innovation needs patents reform -...

Ž .Research Policy 30 2001 403–423www.elsevier.nlrlocatereconbase

Innovation needs patents reform

William Kingston)

School of Business Studies, Trinity College, Dublin 2, Ireland

Received 15 February 1999; received in revised form 1 September 1999; accepted 18 January 2000

Abstract

The international patent system demonstrates many failures to adapt to new ways of producing information. The comingof antibiotics finally forced it to take account of the shift from AindividualB to AcorporateB invention by establishing newcriteria for patentability. However, these favour industries which can operate a AportfolioB approach in their R&D, to thedisadvantage of other industries and smaller firms in all industries. In complex technologies, whose economic importancehas been growing rapidly, patents are now used as much as a bargaining currency to prevent Alock-outB from use ofstate-of-the-art components developed by competitors, as they are as a stimulus to R&D.

Changes which have been proposed to deal with these problems, including empirical supporting evidence, are discussed.These include compulsory expert arbitration of disputes with legal aid for the respondent party in the event of an appeal tothe Courts, an AInnovation WarrantB as a supplementary type of protection, and Ashared-riskB compulsory licensing as apractical way of changing from time to money as the measure of a grant. This would give multiple innovators access toinventions as early as possible, while maintaining or even improving incentives to invest in invention and innovation.q 2001 Elsevier Science B.V. All rights reserved.

Keywords: Patent reform; Dispute resolution; Compulsory licensing; Second-tier protection

1. Introduction

Recent decades have seen a remarkable growth inthe importance of intellectual property. Among thecontributory factors have been the explosion of soft-ware and the use of copyright and patents to protectit; the growth of biotechnology and the patentabilityof its inventions; the development of electronicdatabases and arrangements for protecting invest-ment in them; the multiplication of patents arising

) Corresponding author. Tel.: q353-1-608-1479; fax: q353-1-679-9503.

Ž .E-mail address: [email protected] W. Kingston .

from University research, encouraged by the Bayh–Dole Act in the U.S. and by its imitations elsewhere;

Žand the setting up of TRIPS the Trade-Related.Intellectual Property Section of the World Trade

Organization in 1994, with the objective of bringingabout world-wide enforcement of intellectual prop-erty.

At the same time, there is an increasingly influen-tial body of writing, particularly associated with

Ž .Reichman and Samuelson 1997 , which argues thatthe system is progressively less capable of meetingthe demands which are being made on it. A typicalstatement of this position is the following:

w xT he nineteenth century vision that subdividedworld intellectual property law into discrete and

0048-7333r01r$ - see front matter q 2001 Elsevier Science B.V. All rights reserved.Ž .PII: S0048-7333 00 00090-1

( )W. KingstonrResearch Policy 30 2001 403–423404

mutually exclusive compartments for industrial andartistic property has irretrievably broken down. Thetheory that the classical patent and copyright mod-els coherently address the way intellectual creationsbehave has been discredited by its inability to dealadequately with the behavior of many commer-cially valuable, cutting-edge intellectual creationsŽ .Kronz, 1983, pp. 178, 180 . These recent techno-logical creations account for an ever-growing shareof the gross domestic products of both developed

Žand developing countries Reichman, 1994, p..2500 .

The present article examines the patent systemfrom this standpoint, finds that it is indeed failing todeliver the protection which could be expected fromit, analyses the reasons for this, and discusses someproposed reforms.

2. Some aspects of patent evolution

The tap-root of modern patent arrangements is theindividualistic philosophy of the 18th century En-lightenment. Another less important root goes backthrough the Venetian system to medieval alpine min-ing grants. These gave temporary monopolies toencourage investment of time and effort, as did thevarious arrangements for granting patents for findingnew technology abroad and bringing it home —Aestablishing new manufacture within the realm,B asthe English Statute of Monopolies of 1623r4 put it.Most of these older systems were effectively dead bythe time of the French Revolution, but the doctrineof the Rights of Man on which this was based alsoprovided a new principle on which patents could berevived. This was that since the State had a duty toprotect individual personality in its physical aspect, italso had a duty to protect its extensions in the formof ideas or creative work.

This was the basis of the copyright and patentacts, which were passed by several of the AmericanStates, as well as of the French patent Act of 1791ŽBugbee, 1967, pp. 67–68; Plasseraud and Savignon,

.1986, p. 46 . When it came to having a Constitutionfor the U.S., however, the earlier kind of patentarrangements also had an influence, because the

Ž .justification of the Article 8.1 which gives Congresspower Ato secure for limited times to authors andinventors the exclusive right to their writings anddiscoveriesB is ATo promote the progress of science

Ž .and the useful artsB Federico, 1990 . In ContinentalEurope, the patent Acts, which were passed in mostcountries during the 19th century, followed theFrench model in using only natural rights as theirjustification, but England took a very long time tograft elements from these rights on to its older—Aestablishing new manufactureB — system. Onereason for this is likely to have been the valuablepart patent protection played in the crucial inventionof the industrial revolution, Watt’s separate con-denser for the Newcomen steam engine.

The evolution of patent law has followed a patternof changes which tend to begin in the U.S., and arethen imitated in other countries. This is partly due toindustrialization causing similar problems every-where, but it also owes something, especially in lateryears, to U.S. influence on and through the Secre-tariat of the Paris Convention in Geneva. This Con-vention, which dates from 1883, was the sole regula-tor of the international aspects of patents and copy-right until the World Trade Organization came intobeing in 1994.

2.1. Growth of inÕention by inÕestment

The Anatural rightsB arrangements worked wellenough as long as the Acreations of the mindsB ofindividuals could be turned into reality by them-selves or by their small enterprises. With the comingof large-scale industry, however, and especially whenapplied science became important, they were pro-

Ž .gressively less able to cope cf. Noble, 1980 . Theearliest problem that arose was how to reconcilepatents with the need to employ individuals to inventin research and development laboratories. In theU.S., the power given to Congress by the Constitu-tion provides for the grant of patents only to individ-uals. How then could a firm invest in employingindividuals to work on research and development ifany resulting patents were to be granted to themrather than to the firm? Because the U.S. has aCommon Law system, this could be solved by adecision of the Supreme Court. In U.S. Õs. Burns in

( )W. KingstonrResearch Policy 30 2001 403–423 405

1871, it ruled that employment contracts could validlyinclude a clause providing for any patents granted toan individual as a result of employment to be as-signed to the employer for a nominal fee. Countrieswith a Civil Law system had to change their patentstatute to achieve the same end, and indeed in Ger-many, the great engineer, Wernher Siemens, wentinto politics to bring about the Patent Act of 1877

Ž .there Heggen, 1975, pp. 115–118 .As more and more inventions emerged from in-

vestment in purposive R&D rather than from spon-taneous creativity, it became correspondingly moredifficult to obtain protection for them from a patentsystem which had been established to enable individ-uals obtain rewards from their discoveries. In alandmark U.S. case in 1851, Hotchkiss Õs. Green-wood, the Supreme Court ruled that Asomethingmore than the work of a skilled mechanicB wasrequired if what was accepted as being both usefuland new was also to be patentable. It took nearly acentury to reach a judicial description of what thisAsomething moreB actually was, but in 1941, thesame Court used the expression Aflash of creativegeniusB for it in the case of Cuno Engineering Corp.Õs. Automatic DeÕices Corp. Moreover, it claimedthat this was the level of ingenuity which had allalong been required for patentability.

Between the two cases, the way in which inven-tions emerged had changed almost completely frombeing the result of individual ingenuity to beingproduced by purposive, large-scale investment inresearch and development. The change can even benoted in differences between the legal actors: InHotchkiss, they are individuals; in the Cuno Engi-neering and two 1944 cases which taught similarlyŽMercoid Corp. Õs. Mid-Continent InÕestment Co.

.and Mercoid Corp. Õs. Minneapolis Regulator Co.they are corporations.

Inevitably, the incompatibility between the oldcriterion of patentability and the new method ofinvention showed up in Court decisions. In 1925–1929, one-third of the patents coming before CircuitCourts of Appeal were ruled invalid, but 20 years

Žlater, this had almost doubled Mintz and O’Rourke,.1978, p. 2:218, quoting Congress Hearings . Those

concerned with patents, either as users or as practi-tioners, blamed this on progressively greater judicialstringency, but it is much more plausible that it was

simply the change in numbers of inventions resultingfrom investment. If a Aflash of geniusB is to be arequirement for patentability, it is much easier todiscern it in the output of individuals than of firms.R&D laboratories are staffed by Askilled mechanics,Band those who direct them cannot allow their em-ployees to sit around waiting for Aflashes of genius.BIt is hardly surprising that the Courts found manypatents invalid for lack of these.

2.2. The antibiotics crisis

The mechanical and electrical industries couldlive with a patent system which was unsuited to theprotection of inventions which resulted from invest-ment in R&D, because they have other ways ofobtaining the protection they need, especially throughthe power to exclude competitors that goes withmassive investment in productive assets and in re-sources for marketing. There is one industry, how-ever, which simply cannot do without patents, andthat is chemicals, especially pharmaceuticals. Theformula discovered in the laboratory is identical withthe formula that is patented and with the formula ofwhat the physician prescribes and his patient takes.Once that formula is known, investment in resourcesto make the product delivers only a trivial barrier tocopying, and if the drug is effective, little persuasionis needed for its sale. Because chemical inventionscan be communicated so easily to others, therefore,if patents did not protect them, free-riding wouldensure that investment in research would not bejustified. This is why it was from a revolution in thepharmaceutical industry that the stimulus for revolu-tion in patents eventually came, and the cause ofboth was antibiotics.

