/ «*. o/ TE^ .M414 l^ll'I'liiiiiiiiiill 3 TOaO ODbDlSEM ALFRED P. T WORKING PAPER SLOAN SCHOOL OF MANAGEMENT Funding Strategy Technology Programs Government in Jong-Tsong Chiang WP January 1990 3129-90-BPS MASSACHUSETTS INSTITUTE OF TECHNOLOGY 50 MEMORIAL DRIVE CAMBRIDGE, MASSACHUSETTS 02139 AiVCii. ?; Government Funding Strategy Technology Programs in Jong-Tsong Chiang WP January 1990 1. The author is grateful the to International 3129-90-BPS Institute Systems Analysis (IIASA) and several organizations their financial This paper is in Applied Europe for support and other assistance in doing the field research concerning this topic in a 2. for to appear in Change (forthcoming, number of European countries. Technological Forecasting and Social 1990). A^^O \ K?^^^ Government Funding Strategy Technology in Programs Abstract R&D Industrial in a market economy funding therefore public sector, government to guide private is R&D is a very mainly implemented countries, This paper investigates activities. To reduce 1. incentive in and reaches opportunistic a in number of some preliminary conclusions: behavior and ingrain intrinsic both competition and cost sharing principles should firms, be used concurrently in underwriting firms' R&D projects. Competition principles can be applied across many candidate 2. projects private important tool of the experience of funding national programs industrialized in around the same time or a series of one-of-a-kind projects over longer time horizon. The major 3. that there specific is In practice, 50-50 determine. 5. is the decision is competition" due to few qualified candidates in the appropriate cost sharing level used as a rule of thumb in many is may difficult to countries to making and circumvent "bounded Full cost endorsement for projects firms. "real of competition principles technological fields or in some, especially small, countries. 4. simplify no threat to application rationality." be another "quantum" alternative urged by government but not felt to be very relevant by Introduction The 1980s witnesses technology development programs national One Japan's strategy of industrial collaborative international in this R&D. bodies, public endeavor because in these programs warrants 1970s and emulate to is encourage to funding becomes the most important tool a market special make support of government in R&D economy an industrial R&D is Therefore the funding strategy Nevertheless, attention. it is accurate estimate of the financial activities because of different categorization of budget, indirect support by other subsidies, etc. as For most countries and even to normally very difficult simply hereafter enhancing industrial at Project in the late mainly implemented by private firms. in to and international of these programs' main features VLSI its (referred programs) aiming or international competitiveness. of national proliferation a means of In this research, the focus is only tax credits and on the underlying rationale and strategy, rather than the absolute level, of government direct funding to industrial R&D projects in national programs. Principles in Among Government others, national program. Funding government usually has First it to do three things it a should avoid spreading resources too thinly and instead concentrate them on some priority technology Second, in fields. should identify high caliber teams to undertake the task. Third, strategy has to help attain these goals. To encourage industry to pursue innovation in government can provide funds directions, high commitment to is Government funding should ensure good implementation. it some projects may to share some prescribed R&D cost. But too be very costly and risky. usually done at the expense of other alternatives which may This also have good potential, thus reducing the number of solution routes Too much subsidy from pursued and even chance of success.' government may also entice firms needs from industry and industry To propose ideas and execute Moreover, government may overlook signals and missions carelessly. opinion to may bother to reflect their not industry has no part in the funding. if teams, government can better select process based on competition in one way adopt a screening or another among candidates. This strategy hopefully can also help reduce conscious or unconscious opportunism. 2 competition But it also has some may make opportunism more behavior more difficult to afford. ^ may not exist if there are technological fields. situation may projects, likely Another is that too much and responsible that few qualified candidates the is "real in competition" some "moral hazard" due to "small number" arise."* To ensure the progress. Then One limitations. responsible But most implementation, R&D activities, government can monitor except physical demonstration are difficult to evaluate objectively. concerned contracts are rather incomplete and Furthermore, most full of uncertainty R&D. because of the nature of be unobservable or not easy to detect. reality cannot do much To compensate commitment of that, the more their the do good job own stake. In principle, the higher the In this own indicated above, spill-over effects the appropriability agree with each other. more "serious" resources, the way, firms have "intrinsic" incentive However, the appropriate cost still government would be willing or externalities. to remain to For firms, the firms' (and EC the But both sides may not of the projects. And be more cost share They would commit more resources opposite rationale applies. higi.er sake. own between government and firms sharing formulae decided. as has good reason to hypothesize It firms invest their for their pitfalls funding contingent on the firms' its and "honest" they will be. to other words, government in In firms are not strongly committed to the job. if for the potential government can make may Therefore opportunistic behavior actual financial expenditure is not easy to strictly monitor. In is many national a necessary becomes a for collaboration competitors, which purpose makes appropriate