The Monopoly Factory

Seven years later, a generic drug maker, Endo Pharmaceuticals, applied to the FDA for permission to sell its own, lower-priced version of Oxycontin. Purdue sued Endo, claiming patent infringement. During the trial, Endo’s attorneys argued that Purdue had conducted no clinical studies, and in fact had no evidence whatsoever to support the claim that the drug worked in small doses for 90 percent of patients, an assertion which had been crucial to its patent application. The judge agreed, and invalidated Purdue’s patent, allowing Endo to introduce its own version of Oxycontin in January 2004. By law, however, Purdue was allowed to keep the billions of dollars in monopoly profits it had garnered with patents it should never have won.

How did the patent office wind up agreeing to give Purdue the patent in the first place? It simply took Purdue’s word for it that the assertions it made about its drug were accurate. As Purdue noted archly in its press release about the court decision, “Purdue never made the claim in the patent application that it had done experiments to establish this property and the patent examiner never asked for such information.”

The Oxycontin patent was just one of thousands of instances of lax scrutiny at the patent office in recent years. Paxil and Prozac are two other multibillion dollar drugs that earned monopoly profits for their makers before their patents were struck down. And the office’s errors range far beyond the pharmaceutical industry to include everything from software to biotechnology to e-commerce. Greg Aharonian, a Bay-Area patent consultant who sends out an almost-daily email newsletter on every patent-related development under the sun (an item from February was headed, “Kazakhstan Patent Office Runs Out of Paper”) highlighted one recently granted patent, which included a “graphical traceroute”–a technology used to map online traffic events to physical locations. As he points out, had examiners simply Googled “graphical traceroute,” they would have found, under the first entry, an excellent example of the technology, along with a link to an explanatory paper, published by a different team of technologists in November 1999–more than two years before the patent application was filed.

No one knows how many mistaken patents the office issues each year. But the results of one patent office experiment suggest that, in some areas, the error rate may be as many as half of those issued may be faulty. Faulty patents may be doing immense damage to the economy. They not only allow patent-holding firms to gouge consumers with exorbitant prices, but they also inhibit innovation and research, put a drag on economic growth, and may even create an investment bubble. (See sidebar.)

The cause of the problem is easy to pinpoint. Over the last decade and a half, the patent office has been set up in such a way that it’s an easy mark. The system overwhelms many patent examiners, operates under laws and bureaucratic incentives that favor applicants, and can potentially be hoodwinked by the unscrupulous. A recent Federal Trade Commission report, which laid out these criticisms, concluded that in key industries such as pharmaceuticals, software, biotechnology, and the Internet, the office now “hamper[s] competition that would otherwise stimulate innovation.” For some companies, armed only with dubious claims, the patent office has become not something to fear but a patsy, as easy to fool as those elderly couples who send cash to the Nigerian prime minister’s wife.

Rent control Ever since Thomas Jefferson personally reviewed the first patent application (for an improvement in the making of potash–it was granted), the basic method of the patent office has remained essentially unchanged. Examiners receive applications, then search through the available record to determine whether the invention is genuinely unique.

Patents give inventors the right to reap the fruits of their (often costly and speculative) research. It typically takes years of tinkering and development to turn an innovative idea into a money-maker, and many ideas never pan out at all. Understandably, an innovator needs to know that if he invests time and money in a project, competitors won’t be able to copy his idea the minute it goes on the market, preventing him from profiting. The patent system ensures the innovator that this won’t happen by giving him–if his invention is genuinely novel and useful–a 20-year, government-enforced monopoly. In addition, because the system requires him to disclose the details of his invention when applying for a patent, it encourages the dissemination of technological information that might otherwise be kept a secret.

But it makes little economic sense to award a patent when the animating idea is already in wide distribution. If you were to issue a patent for the nail today, for instance, everyone who wanted to hang a picture or hammer two boards together would have to pay fees to the patent-holder, or else simply abandon all construction. Unjustified patents, then, instead of spurring innovation, simply give patent-holders the right to extract what economists call “rents,” raising the cost of doing business and acting as a drag on innovation and economic growth.