The first of these was penicillin, but this wasnever patented, partly because the authorities inBritain, where it was discovered, considered that itwas wrong that something which was so much to thebenefit of mankind should be a monopoly. The nextone of great importance was streptomycin, and thiswas patented in 1947. Penicillin had undoubtedlybeen discovered through a Aflash of creative genius,Bbut the origin of streptomycin could not have beenmore different. It emerged from Merck’s funding ofthe painstaking examination and testing of more than10,000 kinds of soil microbes for antibiotic activity


by a team led by Selman Waksman of RutgersUniversity, New Jersey; and it started a deluge ofinventions from pharmaceutical firms which wereŽ .and are typically the output of large-scale, routine

Žactivity which is anything but inspired see Kingston,.2000 for a full account .

Nothing could have been clearer at the time thanthat investment in antibiotics R&D could not bemade on the large scale which was desirable unlesspatent protection would be available for its results. Itwas equally clear that the Courts would insist onlooking for the Aflash of genius,B which research forantibiotic inventions on the Waksman model would

Ž .be unable to supply Waksman, 1975 . Conse-quently, there was no option but to change the patentlaw fundamentally. This would require careful draft-ing, because a law which frankly recognized thatinvestment had replaced individuals as the source ofwhat is to be protected, would not have been withinthe power given to Congress by the Constitution.Change could therefore only come about in a waythat forced the reality of inÕention by inÕestmentinto the pretence of inÕention by indiÕiduals.

The year after the streptomycin patent issued, theNew York Patent Bar Association drafted a Bill andwas able to get it introduced in Congress, and this,supplemented by other Bills and pressures, led to thefirst major revision of the U.S. patent law since1870. This was the Patent Act of 1952, and it wasexplicitly intended to make the patent system suit-able for protecting inventions which were the resultof investment in research and development, by killingoff the Aflash of geniusB requirement.

2.3. Defining the A inÕentiÕe stepB

Its first crucial provision from this point of viewwas that APatentability shall not be negatived by themanner in which the invention was made,B so thatthe result of mindless labour became just aspatentable as that of instantaneous vision. The sec-ond was to replace the Aflash of geniusB criterion bymaking it a condition of patentability that the subjectmatter should not be Aobvious to a person havingordinary skill in the art to which it pertains.B Thegap between the patent system and inventions pro-duced by purposive investment, which had beenwidening for more than a century, was now closed.

In the words of a judge who as a patent attorney hadplayed an important part in the shaping of the Act,the result of both these changes is that Along toil

w xstands on an equal footing with flashes of genius BŽ .Rich, quoted in Federico, 1978, p. 1:109 .

From the point of view of how change comesŽabout in intellectual property matters or indeed fails

.to do so , it is worth noting that this Act of Congresswas given its form solely by special interests whichwould benefit by it, firstly the patent attorneys andtheir clients, and secondly, the Patent Office itself,whose bureaucracy would be expanded by growth in

w xpatenting: AThe 1952 Patent Act was written basi-cally by patent lawyers . . . A good 95% of the

w xmembers of Congress never knew that the legisla-tion was under consideration, or that it had passed,

Žlet alone what it containedB Rich, 1978, pp.1:10,.1:11 . There was no mechanism for giving a voice to

those who might be affected adversely by this legis-lation, and none of these did in the event find such avoice.

Of course, all other industrial countries were alsoexperiencing the same progressive incongruity be-tween patent arrangements designed to protect theresults of individual creativity, and inventions whichcame from purposive research and development.They all solved the problem by following the exam-ple of the U.S. and including the inventive step ornon-obviousness criterion in their patent examinationprocedure. Significantly, Japan was the first to dothis in 1959, followed by Sweden in 1967, France in1968, Germany in 1976, and Britain in 1977. Mostimportantly, under the Convention of 1973 whichestablished the European patent, this criterion wasbuilt into examination of applications from the start.As interpreted by Courts in several countries, itmeans that in applying a known technique to adefinable problem, an invention will fail the test ofnon-obviousness if it was Aobvious to tryB what ledto it, or if the trying was done Awith a reasonableexpectation of success,B or if the Achances of success

Ž .were considered worth a tryB cf. Bochnovic, 1982 .

2.4. Use of R&D portfolios

Since it was the pharmaceutical industry’s re-quirements which had brought the issue of protecting


the results of investment in R&D to a head, it ishardly surprising that the U.S. 1952 Act and itsimitations in other countries do in fact give pharma-ceutical inventions very effective protection. So muchis this the case that investments of the order ofUS$100 million can now be rationally made to find,develop, test and market a new drug; that this indus-try is the biggest user of patents in every country;that several countries have given its patents extendedterms to compensate for delays caused by the needfor certification; and that very few disputes in itreach the Courts. What may be considered surpris-ing, since all of this has been built on the non-obvi-ousness criterion, is that this has happened in spite of

w xthe reality that A t he invention of chemical struc-tures which are per se unobvious is an exceedingly

Ž .rare occurrenceB Rowland, 1978 p. 7:201 .This can be understood by analogy with portfolio

theory in finance. Rational investment in a portfolioof risky projects is possible because the risk whichattaches to the portfolio as a whole is statisticallylower than that of any individual project. What savesthe pharmaceutical industry, then, is that the Ascreen-ingB activity which is so characteristic of its researchand development is in fact the operation of a portfo-lio of research possibilities. The research programmeof a large pharmaceutical firm will have several suchAportfolios,B each comprising a large number ofindividual research assignments. It is precisely —and indeed only — because there is a reasonableexpectation of success for each portfolio as a wholethat the firm is able to invest rationally in researchon the large scale necessary. This research will ofcourse follow up a whole series of suggestions fromprior art of advantage to be gained from new combi-nations of it, and there will be a strong expectationof success from somewhere within the portfolio.Such expectation makes rational investment possible.Reverting to the case which was the catalyst for thechange in the patent law, in light of what theyalready knew about penicillin, at the level of theirportfolio of more than 10,000 possibilities Merckwas betting on a near-certainty that Waksman’s teamwould find a valuable antibiotic.

Indeed, the kind of money that it takes to discovera new drug and put it on the market today can onlybe invested on a basis that is pretty close to certainty— in other words, the chances of success have to be

a great deal better than anything which could meetthe inventive step criterion of Anot considered wortha try.B Large-firm R&D is, in fact, rather welldescribed by the words of a British Lord Justice ofAppeal when he excluded from patent protection Aamere business assessment or choice to pursue an

Ž .identified goal by known meansB Hobhouse, 1995 .At the level of the indiÕidual assignments in a

research portfolio, however, a new combination canmeet the criteria of non-obviousness because therewas no Areasonable expectation of successB in tryingthe particular known means which had been used init. In the painstaking examination and testing of soilmicrobes for antibiotic activity from which strepto-mycin emerged, for example, it was a research stu-dent called Schatz who actually first isolated it, andhis name is on the patent for it with Waksman’s. Butsince there was no more Areason to tryB one soilmicrobe rather than another, success could havecome just as easily to any of the other 20 researchassistants in Waksman’s team, and what Schatz foundwas not Aobvious to one skilled in the Art.B

3. Investment and technological complexity

An important aspect of the shift to inventionsresulting from purposive investment in R&D is theincreasing complexity of technologies. Roycroft and

Ž .Kash 1999, p. 262 define a complex technology as:

a process or product that cannot be understood infull detail by an individual expert sufficiently tocommunicate all details of the process or product

Žacross time and distance to other experts. A simpleproduct or process is one that can be understood or

.communicated by one individual .

They have also identified the growing economicimportance of complex technologies. Between 1970and 1995, the percentage of the world’s 30 mostvaluable goods exports represented by complexproducts and processes grew from 53% to 78%Ž .Roycroft and Kash, 1999, p. 8 .

3.1. The role of patents in complex technologies

In contrast to simple technologies such as chemi-cals, for which they are indispensable, patents are


inherently disadÕantageous for complex technolo-gies. The primary reason is that if competing firmshold patents on different components of a complex

Žtechnology, and they fail to cross-license them whichcan happen from many causes, not all of them

.rational development in an entire industry can beŽslowed down or even rendered impossible cf.

.Merges, 1994 .As complex technologies emerged, therefore,

firms devised arrangements to limit such harmfuleffects of patent monopolies, and the most importantpatent-using industries such as automobiles, aircraftand radio, grew on the basis of ApooledB patents.Member firms transferred all their patents to thepool, which in turn allowed all members to use theinformation freely, sometimes subject to a royaltythat was assessed by a committee. However, it isclear that the real pay-off for pool members wastheir freedom to use the technology of all othermembers without the need for negotiation or evencostly litigation, rather than any payment receivedfor their own contributions to the pool. Pools alsofacilitated transfer of non-patentable technical infor-mation between members, as well as the setting ofstandards for components.

