Prizes vs Patents

Summary

If a patent's monopoly is a deliberately-created rent — the steepest, "the-rent-is-the-incentive" end of the rent gradient — the Geoist question is whether the reward to invention can be preserved while the monopoly deadweight loss is removed. The prize-and-buyout literature is the body of economics that takes that question seriously. It does not show that prizes dominate patents in general; its actual lesson is subtler and more useful — the best mechanism depends on what the sponsor knows, and for a class of high-social-value, easily-defined inventions, replacing the monopoly with a public reward can capture most of the deadweight loss at modest cost. This page collects the three anchors: Wright (1983) on the theory of the choice, Kremer (1998) on the patent-buyout mechanism, and the pneumococcal Advance Market Commitment (AMC) as the modern field test.

Wright (1983) — the choice among mechanisms is an information problem

Brian D. Wright's "The Economics of Invention Incentives: Patents, Prizes, and Research Contracts" (American Economic Review, 1983) is the founding formal treatment. Its contribution — as later surveys credit it — is to show that patents, prizes (pay for a specified output), and research contracts/grants (pay for inputs) are three tools whose relative merit turns on asymmetric information between the sponsor and inventors. Gallini & Scotchmer (2002) describe Wright as the one "who gave the first formal treatment of how asymmetric information should inform our choice among incentive mechanisms."[3]

The intuition the literature draws from Wright: a patent lets the market reveal an invention's value and lets inventors self-select which projects to pursue, at the cost of monopoly deadweight loss. A prize or contract avoids the deadweight loss but requires the sponsor to know the value (to set the prize) or the cost (to write the contract). So patents look relatively better precisely when the government is ignorant about value and cost; prizes and contracts look better when the sponsor can define or observe value cheaply — as in military procurement, or a vaccine of known epidemiological worth.[3] Wright's own model is a benchmark in which, under the information structure he assumes, none of the three mechanisms is first-best — which is why the subsequent literature (Kremer; Scotchmer; Chari–Golosov–Tsyvinski) designs hybrids.

Kremer (1998) — buy the patent, then give it away

Michael Kremer's "Patent Buyouts: A Mechanism for Encouraging Innovation" (Quarterly Journal of Economics, 1998) is the most-cited constructive proposal. It opens with the historical precedent:

"In 1839 the French government purchased the Daguerreotype patent and placed it in the public domain."[1]

The mechanism. The government offers to buy patents. To estimate a patent's private value it runs an auction; it then offers to purchase the patent at that private value times a markup equal to the typical ratio of inventions' social to private value. Most bought-out patents are placed in the public domain — eliminating monopoly pricing and the incentive for "rent-stealing duplicative research" — but a randomly-selected few are actually sold to the high bidder, which is what gives auction participants the incentive to bid honestly.[1] Kremer argues a markup "of at least twice" the private value is enough to push invention incentives toward the social optimum, and cites evidence that the gap is real: using 1995 CPS income data, he calculates that "the social value of new pharmaceuticals is 2.7 times the profits that would be extracted by a monopolist who could not price discriminate."[1]

Why it is Geoist in structure. A patent buyout is the IP analogue of a rent tax done in reverse: instead of leaving the monopoly rent with the holder, the state pays the inventor the incentive-necessary reward and returns the knowledge to the commons, converting a private monopoly into a public good while (Kremer argues) sharpening rather than dulling the incentive to invent. It captures the benefit without the monopoly rent — exactly the move ip-rents frames.

The pneumococcal Advance Market Commitment — the modern working example

The Advance Market Commitment is the prize idea's most consequential real-world deployment. Kremer, Levin & Snyder (2020) — Kremer received the 2019 Nobel — review the one pilot AMC, for pneumococcal conjugate vaccine (PCV):

• In 2007, five countries and the Gates Foundation pledged $1.5 billion to guarantee a market for a PCV suited to low-income countries, with firms committing to supply at a capped price ($3.50 per dose) in exchange for a share of the subsidy.[2]
• The design's economic logic maps onto the theory: the price cap limits the static deadweight loss from market power, while the donors' commitment to top up price above marginal cost provides the R&D/capacity incentive and guards against the time-consistency problem (that once a vaccine exists, buyers are tempted to bargain the price down to marginal cost, destroying the ex-ante incentive).[2]
• Outcomes: GSK and Pfizer each committed 30 million doses annually (2010); a third manufacturer (Serum Institute of India) entered in 2019 at a lower price, and prices fell toward ~$2 per dose. The authors report the program is associated with more than 150 million children immunized and an estimated 700,000 lives saved, with PCV coverage in Gavi countries converging to the global rate roughly five years faster than rotavirus vaccine, which Gavi supported over a similar period without an AMC.[2]

Key Findings (verified quotes)

• Kremer's premise: markets under-reward ideas because "if consumers pay only the marginal cost of transmitting ideas, revenues will be insufficient to cover the cost of producing ideas" — the classic public-good failure prizes and patents both try to fix.[1]
• The buyout's twin aims — remove monopoly deadweight loss and out-reward the patent — via a markup: patents could be purchased "at their estimated private value, as determined in an auction, times a markup equal to the typical ratio of inventions' social and private value."[1]
• The AMC's headline: "more than 150 million children immunized, saving an estimated 700,000 lives."[2]

What It Supports

• Intellectual-Property Rents — this is the "capture the benefit without the monopoly rent" toolkit that page points to; the patent-buyout and AMC are its concrete instruments.
• Against Intellectual Monopoly (Boldrin & Levine) — the abolitionist critique gestures at "prizes, direct funding" as the alternative; this page supplies the actual mechanism design and the one field result, and is more careful than the abolitionist case about when prizes beat patents.
• Taxing Tech Rents — extends the same "capture vs. dissolve the rent" logic to platform and data rents.