Application Government funding thus condition for getting subsidy. bait it may more funding of especially otherwise not occur difficult programs, collaboration international) at least in between at all. real or potential This additional theory to determine an level. Competition Principle and Limitations the real world, competition In proposals is based on the evaluation of basically followed by most West European countries investigated in this research: the U.K., Sweden, the Netherlands. This support of industrial requires R&D. some threshold small firms with is is little also Germany's However, effort. Norway, Finland, and long tradition of government to prepare a qualified proposal This process in effect rules out technical reservoir. Once a project is many started, it quite difficult to stop except for very obvious reasons like the termination of firms. ^ However, some evidence shows firms actually spent more money that many than they originally claimed to, because their projects were connected to their central business and many new As research predicted, topics for were stimulated some technological after areas, some time of work.^ there Examples include qualified firms, and real competition does not exist. semiconductor hardware projects materials programs technology programs in in in Norway and Finland. most countries, conventional Finland, and forest-related The main reasons are as follows. semiconductor hardware technology, entry threshold presently there are only a small other more traditional few are number of firms technology programs in is very high this industry. In and In which need rather concentrated and big investment, there are only a few large established is little firms in small countries.^ In some other cases where there enthusiasm from industry, government even has firms to participate in the national programs, to persuade work with them on proposal preparation, and arrange special support from public R&D institutes. the principle of competition In Japan, way. Government intention to applied in a different is support oligopolistic made technology) firms to compete in the global market small number of was not really competitive. projects since the So winners." "potential IT (information bet on a it selection of firms the initial But a series of consecutive national 1960s gave government a chance to better compare and monitor participants' competence and commitment. As a striking example, when Oki failed to come out with a commercial product the New Series Project in response to International the next IBM 370 VLSI Ministry of series, the Trade and Industry (MITI) decided to in exclude it from Therefore competition principle works over Project. time. By contrast, France, a country with strong tradition of government guidance and direct involvement development, overlooked champion strategy" in this many principle fields, another one in semiconductors. It e.g., is in technology and adopted one firm argued that, in "single computers and among others, lack of competition pressure helps explain France's failure in developing a viable semiconductor and computer industry. Application Rationality of Cost Sharing Principle and national Bounded the As some to government funding level, it most governments, and even cases, not surprising that, except is EC Commission, adopted sharing policy in projects mainly executed by firms. government financial backing pilot projects; is In Denmark, the about 40-45% including prototype Sweden, Finland and the Netherlands, about in cost- and half; in the U.K., the original plan of varying percentage depending on the nature of individual projects and of to 50% in most in FRG, across-the-board; in EC programs (e.g., 60% on the average a long tradition of less than half; and ESPRIT, RACE, BRITE and EURAM), 50%. Japan, the situation was more complex: approximately FONTAC, 100% 50% Information Processing System (PIPS), Fifth Generation in The above evidence suggests practice to So level. industry therefore is New in Series, But in in Pattern 40% in Computer. that it quite difficult in real is determine project by project the appropriate cost-sharing to roughly the split the cost evenly between government and "most convenient" way under normal conditions. ^ becomes a the delicate issues to rule of thumb in many capabilities, ha\e little competitiveness, or embrace government tends to countries, It and simplifies be dealt with by "bounded rationality." ^^ However, when the tasks are extraordinarily firms' 50% Super High Performance Computer, 100% in VLSI, and 100% was reduced difficult relative to the immediate impact on the firms' high likelihood endorse the whole cost. of large externalities, Japan's exceptional experience in Super High Performance Computer, PIPS and Fifth Generation Computer Programs as discussed seems support this to Rationale of Full In Japan, second alternative. Funding in Japan PIPS was an add-on project 1971 after the in programming, printing and data transmission High Performance Computer in below details of the special weakness in handling Japanese language recognition began more in 1971 Super previous the in PIPS was highly technical and Project. without a true understanding of implications or its complexity by the various MITI councils and advisory groups. proved to be an extremely difficult general development level Government sponsored the set of tasks, considering the of input-output devices whole cost was then reduced from ¥35 billion was extended two more years from to It in project, this ¥22 the at time. budget but the and the execution billion, Upon the planned eight years. its completion, PIPS did not meet any real commercial marketing criterion. standpoints. The evaluation was mixed, depending on One failure but with some high remarks for the later Fifth actual study said its MITI considered PIPS technical successes.' ^ U.S. different as a OTA whole gave it to be a relatively long-term impact including laying groundwork for Generation Computer Program. '^ outcomes, Japan's government found it Regardless of its very difficult to persuade