and NIST. The

problems are considerably less when the indus-trial partners to cooperative agreements are mem-bers of consortia, and the technologies beingdeveloped are considered generic or pre-competitive.

A second field of conflict is the issue of U.S.preference. A central goal of R&D partnershipsbetween government and industry is to improveU.S. competitiveness and thus promote economicgrowth and rising standards of living. Accord-ingly, there is a strong public interest in seeingthat publicly financed innovations are used inways that directly benefit the U.S. economy. TheFederal Technology Transfer Act of 1986, whichauthorized GOGO labs to sign CRADAs withindustrial partners, directed the labs to ‘‘givepreference to business units located in the UnitedStates which agree that products embodyinginventions made under the [CRADA] will bemanufactured substantially in the UnitedStates.’ ’34 Taking its cue from this law, DOEwrote into its model CRADA a requirement, notjust a preference, for U.S. manufacture.

The realities of international ties betweenbusinesses have forced departures from thisrequirement. The first major exception was in theumbrella CRADA with the Computer SystemsPolicy Project (CSPP), which represents U.S.computer systems manufacturers; in this CRADAthe requirement was rewritten to cover R&D only,not manufacture. CSPP insisted that existingnetworks of manufacturing, R&D, and cross-licensing among computer companies of allnationalities made the requirement for domesticmanufacture impossible. Other companies subse-quently began to demand the same terms and inFebruary 1993 DOE modified its stance, saying itwould consider case-by-case exceptions wheresubstantial U.S. manufacture is shown not to befeasible, and where industrial partners committhemselves under contract to appropriate alternate

3415 us-c, 371q~)(4)@)o35 Me~~~~~d~frOmu.S, De,p~entofEnqyto progr~s~e~~icers mdFi&J offiwlvla.nagers, ‘Restatement of Departmental

Technology Transfer Policy on U.S. Competitiveness, ” Feb. 10, 1993.

.——

benefits to the U.S. economy.35 The general rule

remains to demand U.S. preference; if industrialpartners ask for exceptions they bear the burdenof showing in detail why it should not apply.

This probably does not settle the matter.Controversy seems bound to arise when a tech-nology developed under a CRADA yields asuccessful commercial product that is manufac-tured abroad, possibly by a foreign company.Whenever foreign companies exploit an Ameri-can technology in a high-tech field, there arethose who regard this as a failure of public policy,and the condemnation is likely to be still strongerif the technology was developed in part withpublic money. This view, though understandable,is simplistic.

First, it has always been hard to stop thediffusion of technology, even 200 years ago at thedawn of the industrial age. Today, with rapidcommunication and increasing worldwide tradeand investment, the tendency toward technologydiffusion is far stronger and to a great extent isbeyond the control of governments. Second, andless well-known, is the fact that U.S. fins’ abilityto use access to technology as a bargaining chipin negotiations with foreign firms and govern-ments can be a powerful advantage. That advan-tage can, in the end, work to the benefit of the U.S.economy and standard of living. For example, theability of General Electric’s Aircraft Enginedivision to sell jet engines to European airlinesmay well hinge on adding some value in Europe,and that in turn may mean licensing some of GE’stechnology to a European partner. The Europeancompany gets some of the manufacturing workand some of the know-how, but the Europeansales also create good jobs and technologyadvance in the United States.

The issue of U.S. preference does not simplypose a private interest against a public one, Twoconflicting public interests are also involved: thebenefits of government/industry R&D partner-ships on terms industry finds useful vs. thebenefits of keeping manufacturing jobs at home.


One more major difficulty has bedeviled DOE’sCRADA negotiations: who is liable in casesomeone sues for injury from a commercialproduct based on technology developed under theCRADA? DOE’S initial answer, contained in itsfirst model CRADA, was that the industrialpartner must reimburse the lab or government forany damages awarded; in other words, the com-pany bears all the liability, no matter who is atfault. So many companies found these termsunacceptable that DOE changed its position, andits policy guidelines now exempt the industrialpartner from liability when the damages are dueto the negligence of the lab.

The new formula is not entirely satisfactory toindustry. In case of a suit, it may be difficult forthe partners to sort out responsibility for damages.DOE is considering whether it might be simplerto leave out any reference to liability in CRADAsand let the courts determine who is at fault. Thisissue is probably best seen as part of the largerproduct liability problem that plagues some ofAmerica’s industrial sectors, and is most likely tofind satisfactory solution as part of a broaderresolution.

H The Longer Term Future of theWeapons Laboratories

The discussion so far has assumed that the labswill continue to exist in recognizable form,though they may change in goals, emphasis, orsize. However, many people are asking morefundamental questions about the labs. The DOEweapons labs had their origin in the atomicweapons program of World War II, and after-wards expanded their goals, first to peaceful usesof nuclear energy, then to energy supply and usemore broadly, including the environmental con-sequences of both. More than at any time sincethey were created, insistent questions are arisingabout what national purposes the labs ought toserve and what size and shape is appropriate tothose purposes. Assuming, for the sake Of argu-ment, that the labs have exceptional capacities to


work in harness with industry to advance com-mercially promising technologies, and that theycan work out effective ways of doing so, are theyalso reasonably efficient institutions for thepurpose? What part do they have in a coherentU.S. Government technology policy?

Three divergent points of view have begun toemerge. First, drastically shrink and restructurethe whole DOE laboratory system, perhaps givingthe job to a commission like the military baseclosing commission. Second, maintain and rein-force the labs’ traditional focus on nuclear andenergy technologies. Third, give the weapons labsmajor new civilian missions, including bothpartnerships with industry and new or enlargedprograms directed to public purposes (e.g., envi-ronmental protection). Although there are over-laps in these differing positions, they do representthree distinct evaluations of the labs’ potential.

SHRINK THE DOE LABORATORY SYSTEMThere is little written or formal expression of

this point of view, but some in Congress (espe-cially in committees concerned with governmentoperations) and in the university/industry re-search community put it forward quite forcefullyinformally. They are dubious that DOE labs havea useful place in developing commercial ordual-use technologies-or perhaps even in theirtraditional fields of energy and nuclear power,except for a much circumscribed weapons mis-sion. The criticisms are twofold. First, the weap-ons labs are too imbued with the culture ofnational security and a reward system that pro-motes weapons experts to fit in the civilian world.Second, the labs and the contractors who operatethem are not held properly accountable for theiruse of public funds, and use the money ineffi-ciently.

The first objection might perhaps fade if theweapons labs were to show in a few years’ trial

that they can in fact work productively withindustry. The second is more difficult. Histori-cally, the labs’ parent agencies (DOE and itspredecessors) have given the contractors anddirectors of the labs an unusually free hand inmanagement. On the other hand, the labs havebeen subjected to a good deal of congressionalscrutiny on management issues. It is outside thescope of this report to evaluate the prudence orefficiency of the labs’ management (or of any oneof them; very likely there is a range, with somebetter managed than others).36 Nevertheless it iscertainly true that for their national defense workthe labs have been showered with funds andequipment as few other government institutionshave been. This largesse may have contributed tohabits of inefficiency. If the weapons lab budgetsdecline significantly—as they had not yet done asof fiscal year 1993--financial stringency mightforce greater efficiencies. It is useful to remem-ber, however, that the government’s historicgenerosity and flexibility in funding for the DOElabs have contributed to what is generally thoughtto be their core strengths: multidisciplinary teamsof high professional caliber combined with su-perb leading edge equipment.

