El Mat, I found this equally -relevant and -timely piece of information somewhere on the 'Net, and I thought that I would share it with you:
=====
In, I think it was 1975 or '76, Business Week, for the first time, published company research and development spending. I used that data to compare the R&D expenditures of 42 companies that had been subjected to compulsory licensing decrees to the entire 679 company Business Week survey group.
I found, essentially, no evidence of an adverse compulsory licensing effect on firms' 1975 research and development spending.
I might note a very weak strand of parallel evidence that mirrors these findings. And that is during this period in the 1940's and 1950's when antitrust was subjecting a large number of firms to compulsory patent licensing decrees, U.S. productivity growth was extraordinarily strong.
After about 1970, the use of compulsory licensing declined, productivity growth for the next two decades was weak. I would not suggest that there's a clear cause and effect here. All I can say is that the extensive compulsory licensing of patents did not prevent the United States during the 50's and 60's from enjoying a period of extraordinary productive growth.
Now, after we did our 1958 studies there have been a large number of other studies of this same phenomenon by scholars. The next was by Z. Aubrey Silberston and Taylor of the United Kingdom in a book The Economic Impact of the Patent System, published in 1973. Their work was based on interviews with 32 British companies. They asked a whole large number of counter-factual questions, one of which was: Suppose you could get no effective patent protection on your inventions. To what extent would your research and development be reduced?
The weighted average response from these 32 respondents was an 8 percent reduction; although, 6 of the 32 companies would have reduced their R&D by more than 20 percent in the absence of any patent protection.
Edwin Mansfield did a series of surveys of the impact of patent protection. In his first study, he surveyed 100 U.S. firms and asked: What fraction of their 1981-1983 inventions would not have been developed without patent protection?
The weighted average fraction of inventions that would not have been developed without patent protection was 14 percent; although, there were, as one would expect, some outliers. In drugs 60 percent of the inventions would not have been developed. On the other hand, in fields like office equipment, motor vehicles, rubber products, and textiles the impact of not having any patent protection was estimated at zero.
By far, the most ambitious study of this phenomenon was by Richard Levin and associates at Yale University. Levin is now the President of Yale. I guess working on patents is good for your professional development.
Levin et al. sent very extensive questionnaires to -- received very extensive questionnaires from 650 U.S. research and development managers. Among other things -- and their survey is much richer than I can possibly illustrate today -- but they asked, among other things, the effectiveness of various means for securing the competitive advantages from technologically new products.
And basically they found that patents were the least important means of securing the benefits, the competitive advantages from new products among a number of possible instruments. In other words, their finding was essentially the same as that which we found in our study, Patents and the Corporation. Much more important to the corporate respondents were being first in the market with an innovation, providing superior sales and service, and moving quickly down the learning curve. These, on average, were much more important than patent protection.
There were, of course, some exceptions. Drugs was -- the pharmaceuticals industry was an exception, so was the very similar agricultural chemicals industry where patent protection was considered very, very important to securing the competitive advantages from a new product.
Even in drugs, however, there are other kinds of first mover advantages. In particular, there is a reputation effect from being the first into a market, whether or not one has patent protection.
Now, this was a very revolutionary discovery. The FTC was responsible for making it in a study by Ronald Bond and David Lean, published in the late 1970's. They studied a group of pharmaceuticals that did not have patent protection and a group that did have patent production and essentially found that the structure of the industry and the pricing behavior was very similar between these two groups. Having been the first mover with a non-patented drug allowed the first mover to maintain a 30 percent market share and to charge premium prices.
So even in pharmaceuticals where patents are most important of all, one finds that they are not the only deterrent to rapid imitation.
Now, again, there are exceptions. Drugs is the classic exception. And the reason is as follows: To develop -- first of all, to find an interesting molecule and then to carry it through the series of tests required by the Food and Drug Administration, costs, on average, upwards of $100 million dollars.
Almost all of that expenditure is for finding out whether something works and whether it is safe. Once you found those things out, you've acquired the kind of information that then becomes a public good, which, in the absence of patent protection, anyone could copy very easily and bring out "me-too" generic substitutes.
The copying, on average, costs probably a million or so dollars of R&D. Finding the original drug costs upwards of $100 million. And so this is an industry in which to protect these very large front-end investments patents turn out to be quite important.
Software is similar. I had a very nice quotation from Bill Gates from the book "Hard Drive," pages 102, 103. I'll simply submit it for the record rather than read it.
Let me just tell you about one other area where copying is very easy in the absence of some sort of intellectual property protection; and that is, of course, music. The American Association of Composers and Publishers had a major litigation a decade or so ago to try to determine what royalties the television stations would pay for the use of ASCAP copyrighted music.
Among other things they were going to have Aaron Copeland as a witness until, in deposition, Aaron Copeland observed: "Well, I would pay people to listen to my music."