3.2. Patents as a trading currency

Perhaps inevitably, formal patent pools becamesuspect as aspects of anti-competitive behaviour intheir industries, and in the U.S., Federal Anti-Trustpolicy brought all but the least important of them toan end. However, firms in complex technologies stillfaced the same problem, which is their need tomaintain access to and freedom to use state-of-the artknowledge. Delay in obtaining access to a compo-nent of a complex technology is of course especiallydamaging to firms in technologies with a short life-cycle, where the first firm to the market has everychance of setting the standards and even of lockingthe others out. Since they cannot be members of aformal pool, firms now patent as many componentsor sub-components of a technology as they can, so asto build up the strongest possible position for cross-licensing. The motivation for their extensive use ofpatents is therefore quite different from that of firmsin simple technologies. In the latter, the emphasis

Žmay be said to be primarily offensive to prevent.others from using the invention ; in complex tech-

Žnologies it is primarily defensive to avoid being.denied the use of an invention .

This approach by firms in complex technologiesmeans filing for patents on large numbers of incre-mental innovations. On the face of it, most of theseshould not qualify for patent protection, since it is inthe nature of incremental changes that they emergein an organic way, out of what is there already. Eachchange points toward the next, and is correspond-ingly Aobvious to one skilled in the Art.B Indeed,this is one of the identifying marks of the normalinnovation trajectory in a complex technology, char-acterized by incremental improvements. It is possibleto speculate, therefore, that it is only because theyare part of a complex technology, which itself is notunderstood as a whole, that such small changes areable to meet the non-obviousness criterion forpatentability.

The widespread use of patents as bargaining chipsin complex technologies has logically developed intoarrangements between competitors to pursue someinnovation projects together, at least in the earlystages. In the U.S., there is even a National Coopera-tive Research Act to control these, and 450 collabo-rations were registered under it for the period 1985–1994. Nearly half of the more than 500 U.S. indus-tryruniversity research institutes have up to five

Ž .firms as members Mowery, 1999, pp. 44–45 . Somecompanies sign broad trading agreements with com-petitors that cover periods of several years. Thisoften leads to common industry standards, definedby patents, even to the point where the inventors of atechnology may even offer low- or no-cost licenseson their patents when they want to have their designused broadly or universally.

3.3. Are patent pools more A pro-competitiÕeB?

The use of patents as trading currency shouldhave overall economic benefit because it reducestransaction costs, and indeed, at the limit, no firmcould take the risk of innovating a new complextechnology unless it was sure it could trade in thisway. It should be noted, however, that widespreaduse of patents as a mechanism to deal with Anti-Trust


disapproval of patent pools has probably increasedboth barriers to entry and transaction costs, contraryto the Authorities’ intentions. If a pool is in exis-tence, firms in it will only patent enough inventionsto maintain membership of the pool, or the inven-tions, which they think are significant, and thereforelikely to earn them some automatic royalties fromother members. Since they do not have to fear beinglocked out from using improvements made by otherpool members, they are under no pressure to patenttheir own minor Ainventions,B to build up bargainingpower.

If using a pool is outlawed, however, firms areforced to patent everything which could possiblyhelp them in the slightest way against being lockedout by a competitor’s patent from an incrementalimprovement which they might want to use in thefuture. Consequently, patent filing, prosecution andmaintenance costs are much higher than if a poolexisted, and to these must be added the costs of

Žnegotiating cross-licenses with competitors also ab-.sent in a pool . Moreover, by definition, complex

technologies are large-firm technologies. Such firmshave little need for patents to protect the incrementalresults of their R&D, because they can do this byhaving the productive resources to be first into amarket with an improvement, and then using market-ing on a large scale to fend off rivals. Consequently,it is doubtful if the level of R&D investment incomplex technologies would be significantly reducedif the patent system were to be abolished altogether.

It is even questionable whether granting of patentsfor the ways in which they are used in these tech-nologies is in line with the patent system’s objectiveof Apromoting the progress of science and the usefularts.B This is because of the barriers to entry theyunderwrite, which are analogous to the Acovert car-telsB which exist in markets where a comparablystrong barrier to entry is provided by very large

Ž .advertising expenditures cf. Kingston, 1984, p. 66 .De facto patent pools of the anti-competitive typehave returned through the back door. Strategies ofAsaturation patentingB designed to slow down oraltogether prevent competitors from exploiting alter-native technological trajectories must surely beagainst the public interest. So is the threat of heavylitigation costs, which are often used against finan-cially weaker competitors.

4. The high cost of resolving patent disputes

Such threats, which do not have to be explicit tobe effective, reflect further major distortion in theoperation of the patent system. Patents should be asource of market power for firms which otherwiselack it, but this objective is defeated if enforcementcosts are too high. Before the introduction of theinventive step test, much litigation was either avoidedaltogether or aborted at an early stage by patentees,because of their perception that their chance of ob-taining a ruling that a patent was both valid andinfringed was so low. The much better prospects forpatentees from the replacement of the Aflash ofgeniusB criterion by the inventive step, both in-creased patent numbers and encouraged patentees totake infringers to Court.

In the U.S., this movement was undoubtedly ac-celerated by the designation of the Court of Appealsfor the Federal Circuit in Washington, DC in 1982,as the forum for all patent cases. Before this, in-

Žfringers could practise Aforum-shoppingB as theycan still do in Europe under the Brussels and Lugano

.Conventions whereby they could arrange for a caseto be taken in a jurisdiction likely to favour them.This was a considerable deterrent to the owners ofpatents to appeal a case from a District Court deci-sion. There is little doubt that patentees have faredmuch better from the CAFC, which has even led to aperception in some quarters that this Court has beenresponsible for lowering the standard of patentabil-ity. It is far more plausible that it has done no morethan accept the logic of the need to protect inventionresulting from investment, and not just from theindividual creativity towards which earlier patentarrangements were directed.

These changes caused a marked growth in patentapplications and an estimated 50% increase in litiga-

Ž .tion during the 1980s Merz and Pace, 1994 withinevitable rapid inflation in litigation costs. Lernerhas assembled data from several sources to estimatethat in the U.S. these costs are now more than aquarter of what is spent on industrial basic researchŽ .1995, p. 470 . Just how serious this problem is, hasbeen officially recognized in a recent Report of theU.S. Advisory Committee on Patent Law Reform.No fewer than 36 out of 197 pages of this aredevoted to the cost of enforcing patent rights, not


only financially, but in terms of delay, which itconsidered to be particularly urgent:

the delays and complexity of modern patent litiga-tion have provided overly aggressive lawyers withan ample assortment of opportunities to stall, delayand harass. The net of these factors is that patentlitigation has become an increasingly inefficient,ineffectiÕe and undesirable means of resolÕing

Žpatent related disputes 1992, p. 78, emphasis.added .

The Committee’s main conclusion could not havebeen more damning of the present situation in theU.S. It feared that

unless the problems of cost and delay in patentlitigation are addressed now, the central purposeof the patent system to provide an effective incen-tive for development and commercialization ofnew technology will be seriously eroded. Such anerosion could well prove a threat to the very

Ž .existence of the patent system . . . 1992, p. 76 .

In addition to the measurable costs of litigation,there are many others that are very substantial andthat certainly have important social implications.These are the burdens in terms of distraction, diver-sion of energy, and misdirection of creativity thatany intellectual property dispute currently imposeson innovatory firms.

4.1. Particular disadÕantages of smaller firms

High litigation costs are particularly destructive ofthe contributions to innovation that smaller firmshave proved they can make. It is obvious that themeasurable costs of prosecuting or defending anaction for patent infringement are far beyond theresources of all but the largest firms, apart from thefact that the burden of the costs that cannot be

Žmeasured such as distraction from more immedi-.ately paying tasks falls most heavily on smaller

ones. Smaller firms are especially inhibited frominvesting in innovation through fear of cripplingcosts to defend any intellectual property they might

Ž .own for actual cost levels, see Bouju, 1988 .Neither do they have the resources to enable them

to use the AportfolioB approach, which virtuallyguarantees that the results from large-scale invest-

ment in research and development in certain tech-nologies will obtain patent protection. Inventiveowners and managers in small firms can only affordto try combinations of prior art if they think there issome hope of success, so that their inventions conse-quently often fail the Anon-obviousnessB criteria.Even when they are granted a patent, they often faceintimidation by larger firms in seeking to use the lawto enforce it. Such intimidation, it must be stressed,does not have to be explicit to be effective. Aware-ness of the danger of crippling legal costs can befatal on its own to the use of the classical patentsystem by knowledgeable management in suchsmaller firms.

In public interest terms, these drawbacks to use ofpatents by small firms are serious, because thesefirms have shown that they have quite disproportion-ate ability to invent and innovate. In the U.S., theyreceive less than 4% of Federal support for research,yet they produce more than half of the innovations

Žand get close to two-fifths of all patents State of.Small Business Report, 1997 . These figures

doubtlessly flatter the inventive power of small busi-nesses, because large ones must hold a higher pro-portion of the more valuable patents. Nevertheless, itis a reasonable assumption that substantial economiclosses are being incurred in every developed countrythrough lack of appropriate protection for the inven-tions and innovations of smaller businesses.