The classic example of such “rent-taking” comes from the early automobile industry. In 1879, George Selden, a politically connected amateur inventor and patent lawyer, applied for a patent on a rudimentary car powered by an internal combustion engine. Selden was no genius engineer, and the crude design he submitted to the office simply involved putting a gasoline engine on a chassis to make a moving vehicle–something hundreds of automotive inventors, corporations, and tinkerers across the country were already doing. A car built from Selden’s designs couldn’t have run for more than five minutes.

But Selden had worked with the patent office many times and had figured out how to take advantage of its weaknesses–at the time historians say, the office in chaos. Selden exploited this institutional confusion by repeatedly modifying his application, so that it covered technological advances made by others after the original filing date; this trick is known as a “submarine” patent and, while somewhat more difficult today, is still sometimes attempted.

Sixteen years after the patent application was first filed, despite overwhelming evidence that Selden’s invention was neither unique nor particularly useful, the patent office approved Selden’s request. Just before the turn of the century, Selden sold the patent for a hefty price to a group of automakers, who formed the Association of Licensed Automobile Manufacturers in 1903, allowing them to charge a royalty on every car sold. Ownership of the patent also allowed the consortium to deny licenses altogether to rivals whom they feared might be capable of making better, cheaper cars–which is what they did when Henry Ford, who already had a rudimentary factory churning out a few cars each year, applied for a license. Finally, in 1910, after battling the cartel for seven years, Ford took the matter to court. When he won the following year, the patent was invalidated on the grounds that Selden’s engine design was not novel and that it contributed nothing of social value. For eight years, Ford had put his money and time into fighting the patent. With those resources now freed, the Ford Motor Company turned to new projects, developing the assembly line in 1913, effectively creating the modern mass-production automobile industry, and (in 1914) quickly doubling the starting worker’s salary at the Ford plant to $5 per day–a move which has been credited with helping to create the modern middle class. Had the patent office not mistakenly given Selden his patent, Ford might have made these critical breakthroughs years earlier.

Until the 1929 stock market crash, the federal government’s attitude towards patents had largely favored the rights of patent holders; following the crash, Washington’s attitude changed. Policymakers during the Roosevelt administration believed that patents had been issued in too lax a manner and protected the interests of big business at the expense of consumers. From the 1930s through the 1970s, with New Deal and Great Society policies leading the agenda and the prevailing Washington consensus less favorable to big business, the patent office was kept on a tight leash. By the 1970s, the federal government was circulating a list of “Nine No-Nos,” which warned patent-holders against the use of a wide range of patent-enforcement strategies.

But the economic shocks of the 1970s and the competitive threat that emerged in the early 1980s from Germany and Japan changed the government’s attitude. Economists and business leaders worried that the New Deal patent system was hindering America’s ability to compete. Stronger rights for domestic patent holders, they felt, would make it more difficult for foreign companies, particularly those in Japan, to co-opt American inventions and make them on the cheap, as was happening in the electronics and automotive sectors. These economists and CEOs also argued that stronger patent rights would increase the incentives for American companies to experiment and innovate by permitting them to keep more of the rewards. Congress agreed, and, in 1982, it reassigned all patent cases in the country to the D.C. Circuit Court of Appeals, which had proven to be very aggressive in backing patent-holder rights.

Then, in 1991, under pressure to reign in massive budget deficits, lawmakers passed (and President George H.W. Bush signed) a law that revolutionized the way the patent office does business. Borrowing ideas then in vogue among private sector consultants and CEOs to “reengineer” organizations to make them more “customer-driven,” Congress instructed the patent office, which had always been funded from government revenues, to now pay its own way through fees charged to applicants, and to make the process of winning a patent easier on them.

In many ways, these changes worked exactly as planned. Partly because patents were easier to get and partly because the 1990s were a period of explosive technological innovation, the agency saw a rush of new business. Patent applications more than doubled from 178,083 in 1991 to 355,418 in 2003. The office became self-financing and flush with cash. Congress was even able to tap its surplus pool of user fees to help pay down the federal deficit. The patent office hired new examiners, and in 2003, moved into fancy new headquarters in a billboard property along the George Washington Parkway near the Pentagon.

