Where Einstein Meets Edison

Kremerian Patent Buyouts: Why We Are Still Intrigued After 13 Years and Why They Cannot Work

Kremerian Patent Buyouts: Why We Are Still Intrigued After 13 Years and Why They Cannot Work

Mar 14, 2010

In 1997 Prof. Michael Kremer, then an Associate Professor of Economics at MIT, proposed a radical scheme to replace the intellectual property system in the United States. Eschewing the principle that “Inventors [have] the exclusive Right to their respective Writings and Discoveries” for a limited period of time, as put forth in the U.S. Constitution, Kremerian patent buyouts offer an alternative to the status quo by rewarding innovators the way most individuals and organizations are rewarded: with money, not monopolies.1

Currently the United States’ patent system is an inefficient compromise that under-incentivizes innovation, as the social value of an invention is often worth more than the private value of a twenty-year monopoly, and inhibits technological diffusion by restricting the use of cutting-edge technology through monopolies. As such, debates surrounding the patent system are often framed by a tradeoff between intellectual property rights, the belief that inventors and corporations should hold legally sanctioned monopolies of the creations, and intellectual commons, the belief that inventions should be freely shared public goods. However, Prof. Kremer breaks from this dichotomy by proposing that the federal government purchase each patent for a price that approximates its social value and then release it into the public domain.2 In theory, this would give inventors and corporations the incentive to innovate while not inhibiting technological diffusion. The government would use a second-price sealed-bid auction to determine the private value of the invention, i.e., price the patent, and then pay the patent holder this value multiplied by some constant to approximate its full social value. Since it would be impossible to hold this pricing auction if the winner had no chance of receiving the patent, the government would need to award the winner the patent with some small probability. Therefore, Kremerian patent buyouts could not fully break away from sanctioning monopolies, but they could greatly reduce the number awarded by designating a minimum probability to secure participation in each auction.

Even after 13 years, the possibility of patents without monopolies continues to fascinate entrepreneurs and policymakers alike. By reducing the tradeoffs between incentives for innovation and technological diffusion, Kremer’s system could revolutionize how individuals and organizations approach innovation. But is such a win-win system feasible?

Unfortunately, the answer is a resounding no. At first glance, Kremerian patent buyouts appear to be a panacea, but hidden below the surface are a myriad of theoretical and practical challenges that make Kremer’s system little more than an academic exercise. To his credit, Kremer foresees and address the foremost theoretical challenges to his system; but like many economic theorists, he neglects all practical concerns regarding its implementation.

Despite the elegance of Kremer’s proposal, key theoretical problems remain, including finding the capital to implement the purchases, potential collusion among patent holders and bidders, and the possibility of preventing a patent holder from using his or her invention (since he or she cannot participate in the auction). First, Kremer claims the system could be paid for by increasing taxes because improved consumer goods and lower prices should outweigh the tax increase for the average American (assuming the patents are valued accurately). While increasing taxes is often politically difficult, it does not provide a reason to discard the Kremerian patent buyouts ex ante. It is also clear that if a patent holder were able to collude with a bidder, that bidder could offer a grossly inflated price allowing the patent holder to either be grossly overpaid by the government or to simply purchase the invention back from the bidder. Therefore, this system would require significant anti-collusion measures, including large penalties for being found guilty of collusion and having prospective bidders present all financial and social ties with the patent holder (including a screening process for dummy companies, etc.). While this addresses the core implementation issues, there is still the possible disincentive of preventing a patent holder from using his or her own invention. A simple modification to the model ameliorates this by allowing the patent holder to decline the purchase if the patent is being transferred to another organization instead of the public domain. This way the patent holder can assert that his or her private valuation is greater than that of the auction’s winner (in this case, the patent would remain outside of the public domain either way, so the benefit to society remains the same). While these theoretically resolve many of the implementation issues, Kremer failed to take into account the Yogi Berra effect: “in theory there is no difference between theory and practice. But in practice there is.

An implementation of Kremer’s system relies on the assumption that auctions accurately price patents and that the federal government could calculate the additional compensation patent holders should receive based on the social value of their invention. Unfortunately, neither of these assumptions hold in practice. Auctions often overprice goods because the winner has the highest, or one of the highest, valuations of the good being auctioned. While the federal government could use the auction as a mechanism to determine a value distribution and price the patent accordingly, valuations of patents before the invention has been manufactured or marketed often prove to be very inaccurate. Furthermore, assessing the social value of a single invention ex ante is a nearly impossible task; let alone valuing the social benefit of close to 100,000 patents each year.While Kremer attempts to simplify this process by compensating all patent holders with the same percentage markup over their revealed private value, this would lead to gross inaccuracies, as different inventions differ in the private to social value they provide. Regardless of whether or not a politician would risk making the claim that raising taxes to pay inventors is a social good, the fact that neither of Kremer’s core parameters can be implemented invalidates his proposal.

While Kremerian patent buyouts will not be implemented in the United States for the foreseeable future, we will no doubt continue to see articles written about this radical intellectual property system. The idea that patents and monopolies go hand in hand is something that goes to the very core of our society, as seen in Article 1, Section 8, Clause 8 of the U.S. Constitution, and any idea that decouples these constructs is worthy of careful consideration. While Kremer’s system may not resolve the intellectual property-commons dichotomy, it has shown us that solutions may exist. That insight alone is as worthy of our attention today as it was when he first proposed it 13 years ago.




  1. “The Constitution of the United States,” Article 1, Section 8, Clause 8.
  2. Kremer, Michael. “Patent Buyouts: A Mechanism for Encouraging Innovation.” Quarterly Journal of Economics 113 (1998): 1137–67.
  3. “U.S. Patent Statistics Chart. Calendar Years 1963-2008.” U.S. Patent and Trademark Office, Electronic Information Products Division – Patent Technology Monitoring Team. http://www.uspto.gov/web/offices/ac/ido/oeip/taf/us_stat.htm (accessed on February 25, 2010).