REINFORCE THE LABS’ FOCUS ON NUCLEARAND OTHER ENERGY TECHNOLOGIES

Those who occupy this middle ground regardthe DOE national labs as treasures worth preserv-ing, but consider that several of the labs have lostfocus and should reconcentrate their efforts in thetraditional fields of nuclear power and energy,with their environmental ramifications. Theseviews were stated recently by the Secretary ofEnergy Advisory Board (SEAB) Task Force,appointed by Secretary of Energy James D.Watkins in November 1990 to advise him on “a

36 ~~ ~epm respn~g t. tie expre5S~ interests of the requesting congressional committees and keep~g ~ fid 0~’s

technology-oriented missiom concentrates on the potential technological contributions of the DOE weapons labs to the civilian economy.Analysis of complex management and accounting issues related to the labs is outside the scope of OTA’S assessment.


strategic vision for the National Laboratories . . .to guide [them] over the next 20 years. ’ ’37

The future laid out by the Task Force woulddefine these major missions for the DOE labs:energy and energy-related science and technol-ogy, nuclear science and technology for defenseand civilian purposes, and the fundamental sci-ence and technology that underlie these. For theweapons labs, the Task Force recommended atight focus on nuclear defense (including non-proliferation, verification, and arms control) withwhatever reductions and consolidation are neces-sary in an era of overall reduction of the Nation’sdefense effort. Major new responsibilities forenvironmental cleanup and waste managementwere included, however, for both the weapons andenergy labs. Cooperative work with industry wona cautious endorsement. The Task Force sug-gested that a few flagship labs be designated ascenters of excellence for technology partnershipswith industry, selecting technologies consistentwith their particular missions and devoting asmuch as 20 percent of their R&D budgets tocost-shared projects.

ASSIGN NEW CIVILIAN MISSIONS TOTHE WEAPONS LABS

This approach for more thoroughgoing changehas several versions. One suggestion, proposedby Rep. George E. Brown, Jr., Chairman of theHouse Committee on Science, Space, and Tech-nology, would radically restructure the three bigweapons labs. It would consolidate nuclear weap-

ons design and non-proliferation work at LosAlamos; put verification activities at Sandia andcontinue its responsibilities for engineering thenonnuclear components of nuclear weapons, whilealso making it a center of excellence for technol-ogy transfer; and make Lawrence Livermore acivilian National Critical Technologies Labora-tory, building on the lab’s strengths in materialsscience, computational science, fusion, environ-mental remediation, and biotechnology .38 Brownalso proposed cutting the nuclear weapons RDT&Ebudget from about $2.7 billion a year to half thatlevel over 4 years, and using all the savings forcivilian technology programs in the DOE labsystem. Another suggestion, coming from severalsources, was to devote from 10 to 20 percent, ormore, of the labs’ budgets to cooperative projectswith industry .39

Both these plans would put into the DOE labsan unprecedented amount of money for cost-shared development of dual-use and commercialtechnologies—possibly $500 million to morethan $1 billion a year, depending on the labs’ totalbudgets, with more than half coming from theweapons labs. Compare this with the AdvancedTechnology Program (ATP), operated by NIST,which has the general mission of supportingcommercially promising R&D and awards cost-shared government funding to industry projects,selected on a competitive basis.40 ATP is theclosest thing to a civilian technology agency that

37 ,$ecretq of Enerw Advisory Board Task Force, Fina/ Report, July 1$)92, at~chen~ Memorandum for the chairman and Executive

Director, .%crctary of Energy Advisory Board, from the Secretary of Energy, James D. Watkins, Nov. 9, 1990.38 ~ttcr t. he HO~Orabl~ James D, Watis, Secrew, U.S. Depmrn@ of Energy, from George E. Bm~ Jr., (hlhlll an, U.S. House of

Representatives, Committee on Science, Space, and Technology, Feb. 8, 1992.39 See, forexmple, Council on Competitiveness, ItiUSq US a Customer ojthe Federa/ Luboraton’es (Wstigton, DC; 1992). me COUC1l

is sometimes confused with two other groups with similar names: the President’s Council on Competitiveness, a government interagencycomrnitlcc made up of Cabinet members and chaired by Vice-President Dan Quayle under the Bush Administration; and the CompetitivenessPolicy Council, an independent advisory committee created by Congress and composed of Federal and State officials as well as private sectormcmbers.

40 u~~c tic cooperative ~ctiviti~s at DOE and o~er government labs, &e A’rp program simply provides cost-shared funding for R&D

performed by the industrial partners.

331-050 - 93 - 2 : QL 3


now exists in the Federal Government.41 Its initialfunding in fiscal year 1990 was $10 million; 4years later, in 1993, its funding was $68 million.

The possibility of a sudden infusion of a muchlarger pot of government money for cooperativeR&D than ever before raises several importantquestions. One is whether a lab mission broadlydefined as ‘‘economic competitiveness’ is work-able. Some top officials at the labs fear that suchan imprecise definition of their responsibilitycould lead the labs to scatter their efforts andbecome nothing but job shops for industry. Aparticular strength of the billion-dollar weaponslabs is their depth and versatility, but even theselabs need to focus on technologies that fit theircore competencies best.

A different approach would be to assign to thelabs responsibilities for new missions that areclearly public in their goals and benefits, but alsohave the potential to replace defense activities asgenerators of technology, good jobs, and wealth-creating industries. Although the definition of“public missions ‘‘ is not fixed and immutable,there is general agreement on certain areas inwhich technological progress is important forhuman welfare, but is not likely to attractadequate private R&D investment because it doesnot promise individual companies enough profitto compensate for the risks. Some obviouscandidates are the large, various, and growingfield of environmental cleanup and pollutionprevention; a nationwide communications “su-perhighway;’ revitalized education and trainingthat take full, imaginative advantage of computeraids and networks; and energy-efficient transpor-tation systems that offer the public benefits ofreduced environmental damage and less depend-ence on foreign oil (for more discussion, see chs.7 and 8 and this chapter, below). Public missionscould also encompass such things as support of

advanced manufacturing technologies-an areaof relative neglect for U.S. public and privateinvestment.

It seems unlikely that any one new nationalmission can attract the generous, sustained levelof funding that nuclear defense has received for50 years, but it is possible that some combinationof missions might be sufficient to keep the labs inthe frost rank of R&D institutions, able to drawexcellent researchers and do outstanding scien-tific and technical work.

A question that immediately follows is hownew national missions might be assigned to theDOE weapons labs. The primary national interestis in the substance of the missions themselves,and there are certainly public and private R&Dinstitutions other than the weapons labs—including industry and universities-that couldshare some of the tasks. Other agencies and theirlabs also have abilities that overlap with certainstrengths of the weapons labs. Although someoverlap in R&D is desirable, money and effortcould be wasted if there is no interagencycoordination or strategic planning. A coherent,rational R&D plan for a big new national initia-tive in areas such as environmental cleanup or lesspolluting transportation systems would set clear,concrete goals, milestones, and measures ofperformance, and would parcel out work towhichever government agencies are most fit for it,as well as enlisting university and industrycollaboration. In fields of most interest to indus-try, such as advanced manufacturing technolo-gies, industry guidance and cost-sharing would beessential.

Although coherent planning is unusual ingovernment-supported R&D, there is a precedentin the High Performance Computing and Com-munications Program (HPCCP). The program’sgoal is “to accelerate significantly the commer-

41 As ~ot~, ~~er agencies ~ve R&D pmg~ tit yield res~ts of -t benefit to VariOUS industries, e.g., MI-I, NASA, the Department

of Agriculture. But with the exception of NIST’S manufacturing engineering and standards and measurements labs, Federal agency R&D isdirected toward specific public missions (e.g., health) or to particular industrial sectors identified as important to public purposes (e.g., aircraft,space, agricuhure)--not to commercial goals or competitiveness generally.


cial availability and utilization of the next genera-tion of high performance computers and net-works’ ’42 and allow the private sector to leapfrogover improvements in supercomputers and net-works that would otherwise be gradual andincremental. While HPCCP has encountered somecriticism, it generally gets high marks both fromparticipating agencies and from industry observ-ers. Some planning for other Federal technologyprograms (e.g., advanced materials and process-ing, biotechnology, advanced manufacturing R&D,new energy technologies) is taking place but is inearly stages compared with HPCCP, and theplanning process is laborious.

I Alternative R&D InstitutionsAssuming that the DOE weapons labs achieve

smooth working relationships with industrialR&D partners, are they too big, too expensive andtoo encumbered by their nuclear weapons historyto serve the purpose efficiently? Some havesuggested that a more useful kind of institutionmight be relatively modest regional centers withan unequivocal mission of doing applications-oriented R&D partially funded by industrialclients. Another model is ARPA. This small,free-wheeling DoD agency has a stellar record ofadvancing high-risk high-payoff technologies—not only in strictly military systems such as smartweapons and stealth aircraft, but also in dual-usecore technologies, including microelectronics andcomputer hardware, software, and networks. ARPAdoes virtually no lab work of its own, but usescontracts, grants, and cooperative funding forR&D in private companies and universities.