5. Summary of diagnosis

From the foregoing, it appears that the patentsystem:

v operates very well indeed for firms in the chem-ical industries;

v operates poorly in industries whose firms cannotuse a portfolio approach in their R&D;

v is disadvantageous to complex technologies andis scarcely needed by them for invention protec-tion, but they nevertheless use it heavily toprevent being Alocked outB from access to com-ponents competitors’ technology;

v has very expensive dispute resolution arrange-ments; and partly for this reason,


v it serves small firms, which have most need ofeffective protection for their inventions, particu-larly badly.

6. Responses to patent system failures

6.1. Use of copyright to fill gaps in patent protection

As might be expected, the failure of the patentsystem to protect inventions in certain areas has ledto use of alternative arrangements to try to fill thegaps. The most striking illustration of this is com-puter programs. When patent protection was soughtfor the earliest of these, it was denied by the U.S.Patent and Trademark Office, not on strict legalgrounds, but simply because the Office did not havethe resources to cope with the volume of applicationsit envisaged would follow. The Court of first in-stance rejected the Office’s decisions, but theSupreme Court supported them, so for many years,developers of such programs could only obtain copy-right protection. Recently, this Court changed itsstance, so protection of such programs by patents hasbecome increasingly possible in the U.S., althoughnot elsewhere to the same extent.

Copyright is also being adapted to protect func-tional designs, as under British legislation which willbe discussed below. This can also be seen in aresponse to patent system failure in the U.S. relatingto manufacture of fibreglass yacht hulls. Consider-able originality, skill and investment is needed tooptimize a modern hull configuration, which will bethen be expressed in the AplugB or mould fromwhich multiple copies can be produced. However,this design and research work is not carried on in away which can enable it to meet the Anon-obvious-nessB criterion and get a patent. Since the hullconfiguration is consequently not protected, all afree-rider needs to do is buy a single example of asuccessful design from its originator, shape his ownAplugB from this, and proceed to make and sellperfect copies without having to make any of theinvestment it took to develop the design, or to sharein its risk.

The State of Florida rightly considered that thisdiscouraged investment in boatbuilding and passed alaw protecting vessel designs including their Aplugs.B

However, this was struck down by the SupremeCourt, using the words Athe federal patent laws mustdetermine what is protected, but also what is free for

Žall to useB Bonito Boats Õs. Thundercraft Boats,.1989, p. 141 . The response in this case has been the

AVessel Hull Design Protection Act,B passed by theHouse of Representatives and sent on to the Senatein March, 1998. This is an amendment to the Copy-right Act, and echoes some of the wording of Britishfunctional design protection, e.g., the hull designmust not be Astaple or commonplace.B

The Vessel Hull Design Bill and its discussion inCongress reveal some of the factors which makepatents reform so difficult. In the 19th centuryparadigm, patents protect AfunctionB and copyright

Žprotects Aexpression.B It was essential especially in.light of the explicit Supreme Court ruling that the

new protection should be located in Copyright, thatis, in Aexpression.B To achieve this, the bill containsa clause that Hull Design protection shall not applyto any design which is Adictated solely by a utilitar-ian function of the article that embodies it.B It hasindeed been claimed that Athe perfecting of a type ofobject mechanically is evidenced by its beautyBŽ .Gilfillan, 1935, p. 8 but this is hardly what is atstake here: like an aircraft’s wing, it is function andnot aesthetics which determines the shape of a ves-sel’s hull.

As expressed in the House discussion, the U.S.Patent and Trademark Office did not have a formalview on the bill, Abut as a general policy, they prefer

Ž .not to enumerate subgroups of patentsB Scott, 1998 .The clear implication of this statement is that theOffice likes to think that the existing patent system isable to provide protection for all technologies, anattitude they share with Patent Offices in other coun-tries. It is manifestly not correct. As explained above,the system is strongly biassed in favour of technolo-gies whose R&D can use the portfolio approach,primarily the pharmaceutical industry. It gives corre-spondingly poor protection to firms that cannot usethis approach to meet the inventive step criterion,above all, the small firms whose inventions most

Ž .need patent protection. As Thurow 1997, p. 103has argued in the wider context,

The world’s one-dimensional intellectual propertysystem must be overhauled to create a more differ-


entiated one. Trying to squeeze to-day’s develop-ments into yesterday’s system of intellectual prop-erty rights simply won’t work. One size does not fitall.

6.2. A Second-tierB patent protection

Several other attempts to fill gaps in protectionresulting from the way in which the patent systemhas evolved are categorized as Asecond-tierB patentprotection. These often build upon existing arrange-ments for Autility modelsB or Apetty patents,B now in60 countries, of which at least 12 have been estab-

Ž .lished since 1980 Janis, 1999, p. 153 . Britain hadlegislation for utility models as early as 1843 andGermany introduced its system in 1891. Both wereintended to protect the external configurations of, forexample, handtools, and were not concerned with thefunctional aspect of these. The British system lapsed,and German Gebrauchsmuster protection was usedonly to a limited extent. However, in 1987, Britainintroduced functional design protection on a copy-right rather than a patent model in that it is auto-matic, requiring neither application nor fees, and ofcourse there is no examination. The criterion ofnovelty is that the subject matter is Anot common-placeB and protection applies to first marketing withinthe European Union as well as to first design. Theterm is 15 years from design or first marketing,whichever is shorter, and during the final 5 years ofthe term, compulsory licensing applies, with theComptroller of Patents as arbitrator if the partiescannot agree royalty terms.

In 1979, Australia introduced Apetty patentB pro-tection, giving a short-term of protection withoutexamination, but with the same Anon-obviousnessBcriterion as regular patents. This feature ensured thatit was scarcely used at all. In 1997, proposals toremedy this were brought forward in the form of anAInnovation Patent,B which would give protectionfor 8 years without examination, with a much re-laxed novelty criterion.

In 1990, Germany upgraded its utility model ar-rangements so as to make the range of what they canprotect the same as regular patents, and in 1992,Ireland introduced a Ashort-term patentB to be grantedfor 10 years, with application but without examina-

tion unless requested. The criterion of novelty, thatthe claimed subject matter should not Aclearly lackan inventive step,B is substantially that of the British1949 Act, before that Country adopted examinationfor non-obviousness in 1977. Ireland’s examinationof applications for regular patents has looked for aninventive step since 1964.

The authorities of the European Union have beendistracted from concern with the broader inadequa-cies of the present patent system by their difficultiesin bringing the Community Patent Convention intolife. This is intended to give a unitary patent for allmember-countries, with a single Court to deal withdisputes. Nevertheless, in 1995, they started to pro-duce proposals for harmonizing utility model protec-tion throughout the EU. In their latest form, these arefor a 6-year term with two possible renewals each of2 years, with no examination of applications. Biolog-ical inventions and chemical processes would beexcluded from the scope of protection, as wouldcomputer programs. The alternative to the non-obvi-ousness criterion of regular patents is that the sub-

Ž .ject-matter should disclose either a particular effec-tiveness in terms of, for example, ease of application

Ž .or use, or b a practical or industrial advantage.

7. Further reforms needed

The evident motivation for much of the develop-ment of second-tier protection is to try to repairsome of the faults in the regular patent system, butwhat has so far been proposed for second-tier patentarrangements falls well short of what is needed.However, three broad changes, two of which alreadyhave support from empirical research, could go along way towards making regular patents and sec-ond-tier protection into an effective combination,capable between them of operating equitably and

Žapplicable to all technologies. These changes in.order of likely difficulty of introduction are:

v Compulsory arbitration of disputes, with legalaid for the respondent party in the event of anappeal to the Courts from an arbitration deci-sion;

v Direct protection of innovation, in the form ofAInnovation Warrants,B as the best possible sec-ond-tier system, to provide protection in tech-


nologies where the portfolio approach cannot beused well enough in R&D to obtain validpatents; and

v Changing from time to money as the measure ofthe patent grant.

8. Reducing the cost of resolving patent disputes

8.1. Penalty of using ordinary courts

An important reason for the high levels of patentlitigation cost is use of the same Court arrangementsfor settlement of what are at heart technical issues, asfor non-technical disputes. Even in a specialist courtwith some judges who have a scientific or engineer-ing background, such as the Court of Appeal for theFederal Circuit in Washington, DC, the judges stillhave to be educated in the specific technology of

Žeach case by teams of attorneys the most highly.paid teachers in the world . This education also has

to be provided in duplicate to the judges by theseexpensive mentors on behalf of both parties to adispute.

This AeducationB of judges rarely reaches thelevel where even the most self-confident of them feelable to pit their own technical assessment againstreally expert witnesses. Consequently, the advantageis with the party that can afford to have the most

Žauthoritative witnesses who also tend to be the most.expensive on its side. As soon as the possibility of a

dispute between pharmaceutical firms appears on thehorizon, for example, a race begins between them toretain the services of the most prestigious academicsin the particular field. It is surely significant that the

ŽU.S. Commission on Patent Law Reform 1992, p..95 explicitly urged measures to reduce the use of

experts and to diminish their role as advocates inpatent cases, recommending that judges require Aastrong showingB of need before authorizing testi-mony from expert witnesses.

