There are almost as many ideas for reforming the patent system as there are patents. What most have in common, however, is their acceptance of the standard patent paradigm: monopoly as a reward for innovation. This is probably no coincidence, since the Copyright Clause of the Constitution apparently authorizes Congress to grant only limited monopolies and not more exotic compensation. Ignoring constitutional concerns, however, we can try to imagine how a patent system might function without monopolies.

Given most economists’ near-innate aversion to monopolies, it is perhaps unsurprising that a proposal for a monopoly-free patent system came from Michael Kremer, economist at Harvard. Kremer’s idea centers on patent buyouts, in which the government pays an inventor in return for the inventor relinquishing her patent rights. The government subsequently puts the purchased patent in the public domain. Kremer’s paper details the specifics of his plan. Fellow economics professor Steven E. Landsburg of the University of Rochester has an informal introduction here.

Patent buyouts are not an entirely novel concept. The French government famously bought the rights to the daguerreotype photographic process for the public domain by paying lifetime pensions to the two inventors. The inventors had been unable to license their invention in the usual way and turned to a friend with political connections. The daguerreotype subsequently became the dominant photographic technique until patents on better processes like Henry Fox Talbot’s calotype expired. Coming a little closer to home, Kremer also mentions several southern states’ buyouts of Eli Whitney’s cotton gin patent, which Whitney agreed to in large part because Southern courts had not been keen on enforcing his rights against local infringers anyway. In both cases, governments more or less successfully used patent buyouts to remedy failures in the conventional patent system.

Kremer’s innovation comes in using public auctions for determining the buyout price. Kremer’s patent system operates normally up until the patent is issued, at which point the public is invited to bid on it. In most cases (Landsburg suggests 90% of the time), the government pays the inventor the amount of the highest bid and the patent goes to the public domain. To give bidders incentive to bid meaningfully, the highest bidder pays the inventor and secures the normal patent rights in some randomly selected fraction of cases. Kremer also suggests the government pay a fixed markup in addition to the highest bid (to account for the patent’s social value) and describes some systems for preventing inventors from colluding with bidders to inflate prices (e.g. third-price auctions, where the winner pays the third-highest bid).

There are several appealing aspects of Kremer’s system. The public domain would obviously expand substantially, reducing expenditures on patent licensing and litigation by perhaps 90%. Independent inventors would also benefit, since they would be relieved of the substantial burden of selling or manufacturing their invention after it had been patented. Consider for instance the “Jaromatic,” a patented universal jar and bottle opener invented by a 68-year-old retired man who is currently offering all the rights to his invention for $250,000. He has had no takers so far, and consumers remain an automatic jar opener poorer for it. Certainly there are other, similar patents currently owned by inventors without the desire or wherewithal to license them. On the other side, manufacturing firms currently face the daunting prospect of trying to negotiate patent rights individually with independent inventors. An auction-based patent system would give inventors the compensation they desire and manufacturers the designs they need while sparing both parties the costs of negotiation.

Looking beyond the economic idealizations reveals potentially fatal flaws for an auction-based patent system, however. The first question is where the government will get the money to purchase 90% of the nearly 200,000 patents granted each year. Landsburg waves his hands at this, suggesting that tax revenues can pay for the buyouts because goods will be cheaper on average as a result of fewer patents. Whether or not that claim is right as an empirical matter, imposition of a patent tax is probably politically impossible.

Doubts also exist over the ability of an auction to accurately price patents. Although economists like Kremer and Landsburg seem to take market efficiency as an article of faith (skeptics might call it dogma), there is reason to believe that the value of new patents is essentially unknowable. The full utility or commercial value of an invention might not manifest itself for several years, which is one reason the patent term is as long as it is. Kremer’s system requires bidders to forecast value based on the relatively meager information available when a patent is granted, a difficult if not impossible task that could lead to wildly inaccurate buyout prices.

Finally, a bought-out patent cannot be easily challenged if it is later found to be invalid or too broad. The current patent examination system (which Kremer’s proposal leaves unchanged) depends heavily on post-grant litigation to check improper and over-broad patents. Once a patent has been bought out and placed in the public domain, however, there is little incentive for any party to challenge it. Even if someone wanted to challenge a bought-out patent, what could the remedy be? The inventor could not be expected to return part of her compensation, since she might have received it years ago. Given current concerns over the quality of patents granted by the USPTO, an auction-based system would almost certainly result in the government overpaying for many patents that do not cover truly innovative inventions. Indeed, one might expect dubious patent applications to rise in an auction-based system. Without the threat of subsequent litigation, malicious inventors need only slip their applications past the examiner to collect their reward at the government’s expense. Coupled with the difficulty of valuing recently granted patents, this opportunity for fraud is a serious defect of an auction-based patent system.

For all its faults, Kremer’s auction-based system still provides an interesting though experiment on patent reform. At the very least, it should encourage people to consider patent reforms with broader reach and a more creative bent. Even if auctions and patent buyouts aren’t the answer to our present patent woes, Kremer’s proposal forces us to question the oft-presumed necessity of the connection between patents and monopolies.