industry to pool funds for PIPS. In the Fifth Generation Computer Program, the ultimate goal develop a radically new kind of computer based on revolutionary is to breakthroughs it in computer architecture and successful to a reasonable degree, is it intelligence. artificial may put Japan at the leading edge of computer and information technology, a step further from present In fact, MITI's attempt in this position. of its creativity deemed as to covered by industry. a natural extension As until recently are budget ¥100 billion to total The Ministry of Finance difference which the computer However, private industry was half of the planned make up its also aims at It such vague and seemingly unrealizable goals. also refused is and original innovation, which Japanese strength. commit reluctant to generation program long-term series of national initiatives. stimulating not and the fourth microelectronics strong If was supposed MITI ended up being a result, to be solely responsible for the total funds ¥50 billion as promised. ^^ For the Super High Performance Computer Project, the rationale for government 100% support available. in not clear according to the information But there are some clues. 1962 was widely perceived as a Fujitsu, not is Oki and run.i'* Meanwhile, powerful 360 project in NEC series. 1966 might be due to this predecessor failure. The FONTAC parts launched developed by did not connect properly, and the machine could IBM in 1964 introduced its much more This serious situation triggered this national to quickly membership increased Its to respond to IBM's new threat, and include all six its major manufacturers. urgent condition that MITI decided to underwrite the Super High Performance Computer Project, It fully Remarks Concluding programs, government assumes a direct national In developmental (as opposed to regulatory) industry asymmetry innovation intrusion than programs. ^^ of vested constituencies, with taken this international bodies like arm's length. In real imposed by "bounded the experience in seems advisable more ^ to correct little in many in "market failure," but EC Commission practice, it to in support industrial R&D and threat of "opportunism." shows countries the sense that it that strategy this at intrinsic Though may become financing designs in the future. Notes 10 a But funding strategy R&D helps guide industrial sense in the subsidized firms. rather simple funding and even however, there are many limitations, rationality" many countries of national importance and ingrains delicate government deep consideration, 6 into challenge, the competition and cost sharing widely adopted are typified by the failure," "white elephants."^ also create responsibility many government-led groups like politicians, bureaucrats and interest also is To cope this general This information "government if the involvement may not only do principles government. an important cause of failure in is technology direction in has better knowledge about the commercial prospects of a technological may But role. incentive far in the and from perfect, starting point for ^ A comes from better selection a wider variety is advocated in Nelson and Winter (1982). 2 Using competition opportunism to reduce is one of the main arguments by Williamson (1975). 3 This 4 Williamson (1975). 5 This is is Dutch experience expressed November 1989. 6 This confirmed in December 1989. ' This is is pp. " in in May EC Hague programs in in Brussels 1989. Anchordoguy (1988), p. 221 and Nelson (1984), 45-47. This rationale EC is confirmed Commission, etc. in 1989. 10 refers "Bounded limits of human formulating rationality" actors. As the Nordic countries, in several November and December to OTA interviews in April, the cognitive May, and computational and solving complex problems, processing information, (1957). Bloom the Netherlands, a result, individuals face limits in and anticipating the consequences of 12 interview in interviews concerning April and See, for example, the U.K., 11 in an (1988). based on interviews with program leaders or coordinators in Nordic countries S made by Keck a different interpretation (1984), pp. 64-70, 76-80. (1983), p. 417. 11 their own decisions. See Simon 13 The Fifth Generation Computer Program has been widely Succinct analysis can be found in discussed. Levy and Samuels (1989), pp. 66-68; Bloom (1984), pp. 71-72; Anchordoguy (1989), pp. 188- 189. 14 Levy and Samuels (1989), 15 Keck 16 For "government failure" see Wolf (1979). 17 "White elephants" refer p. 64. (1988). to projects which fail the initial expectation of a practical value greater than cost, or more precisely, opportunity cost. 12 Bibliography Anchordoguy, Marie (1988), Computers Cambridge, Bloom, MA: Inc.: Japan's Challenge IBM, to Harvard University Press (forthcoming). MITI as a Policy Instrument Justin L. (1984), Japan's in the Development of Information Technology, Cambridge, MA: Harvard University Center for Information Policy Research. Keck, Otto (1988), "A Theory of White Elephants: Asymmetric Information in Government Support for Technology," Research Policy, Vol. 17, pp. 187-201. Levy, Jonah D. and Samuels Richard Innovation: Japan," Paper J. (1989), "Institutions and Research Collaboration as Technology Strategy MIT MIT-Japan in Science and Technology Program Working 89-02. Nelson, Richard R. (1984), Hight-Technology Policies: A Five-Nation Comparison, Washington, D.C.: American Enterprise for Institute Public Policy Research. Nelson, Richard R. and Winter, Sydney G. (1982), Theory of Economic Change, Cambridge, Simon, Herbert A. (1957), Administrative York, OTA An Evolutionary MA: Belknap Press. Behavior, 2nd edition. New NY: MacMillan. (U.S. Congress Office of Technology Assessment) (1983), International Competitiveness in Electronics, Washington, D.C.: U.S. Government Printing Office. Williamson, Oliver E. (1975), Markets and Hierarchies: Analysis and Antitrust Implications, New York, NY: Free Press. Wolf, C.jr. (1979), "A Theory of Nonmarket and Economics, Vol. 22, pp. 107-139. Failure," Journal of Law \ Date Due ^ ^ (, Lib-26-67