It's hard to compete for attention with the drama of the big bailout and this presidential race. But bubbles are news, at least when they burst. The patent bubble hasn't burst, but judging by last week's Wall Street Journal story on Intellectual Ventures, it's getting pretty big.
The brainchild of former Microsoft CTO, Nathan Myrhvold, Intellectual Ventures has reportedly amassed $5 billion in capital and a portfolio of over 20,000 acquired patents -- and it's looking for more. From the perspective of the tech sector, Intellectual Ventures combines two questionable business models, the patent troll and the pyramid scheme, in a form that evokes Wall St.'s cleverness in designing glitzy vehicles for esoteric assets.
IV does not create or market products, so it is invulnerable to the patents of others. It looks like a patent troll, because it makes money from "being infringed." But IV has a new twist: The companies that settle not only pay license fees but are induced to invest in IV, thereby providing the capital to acquire more patents, set up new licensing funds, and pursue other companies. IV's licensee investors, including Sony, Nokia, Google, eBay, Intel, and Microsoft, are all sworn to secrecy about their involvement. According to the Journal article, other investors include pension funds and university endowments, and Myrhvold will target smaller companies for licensing fees (and equity) in the future.
The deep roots of the bubble
IV is only the biggest, and clearly the most sophisticated, manifestation of the IT-centered patent bubble that has been growing over many years. The bubble has roots in the decisions of the specialized appeals court and its predecessor dating back to the 1970s and 80s that made patents more powerful, more plentiful, easier to get, and harder to invalidate. The bubble was also fed by appellate decisions in the 1990s that made a vast, undefined range of abstract subject matter patentable -- including software and business methods. At the same time, the Patent and Trademark Office, intoxicated by the decision of Congress to let it fund itself out of the fees it received, let out the stops, and decided that its new mission was "to help customers get patents."
Cheap, plentiful patents made it easy for IT companies to assemble large portfolios, which they used not for exclusivity but for freedom of action, by cross-licensing their portfolios to other portfolio holders. On the other hand, patents were also considered important to startups. Venture capital liked them, in part because they could be sold if the business failed. And IT startups burgeoned in the 1990s -- ironically because the Internet provided a huge, rich, and largely patent-free platform on which anybody was free to innovate without asking permission.
Startups found that a few patents could be helpful in protecting a technological niche, but were of little value in gaining freedom of action in a portfolio-dominated market for complex products. For big companies, their accumulated portfolios provided an arsenal against patent attacks, as well as against any startups that might poach on the company's turf. In effect, portfolios enabled incumbents to maintain market share, and the increasing use of "balancing" payments to account for differences in portfolio size gave patent departments ammunition to argue for more and more.
This wasn't the way the patent system was supposed to work. Legend preserves the idea of a seminal, patented invention that the inventor is free to nourish into a product. That model may still apply in some fields, in some cases. But since a patent is only a right to exclude others, you really need a trading portfolio to access the technology required for a complex product. Still, it was possible to justify the portfolio system with the argument that "net users" (those that on balance made use of the technology of others) should pay "net innovators" (those that had big portfolios).
But there was a major flaw in this model: Most startups fail. And when they fail, their patents are sold off along with everything else.
As a result, patents were liberated from going (or at least, hopeful) businesses where they were held for protection. Instead, they were bought by speculators keen to find the buyers best positioned to "extract value" from the patent. In other words, trolls.
Can it be fixed?
An unintended byproduct of portfolio building, trolls revealed the downside of a hyperactive patent system. The reform package in Congress has modest measures aimed at reducing the rare but astronomical burden of full-blown litigation. But even these modest remedies are facing opposition from three distinct directions:
-- upstream inventors, universities, and specialists who experience only the benefits of patents without the liabilities;
-- the patent bar, who benefits directly from increased demand for patents, as well as the friction within the system; and
-- a whole sector of the economy, centered on pharmaceuticals, where the value of patents actually corresponds meaningfully to the value of real products.
The battle has been fought to a standstill. The patent ecosystem has become too complex and diverse for consensus. Moreover, much of it has gone underground and "off the books" -- into threatening letters, secret settlements, massive cross-licenses, and a wide variety of privately pooled interests and funds. All of this is beyond the grasp of a Congress that has a hard enough time coming to grips with the doctrinal complexity of patent law and reported cases. Nobody wants to bite the bullet and say, if we're going to regulate innovation with a crude tool like patents, we should do it in a forthright and transparent way that takes economic factors and fundamental industry differences into account. Instead, we hear constant incantation about the need to find a solution that works for everybody. (Remember that "everybody" includes speculators that have relied for decades on bad judicial decisions in order to put their children through college.)
Just like the market bubble of the 1990s and recent real estate bubble, it's hard to tell a bubble when it's growing. And too many stakeholders, including politicians, benefit from widespread belief that a growing bubble is grounded in reality and will continue indefinitely. You have undoubtedly read some of thousands of guest columns by patent attorneys explaining how you, everybody, needs patents in an economy driven by innovation. When Microsoft announces that it's going to increase its patenting 50% next year, everybody wonders, "why aren't we doing that?"