8.2. AlternatiÕe dispute–resolution methods

Because of the scale of litigation costs, a series ofalternative dispute resolution means has developed,in attempts Ato resolve disputes more quickly, in aless costly manner, and in a way that will reduce theburden on the judiciary — without sacrificing jus-

Ž .ticeB Creel, 1987 . The most important of these isexpert arbitration, and within the last few years, forexample, the World Intellectual Property Organiza-tion has set up arrangements for arbitration of dis-putes, but this has been very little used. The obviousreason is that arbitration is Õoluntary, so that it canbe used only when both parties agree on it as a wayof settling their dispute. In turn, such agreementinvariably reflects comparable levels of economicstrength. Arbitration is not used in intellectual prop-erty cases where the cost of litigation intimidates —and is indeed intended to intimidate — attempts by aweaker party to obtain justice. Deliberate policydecisions to infringe patents and to use the courtsystem to postpone the day of reckoning until adominating market position has been built up arefrequent. Large firms have enormous opportunities atpresent for imposing burdens on their opponents inlitigation in terms of delays and discovery actions.As an example, when E.M.I. invented and innovatedthe brain scanner — Athe most important advance indiagnostic technique since X-raysB — their patentswere deliberately and quickly infringed by GeneralElectric and Ohio Nuclear.

Although E.M.I. did win its actions long after-wards, the royalties which it obtained did not re-motely compensate it for having been driven out ofthe business that it had pioneered, an event thatoccurred within 7 years. There can hardly be anyarea in which the dictum that Ajustice delayed isjustice deniedB applies with more force than in intel-lectual property, because lead time is such an effec-tive alternative way of protecting investment in inno-

Ž .vation cf. Levin et al., 1987 .Since at present a patent is no more than a license

to litigate, a great need exists for some means ofŽresolving disputes that will be swift especially be-

cause the law’s delays hurt smaller firms more than.large ones and will also eliminate differences in

financial resources between the parties as far aspossible. If Awar is too important to be left to thegenerals,B then equally, patents are too important tobe left to lawyers and judges.

8.3. Compulsory arbitration

These problems could be solved by compulsoryŽarbitration with a special feature added Kingston,


.1995 . This would be carried out by an expert orexperts selected on an ad hoc basis for each case outof the technological and academic community. Suchpeople would need no education in the state-of-the-art, could deal with the parties without intermedi-aries, and could deliver a quick decision. Persuasiveempirical support for this approach has emerged as aby-product of research in the AInterferenceB recordsof the U.S. Patent and Trademark Office. Interfer-ences exist because of the explicit provision in theConstitution, referred to earlier, that patents andcopyright can only be granted to indiÕidual inven-tors and authors.

This provision makes it necessary for the Officeto have procedures to ensure that a patent is grantedonly to whoever can be identified as Athe true andfirst inventor.B Consequently, when it is noted that

Ž .two or more applications that might possibly be forthe same inventive entity have been filed, an Ainter-ferenceB is declared, and all parties must provideevidence as to their respective dates of Aconceptionof the inventionB and of their efforts to reduce it topractice. The Board of Patent Appeals and Interfer-ences, which decides on the evidence submitted, isrecruited from the most experienced members of thePatent Office’s Examiner Corps.

Interference procedure, therefore, is precisely asystem of compulsory arbitration of patent disputesby experts. It is an actual large-scale working prece-dent or model for the means of resolving suchdisputes now proposed. The Board delivers between40 and 50 final decisions each year. Just over one-third of these are appealed to the courts, but only 5%of them are either wholly or even partially reversedŽ .Calvert and Sofocleous, 1989, 1992, 1995 .

Another valuable feature of the interference modelis the high proportion of cases that are settled volun-tarily. Out of a typical 61 cases studied, for example,no less than 14 ended in agreement between theparties. It is not possible to be certain of the condi-tions that generally apply in such agreements, but afew cases where agreement texts have been found inthe files suggest that — as one might expect — theparties cross-license their inventions. The result isthat any new technology arising from these caseswill reflect competitive development effort by atleast two firms. Such a technology can be expectedto advance more quickly than others, because each of

the parties would explore its own particular Atrajec-toryB of improvement possibilities.

All in all, therefore, in terms of how few of itsdecisions are successfully appealed to the courts andof how many of its cases reach voluntary settlement,the performance of compulsory expert arbitration inthe interference procedure is a very good augury ofhow well a similar system could work for the settle-ment of disputes after grant of exclusive privilege.Any arrangement that works 95% of the time, afterall, is giving at least as good a result as we areentitled to hope for in human affairs.

Extending compulsory arbitration to post-grantdisputes would depend on being able to find individ-uals who need no education in the state-of-the-artand who are also completely independent of theparties to a dispute. The record of the AdministrativeJudges in the U.S. Patent and Trademark Officeshows that they would be eminently capable ofperforming this task. There is no reason to think thattheir opposite numbers in, e.g., the European PatentOffice, would not be equally able to deal with post-grant issues. In order to stress arbitrators’ indepen-dence, the special Board might be administrativelyand even physically separated from the Patent andTrademark Office. To ensure that the arbitratorswere always up-to-date in the state of their particularArts, they might be seconded from their Office to theBoard only for limited periods, or even for particularcases in relation to which their expertise is particu-larly relevant.

8.4. Appealing arbitration decisions

Since any form of arbitration must allow forappeal to the courts, it might be considered that thisproposal only pushes the problems of cost and intim-idation back a stage. This objection could be dealtwith by providing legal aid for the party which hadaccepted the arbitrator’s decision, i.e., the respon-dent in an appeal case. It is most unlikely that thiswould cost very much. No small- or medium-sizefirm would ever appeal, both because it would lackthe resources to litigate and because to do so wouldgive a gratuitous advantage to its opponent by shift-ing the ground of the battle to where the latter islikely to be stronger. Large firms would also seemany convincing reasons for not appealing, once


they were no longer able to bring their financialadvantage to bear, including the following:

v Courts invariably give a lot of weight to theoutcome of expert arbitration, as the Interferencefigures show. The odds for an appellant must there-fore be against winning the legal battle, having lostthe arbitration.

v In present circumstances, even the threat oflitigation will almost certainly force a small firm tocapitulate. It is quite a different matter if such a firmwill be provided with the resources to defend itselfin court because it has accepted the arbitration award.Large firms will be reluctant to fight an opponentwho has now effectively been made Atheir ownsize.B

v At present, intimidation is not evident to thepublic. With the arrangement proposed, there wouldbe more transparency. A large firm would likely bereluctant to have it known that it is using its financialstrength against a smaller one after an independentarbitrator has ruled in favour of the latter.

v Most litigation is entered into in the expectationthat it will end in compromise. Since the firm thathas accepted the expert arbitration would not have topay any subsequent legal costs itself, it would beunder less pressure to settle the case out of court.This would greatly add to the AdownsideB of thepossible outcomes that a prospective appellant wouldhave to take into account when deciding whether ornot to appeal.

v Quick decisions, such as could be obtained bycompulsory expert arbitration, may be found to beeven more valuable to large firms than to smallerones because their range of innovative activities islikely to be correspondingly wider. Consequently,they may see it as being in their interest to supportthe proposed system, specifically by adopting a for-mal policy of not appealing decisions to the courts.

8.5. Paying for arbitrations and legal aid

Any costs of the proposal should be regardedexactly as a counterpart to the necessary expense ofconventional policing of other kinds of property.Every property right system involves policing by theState. If patents are to be real rights of property,therefore, the State cannot escape having to spendmoney on their protection.

The near-certainty that no small firm would ap-peal from an arbitration, and the likelihood that largefirms would be cautious about doing so, supportedby the extremely low proportion of successful ap-peals to the Courts from Interference arbitrations,also have another implication. This is that commer-cial insurance cover against costs, in the first in-stance of appeals, and subsequently of arbitrationsalso, might be obtainable. Patent litigation insuranceup to now has been merely cosmetic, and largelyuseless to small firms, because it is so hedged aboutwith limitations, but if the figures were seen tojustify it, it would not be long before some insurerwould make it available. In that event, the legal aidaspect of this proposal might become redundant.

8.6. A Õoluntary patent pool?

An interesting recent response to the excessivecost of litigation in patent disputes is the NationalPatent Board, which was established in the U.S. onthe initiative of Procter and Gamble. At the time ofwriting, this has 40 members, all large firms, whichagree that any dispute between them will be dealtwith by alternative resolution means, such as media-tion or arbitration. If a member firm decides tolitigate the issue subsequently, it must then assumeall the costs of the earlier procedures.

Research for the European Commission is cur-rently investigating a variant of this for small firms.This would in effect be a voluntary patent pool. Afirm which joined would accept compulsory arbitra-tion of any dispute with another member. Members’fees would provide the pool’s Direction with re-sources to undertake litigation in cases where anon-member is in dispute with a member and refusesarbitration. It is believed that as a track record ofsuch litigation was built up, intimidation would bereduced correspondingly. A potential infringing firmwould not know whether or the dispute in which it isinvolved would be taken up for litigation by thepool’s Direction, but it might well consider that therisk of this was enough to make it prudent to negoti-ate a license.

A particular advantage of a voluntary pool, ofcourse, is that it could be established without legisla-tion. Also, because of the size of its member firmsand the diversity of their markets, there could be no


question of such a pool being judged as anti-compe-titive.