Counter-position / Honest Limits

Prizes are not a free lunch, and the same literature says so:

• The sponsor must know or define value. Per Wright, prizes and buyouts only beat patents when the sponsor can value the invention cheaply; for the great mass of inventions whose worth only the market can reveal, the patent's decentralized value-discovery is hard to replace. This is why Kremer keeps an auction — to borrow the market's information — rather than have a bureaucrat set the price.[1][3]
• Time-consistency / expropriation. Kremer notes that Wright (1983) and Scotchmer (1997) argue "the potential of prizes is [limited because] the authority awarding prizes might be tempted to expropriate inventors by offering inadequate prizes" — the credible-commitment problem the AMC's binding legal contracts are designed to solve, but which a discretionary prize fund reintroduces.[1]
• The AMC lacks a clean counterfactual. The authors are explicit: "we do not know, absent an AMC, whether and when vaccines would have been developed, how much push funding would have been spent, or what prices would have been set."[2] Critics add that the pneumococcal AMC largely accelerated adoption and scale-up of vaccines already in late-stage development, rather than inducing a de novo invention — so it is stronger evidence for prizes as a deployment/manufacturing incentive than as a substitute for the patent's research incentive.
• Manipulation and administration. Kremer himself flags that inventors "could bribe auction participants to submit high [bids]," requiring anti-collusion design; every prize scheme carries analogous administrative and gaming risks.[1]

The net reading, consistent with ip-rents and the innovation objection: prize and buyout mechanisms are a real, sometimes-demonstrated way to preserve the incentive while returning knowledge to the commons — best-evidenced for well-defined, high-social-value goods like vaccines — but they do not abolish the information problem the patent solves, and the choice among patents, prizes, and contracts stays a case-by-case design question, not a settled verdict for prizes.

See Also

• Intellectual-Property Rents — the concept these instruments act on
• Against Intellectual Monopoly (Boldrin & Levine) — the abolitionist case these mechanisms answer more carefully
• Objection: Taxing quasi-rents kills innovation — why the incentive cannot simply be confiscated
• Taxing Tech Rents — the platform/data analogue of capture-vs-dissolve
• Spectrum Auctions — the sibling auction-to-reveal-value instrument
• Economic Rent · Quasi-Rent · Geoism

Sources

1. Michael Kremer (1998), "Patent Buyouts: A Mechanism for Encouraging Innovation," Quarterly Journal of Economics 113(4), 1137–1167. DASH PDF · QJE — used for the Daguerreotype precedent, the auction-plus-markup buyout mechanism, the "markup of at least twice" and social-value-2.7×-monopoly-profit figures, the public-good premise, the Wright/Scotchmer expropriation caveat, and the bribe-the-auction manipulation risk. Full text fetched and read this session (2026-07-11).
2. Michael Kremer, Jonathan D. Levin & Christopher M. Snyder (2020), "Advance Market Commitments: Insights from Theory and Experience," AEA Papers and Proceedings 110, 269–273; NBER Working Paper 26775. NBER PDF — used for the $1.5bn pledge (2007), the price-cap-plus-top-up design and its time-consistency rationale, the GSK/Pfizer/Serum Institute supply commitments and price path, the "150 million children immunized / 700,000 lives saved" figures, the ~5-year-faster-than-rotavirus coverage comparison, and the "we lack a valid counterfactual" caveat. Full text fetched and read this session (2026-07-11).
3. Nancy Gallini & Suzanne Scotchmer (2002), "Intellectual Property: When Is It the Best Incentive System?," in Innovation Policy and the Economy, Vol. 2 (Jaffe, Lerner & Stern, eds.), MIT Press. NBER chapter PDF — used for the authoritative characterization of Wright (1983) as the first formal treatment of mechanism choice under asymmetric information, and the patent-vs-prize observability/value-definition tradeoff. Fetched and read this session; stands in for Wright's own 1983 AER paper, whose free scan (KEI copy) is image-only and not machine-readable.
4. Brian D. Wright (1983), "The Economics of Invention Incentives: Patents, Prizes, and Research Contracts," American Economic Review 73(4), 691–707. KEI scan · RePEc — the founding paper; citation and pagination verified this session, argument characterized via Gallini & Scotchmer (source 3) and Kremer (source 1) because the only free copy is a non-OCR'd scan.