THE FRAUNHOFER MODELGermany’s Fraunhofer Society (Fraunhofer

Gesellschaft, or FhG) has been proposed as a

model for small-scale R&D institutions workingin harness with industry. It is the smallest butprobably best known and most admired of Ger-many four major publicly funded researchinstitutions, which are managed and funded byBMFT, the science and technology agency. TheFhG consists of 47 regional institutes withcombined budgets of about $375 million a year;about 30 percent of their funding comes fromcontracts with industry, another 30 percent fromgovernment contracts, and most of the rest fromnational and state government grants. The FhG’sclear mission is to promote innovation in civiliantechnologies and rapidly transfer research resultsto industry. The institutes put their efforts intoapplications-oriented R&D, often focused on theneeds of regionally concentrated industries, andforge links between universities, industry associ-ations, and individual companies.

There is little parallel with the FhG in theUnited States. Federal support of regional centersworking with local industries on application-oriented R&D and technology demonstration hasscarcely existed, but a new program calledRegional Technology Alliances (RTAs) maydevelop into that kind of system. Authorized infiscal year 1992, the RTAs received their firstfunding in fiscal year 1993, at the very substantiallevel of $97 million. This new program was partof a $1-billion defense conversion package toencourage technology development and diffusionin both defense and civilian sectors, but the lawstrongly emphasizes national security goals, andthe program is lodged in DoD, managed byARPA. This might constrain the RTAs fromdeveloping the frankly commercial character ofFhG.43 However, in planning the program, ARPAformed close cooperative ties with NIST, DOE’sDefense Programs, NASA, and the National

42 Fede.~ cOOrd~fig cOmcll for Science, Engineering ~d TcckoIogy, Grand challenges: High perjiormance cOmpUfi?lg and

Communications, a Report by the Committee on Physical, Mathematical, and Engineering Sciences, to Supplement the President’s Fiscal Year1992 Budget (Washington, DC: OffIce of Science and Technology Policy, n.d.), p. 2.

43 Intcr~~t~gly, the ~G fomd it5 ewly support from the milit~, but ~s long sin~ outgrown that identity. Today, Ody 7 Of tk 47 FhG

institutes perform primarily military R&D.


Science Foundation, and each was expected totake some of the responsibility for this and otherdefense conversion programs.

Assuming the RTAs succeed in forming linkswith commercial companies, they might fill animportant niche in U.S. cooperative R&D. Theywould not be suited, however, to undertakinglarge-scale, long-term projects with a strongpublic purpose. Nor does it seem feasible for DOElabs to remake themselves on the FhG model(though that suggestion has been aired). Althoughsome of the labs (Sandia in particular) havealready demonstrated some ability to work withsmall companies in adapting lab technologies tothe companies’ needs, the labs’ main strengths—technical talent in depth, multidisciplinary teams,expensive state-of-the art equipment—seem moresuited to big projects.

ARPAARPA has attracted even more attention as a

model for government-supported R&D. Throughits 35 years of existence, ARPA has gained areputation for rapid, flexible decisionmaking, andfor placing its bets intelligently. At times it hasbeen a major player in promoting advanced dual-use technologies and has fostered the develop-ment of industries whose main markets werecommercial but that also could be an importantsource of supply for DoD. At other times, politicalpressures have confined ARPA more narrowly tostrictly military objectives (see ch. 5).

The pressures today are running the other way.With defense budgets declining, DoD has morereason than ever before to emerge from thedefense procurement ghetto, and buy more fromthe civilian sector. The advantages are twofold:prices are usually lower on the commercial side,and very often commercial technologies are moreadvanced--especially in computers and telecom-munications. After at least a partial eclipse in the1980s, ARPA has reemerged as a premier dual-use agency.

Despite the apparent divergence of military andcommercial products (no one needs a stealth jet

transport), critical technologies embodied in theseproducts —advanced materials, semiconductors,software-are converging. Five of ARPA’s 10offices direct their research toward core technolo-gies in electronics, microelectronics, computing,software, and materials, and they control 80percent of the agency’s budget. Moreover, theyare putting more emphasis than ever before onmanufacturing process technologies. Many of theagency’s projects in this area are cooperative,partly funded by industry. ARPA typically prefersto work on these projects with commercialcompanies or commercial divisions of companiesthat also do defense work. The advantage forARPA is that the company will support continueddevelopment of the technologies through itscommercial sales, while serving as a source ofsupply for DoD. The broader economic advantageis wide diffusion of the ARPA-supported technol-ogies and superior commercial performance.

ARPA is so highly regarded as a promoter ofadvanced technologies that, while the rest of thedefense establishment faced shrinking missionsand budgets, ARPA received a huge jump infunding in fiscal year 1993, from $1.4 billion to$2.25 billion; this included $257 million for s&defense conversion programs for codevelopingdual-use technologies and supporting manufac-turing process technologies and education. Inaddition, in recent years Congress has mandatedARPA funding for specific dual-use programs,beginning in 1987 with the unprecedented 5-year,$500-million funding for Sematech (the semicon-ductor manufacturing consortium, cost-sharedwith industry), and continuing on a smaller scalewith programs in high-definition systems, ad-vanced lithography, optoelectronics, and advancedmaterials.

Besides all this, the defense conversion legisla-tion for 1993 gave ARPA some entirely newresponsibilities in areas with which it had noexperience. These are the Defense ManufacturingExtension program, which will contribute to thecosts of State and regional industrial extensionprograms for small and medium-size manufactur-


ers; the Defense Dual-Use Extension Assistanceprogram, aimed at helping defense companiesdevelop dual use capabilities; and the RTAsdescribed above. Each of these programs wasfunded at $97 million; for all of them, includingthe RTA, ARPA formed a joint TechnologyReinvestment Project with four other Federalagencies to plan and oversee the programs.

ARPA is becoming, de facto, a dual-usetechnology agency with a wide range of responsi-bilities. Congress expressed its intention to for-mally give the agency a dual-use mission bydropping the word ‘‘Defense” from its title,restoring its original name of Advanced ResearchProjects Agency; in February 1993, PresidentClinton directed DoD to make the change. Con-gress has stopped short of naming ARPA as theNation’s lead agency for technology policy, andthere is support in Congress, as well as Adminis-tration backing, for much higher funding for thesmall civilian technology development and diffu-sion programs lodged in NIST.44 ARPA, with allits cachet of success in dual-use technologies, isstill a defense agency with the primary mission ofmeeting military needs. Despite the many over-laps in technologies having both defense andcommercial applications, the match is by nomeans complete, nor are priorities necessarily thesame.

1 Coordinating Institutions forNew Missions

Whether new missions for the weapons labs aredefined as supporting U.S. competitivenessthrough R&D partnerships with industry, or astaking part in new national initiatives for publicpurposes, with collateral benefits for competitive-ness, the question of strategic planning becomesmore insistent the more money is involved. AtDOE headquarters, the managers of Defense

Programs have felt the need to impose a strategicplan on a cooperative program funded at $141million. If the amounts available to the DOE labsfor industrial partnerships were to rise to $500million or $1 billion, as is implied by somecurrent proposals for the labs’ future, the prob-lems of managing such a big, visible programwithout order, priorities, and interagency coordi-nation could become still more apparent. Ofcourse, if lab/industry partnerships were managedat the lab level on a first-come-first-served basis,most would likely concentrate on critical technol-ogies, simply because these are of greatestinterest to both public and private partners. It isdoubtful, however, that uncoordinated, individualprojects would advance critical technologies aseffectively as a well-planned multiagency strat-egy, such as the HPCCP.