The fact is that there are other incentives for intellectual endeavor that drive people to create easily duplicated works. If one wants to look back into history, one finds that copyright for music to benefit composers did not exist until late in 18th century England. That came as a result of a lawsuit brought by the son of Johann Sebastian Bach, Johann Christian Bach, before the British court system. Only after that point did composers get the direct benefit of musical copyright.
Well, that's the story that our research reveals. There are -- I think there's one caveat that needs to be stated here. And that is -- well, let me repeat the main conclusion, which is for large corporations, there are lots of other incentives to engage in R&D. If you don't keep running on the treadmill, you're going to be thrown off. That's the basic story.
On the other hand, the spectacular successes that sometimes come from patented products may provide a sort of demonstration effect and lure to other smaller firms that would like to make it big. And it's just conceivable that if it weren't for these very large rewards that had been protected by patents we would have less entrepreneurial activity in the start-up of small, high-technology companies.
I do have some slides, and I think they're worth showing, and then I will withdraw.
What I want to say is that the distribution of rewards to technical innovations is what economists call "highly skewed."
Here are data by Henry Grabowski and John Vernon on the rewards, the surplus of sales revenue minus production and distribution costs over or under research and development costs for 100 new drug products introduced into the United States in the 1970's. These were successful drugs in the sense that all of them made it through FDA.
To summarize very briefly, 7 out of 10 did not earn a discounted present value of surplus of production of sales over production and distribution costs. 7 out of 10 didn't earn back their R&D costs.
The 8th decile broke even.
And then there were two groups, Groups 1 and 2, the so-called blockbusters, that achieved rewards far in excess of their R&D
I have been doing a more general study that suggests that this phenomenon is probably nearly universal, that is to say, a relatively few winners offset the losses of large numbers of losing R&D investments.
Let me just illustrate with one other chart. The solid line in graph one here is a time plot of the number of companies in which the very first U.S. high-technology venture capital group, American Research and Development, invested. You see that by the late 1960's, ARDC was invested in roughly 45 different companies.
Now, the lower panel shows the stock value of ARDC's portfolio over time. And you see in the late 50's, ARDC's portfolio rising somewhat in value as high voltage engineering and a couple of other successes come in. And then, in 1965, Digital Equipment goes public; and suddenly the value of ARDC's portfolio skyrockets. The lower line shows the value of ARDC's portfolio without DEC. And you see that, basically, almost all of ARDC's success came from investing in a single company, Digital Equipment Company.
Well, what this suggests is that one needs to be a little bit cautious. One shouldn't simply rampantly go out and trample on intellectual property. It's important to have those occasional winners. And probably the main reason why it is important is that it provides a lure to lots of little people willing to make that high-tech investment.
Now, I think I'll just stop here. I'm going to be on this afternoon, and I'll pick up the rest of my testimony this afternoon.
COMMISSIONER VARNEY: Thank you, Professor.
Let me turn first to my colleagues for questions, if they have any.
I have one question, and then I will give it to Susan.
It's interesting, your conclusions. What I'm wondering, if you take out the drug companies and the other exceptions, you know, we see a lot of consent decrees here, where the parties are absolutely delighted and willing to settle a potential case; and essentially they involve cross-licensing of patents or other intellectual property.
In your opinion -- I know it's maybe not easy to answer because I'm going to ask it in the abstract. But when we're not dealing in the drug companies, are we potentially missing the mark when we're trying to keep competition healthy when we see a three to two overlap kind of merger and our solution is to require a broad cross-licensing of the patents?
I mean the import of what you said is that the patents really aren't all that important. And I've been wondering, maybe we're not quite getting the right answer when we're allowing these -- as you know, we have seen a proliferation of consents in the last few years where we've cross-licensed patents.
What are your general thoughts?
MR. SCHERER: Well, one of the reasons why in the 40's and 50's roughly 100 antitrust suits were settled through compulsory licensing of patents is that that was viewed by the respondents, in at least the consent cases, as a much less stringent remedy than the alternatives of litigating and possibly facing major divestiture.
So, in general, most firms consider it a milder remedy than significant break up of operating units, divestiture of operating units.
Your question implies the question: Are you getting too little? Should you go for more?
And I don't think the simple fact that patents aren't very important to these firms necessarily implies that you're getting too little in the way of settlement.
The issue is: Is that sufficient to restore the degree of competition that you want?
Now, it's not an easy question to answer. I can give an "on the one hand" and "on the other hand" -- despite Harry Truman's dislike for economists to do that sort of thing. On the one --
COMMISSIONER VARNEY: But that is precisely the point: It's not that you're getting too little. It's that: Is what you're getting ensuring a restoration of competition or a maintenance of competition?