The 1991 decision to make the PTO pay for itself, however, has created a series of perverse incentives that encourage the office to approve undeserving applications, and has made it easier for applicants to game the system.

Because each new application now brings in a $380 fee, the agency has an incentive to approve those patents sending a signal to the market to apply for more. Additionally, patent-holders pay annual maintenance fees for the first 12 years of a patent’s life, meaning that each approved patent brings in a total of over $3,000 to the office. So every patent issued means a bigger budget for the patent office, and helps to guarantee that Congress will continue to look kindly on the office. “It’s like telling the Treasury Department, go call the Bureau of Engraving and Printing and tell them that they’re gonna get paid by how many twenties they print,” says David Martin, who runs M-CAM, an intellectual-property consulting firm based in Charlottesville, Va., and has testified frequently before Congress about the patent system.

Dan Ravicher, of the Public Patent Foundation, a non-profit legal organization, agrees. “At the agency level, if you want to increase the number of people applying for a patent you don’t want a reputation for being tough on applications,” he says. “You want a reputation for being a rubberstamp. And that’s pretty much what the patent office is now.” The agency denies that management encourages examiners to approve applications. But one examiner told The Washington Monthly that he had been told by a manager, “We’re not the rejection office … if you can’t figure out what’s going on, don’t reject it.” Another examiner agrees: “That’s where the push is coming [towards allowing more applications], because allowances bring maintenance fees.”

The patent office, operating under these institutional incentives to push more patents out the door, has set up a system that encourages individual examiners to green-light more of the applications that cross their desks. The first of these individual incentives stems from the fact that the examiners are overworked. Even with increased number of examiners, and their relocation to a more comfortable environment, the patent office hasn’t been able to come close to keeping pace with the mushrooming number of patent applications. In 1983, the office received 87 applications per examiner. That figure has since risen steadily, and in recent years has generally been over 100.

That increase would be manageable for the office were it not for the fact that, over the same period, applications have become vastly longer and more technically complex, as software and information technology have come to account for an ever-larger proportion. “Whereas in the old days it was extremely rare that you got a case of 100 pages, it’s now pretty common. Cases have just gotten bigger,” says Ron Stern, a veteran examiner and the head of the patent examiner’s union. As then-commissioner James Rogan noted in testimony before a House subcommittee in 2002, “The increasing volume and complexity of our workload poses serious issues for the patent office. Some might even use the word ‘crisis.’”

While the weight of the workload gives examiners a passive incentive to rubber-stamp patents, the office’s worker evaluation system gives them an active one. The size of an examiner’s bonus is determined in part by the number of “counts” he amasses. Examiners gain one or more counts each time they open and close a new case. But when examiners reject patent requests, applicants typically adjust the claim and file a “continuation,” denying the examiner a count. So, an examiner concerned about his bonus has a strong incentive to approve the application. “There’s a gaming of the system,” says Harold Wegner, a former examiner who now works as a patent attorney with the law firm Foley and Lardner. “You can get a stack of applications and just allow patents and get your [counts] that way.”

These difficulties are made even worse by a human-resources problem: For all the profit-chasing tendencies of the new patent office, it’s still stuck with civil service salaries for its examiners. After four or five years on the job, many examiners still make below $70,000 a year, not enough for the office to compete with intellectual property law firms or corporations who covet the examiners for their inside knowledge of the system and often pay twice as much in salary, patent attorneys say. “We have been the training ground for the private sector,” Stern, the union official, adds.

Having been lured away from the patent office, these newly-minted IP consultants and attorneys have the insider knowledge to exploit the weaknesses in the system on behalf of their applicant clients. They know, for instance, that the sheer number of applications and the complexity of the technical topics involved ensure that examiners are unlikely to find every pertinent piece of “prior art”–documents showing the state of innovation in the field, and from which an examiner can determine whether the purported invention is truly novel.