There is no U.S. Government agency with aclearly defined responsibility for managing tech-nology initiatives that span several agencies. Thecommittees of the Federal Coordinating Councilon Science, Engineering, and Technology(FCCSET) in the White House Office of Scienceand Technology Policy (OSTP) are the nearestapproximation, but they have generally operatedas consensus groups with no real locus ofdecisionmaking authority. Other Nations do haveinstitutions that guide technology initiatives,usually in a science and technology agency.Germany has its Federal Ministry of Research andTechnology (the Bundesministerium furForschung und Technologies, or BMFT) and Japanhas its Science and Technology Agency. Also, theJapanese Ministry of International Trade andIndustry (MITI) contains another technologyagency, the Agency of Industrial Science andTechnology.

45 Both have many technology policyresponsibilities, including funding and oversee-ing R&D laboratories that contribute to civilian

44 Bills in ~e Hou~~ and Scmtc ~ he Iosrd con~css (s 4 and H.R$ 820) would geafly ~creme fuding for NIST’S IWilNlfiiChUiIlg

technology centers and tbe Advanced Technology Program. President Clinton has proposed similar measures.M Japan’s Science and Tec~ology Agency bd a budget of 522 billion yen ($3.9 billion) fi 1991; ~~’s Agency of ~dus~~ Science ‘d

Technology bad 117 billion yen ($870 million). The German BMIT had a 1992 budget of 9.4 million DM ($4.4 billion).


technology development, often with substantialparticipation and support from industry.

SUMMARY OF POLICY ISSUESAND OPTIONS

While military needs will continue to consumesizable government resources for R&D, DOEweapons labs may soon face significant reduc-tions in funding. There are plenty of claims formoney not spent on development of nuclearweapons. An obvious candidate is deficit reduc-tion. In the long run, a smaller burden ofgovernment dissaving could contribute to moreprivate investment, and to the growth prospects ofthe American economy. Accordingly, deficitreduction will be a policy priority for Congressand the Administration over the next few years.

Deficit reduction is only one of the claims onwhatever resources are saved through reducedweapons development. There are plenty of others,from improved education and health care tosupport for the newly democratic but strugglingregime in Russia. There are also persuasivearguments in favor of stronger government back-ing for American industry competitive perform-ance since R&D-traditionally part of the foun-dation that supports U.S. competitiveness—shows signs of weakening.

There is substantial support both in Congressand the Clinton Administration for cooperativeR&D partnerships between government and in-dustry, including cost-shared agreements be-tween companies or consortia of companies andgovernment laboratories. Those who favor lab/industry collaboration share the conviction thatnow—at a time when R&D is flat but competitiveindustries rely more than ever on knowledgeintensity-is not the time to cast away technologyresources that have taken decades to build up.Rather, every attempt should be made to use themin ways that contribute directly to the civilianeconomy. This does not preclude cutting theweapons labs to a size appropriate to their newdefense missions, which will largely be non-

proliferation, safety and security of nuclear stock-piles, and decommissioning of excess weapons,though some nuclear design capability will bemaintained. It does require prompt action to solveproblems that are hindering cooperative R&D.

This positive point of view is not universal.There is a strongly held opinion that all DOE’snational labs—the multiprogram energy labs aswell as the weapons labs—have lost their originalfocus, which was to promote peaceful and mili-tary uses of atomic power, and are now anextravagance the Nation can ill afford. Theywould like to see the lab system given ruthlessscrutiny, possibly leading to closure of some labs,downsizing of others, and redirection of govern-ment R&D spending.

For the longer term, survival of the DOE labsystem may depend on the labs’ success infocusing on new missions that provide clearpublic benefits. The weapons labs built theirexcellent staffs, equipment, and technologiesaround their core public mission of nationaldefense (and to a lesser extent, energy technolo-gies and the science underlying them). Peacetimepublic missions could include a larger and moreexplicit interest in promoting industrial competi-tiveness, but the grounds for supporting nationallabs with the taxpayers’ money are more compel-ling if the labs’ missions feature public benefitsthat the market is not likely to supply.

I Options to Shrink the DOE LaboratoriesThose who consider the weapons labs too big

and their culture too remote from that of privateindustry to contribute effectively to competitive-ness see the present moment as a good one torationalize, downsize, and consolidate the labs.Many would include all the DOE’s multiprogramnational labs (and possibly other Federal labs aswell) in the scrutiny. But it is the weapons labs,with their lion’s share of DOE R&D funding andthe obvious change in their mission, that aregetting the most attention.

I-Summary and Findings 131

Policy Option 1: Cut the labs’ budgets to fit thescope of scaled-back weapons functions.Through their regular budget and appropria-

tions functions, Congress, the Administration,and DOE are already engaged in cutting backnuclear weapons activities at the labs. However,the cuts may be fairly small and gradual as thelabs take on expanded nondefense functions,especially in environmental cleanup and energyprograms.

Policy Option 2: Create a LaboratoryRationalization Commission.Should Congress decide to thoroughly restruc-

ture and downsize the weapons and other DOElabs, it may wish to create a Laboratory Rational-ization Commission composed of experts fromDoD, DOE, the private sector, and other appropri-ate institutions to recommend how to manage thecuts, organize the work remaining to the labs, andmake any necessary improvements in lab man-agement. To do this with care and forethoughtwould inevitably take time. It is likely that thecommission’s recommendations would take atleast a year or so to formulate. This argues forpostponing any deep cuts or major reorganiza-tions while the commission is at its task, andmeanwhile working to improve the technologytransfer from the labs, including the CRADAprocess.

I Options to Improve Technology TransferFrom the DOE Weapons Laboratories

A second approach is to make the talents andresources of the weapons labs more readilyavailable to private firms. This approach is notincompatible with reduced funding for the labsand might even be combined with a strategy ofthoroughgoing restructure and downsizing of thelabs, should Congress choose that option.

The months that it usually still takes toconclude a CRADA with the weapons labs is areal threat to the effort’s success. There is nosimple answer to speeding up and simplifying theprocess. Some laboratory people and many repre-

sentatives of companies that have tried to negoti-ate CRADAs with DOE favor giving moreauthority to lab directors. They believe, probablycorrectly, that this would hasten the process,especially if the labs had the power to spenddesignated funds from their R&D budgets forCRADAs rather than redesigning ongoing pro-jects to include cooperative agreements withindustry.

There are several criticisms of this approachthat deserve to be taken seriously. A major one isthat with the funds for CRADAs in DOE’sDefense Programs so large, it makes some senseto take a strategic approach to lab/university/industry partnerships, concentrating resources oncritical technologies and minimizing overlaps.Second, there is the question of trust. The view ofsome at DOE headquarters is that the directors ofGOCO labs maybe too willing to compromise thenational interest in order to find industry partners,to avoid deep budget cuts in a time of changingmissions and uncertain funding. Furthermore,some in Congress have little faith in the dedica-tion of some of the labs’ contractors to putting thenational interest first. If lab directors are givenmore authority over CRADAs, fear of congres-sional investigations might stall the process.Finally, the division of congressional responsibil-ity for DOE authorizations (energy and naturalresources committees authorize energy program,and armed services committees authorize defenseprograms) complicates legislative guidance onfunding and managing technology transfer.

In short, there is little consensus among experi-enced, knowledgeable people on how to stream-line the CRADA process while getting the mostout of it in technologies that advance the nationalinterest. The lack of a U.S. Government coordi-nating agency for technology development anddiffusion programs makes the uncertainties moreacute. Greater coordination might be initiated inthe new Administration, which seems committedto a more active government technology policythan the previous administrations but, at thiswriting, that is unknown.


The specific policy options that follow aremostly confined to short-term issues of makingthe new process of industry/lab cooperative R&Dprojects work more smoothly. Broader issues,including the longer term future of the labs, theirpossible role in R&D support for new nationalinitiatives, and coordination of government-widetechnology policy, are discussed in more generalterms. Government-supported R&D has entered agenuinely new era, and all the issues involvedcannot be solved at once. In the face of theuncertainties, the options proposed here should beregarded as experiments, and results should bemonitored. This does not imply that experimentsshould be tentative, or that monitoring shoulddevolve to micromanagement. Congressional mon-itors should remember that the labs will needfreedom to experiment that positive results taketime, and that failures are part of any high-riskundertaking.

Policy Option 3: Shorten the process ofinitiating CRADAs.