MR. SCHERER: Now, in my monograph, the economic effects of compulsory patent licensing, as I indicated in my testimony, I did a study of whether these compulsory licensing decrees had led to a reduction in research and development. And the answer is, relative to peer groups, no.
I also did a study of whether these compulsory licensing decrees had led to significant changes in market structure in four firm concentration ratios relative to all the other variables that we know affect industry structure.
Now, let me say that our models predicting industry structure are very weak. They leave an awful lot unexplained. But when I put these compulsory licensing decrees into the equations predicting changes in industry structure, I found that the compulsory licensing decrees had essentially no impact so that maybe structurally, unless you're very, very careful in antitrust settlement, a merger, say, case settlement, unless you're very, very careful and make sure that you really are opening up some technology in a significant way, you're probably not having much structural impact.
On the other hand, in my testimony this afternoon, I'll talk about the Xerox case. That decree, I think, had a very major structural impact.
There were major structural impacts also in the DuPont-ICI cases of the 1940's, in the General Electric case of the 1940's.
There are cases, as the Hollabaugh-Wright study for the Senate shows. There are cases where compulsory licensing decrees did have very significant structural impacts.
MS. VALENTINE: One quick follow-up on that, because on that very study that you did on the firm and patenting, did your survey of compulsory licensing cover pharmaceutical decrees?
MR. SCHERER: There weren't very many. As I remember, there were only two pharmaceutical industry compulsory licensing decrees, a couple of antibiotic decrees. Tetracycline and one other antibiotic.
It did include those cases, but they're a small minority of the total sample and, therefore, statistically would not be found to have significant results.
MS. VALENTINE: Even if it had shown a negative impact --
MR. SCHERER: I don't think it did have a negative impact.
MS. VALENTINE: -- a follow-up question --
MR. SCHERER: I happen to know the pharmaceutical industry case pretty well. I worked on the FTC Tetracycline case for the respondents. And my strong impression is that the respondents continued to do research and development very aggressively and were driven particularly in their R&D on antibiotics, not by the consequences of these antitrust decrees but by a different kind of competition, the competition of mutation of microorganisms and the development of selective resistance on the part of organisms in the human body.
COMMISSIONER VARNEY: Susan.
MS. DeSANTI: I just have just a couple of questions.
This is very helpful, you're bringing in empirical work from the 50's, the 70's, and the 80's. Do you see anything in the 90's in the way that competition is working or in how markets are working that would lead you to believe that you would likely get different answers if someone were to do similar studies for the 1990's?
Is there anything in particular that you see?
MR. SCHERER: Well, yes, indeed, I do. And fortunately we will get an answer to your question within the year.
The Levin et al. survey is being replicated, not only for the United States but also for several other nations, including Japan, Germany, Switzerland, England, and some others by a group headed by Wesley Cohen, C-O-H-E-N, at Carnegie Mellon University. They are very near completing the analysis of their survey responses; and so we will be seeing, I'm sure, within the year the first articles coming out of this replication of the Levin study.
My expectation is that it will show differences. The reason it will show differences is that the U.S. patent law has been radically revised in the last 15 years. It wasn't the intent of Congress to revise the patent law except in certain process patent cases; but the creation of the appellate court for the federal circuit, in fact, led to very, very important, substantive changes in the law, a consequence entirely unintended by the Congress, to repeat.
The patents have been strengthened greatly as a result of those substantive law changes. There have been much larger awards now in patent infringement cases. Firms are conscious on the one hand that if they have a good patent, it's a powerful instrument; and they're using it in the courtrooms to try to keep people off their turf and collect royalties from them. The largest approved settlement was Polaroid v. Eastman Kodak, $700 million of damages award.
On the other hand, the impact of this change is that smaller firms, and even some rather large firms trying to develop a new product, are essentially finding themselves in a situation like walking through a mine field: There are lots of unexploded patents out there, and you just might step on one and have your corporate leg blown off. I don't think we know fully what the consequences of this change in the law will be. But I find it a rather scary situation, to be honest.
MS. DeSANTI: Thank you.
I'll save my other questions for this afternoon.
COMMISSIONER VARNEY: Terrific.
I would like to introduce our next panelist. And as she concludes, I would also invite our panelists to ask questions if they have any.
We are delighted to welcome Esther Dyson this morning. Ms. Dyson is President of EDventure Holdings, a small, diversified company focused on emerging information technology worldwide and on the emerging computer markets of Central and Eastern Europe.
Ms. Dyson is very active in industry affairs. She is the Chairperson of the Electronic Frontier Foundation and a member of the U.S. National Information Infrastructure Advisory Council. She also Co-Chairs the NIIAC's Information Privacy and Intellectual Property Subcommittee.
Ms. Dyson has written extensive articles on various industry issues for Harvard Business Review, Forbes magazine, WIRED magazine, and many others. She has lectured extensively on numerous industry topics.