And the law, perversely, helps those applicants who try to sneak through inventions that are not genuinely new. It only requires that they cite the prior art that they are aware of, rather than compel them to make a good faith search. Skilled applicants thus frequently conduct only cursory prior art searches of their own, submitting just a handful of examples, allowing them to make as broad a claim as possible in their application and forcing overworked examiners to search for all relevant technologies themselves. “I can’t go through every single journal,” admits one current examiner.

Meanwhile, some applicants employ the opposite strategy, dumping reams of prior art–some of it of questionable relevance–on the examiner in order to prevent him from finding the pertinent examples that might limit the scope of the claim. “After I did a final rejectionthey submitted over 200 patents and 100 foreign documents for me to consider,” says the same examiner. “When you’ve got 100 things, it can sometimes just be overwhelming.”

That’s particularly true when you consider the amount of time examiners are given to deal with these hundreds of documents. According to the FTC report, each examiner has “from eight to 25 hours to read and understand each application, search for prior art, evaluate patentability, communicate with the applicant, work out necessary revisions, and reach and write up conclusions.” Says one former examiner who left the patent office last year, “We were under a lot of pressure to get the patents through the system, so as a result, the quality of examining them goes down. You’re not spending as much time as you should because you basically just want to get them through.”

The result is a system that has approved ever-higher percentages of applications. According to a 2002 study of patent office data by attorneys Cecil Quillen, and Ogden Webster, and Richard Eichmann of Cornerstone Research, the success rate for patent applications rose from 69 percent in 1984 to 86 percent in 2000. In other words, the patent office now grants applications to more than eight out of 10 applicants.

The patent office argues that it is simply carrying out the will of Congress. Brigid Quinn, an agency spokeswoman, says that examiners are required by law to grant applications, unless they can find evidence to reject them. “So the office is doing exactly what the law requires that they do,” she says. But in fact, as David Martin, an intellectual property consultant, points out, the law as written only says that an applicant “may obtain” a patent if sufficient evidence can’t be found to reject it. It doesn’t require that the office grant a patent. When the patent office created its Manual of Patent Examining Procedure (MPEP), an internal document used to establish uniform standards of examination, it wrote, “shall obtain,” rather than “may obtain.” That change has allowed the office to act in its own self-interest by treating applicants as favorably as possible, while simultaneously claiming that its hands are tied by the law. “That subtle substitution of ‘shall’ for ‘may’ is one of the key reasons why we have the problems we have right now,” says Martin.

The patent office denies that the quality of its patents has declined. And, in fact, there are no comprehensive measures of patent quality. But an internal agency experiment gives an instructive glimpse. In 2001, the office created a new set of rules for reviewing “business method” patents–a relatively new category of patentable material, covering non-technical innovations, such as Amazon.com’s “one-click” ordering system (the patent protects the idea, not the technology necessary to carry it out). The complexities of these patents were causing problems for many examiners, so the office instituted a “second-set-of-eyes” system, which simply required an additional examiner to review each application. Under the new system, the allowance rate for business method patents was quickly halved. That suggests patents in other complex fields, such as bio-technology, given without a “second set of eyes,” may have similarly higher error rates.

The countless faulty patents that are pumped out through this process can be extremely damaging to the American economy. They force consumers to pay more than they otherwise should. They make the patent office an easy mark for those who would manipulate the system–and indeed have given rise to a whole new category of scammers: “patent trolls.” (See sidebar.) Some economists think that all the extra patents in corporate portfolios have artificially enhanced the value of those stocks, creating what they call a “patent bubble,” with potentially devastating results for many investors. Worst of all for the economy and for society as a whole, a faulty patent system can create substantial obstacles to basic, critical scientific and technical innovation.

Consider, for instance, the area of breast cancer research. In the early 1990s, a group of American and European genetic scientists, who had been working together to identify and sequence the genes whose mutations had been linked to breast cancer, began to believe they were on the verge of a collective breakthrough. By 1990, the group had significantly narrowed down the location of one of the key genes, BRCA1. The consortium reported that it was “very close” to finding BRCA1, with one researcher declaring that “it will probably be luck, whoever finds it.”