Several actions could be taken under thisumbrella (see ch. 2 for details). For example,Congress might wish to shorten the time allowedfor DOE field offices to approve CRADA docu-ments; or it might eliminate separate approvals,first for the joint work statement and next for theCRADA itself-a two-step process that can takeup to 120 days.

Another option in this connection is to giveDOE an exemption from the Freedom of Infor-mation Act (FOIA) covering proposals for coop-erative R&D. In describing proposed researchprojects, companies often include informationthat they wish to keep out of the hands ofcompetitors (including foreign companies). TheDOE labs are protected from FOIA requests to seethe proposals, but DOE headquarters is not. Thelabs and their industry partners have on occasionremoved or marked proprietor-y information fromproposals before sending them to headquarters forreview, but this adds delay and aggravation to theprocess. NIST has, and uses, a FOIA exemption

for proposals it receives for R&D projects in itsAdvanced Technology Program. Congress mightwish to give DOE the same authority.

Policy Option 4: Reallocate CRADA authority.

Another option would be to direct that thescreening process Defense Programs has estab-lished be shortened or dropped. Much of the delayin getting CRADAs out of the weapons labs is dueto DP’s coordinating process, which involves acall for proposals and then a two-step evaluationof the proposals. All this takes place before thesubmission of work statements or CRADAs to thefield offices. The purpose, as noted, is to mini-mize overlap, assure complementarily of projects,and determine the fit with the strategic goals ofDPs cooperative R&D program. But the effect,inevitably, is delay. DP aims to keep the wholeprocess-its review plus the CRADA negotiation—to no more than 6 months, with the eventual goalof 4. In practice, in the last half of 1992 the DPprocess by itself was taking 5 or 6 months; withthe addition of another 1 to 3 months at the fieldoffices, the total time to initiate CRADAs proba-bly exceeded 6 months for most CRADAs. Thiscounts only the time after lab and outsideresearchers have spent time defining a piece ofwork together.

Suggestions have come from several quartersfor delegating CRADA authority to the labdirectors, This could weaken or undermine thesystem DP has set up to impose a coordinationand strategic goals on cooperative agreements.Also, it could mean a change in the law; NCTTAexplicitly requires GOCO labs to obtain parent-agency approval of both the joint work statementand the CRADA. Two variants of the option areas follows.

●

●

Option 4a: Give lab directors greater discre-tion in allocating budgets to technologytransfer. This would not necessarily requirea change in the law.Option 4b: Give GOCO lab directors fulllegal authority to execute and fund CRA-DAs. This would require a change in the law.


Some compromise choices, also requiring legisla-tive

●

●

change, might also be considered.

Option 4c: Give the lab directors authorityto conclude CRADAs of a certain size (up toas much as $1 million, say) without DOEoversight, or on the same terms as the GOGOlabs (30 days for parent agency disapproval).Option 4d: Put up to one-half the fundsavailable for CRADAs at the disposal of thelabs, reserving the other half for a morestrategic program managed by DOE head-quarters and requiring agency approval;these projects would be national in scope andthe labs would submit competitive propos-als, as they do in the present DP scheme.46

Policy Option 5: Require that DOE allocate acertain percentage of the labs’ budgets totechnology transfer.

This proposal is gaining currency. In theirFebruary 1993 statement of technology policy,President Clinton and Vice-President Gore statedthat all DOE, NASA, and DoD labs that can makea productive contribution to the civilian economywill be reviewed, with the aim of devoting 10 to20 percent of their budgets to cooperative R&D.47

Congress had previously expressed support forthe idea.48 In 1992, the portion of the weaponslabs’ budget funded by DOE programs was about$2.7 billion;49 10 to 20 percent of that wouldamount to $270 to $540 million in the weaponslabs alone—assuming that their present levels offunding continue.

Although there is some concern that the 10 to20 percent target is unrealistically high, theconcern is probably misplaced. In fiscal year1993, when DP had $141 million set aside forCRADAs (mostly in the three big weapons labs),there were many more proposals than could befunded; that amount was more than 5 percent ofthe weapons labs’ total DOE funding for 1992 andnearly 9 percent of their DP funding. Anotherconcern is how such a scheme would work its waythrough Congress. It could prove tricky, sinceDOE’s authorizations are handled by two com-mittees in the Senate and four in the House ofRepresentatives; appropriations are handled bytwo subcommittees of each chamber’s Commit-tee on Appropriations.

Policy Option 6: Establish stronger incentivesfor technology transfer.

Incentives might compensate for difficultiesthat now stand in the way of lab researchersspending time on technology transfer projects. Intheir annual planning process, DOE and the labsdecide on the projects the labs will work on in thefollowing year. Once the plans are in place, labresearchers find it hard to devote more than a fewdays to planning cooperative work with outsidepartners; they have to account for their time quitestrictly. The lab’s overhead account is the onlyplace to charge for time spent in planning jointR&D, and there are many claims on that account.When researchers spend time planning coopera-tive work, it is often their own time, on nights and

46 something like this so percent SOIUtiOn was proposed by Albert Narath, President of Sandia National Laboratories, in hetigs &fore the

U.S. House of Representatives, Committee on SrnaU Business, Subeornmittee on Regulation Business Opportunities, and Energy, “Reducingthe Cycle Time in Lab/Industry Relationships,’ Dec. 4, 1992. While supporting DOE’s role in approving CRADAS, Narath also made a casefor greatly streamlining the process.

47 c *TwhnOlOgy for America’s Economic Growth, ” op. cit., footnote 15. A variant is the suggestion from the Secretary of Energy AdvisoryBoard Task Force that certain labs in the DOE system be designated as technology partnership “centers of excellence, ’ and devote up to 20percent of their budgets to the purpose. Somewhat inconsistently with its recommendation that the weapons labs confiie their activities tonuclear defense, the Task Force suggested Sandia as well as Oak Ridge as candidates.

48 ~ its repo~ on be fix- yea 1$)93 Dor) ~u~onmtion bill, tie semte co~ttee On AXIII@ Sel-vices directed DOE to set a god of

allocating 10 percent of the Defense Programs R&D budget to technology transfer. U.S. Senate, Committee on Armed Services, NationalDefense Authorization Act for Fiscal Year 1993: Report, report 102-352, to accompany S. 3114.

@ Theti total budget was $3.4 billio~ but about $700 million was Work for Others (WFO), mostly the Department of Defense. A fewCRADAS have been funded by WFO, but most CRADAS currently come from DOE program funds.


weekends. This constraint, combined with luke-warm enthusiasm for technology transfer on thepart of some of the labs’ middle managers, canslow or abort potential CRADAs.

The law already encourages technology trans-fer by providing that 15 percent of the royalties ofany patent licenses may be awarded to theindividual lab researchers who developed thetechnology. This incentive is chancy and ratherremote, however. Top managers at the labs couldinstitute more immediate rewards. These mightinclude giving to project managers active intechnology transfer extra staff positions or acoveted piece of lab equipment. The lab managersmight make technology transfer a more promi-nent factor in employees’ performance ratings.None of these measures would require congres-sional action, but might be encouraged in over-sight hearings.

Congress might wish to take more directaction, as in the following two suggestions:

Option 6a: Direct that part of the labs’ over-head account be allocated to pre-CRADAdevelopment of proposals of joint work.Option 6b: Establish a governmentwide setof awards for effective technology transferfrom Federal laboratories. Awards of thiskind, if sparingly used, can be surprisinglyeffective. 50

Policy Option 7: Reassess definitions ofnational interest in the context of technologytransfer.

Private industry creates most of the Nation’sjobs, value added, and technology development.It is clearly in the national interest for Americanfins, and foreign firms that do business here, toprosper. However, the match between nationalinterest and corporate objectives is not perfect. Inthe context of cooperative R&D agreements,three issues that have generated conflict, legal

wrangling, and delay are U.S. preference for R&Dand manufacture, disposition of intellectual prop-erty rights, and liability for damages.