And we are delighted to welcome you for what I understand is your maiden testimony in front of a government agency. We hope this doesn't get you a whole lot more. It may.
MS. DYSON: Thank you very much. Good morning.
I made a deal with the -- whatever you call them -- the esteemed people of the FTC, which was: I'm going to speak very briefly. I'm not an economist, and I haven't done a lot of surveys. And I would encourage you to ask questions, partly because, as you mentioned, this is the first time I've done this; and I'm not really sure what's relevant. I didn't quite get it. So I'm going to try and tell you what I know about and then see if you can help me figure out what I know that's relevant.
But I would like to say in starting, I was thrilled to hear your testimony, because much of what you said really agrees with what I'm about to say, which is how little intellectual property as property matters in many of these high-tech industries.
I had always thought there was quite a strong distinction between the value of patents, which I considered to be property and probably more valuable, and the value of copyright especially in the software industry, which is what I'm going to deal with here.
What I'd like to do is almost tell you just a few anecdotes that I think illustrate some of the broader principles.
What I'm going to say is not that I think copyright law is irrelevant and certainly not immoral, but that in many cases it -- well, maybe "irrelevant" is a good word; but the point I'm trying to make is that there should still be a strong connection between the creator of the intellectual property and the property itself but that that connection is rarely likely to lead to financial gain in itself.
As Dr. Scherer said, it's more likely the reputation effect, the attribution, the brand name, those are extremely valuable; and, therefore, for a creator, it's very important to have that identity and that creative activity attributed to the creator.
But selling the property or licensing it is rarely going to be the way to make money. The way you make money is going to be the services, the support, subscription revenues, that kind of thing.
And now let me go to a country where intellectual property laws -- there was a lot of argument over whether they existed. They were called author's rights. And it was much closer to this notion of, you had the right to have the work attributed to you, but you got paid a salary by the state; and the intellectual property was not really yours. And this is Russia.
A friend of mine ran the Clipper software business in Russia at a time when there was really no copyright law to speak of. And what he did was he went out and he promoted Clipper, and he started user groups, he supported developers, he ran training course, he had a hot line; and the products became tremendously successful. But people were not really buying the software. They were buying the infrastructure he had built around it. It was very easy to go out in the street or at least into some little garage and buy a copy of Clipper from somebody else for $2, but people would pay the $500 or $1,000 in order to get the support and service around the product.
And so what they were paying for was not the intellectual property itself but on the one hand the brand name, the assurance that this was the real thing, and on the other hand the infrastructure built around the product.
Now let's take the case of NetScape in the U.S., which is on everybody's mind. They have become a huge success -- if not financially, although they eked out a profit -- in the stock market by gaining 70 percent market share of the Internet browser market in a matter of months. They did this by putting their software on the Net so that people could download it for free.
That was really great, but it didn't bring in a lot of revenues. And they have this kind of strange policy that you can have it 30 days for free and then you're supposed to pay. And amazingly enough, a lot of people do pay and a lot don't. But what is going to make Net Scape successful in the long run is their ability to continue to be ahead of the competition. They will not sue people for basically cloning the capabilities that their product has. They understand that if they can't come out with something better every five or six months that will then be copied by the market in the next five or six months, they're going to be dead, because just as quickly as they could get 70 percent of market share, they can lose it.
So their value-added is, number one, their brand name; and, number two, their ability to continue to produce new versions, revisions, of the software and, second, it's going to be the support around it.
And I actually know this very well personally because I was trying to buy a copy of their server software, which is the stuff you are supposed to pay for, for a company I'm associated with in Poland. But we could not get them to answer the telephone. We could download the software, but we couldn't get the support. And so we decided not to buy it.
I talked to Jim at Net Scape, and that situation may have changed. But the point here is, it was not the software we wanted to buy it was the support around it.
And I just wanted to make one other point with regard to this notion of it takes $100 million not so much to develop a drug but to find out what works and what doesn't.
And what Net Scape is doing, which a lot of other companies are doing, is they're now moving their quality assurance out of house. If it costs maybe $10 million or something to QA a software product, they're now letting their customers do it.
This all started kind of inadvertently, partly with data programs but partly with Intel, which inadvertently let its customers QA its product. And because they weren't doing it on purpose, it kind of gave them a black eye.
But what Net Scape and other companies now have is something they call the "Bugs Bounty" program, where the customers do the QA for free. And, interestingly, of course, if you're a marketing person you say, well, somebody who's corrected a bug is going to be much more loyal to my software than somebody who just got it for free, somebody who's invested in it.
So the balance of power in the whole arrangement between producers and consumers changes at least in the software business. There's no longer producers and consumers. It's a much more two-way street. It's much more interactive. And companies are going to be paid for intellectual service and process not for intellectual property.
continued at:
ftc.gov ;-) |