One member of the consortium, University of Utah geneticist Mark Skolnick, was not inclined to wait on luck or collaboration. Skolnick founded a private company, Myriad Genetics, based in Salt Lake City. It was Myriad that took the final step to sequence BRCA1, using detailed genealogical tables kept by local Mormon families. The company designed a test for mutations, then quickly applied for, and was granted, patents covering how the gene could be used.

The broad language of these patents gave Myriad almost total control over how–and whether–researchers and clinical practitioners could study, and test patients for, the gene associated with a disease that affects around one in eight American women. In 1996, one researcher, the University of Pennsylvania’s Arupa Ganguly, who had developed a test for breast cancer sanctioned by the National Cancer Institute (NCI) got a cease-and-desist letter from Myriad informing her that her work fell under their patents. Worried about the high cost of bringing the case to court, the university’s lawyers decided not to challenge the claims–even though Myriad’s patents have been overturned in Europe, and even though several leading experts have argued that they wouldn’t stand up on appeal. After Ganguly had moved on to other projects, the National Institutes of Health signed a licensing agreement with Myriad, allowing NCI to continue testing, but requiring it to send all test samples to Myriad’s Utah lab for analysis.

With no competition to control costs, Myriad can effectively set the price for its test. Women who would have been charged $1,600 by Ganguly’s lab were immediately forced by Myriad to pay around $2,400. “If someone else develops a test that costs $50, that test won’t come to market,” notes Dr. Steven Narod, a leading breast cancer researcher at the University of Toronto.

But higher prices aren’t the only problem. Myriad’s broad patent has effectively kept any other researchers from trying to make the test more accurate at catching cancers. Debra Leonard, a professor of pathology and laboratory medicine at Cornell, has been the target of numerous gene patent infringement claims by Myriad and others for her work on genes linked to Canavans, a rare disorder that causes severe neurological dysfunction in infants, and other diseases. “If you had a disease, would you want one company–with a commercial profit orientation–making all the observations about your disease in the testing diagnostic arena?” she asks. “I wouldn’t.”

Reformers propose various ways to fix the whole mess of problems that has ensnared the patent system. The most obvious fix would be to make the government subsidize the patent office again, instead of having the office pay for itself out of its own application fees. That would eliminate some of the financial incentives to green light patents. Given the mounting federal budget deficit, however, it’s hard to see Congress suddenly turning down a source of revenue. Eliminating the count system, which encourages examiners to allow applications, is another obvious fix. The agency might still have an incentive to push through applications, but at least examiners would be somewhat insulated from that same pressure. Another wise move would be to hire enough examiners so that all complex patents are subject to the “second set of eyes.” (There is money in the current patent office budget to hire more examiners, but probably not enough.)

More substantively, there’s widespread agreement that attempts to challenge a granted patent need to be made more quickly and easily. That means introducing mechanisms allowing third parties to come forward after a patent has been granted with evidence of prior art that the examiner should have considered. Most versions of this plan call for a system of “post-grant review,” creating a mini-trial for each patent being questioned, which would move the appeals from costly and time-consuming trials in the courts to a quicker and cheaper internal office review. (The cost of challenging a patent in the courts averages $1 million; hearings under the right kind of post-grant review system would cost a fraction of that.) The challenger would bring evidence–in the form of prior art–against the patent, the patent-holder would defend the patent, and the examiner would act as a kind of judge.

Serious reform of the patent system threatens the financial position of some of America’s largest corporations, and key industry trade groups such as the Pharmaceutical Research and Manufacturers of America remain adamantly opposed. Still, there are some signs that the resistance to reform is beginning to weaken. Some leading high-technology firms such as Google are calling for major changes in the way patents are issued. The FTC’s 2003 report has helped triggered a new round of hearings and conversation on Capitol Hill. And the growing technological sophistication of America’s economic competitors has policymakers in Washington searching for new ways to spur innovation here at home.