A strict requirement for U.S. manufacturingcould drive many potential partners away fromthe labs, possibly leaving only smaller companieswith few international ties and limited R&Dresources of their own to match lab contributions.Moreover, requiring U.S. preference might evendeprive some companies of their best shot atcommercializing advanced technologies. A broadportfolio of technologies, including those devel-oped in partnership with the labs, is a distinctadvantage to a U.S. company negotiating with aforeign company for access to its technologies.The most reasonable course may be to choosesomething less than an ideal outcome and acceptthe discomfort.

● Option 7a: In relevant legislation Congresscould either insist on U.S. preference, under-standing that many industrial partners willopt out; or permit a form of preference thatcompanies can comfortably handle, as in theumbrella CRADA that DOE signed withcomputer systems companies, which re-quires only that companies perform substan-tial R&D, not substantial manufacturing, inthe United States. The latter option wouldaccept the possibility that this Nation mayeventually import products based in part onAmerican publicly funded R&D.

Another choice is to establish a general principleof U.S. preference, but to make exceptions caseby case. This could be done in one of severalways:

. Option 7b: Congress could direct agencieswith cooperative government R&D pro-grams to grant exemptions from U.S. prefer-ence only when industrial partners show thatsubstantial manufacturing in this country is

50 ~ ~=ple is the ~co~ B~dl-idge Natio~ Q~ity Aw~d, ~:r~ted by Congress in 1987 and awarded each year to a few companies

or organimations that bave benefited the Nation through improving the quality of their goods and services. Hundreds of companies apply forthe award each year, even though bidders must go through a rigorous self-examina tion merely to apply.

—

not feasible, and they commit themselves toproviding alternative benefits to the U.S.economy. As noted, DOE has adopted apolicy along these lines.

. Option 7c: Congress could establish a U.S.Preference Review Board to make case-by-case decisions on exceptions to the U.S.preference rule for any agency with cost-shared R&D projects with industry. Con-gress might consider empowering OSTP toexercise this function, or creating a smallindependent agency to consider U.S. prefer-ence issues governmentwide. The boardwould have to pursue the dual aims of actingswiftly but avoiding rubber-stamp approv-als.

Both these last options are inclined to causedelay. Having a governmentwide board makethese decisions might well be more unwieldy thanleaving it to the agencies, though there wouldprobably be more consistency in the decisions.Another disadvantage is that the board’s deci-sions might please no one. It has certainly beendifficult for officials in the Commerce, State, andDefense Departments to agree on control overexports of technologies that might, if allowed,threaten U.S. national security but, if forbidden,unnecessarily harm U.S. commercial interests.

The same kind of conflict, and possibly thesame kind of resolution, exists for intellectualproperty rights. This is an unsettled area in DOECRADAs, and is the subject of much hardbargaining between the labs and their industrialpartners and consequent delay. Possibly, settle-ment of some of these issues may evolve withmore experience, but differences among indus-tries, and among companies within industries, arelikely to remain. Congress may wish to empha-size one side or another of the intellectualproperty issue and live with the consequences. IfCongress chooses to support the public purpose ofwider diffusion, fewer companies may be inter-ested in partnerships; if it chooses to givecompanies more protection, the public return on


taxpayers’ investment may be more limited, or atany rate less direct.

DOE turns over to GOCO lab operators most ofthe authority for settling with industrial partnerson the disposition of intellectual property rights,subject to the government’s right to use theintellectual property for its own purposes. Con-gress may wish to provide some guidance thatwould more clearly define the scope of negotia-tions over intellectual property.

. Option 7d: Congress might choose, in theform of resolution or law, to provide guid-ance that discourages the grant of exclusivelicenses that have a broad field of use, or thatlimits the time during which exclusivelicenses prevail. Alternatively, Congress mightencourage DOE and the labs to accommo-date companies’ desires for broader intellec-tual property rights.

One further problem is that some companieshave run into frustration and delay in CRADAsinvolving more than one DOE lab because eachnegotiates terms separately, and makes differingdemands in such areas as intellectual propertyrights and U.S. preference. DOE’s recent guid-ance to field offices on U.S. preference shouldmake for more uniformity and predictabilityamong the different labs on this issue, but thepotential for inconsistency among labs remains inthe handling of intellectual property. ThoughDOE has given GOCO contractors most of theauthority over disposition of intellectual propertyrights in cooperative agreements, it can stillexercise oversight and provide guidance.

. Option 7e: DOE might, through technicalassistance and policy guidance, encouragethe labs to harmonize the terms of theiragreements with industrial partners, espe-cially in multilab projects. Through over-sight, Congress could encourage such actionby DOE, or alternatively require it by law.

Another national interest issue is liability.There may be some practical possibilities ofagreement on this issue that would suit both the


government and private companies. Both per-ceive that damage claims are becoming moreburdensome, and both would no doubt welcomesome general limitation on liability. However, nopolicy option is proposed here, as OTA has notdone extensive analysis of the product liabilityissue.

Policy Option 8: Measuring the value ofcooperative research and development.

Assuming that the CRADA process can bemade to work more smoothly, a longer term

question will be how to measure the value of theagreements. Success cannot, of course, be meas-

ured overnight. Nor is it easy to establish mean-ingful measures of success for R&D projects,especially from the standpoint of social returns.Economic results such as numbers of jobs createdor value added are hard to trace with any precisionto R&D; other factors are too important.

A practical measure of success, after 5 years orso of experience, might be the continued orgrowing interest by industry in submitting pro-posals for cooperative work. If companies, whichhave their own internal measures for success ofR&D investments, continue to put money andeffort into the projects, it is fair to conclude that

they consider the ventures worthwhile. In thelonger run, cooperative R&D projects may bejudged by the general measure of whether they aredeveloping technologies that form the basis forcommercial production, keeping in mind that

there must be allowance for failures as well assuccesses in any program of high-risk, potentiallyhigh-reward R&D.

Evaluation of the results of public R&Dinvestment may have to be largely judgmentalrather than precisely quantitative. That does not

argue against making the attempt. If after a fairtrial period the labs’ cooperative R&D is judgedto be seriously disappointing, it would makesense to shift money to other R&D performers.Congress might direct the Secretary of Energy todevelop an evaluation procedure for cooperativeR&D. Alternatively, OSTP might be directed to

develop evaluation procedures for all government/industry cost-shared R&D.

The options laid out above are mostly aimed at

streamliningg the CRADA process. In some cases,the streamlining comes at the expense of mini-mizing strategic guidance at the DOE headquar-ters level, as Defense Programs is now attemptingto provide. Given the large size and scope ofDOE’s R&D program, a screening process andstrategic direction make a good deal of sense—still more so if DOE takes part in governmentwideinitiatives to advance certain technologies. Thedownside is that DP’s internal screening prolongsthe CRADA process, trading oversight for fasteraction. A middle course may be possible, givinglabs more direct authority over a portion of thefunds available for CRADAs, or over CRADAsbelow a certain size.

In the short run, it might be worth sacrificingsome coordination and strategic direction in theinterests of getting the program working whileindustry interest is high. In the longer run, onceDOE, its field offices, and its laboratories becomemore accustomed to cooperative R&D, it may bepossible to set priorities for CRADAs and othercooperative work that fit within strategic initia-tives without months of delay in selecting pro-posals.

H The Longer Term Future of theDOE Weapons Laboratories

Most of those who see a national role ofcontinuing significance for the labs considercooperative work with industry an importantthough not necessarily central part of their future.Thus, the future of the labs will depend in part ontheir success in making the cooperative processwork. In thinking about the long-term future ofthe labs, however, cooperative R&D and otherforms of technology transfer should not beconsidered in isolation. The option of making atleast one of the weapons labs into a center forcooperative development of critical technologieshas been floated, but it has some important


drawbacks. The weapons labs built their emi-nence by their work on public missions ofnational importance, primarily defense. The tech-nologies and talents that private companies arenow eagerly pursuing are the legacy of thatmission. A national mission of ‘‘economic com-petitiveness’ seems an unlikely replacement,because it is so diffuse.51 The fear of lab officialsthat labs with such a mission could becomenothing but job shops for industry is probablywell-founded.