9. An Cinnovation warrantD

A proposal for this was explored in detail for theŽ . Ž .European Commission 1996 Kingston, 1987 and

research since then has refined its approach to pro-tecting innovation directly, instead of indirectly,through whatever protection a patent is able to giveto its related invention. It is a type of Asecond-tierBpatent protection, but it will be shown below that it

Ž .shares none of the disadvantages which Janis 1999has found in other types of this, either actual orproposed. Its objective is to provide protection forany kind of information that results from investment,in the following ways:

v The criterion of novelty would be the strictlycommercial one of Anon-availability in the ordinarycourse of trade.B Anything, not just technology,which is not so available, but could be made avail-able by investment, would be entitled to protection.Although the principle of granting protection forinvestment in innovation per se without any Ainven-tive stepB or creativity requirement was consideredradical in 1987, it now underpins the EuropeanUnion’s 1996 Database Directive.

v Applications would be published immediatelyon the Internet, and their examination would relyheavily on information resulting from pre-grant op-position.

v The grant of exclusive privilege would be irre-Õocable for a period, to make it a sound basis forinvestment. Who would take the risk of drilling foroil if any license granted could be made worthless asa result of new knowledge of the seismic structure ofthe allotted block becoming available? Yet, this is ineffect all the security a patent gives to an investor atpresent, since no Patent Office can guarantee that itssearch is absolutely comprehensive, and judicial de-cisions cannot be forecast with any certainty.

v The term of the privilege would vary accordingto the risk of the investment, i.e., whether it relatedto a radical innovation or to an incremental innova-tion or to something in between. It would also varyaccording to the meaning of the level of risk for theapplicant. A proposal requiring investment that could

bankrupt a small firm in the event of failure, forexample, might scarcely be noticed in the accountsof a multinational. Terms in general would be signif-icantly shorter than for patents at present, but froman investor’s point of view a short term of reallyeffective exclusivity is worth a lot more than anominally longer one that is uncertain. Replacementof the measure of the grant of exclusive privilege interms of time by a money measure, to be discussedin a later section, would be especially appropriate.

9.1. Warrants tested against Janis’ arguments

9.1.1. Presumption of ÕalidityJanis stresses that it is characteristic of second-tier

arrangements that applications are not examined. Theresulting lack of any presumption of validity elimi-nates speedy interlocutory injunctions against an in-fringer. Some regimes provide for examination whencalled for, but this still involves delay. Inability toobtain injunctory relief can be disastrous for a smallfirm, especially if the infringer has strong marketingresources.

Innovation Warrants, in contrast, would have sucha presumption of validity, both because applicationswould be examined, and because of the nature oftheir novelty criterion, which is based on fact ratherthan judgement. This would be helped by pre-grantopposition, in spite of the bad reputation whichattaches to this procedure. The U.S. has never had it;in Japan, it is notorious as a non-tariff barrier toimports; and industry’s view of the European PatentOffice version is that it Amust be judged a failure.An unacceptable number of years often elapses be-fore a final decision, which is then often the begin-

Ž .ning of national Court proceedingsB Yorke, 1999 .An aspect of these delays is the unwillingness of thatOffice to admit that Aone size does not fit all,Breflected in the domination of its procedures by therequirements of its biggest user, the pharmaceuticalindustry. Since this industry has to carry out lengthytests for opposition purposes, periods for responsewhich are appropriate for it, are far longer than isneeded for other technologies.

With a novelty criterion of Anon-availability inthe ordinary course of trade,B the factual informationneeded for a decision is easily collected. Moreover,for this purpose, the best imaginable search engine is


the vigilance of firms likely to be affected competi-tively by grant of a Warrant. Publishing new applica-tions on the Internet would facilitate their ease ofmonitoring. The combination makes it possible forboth opposition and examination periods for a War-rant to be very short indeed, and for its grant to beirrevocable, at least long enough for a larger firmwhich takes a license on the invention, to gain AfirstmoverB advantage on the market for it. Since it is inthe interest of society that there should be manyinnovating firms, the irrevocable part of the Warrantterm might be longer if a small firm is exploiting theinvention itself, since seed and venture capitalistswould require this.

9.1.2. ATacitB and A explicitB knowledgeŽ .Janis 1999, p. 171 describes the criteria for

grant of protection in second-tier patent systemsgenerically as Asoft obviousness,B and foresees muchdifficulty in resolving disputes over these in practice,especially in the case of the proposed EU UtilityModel. The same factors as would give the Innova-tion Warrant its presumption of validity would alsogive it an advantage in this respect. The criterion for

Žgrant of a Warrant Anon-availability in the ordinary.course of tradeB leads to protection of tacit informa-

Žtion defined by Polanyi, 1967 as Awhat we know.but cannot tellB as well as the explicit information

which is all that the classical patent system tries toprotect. Tacit information is generally discussed interms of the practical knowledge and skills of thosewho make a new product, which are only transfer-able with difficulty, when indeed they are transfer-able at all. A point to note is that this includesknowledge of what does not work as well as whatdoes. Both these types of tacit information embodiedin a new product have been obtained as a result ofthe investment and risk of the originator, and conse-quently deserve to be protected just as much as therelated explicit knowledge.

9.1.3. A AnticommonsB propertyA third point made by Janis against second-tier

Ž .patents is the danger of what Heller 1998 callsAanticommons property.B This is the opposite offreedom of access to a resource, which leads to itsoveruse and consequent destruction, as in the case of

the Newfoundland cod fishery. In an Aanticommons,Bso many individuals have rights that the resourcecannot be effectively used at all. If these rights arepatents, then the AblockingB effect already discussedin relation to complex technologies can come into

Ž .being cf. Lemley, 1997 . Resources will then beunderused because of Aintellectual property rights infuture products or by permitting too many upstream

Žpatent owners to stack licensesB Heller and Eisen-.berg, 1998, p. 699 .

There are two safeguards against this in the Inno-vation Warrant proposal. The first of these is thatsince the function of the Warrant is to make itpossible to invest rationally in innovation, it lapses ifthat investment is not made; this makes perverse

ŽAblockingB i.e., in order to exact ransom from afirm which is committed to make such an invest-

.ment impossible. Secondly, the monopoly periodwould be relatively short, and would be succeededby a further period when compulsory licensing wouldapply, as in the British functional design protection.In relation to a similar proposal to protect computer

Ž .programs Reichman 1994, p. 2540 finds that Aperi-ods of 2 or 3 years of lead time have particularappeal.B

9.1.4. Smaller firmsŽ .Fourthly, Janis 1999, p. 179 calls attention to

the extent to which interest in second-tier protectionis caused by concern that the inventions of smallerfirms do not have access to protection from theregular patent system. This is, of course, perfectlytrue, and Janis is also right to point out that from itscost aspect, far too much attention has been paid toacquisition costs. The potential costs of enforcementare greater by an order of magnitude, so that it is onthese that attention should be focused. The specialform of compulsory arbitration discussed in Sections8.3 and 8.4 above is the appropriate response to thisproblem, and has been part of the Innovation War-

Žrant proposal from the outset cf. Kingston, 1987, pp..328–331 .

9.1.5. Other comparisonsFinally, Janis explains why second-tier patents are

unable to justify the claims made for them that theywould give protection to non-traditional technologies


Ž . Žp. 190 and to products with short life-cycles p..188 and that they are more appropriate to develop-

Ž .ing countries than regular patents are p. 194 . TheŽInnovation Warrant’s novelty criterion Anon-avail-.ability in the ordinary course of tradeB clearly ap-

plies as well to non-traditional technologies as totraditional ones. Internet publication of applications,combined with that novelty criterion and the shortperiod allowed for opposition, would use up less ofany product’s potential life-cycle than any otherarrangement. For developing countries, the Warrantshares with other second-tier systems the advantageover patents that adaptation of its general approach

Žto local requirements is not restricted by TRIPS theintellectual property aspect of the World Trade Orga-

.nization Agreement .

10. Money instead of time as the measure of apatent grant

The third proposal is the most radical, but corre-spondingly promises most benefit. It is to change theway patent grants are measured from time to money.For inventions that result from investment, time is anextremely poor measure — the proper one can onlybe money. No doubt, in the early days of patents,any measure other than time was out of the question,since accounting techniques were undeveloped. ButKelvin’s dictum that Awe advance according to theprecision of our measuresB is not only true fornatural science: to persist with such a poor measureas time for patents is simply to ignore all theachievements of accountancy.

It can only be by sheer accident that any period ofŽtime which is the same for all technologies as the

.20-year patent term is today could provide just theincentive that is required to bring about a particularinvestment at high risk. As will be seen below, manyof the problems of intellectual property rights, espe-cially for new fields such as biotechnology andinformation processing, are actually caused by usingtime as the measure of a patent or copyright grant.The techniques are now available to change themeasure of the grant of exclusive privilege from timeto money, without reducing incentives to invest inR&D, and with strengthened support for innovation.