Fortunately, the Supreme Court has recently stepped in to overrule some of the excesses of the patent appeals court -- in particular, its low standard of obviousness and automatic injunctions (even against complex products with one infringing function). Facing a backlog of 1,200,000 patent applications, the PTO itself has become more demanding. It no longer claims to be in the business of "helping customers get patents." But powerful forces have been set in motion in the market -- with no direct means of control, or even monitoring.
Whose reality? Measuring the gap...
Using established means of indirect measurement, economists James Bessen and Michael Meurer calculate that in 1999 the total world-wide value of patents in force for U.S. public companies in computer and communications technologies was $4 billion (in 2008 dollars). A very modest figure compared to some $140 billion for chemicals and pharmaceuticals, areas where the patent is considered to work well.
But wait. Going by the identified licensee/investors, IV's portfolio is heavily concentrated in computer and communications technologies, and its assets are $5 billion. As big as IV's portfolio is, there are hundreds of thousands of other patents out there -- 200,000 in software patents alone. Is IV simply sitting on a pile of overvalued patents?
The short answer is no.
Patents have been touted as the currency of the knowledge economy. Over the past ten years, an exuberant cottage industry has grown up around valuing, exchanging, monetizing, and investing in patents as assets. IV has a lot of company; there may well be another $15 billion out there in patent funds and licensing firms hanging over the IT sector. And that's not counting the hundreds of thousands of patents currently "locked up" in the portfolios of IT producers.
We do know -- but only when we stop to think about it -- that patents are not conventional assets, despite the fact that many people would like them to be. They are not even a right to use the technology. They are only a negative right to stop others from using it.
So a patent may look like an asset in your hands, but it is a liability, present or potential, to everybody else. And this cuts both ways: If you're going to count your patent assets, you should also come to terms with the patents you may be facing. In an environment where there are thousands of patentable functions in a product or service, along with large numbers of innovators working independently and getting their own patents, there will be many, many more patents that belong to others than belong to you (even if you are Microsoft or IBM). Not only can one patent shut down your entire product line or service, you will also be liable for past use, even when you were unaware of the patent. Neither independent invention nor innocent infringement is a defense.
How could this go unnoticed? How can IV's patents alone appear to be worth 25% more in 2008 dollars than all patents were worth to the entire sector in 1999 at the height of the Internet boom?
Bessen and Meurer's calculus of costs, and to a large degree, their calculus of patent value, is based on public companies -- i.e., companies that produce real products for real people in real competitive markets. IV and virtually all patent licensing firms are private companies who do little but license patents.
Myhrvold himself reveals the answer to our conundrum in his Wall Street Journal interview: "The world has lots of inventions it doesn't know what to do with. We think we do. So we go do the classic private equity thing and say, it is worth more in our hands than yours."
He's right: Like the classic troll, he can wield his patents without fear that someone will counterclaim with their patents, so patents in his hands are worth far more than in the hands of a producing company that needs cross-licenses from others.
Measuring the unmeasurable
What is the real value of those patents? Even big boys with portfolios can't afford to find out. As Myrhvold says: "Now suppose we hypothetically approach somebody or they approach us, and we have 1,000 patents in their area. I say, I can't afford to sue you on all of these, and you can't afford to defend on all these."
So nobody can afford to find out, but it's somewhere between modest value in the portfolio of a big producer and maybe 20 times as much in the hands of a licensing specialist -- especially an experienced and sophisticated troll who knows to extract value from a variety of sources. Free market economics says that patents will eventually end up with the trolls (the "highest and best use" for those familiar with real estate appraisals).
There you have it: Arbitrage in exotic assets. A system that values legal instruments over real products. And a context-dependent and uncertain value for the legal instrument itself. No better than mortgage-backed derivatives. And maybe a lot worse.
Fixing the blame
Myhrvold blames the producers -- "a culture of patent infringement that he says has let powerful Internet and tech companies steal other inventors' intellectual property."
"Some of them are committed infringers, they're complete pirates," he says, declining to be specific. Myrhvold was more charitable in testimony before Congress when it was clear that he was implicating his former employer -- where developers were merely advised (by lawyers) to avoid reading patents.
What he does not mention was the original cause: a patent-happy culture in Washington reflected in far too many government-granted patents -- far too many for engineers to assimilate, and an appellate court that imposed harsh penalties for those unlucky enough to learn that they might be infringing.
We can be grateful to IV for making the contours of the bubble visible. But when will the bubble burst and how? What happens when some big, precarious IT firm fails (I won't be specific either), flooding the market with tens of thousands of hitherto "undervalued" patents? Who will be left holding the bag?
One possibility is that the Supreme Court will revisit the issue of patentable subject matter and make it clear that the Federal Circuit's permissive policies were not grounded on Supreme Court precedent. That would deflate much but by no means all of the speculation, and patent holders and attorneys would undoubtedly beseech Congress for rescue.
The other possibility is that the bubble is not in speculative instruments but in the tech sector: That all those going businesses with real customers are overvalued. Why? Because they have undervalued liabilities owed to outsiders armed with powerful legal instruments and seductive, irresistible incentives for attracting capital.
In that light, Intellectual Ventures looks positively pioneering: a high-concept fund that allows investors to hedge against their own destruction.