NEW PUBLIC MISSIONS

There is no lack of candidates for new publicmissions that might take the place of a muchreduced national defense mission and spend atleast part of a “peace dividend. ’ Not forgettingthat deficit reduction will claim a high priority,there are also strong arguments for new publicinvestments to strengthen the foundation of thecivilian economy and mitigate the economic andtechnological losses from defense cuts.

In choosing amongst a number of worthy newnational initiatives, one factor to keep in mind istheir ability to match the benefits the shrinkingdefense effort has conferred on the Nation (ex-cepting, of course, the ability to defend the Nationmilitarily). Foremost is the capacity to meet aclear public need-one that the commercialmarket cannot fully meet but that is well under-stood and broadly supported as essential to theNation’s welfare. In meeting such a need, thedefense complex also created other public bene-fits. It supported a disproportionate share of theNation’s R&D, some of which had such importantcivilian applications that whole industries werefounded on them. It provided many well-paid,high-quality jobs. It provided a large market—often the crucial first market-for technologicallyadvanced goods and services. A final factor,

though not a determiningg one, in choosing amongnew national missions is their ability to makegood use of valuable human, institutional, andtechnological resources formerly devoted to de-fense purposes-such as those in DOE weaponslaboratories.

NEW INSTITUTIONS

the President, his Cabinet, andon new national missions, set

NEW MISSIONS,

If and whenCongress settlepriorities, and establish finding levels, the nextquestion is who will carry them out. Whateverinitiatives are chosen, it seems likely they willinvolve many agencies, universities and nonprofitinstitutions, and hundreds, maybe thousands, ofprivate companies. While there are immediatequestions of how to deal with the changing sizeand missions of DOE weapons laboratories andsome DoD laboratories and test facilities as well,the answer probably is not to assign any of them,a priori, the leading responsibility for a majornew public mission. The job calls for manage-ment and coordination at a broader level than thatof individual R&D institutions.

Lacking a technology agency at Cabinet level,such as many other nations have, the U.S.Government has recently relied on OSTP in theExecutive Office of the President for whatevercoordination of government R&D programs hastaken place. Within OSTP, the job has gone tointeragency FCCSET committees. As noted, thecommittees have had no clear decisionmakingauthority. Moreover, at times their influence hasgone into complete eclipse, as in the early tomid-1980s when the Reagan Administration sawno need for a government technology policy. Asan agency in the Executive Office of the Presi-dent, OSTP is especially subject to the prevailingoutlook in the White House. It also lacks continu-ity; often it is staffed primarily by detailees from

5 1 NO(C, howev~~, tit ~Omc U,s. @vernmentR&D i n s t i t u t i o n s ~ve successfu~y ~ected ( h e i r

industries. Examples are the aeronautics R&D program and facilities of NASA (’growing out of the support provided by the National AdvisoryCommittee on Aeronautics, or NACA, from 1915 to 1958) and the cooperative research program of the U.S. Department of Agriculture, States,and land-grant colleges, dating back to the 19th century.


executive branch agencies and l-year fellowsfrom professional scientific organizations. On theother hand, in an Administration interested intechnology policy, OSTP could play a particu-larly influential role, since multiagency policycoordination is usually considered a specialresponsibility of White House offices.

Other ideas are to transfer some DOE labs, andpossibly other Federal laboratories, to a differentor new agency with overall responsibility fornational technology policies and programs. Thesemight include application-oriented R&D pro-grams, such as Regional Technology Centers, andtechnology diffusion programs, such as industrialextension services, as well as multidisciplinary,science-based R&D programs. Several bills inpast Congresses have proposed to create anagency or Cabinet-level department for the pur-pose.

Alternatively, an existing agency might beadapted to the purpose. MST, which houses theAdvanced Technology Program, a small technol-ogy extension program (Manufacturing Technol-ogy Centers), and the Baldridge Award, as well asits own laboratories, has been suggested as apossibility. ARPA, with its fine reputation as afunder of long-term, high-risk dual-use technolo-gies, has attracted still more attention. It controlsmore R&D funds for dual-use technologies thanany general purpose civilian agency, and thedefense conversion legislation of 1992 gave itnew responsibilities in technology diffusion.Still, despite the interest in reaffirming its dual-use character, ARPA is not likely to be given theleading responsibility for overall U.S. Govern-ment technology policy, because it is frost of all adefense agency answering to defense needs.52

NEW NATIONAL INITIATIVES

Of the possible choices for new nationalinitiatives that meet public needs, some of themost persuasive could not only promote advancedtechnologies and foster the growth of knowledge-intensive industries, but do so in environmentallybenign ways. Environmental protection itself isan obvious candidate; this very broad categoryincludes cleanup of hazardous wastes from pastactivities, management of wastes currently beinggenerated, end-of-pipe pollution control and,perhaps most promising, clean technologies thatprevent pollution. Public support for environ-mental improvement in this country is strong andgrowing. Global environmental issues too arerising to the top of the policy agenda, fed byconcerns over global warming, the ozone holeover the Antarctic, acid rain from industrializedcountries, and deforestation and species lossthroughout the world.

Part of the drive for pollution preventioncenters on energy. World demand for energy isexpected to continue growing well into the nextcentury, especially in the developing world.Technical progress in the last decade raises thepossibility that nonpolluting or less-pollutingrenewable energy sources may be able to meetmuch of this demand. There are special opportu-nities to substitute more environmentally benignforms of energy use in the United States, becausewe are such disproportionately large consumersof energy, especially in auto and air transport.

Energy-efficient transportation is a theme thatis often proposed for new national initiatives.53

New forms of transportation-both advanced railor guideway systems and cars that use new typesof energy—are centers of interest. These systemsnot only offer the public benefits of reduced

52 me que5tion of where to lodge respo~ibility for technology policy or for broad initiatives related to U.S. competitiveness is diSCUSSedin some detail in U.S. Congress, Office of Technology Assessment, Competing Econonu”es: Amen”ca, Europe, and the Pacific fi”m,OTA-ITIW98 (Washington DC: U.S. Government Printing Office, October 1991), ch. 2. See also John Alic, et al., Beyond Spinofl: iUiZitaryand Commercial Technologies in a Changing World (Bosto~ MA: Harvard Business School Press, 1992), ch. 12.

53 ~esid~t Clhton and vicefi~ident Gore included in their program for technology initiatives one to help industry develop nOr@hU@

cans that run on domestically produced fuels. “Technology for America’s Economic Growtlq ” op. cit., footnote 15,


pollution and lesser dependence on foreign oil,but might also provide economic benefits thatdefense once bestowed on the economy. Inaddition, some might use technologies and skillsformerly devoted to defense purposes. As anexample of one such initiative, new transportationsystems are considered in this report from theviewpoint of their potential to replace benefitsdefense formerly provided. This report does notaddress transportation policy broadly; other OTAstudies have analyzed many of the relevantquestions, including the degree of greater energyefficiency and reduced dependence on foreignsources of oil that various transportation alterna-tives might offer, as well as issues such asadequate capacity and convenient connectionsbetween highway, air, rail, and water transport.

Less polluting or nonpolluting personal vehi-cles look promising as an area of industrialgrowth, a driver of advanced technologies, apotential provider of good jobs, and a user oftechnologies and skills no longer needed for ColdWar purposes. Americans have historically cho-sen the automobile as their means of transport,and much in this country (e.g., the interstatehighway system, cities that sprawl out intosuburbs) favors its use. Electric vehicles (EVs),which depend completely or substantially onbatteries for propulsion, could have some near-term market potential in meeting stiffer air-quality standards. California has mandated that 2percent of vehicles sold in the State by 1998, and10 percent by 2008, must have zero emissions,and some other States (New York, Massachu-setts) are following suit. EVs are at present theonly cars able to meet that standard.