10.1. Need for rapid innoÕation diffusion

There is now persuasive evidence that progress inany field of technology is made most rapidly whenseveral firms are competing to capture a share of anew market, and to widen the scope of application ofan invention, through making incremental improve-ments along overlapping or competing trajectoriesŽ .cf. Merges and Nelson, 1990, p. 908 . This evidencestrongly suggests that early freedom to use inven-tions should be an essential component of modernintellectual property. Complex technologies do infact achieve this through their use of patents as atrading currency, as the growth in their importance

Ž .reported by Roycroft and Kash 1999 confirms. Onthis point, it is ironic that if the existing patentsystem actually worked in practice as it is supposedto do in theory, there would be even less innovationthan there is. Empirical research has shown thatŽ .except for chemicals in contrast to a patent’s nomi-nal 20-year monopoly term, the average period ofeffective protection is no more than 3 years. Worsestill, the cost to the imitator of producing a competi-tive product within that period, has been shown to beless than what the invention had cost the originating

Ž .firm Mansfield et al., 1981; Levin et al., 1987 .Instead of relying for innovation diffusion on the

failure of the patent system to deliver the protectionit promises, therefore, it would clearly be better touse a compulsory license system to get the processof diffusion started in an orderly way and as early aspossible. This would open the way for competitivefirms to explore several different trajectories of in-cremental change, to the undoubted benefit of prod-uct quality, employment and economic growth.

The idea of compulsory licensing is anathema tousers of the patent system, who see it as inimical totheir chances of gaining big enough rewards to jus-tify the risks they take. Their views must be takenwith the utmost seriousness, since if any compulsorylicensing arrangement meant reducing the incentiveto invest in inventing, there would be fewer or evenno inventions to diffuse. The problem is therefore toreconcile the need for the quickest and widest possi-ble diffusion of inventions to the market, with main-taining, indeed improving, incentives to invent andinnovate. This requires a quite new kind of compul-sory licensing.


10.2. A Shared-riskB compulsory licenses

Ideally, any exclusive privilege should last untilits possessor has made profits from his innovationwhich amount to some socially acceptable multipleof the investment which was made in it at high risk.However, ex post measurement of this kind has beenfound to be impractical. Measuring ex ante, i.e.,establishing what it costs to produce an invention orinnovation, is much more practical and less suscepti-ble to fraud. If obtaining protection is going todepend upon keeping precise records, firms will

Ž .make sure that this is done. As Merges 1992, p. 55has pointed out, proving such costs

will not be difficult or burdensome. Patent appli-cants and patentees collect this information anyway

Ž .for a variety of reasons, including 1 tax benefits,Ž . Ž .2 internal cost accounting, 3 use in project

Ž .evaluation, 4 use in licensing negotiations and thelike. Patentees appear to have no trouble showingresearch expenditures at the damages stage of apatent infringement suit, and . . . such informationhas been introduced in some cases to show thenonobviousness of the invention involved. Simplyadding one more reason to collect data on the costof a research project does not appear to pose amajor problem.

It is also the case that analysts of high-tech stockson the Nasdaq market are increasingly paying atten-tion to pricerR&D ratios, which in itself must beforcing innovatory firms towards more preciserecording of their investments in research.

Availability of such data would open the way toallowing a competing firm to share the use of anyinformation generated by an inventorrinnovator,provided that it also shares retrospectively in theinvestment which the originator had to make in orderto bring the information into being, as well as therisk of that inÕestment. This could be done by mak-

Ž .ing the payment for a compulsory license a amultiple of the R&D investment the originator had

Ž .made, and b a once-off, capital sum rather thanroyalties.

Ž .The reason for a is of course to take account ofthe risk which had been taken in making the invest-ment, and the multiple must be such as at least to

Ž .maintain present incentives to invest. That for b is

because of the principle of sharing retrospectively inboth the investment and the risk that brought theinformation into existence. That investment is now asunk cost for the originator. Fair sharing in it by alicensee should not depend on that licensee’s futuresuccess in using the information, as would be thecase with royalty payments. By reducing the numberof potential competitors who would take such alicense, this condition would also enhance the origi-nator’s return on investment.

Socially acceptable AmultiplesB of the investmentwhich had been made to produce the informationthat is to be the subject of the license would beofficially prescribed for capital payments before-hand. They could be varied according to the needsand risks of different industries and according to thestage of invention or innovation reached when thelicense is taken out, but of course, there would be noquestion of adjusting the multiple to individual firms,any more than the patent term would be adjustedtoday.

10.3. Empirical support

To investigate how this proposal could work inpractice, the risks and amounts of investments of23,000 cases in the U.S. Small Business InnovationResearch Programs, stage-by-stage from ideas to

Žproducts on the market, were calculated Kingston,.1994 . The results indicate the feasibility of working

out AmultiplesB that would combine better incentivesfor investment in inventing and innovating than thoseoffered by the patent system now, with the earliestpossible and widest diffusion of the results.

10.4. Some adÕantages of measuring by money

Ž .Mazzolini and Nelson 1998, p. 274 have pointedout that recent growth in interest in intellectual prop-erty is associated with a general view that Ais heavilyweighted toward the proposition that strong and broadpatent rights are conducive to economic progress.BThey urge caution in respect of this proposition, andit is interesting to note how far using money insteadof time as the measure of protection can provideanswers to their objections. AShared-riskB compul-sory licensing would combine the possibility of large


rewards for an originator with immediate access toinformation by followers. Those who made improve-ments on an original breakthrough would not beblocked by the originator from putting it on themarket, nor would the originator be blocked fromusing the improvement, both being recompensed ap-propriately by the AmultiplesB for their respectivecontributions.

Another consequence is that there is likely to bemuch less infringement, since competitors will calcu-late that it is in their interest to buy a Ashared-riskBlicense and get into production quickly, rather thantake the chance of losing early mover advantage inthe market as well as an infringement action. Itwould also be reasonable to expect the Courts totreat infringers more severely than they do at present,because of failure to exercise an option which hadbeen open to them.

For complex technologies, Ashared-riskB compul-sory licensing would provide all the advantages andnone of the disadvantages of a patent pool. No firmwould ever find itself locked out from use of anycomponent which it needs. The waste and otherharmful effects of policies such as Asaturation patent-ingB by firms in complex technologies, which ofcourse could not even have been envisaged when theoriginal, Anatural rightsB patent system was formu-lated, would be eliminated.

10.5. More incremental innoÕation

A single patent of the present type, if it has broadscope and with its monopoly measured by time, caneven enable a firm to deter others from trying to

Žinvent Ain its neighborhood.B Lerner’s 1995, p..463ff empirical research has shown, for example,

how in biotechnology small firms are inhibited frompatenting in areas already colonized by larger ones.This will also mean that there will be a narrowerrange of incremental innovation than there could be,since this type of innovation depends on the competi-tive efforts of several firms to satisfy the full rangeof a market’s requirements. Neither of these disad-vantages would apply to Ashared-riskB compulsorylicensing arrangements. Any rival firm would be freeto work the invention by retrospectively sharing boththe investment made by the originator and its risk

through making the prescribed capital payment for alicense. Nor would this remove an advantage whichhas been claimed for present arrangements, which isthat others may be forced to work on quite differentalternatives. With the proposed change, any competi-tor would remain free to do this as an alternative topaying for a license.

Ž .Kitch’s 1977 Aprospect theoryB of patents ar-gues that broad patents are necessary if there is to beorderly development of an invention which opens upa whole range of follow-on inventions. It is claimedthat effective control of these by the holder of theoriginal master patent reduces waste in terms of theefforts of a number of inventors who work in theprospect of a profitable invention, but find that theyare beaten in the patent Arace.B Just as what is nowproposed would deal with the problems of Ablock-ingB patents, it would also achieve the objectiveslisted by Kitch very efficiently, since licenses to useany master patent would be available to all thosewilling to share retrospectively in both the origina-tor’s investment and its risk. The wastes associatedwith patent AracesB would be eliminated.

10.6. Implications for uniÕersity research

In recent years, there has been remarkable growthŽin patenting of the results of university research cf.

.Mowery, 1999, p. 41 . This is causing concern lestintellectual property rights should be granted onwhat used to be considered as AscienceB and there-fore intrinsically in the public domain. The unease isall the stronger because of the extent to which suchresearch is publicly funded. It is widely held thatinformation which is the result of such funding,should be made available as extensively as possible.Without reversing the encouragement now beinggiven to universities to use intellectual property,generous licensing of it is clearly desirable. Indeed,the world’s biggest royalty-earner from universityresearch, the Cohen–Boyer biotechnology patent, hasbeen very widely licensed. This, of course, is exactlywhat would happen with all inventions if the changesnow proposed were in force.

A question of particular importance for scientistsis, how early in the process of information genera-tion can protection legitimately be introduced? Com-


pletely free access to the earliest stages has tradition-ally been the norm, and those who work in basicscience are naturally on their guard against anyerosion of this freedom by those who would extendthe range of proprietary protection. The changesproposed would protect freedom of access, becausethey would eliminate monopoly rights. Without erod-ing incentives to produce information, they wouldmake access to and use of that information contin-gent only on readiness to share in the effort that wasresponsible for it. Such arrangements could hardly beobjected to by anyone who believes that useful infor-mation should be both as plentiful and as freelyaccessible as possible.