Battery EVs will probably fill most of the earlydemand for ultra-clean cars, and they are emi-nently suitable for some niches (e.g., PostalService or other in-town delivery vehicles);however the market for them may turn out to belimited. Vehicles powered by a combination of

fuel cells and batteries are currently less advancedthan battery EVs, but in the long run could be themore successful technology if they are moreeasily able to provide the range and quickrefueling that battery EVs are struggling toachieve. Still, fuel cell technology for automo-biles is immature and unproved; whether afforda-ble cost and reliability can be achieved is not yetknown. Both battery and fuel cell EVs facecompetition from other kinds of less pollutingvehicles, many of which are better developed, arecontinuously improving, and require much lessnew infrastructure. Alternative less pollutingfuels for vehicles using the time-tested interna-tion combustion engine include methanol andethanol, natural gas, and reformulated gasoline.Moreover, although battery and fuel cell EVs arethemselves without emissions and do not causelocal pollution, the energy source used to generateelectricity for them may be polluting.

U.S. Government support for the developmentof nonpolluting cars was already underway inearly 1993, but in a limited and uncoordinatedway. The Clean Air Act of 1990 and the stricterCalifornia standards have provided strong impe-tus for industry to develop clean cars, and there issome very modest support for purchase of non-polluting or less polluting vehicles for govern-ment fleets. However, the main encouragementon the part of government is, first, in the field ofregulation, and second, in research, development,and demonstration (RD&D). DOE and the De-partment of Transportation (DOT) both havescattered RD&D projects underway. The biggestof these is in DOE’s Conservation and RenewableEnergy program, which had a fiscal year 1993budget of $60.8 million for electric and hybridvehicle research, of which more than half ($31.5million) was for battery EVS.54 DOT has a$12-million project for cost-shared funding ofconsortia to develop EVs and advanced transit

54 Fuel ~elj R~ @ $12 ~lli~q ~~d h~bfid vehicles, def~ed M hose powered by elec~ci~ Cornbhed with a SXIldl iKlterIld Combustion

engine, got $16.8 million.


systems, related equipment, and production proc-esses.

The U.S. Advanced Battery Consortium (USABC),formed in 1991 as a collaborative effort betweenDOE and the Big Three American automakers, isthe largest government-supported R&D projectfor EVs. It is funded at $260 million over its frost4 years, 1992-96 (there are plans to continue it for12 years); of this, each auto company is providing$36 to $40 million, the Electric Power ResearchInstitute is contributing $11 million, and DOE ispicking up the rest, which amounts to $130million or one-half. USABC has set developmentand performance goals for mid- and long-termbatteries, on a timetable shaped in part by thecoming requirements of the California emissionslaw.

So far, defense conversion (i.e., the use ofdefense talents and resources for new civilianpurposes) has played little part in USABC. It islargely a civilian enterprise, with the Big Threeautomakers running the show from the privatesector side. Sandia is the only weapons labinvolved, but other DOE labs—Argonne, theNational Renewable Energy Lab, LawrenceBerkeley, and the Idaho National EngineeringLab-are participants. Outside USABC, severaldefense firms are using their experience withelectric propulsion systems in building powertrains for electric vehicles. Westinghouse Elec-tric’s electronic systems group, for instance, iscooperating with Chrysler in such a program. TheDOT program for EVs has explicitly tried to enlistdefense resources in some cases. One of its four1992 awards was a $4-million grant to Califor-nia’s Calstart project, a consortium that aims tocreate a new industry providing transportationtechnologies and systems. It includes in itsmembers aerospace companies, utilities, univer-sities, small high tech companies, transit agen-cies, and representatives of environmental andlabor interests.

Key areas in the development of both battery-powered EVs and the fuel cell-battery alternativeoverlap with many technologies developed for

military purposes both by industry and govern-ment labs. These include the handling and use ofnew fuels such as hydrogen; the application ofadvanced materials such as ceramics, plastics,alloys, and ultra-light composites; the use ofcomputers to model manufacturing processes andperformance and thus improve design; the devel-opment of fuel cells, batteries, and ultracapaci-tors; and the use of electronic controls andsensors. The demands of space flight, stealth,undersea operation, strategic defense, and othermilitary and aerospace programs have pushedforward work on these technologies.

Most of the government’s efforts for EVs haveso far been directed toward developing andshowcasing battery EVs in the near future. Thefuel cell-battery alternative has received lessattention. The R&D investment needed for aconcerted, integrated program to overcome theformidable technical challenge is substantial, andwould seem to offer the promise of highly paidscientific and engineering jobs over the next fewyears. If the efforts are successful, they mighteventually support the creation of a new kind ofauto industry with substantial numbers of produc-tion jobs and the advance of many new technolo-gies.

High-speed ground transportation systems(HSGT)---in particular magnetically levitatedtrains-are also often proposed as new initiatives,but here there may be fewer attractions in the wayof new technologies, new jobs, and defenseconversion. These systems may fill the bill formany transportation policy objectives, includingless pollution and less dependence on foreign oil,and they have the additional attraction of lessimpact than highways on the use of land. How-ever, most analysts agree that maglev or highspeed rail systems are probably limited to a fewheavily traveled corridors like the route from SanFrancisco to San Diego, the Eastern seaboard, andparts of Texas, at least if the system is not to relyon ongoing heavy public subsidy. There may beother growth opportunities abroad, but severalforeign companies, having long experience in the

field and historic, generous government subsi-dies, are much better positioned to take advantageof them than fledgling U.S. companies.

Whether HSGT could spur the advance ofhighly innovative, broadly applicable technolo-gies is questionable. There are no breakthroughtechnologies in high speed steel-wheel-on-railsystems, such as France’s Train a Grande Vitesse(TGV) and Japan’s Shinkansen; rather they em-body incremental advances over rail systems thathave evolved over nearly 200 years. Even maglevtrains, long the favorite technology of the futurefor engineering optimists, are not necessarily heldback by technological problems that the ingenuityof the aerospace and defense industries couldsolve so much as the tremendous expense of thesystems, the difficulty of acquiring rights of way,and the tough competition of air and auto travel ina big country with widely separated cities andrelatively low population density. Maglev mightcontribute to the advance of some technologies,such as strong lightweight composite materials,an area in which the defense sector is a leader, butoverall the effects would probably be helpfulrather than crucial. Still, it is unwise to be toodismissive about the technological possibilities.The Japanese maglev system uses low-temperature superconducting magnets, and workfor the system has contributed to cryogenictechnologies with applications in other fields.Possibly, high-temperature superconductivity (HTS)will get a boost from maglev, though this is by nomeans certain since the magnets are a very smallpart of this large system and may not offer enoughadvantages to offset their development cost andtechnological uncertainties. One DOE weaponslab, Los Alamos, and two multiprogram energylabs, Oak Ridge and Argonne, have ongoingcost-shared projects with industry on commercialapplications of high-temperature superconductiv-ity. The application nearest fruition is energystorage devices for electric utilities, to help solvethe problems of peak use.


The hope for large numbers of manufacturingjobs from HSGT initiatives is probably mis-placed. Japan is a premier producer, consumer,and exporter of passenger train cars, but theindustry there (finished cars, freight and passen-ger, and parts) employed fewer than 15,000people in 1990, of whom about 3,000 wereemployed in building 288 cars for the Sh-inkansen. Similarly, about 100 train sets (includ-ing 200 locomotives and 800 cars) were built overa 3-year period for France TGV with a manufac-turing workforce for the rolling stock and parts ofabout 4,000. Most of the jobs involved in buildinga HSGT system are in construction; many of theseare skilled high-wage jobs, but they are temporaryand often create boom-and-bust effects in localeconomies. There may be excellent transportationpolicy reasons for building HSGT systems inparts of the United States, but on the basis of thepreliminary analysis in this report, they do notlook like a very promising replacement for thecivil benefits of defense.

Indeed, there is no one new national initiativethat fills that bill. For example, in the long run,nonpolluting cars might form the basis for a newindustry that would foster technology advanceand create large numbers of productive well-paidjobs (perhaps only replacing jobs lost in theconventional auto industry, but possibly creatingnew ones, if the world market for ‘‘green’ carsexpands). However, such anew industry will takeyears to grow. Eventually, a combination of newpublic and private investments can provide bene-fits that formerly came from defense, and do it inways more directly rewarding to the civilianeconomy and US. competitiveness. Meanwhile,measures that help U.S. workers and firms dotheir jobs more productively and spur local andnational economic growth are the best bet fordefense conversion.

summary and findings - princeton university

Documents