10.7. AdÕantages for biotechnology

The issue of appropriate intellectual propertyrights for useful information, which emerges close tothe frontier with basic research, is raised in an acuteform in patents for biotechnology. This is emphati-cally a field where discoveries are made, not by theindividual activity of the patent system’s origins, butby large-scale, expensive and purposive research ac-tivity, much of which takes place in universities.Commercialization of the results may require the useof a number of gene fragments, so that the holder ofa single patent can be in a classical AblockingB

Ž .position cf. Heller and Eisenberg, 1998 . Two waysin which changing from time to money as measureof the grant of protection could deal with this prob-lem are immediately evident.

v It would provide a basis for ending the battlebetween those who want exclusive privileges in re-spect of biotechnology to be as strong or as weak aspossible. This conflict can be seen in its extremeform in the contrast between one group which holdsthat patents for gene sequences should be issued withbroad claims, and another group which wants distri-bution of all relevant information to be so free aseven to be on the Internet. If Ashared-riskB compul-sory licenses were available, the first group wouldget the incentive for risky investment which it wants,and the second one would see all information able tobe used freely, subject only to the condition ofsharing retrospectively in both the investment andthe risk which produced it.

v It would greatly reduce both the volume oflitigation and the cost of individual actions. One ofthe roots of the thinking of those who are againstpatent protection for biotechnology inventions is thebelief that this type of monopoly is anti-social, cou-pled with a certain fear of it. Opponents of patentsover-estimate the extent to which they actually de-liver the monopoly power that they nominally con-fer. At the same time, in an entirely new field suchas biotechnology, there are indeed grounds for cau-tion, lest whole areas of development in the lifesciences might be in danger of monopolisation byindividual firms. This would also carry with it thedanger of slower development through lack of stimu-lus from competition. Changing from time to moneyas a measure of any exclusive privilege granted, withits corollary, Ashared-riskB compulsory licensing,would result in the multiple sources of innovationwhich optimum development requires. In doing so, itwould also remove the legitimate grounds for con-cern on the part of those who oppose biotechnologypatents in their present form. The fact that the Presi-dent of the U.S. and the Prime Minister of the U.K.have felt it necessary to issue an unprecedented jointstatement on this matter, is evidence of the depth ofthis concern.

11. Conclusion

Legal changes had to be made to try to adjust theadministration of patents to the reality that inventionand innovation now primarily result from investmentrather than from individual creativity. An unintendedresult of these changes has been to make it difficultfor firms in most non-chemical technologies, and allsmaller firms, to obtain the protection they need forrisky investments in innovation. However, reformsare possible, including compulsory arbitration of alldisputes, with legal aid for the respondent party inthe event of an appeal from an arbitration to theCourts; the Innovation Warrant version of second-tierpatent protection; and changing from time to moneyas the measure of the exclusive privilege granted,made possible through Ashared-riskB compulsory li-censing in exchange for capital payments.


References

Bochnovic, J., 1982. The Inventive Step. Weinheim, Basel, VerlagChemie.

Bouju, A., 1988. Patent Infringement Litigation Costs. Longmanfor the Commission of the European Communities, London.

Bugbee, B.W., 1967. Genesis of American Patent and CopyrightLaw. Public Affairs Press, Washington, DC.

Calvert, I.A., Sofocleous, M., 1989. Interference statistics forfiscal years 1986 to 1988. Journal of the Patent and TrademarkOffice Society 71, 399–410.



Commission on Patent Law Reform, 1992. Report. U.S. Govern-ment Printing Office, Washington, DC.

Creel, T.L., 1987. Guide To Patent Arbitration. Bureau of Na-tional Affairs, Washington, DC.

European Commission, 1996: Directive 96r9rEC of the Euro-pean Parliament and of the Council of 11 March 1996 on the

Ž .Legal Protection of Databases, 1996 O.J. L. 77 20.Federico, P.J., 1978. Origins of section 103. In: Witherspoon, J.F.

Ž .Ed. , Nonobviousness — The Ultimate Condition ofPatentability. Bureau of National Affairs, Washington, DC.

Federico, P.J., 1990. Operation of the Patent Act of 1790. Journalof the Patent and Trademark Office Society 72, 373–385.

Gilfillan, S.C., 1935. The Sources of Invention. University Press,Chicago.

Heggen, A., 1975. Erfindungschuts und Industrialisierung inPreussen, 1793–1877. Vandenhoeck und Ruprecht, Gottingen.¨

Heller, M.A., 1998. The tragedy of the anticommons: property inthe transition from Marx and markets. Harvard Law Review111, 621–688.

Heller, M.A., Eisenberg, R., 1998. Can patents prevent innova-tion? The anticommons in biomedical research. Science 280,698–701.

Hobhouse, L.J., 1995. Biogen. Reported Patents Cases 25, 114.Janis, M.D., 1999. Second tier patent protection. Harvard Interna-

tional Law Journal 40, 151–219.Kingston, W., 1984. The Political Economy of Innovation. Marti-

nus Nijhoff, Dordrecht.Kingston, W., 1987. Direct Protection of Innovation. Kluwer

Academic Publishing, Dordrecht.Kingston, W., 1994. Compulsory licencing with capital payments

as an alternative to monopoly grants for intellectual property.Ž . Ž .Research Policy 23 5 , 1275–1289, November .

Kingston, W., 1995. Reducing the cost of resolving intellectualproperty disputes. European Journal of Law and Economics 2Ž .1 , 85–92.

Kingston, W., 2000. Antibiotics, invention and innovation. Re-search Policy, in press.

Kitch, E.W., 1977. The nature and function of the patent system.Journal of Law and Economics 20, 265–290.

Kronz, H., 1983. Patent protection for innovations: a model.European Intellectual Property Review 5, 178–183.

Lemley, M.A., 1997. The economics of improvement in intellec-tual property law. Texas Law Review 75, 990–1084.

Lerner, J., 1995. Patenting in the shadow of competitors. Journalof Law and Economics 38, 463–495.

Levin, R.C., Klevorick, A.K., Nelson, R.R., Winter, S.G., 1987.Yale study of R&D appropriability methods. Appropriating theReturns from Industrial Research and Development. Special

Ž .issue on microeconomics no. 3 of Brookings EconomicPapers, pp. 783–831. Washington, D.C.

Mansfield, E.M., Schwartz, E., Wagner, S., 1981. Imitation costsand patents: an empirical study. Economic Journal, 904.

Mazzolini, R., Nelson, R.R., 1998. The benefits and costs ofstrong patent protection: a contribution to the current debate.Research Policy 27, 273–284.

Merges, R.P., 1992. Uncertainty and the standard of patentability.High Technology Law Journal 7, 1–70.

Merges, R.P., 1994. Intellectual property rights and bargainingbreakdown: the case of blocking patents. Tennessee LawReview 62, 75–106.

Merges, R.P., Nelson, R.R., 1990. On the complex economics ofpatent scope. Columbia Law Review 90, 839–916.

Merz, J.F., Pace, N.M., 1994. Trends in patent litigation: theapparent influence of strengthened patents attributable to thecourt of appeals for the federal circuit. Journal of the Patentand Trademark Office Society 76.

Mintz, H.H., O’Rourke, C.L., 1978. The patentability standard inhistorical perspective: AinventionB to section 103 Anonobvi-

Ž .ousnessB. In: Witherspoon, J.F. Ed. , Nonobviousness — TheUltimate Condition of Patentability. Bureau of National Af-fairs, Washington, DC.

Mowery, D.C., 1999. America’s industrial resurgence: how strong,how durable? Issues in Science and Technology 15, 41–48,Ž .Spring .

Noble, D.R., 1980. The corporation as inventor: patent law reformand patent monopoly, America by Design. Oxford Univ. Press,Oxford, Chap. 6.

Plasseraud, Y., Savignon, F., 1986. L’Etat et l’Invention. Paris, LaDocumentation Francaise.

Polanyi, M., 1967. The Tacit Dimension. Anchor Books, GardenCity, NJ.

Reichman, J.H., 1994. Legal hybrids between the patent andcopyright paradigms. Columbia Law Review 94, 2432–2558.

Reichman, J.H., Samuelson, P., 1997. Intellectual property rightsin data? Vanderbilt Law Review 50, 52–166.

Rich, G.F., 1978. Congressional intent — or, who wrote theŽ .Patent Act of 1952? In: Witherspoon, J.F. Ed. , Nonobvious-

ness — The Ultimate Condition of Patentability. Bureau ofNational Affairs, Washington, DC.

Roycroft, R.W., Kash, D., 1999. The Complexity Challenge:Technological Innovation for the 21st Century. Francis Pinter,London.

Rowland, B.I., 1978. Obvious to try — a nonstandard ofŽ .patentability. In: Witherspoon, J.F. Ed. , Nonobviousness —

The Ultimate Condition of Patentability. Bureau of NationalAffairs, Washington, DC.


Scott, R.G., 1998. Congressional Record of debate on Bill H.R.2696, March 18, 1998. United States Government PrintingOffice, Washington, DC, H1243, Representative of Virginia.

The State of Small Business: a Report of the President, 1997.United States Government Printing Office, Washington, DC.

Thurow, L.C., 1997. Needed: a new system of intellectual prop-erty rights. Harvard Business Review 75, 95–103, Septem-berrOctober.

Waksman, S.A., 1975. The Antibiotic Era. The Waksman Founda-tion of Japan, Tokyo.

Yorke, B., 1999. Patenting in the Next Millennium. Paper read tothe European Union Patinnova Conference, Thessaloniki.Mimeo., copy on file with author.

innovation needs patents reform -...

Documents