In my second post, I want continue my discussion of reverse settlements. Recall that the basic argument against reverse settlements is that they extend the duration of a pioneer drug company’s patent beyond what it might expected to be if there were no settlement. (Elhauge and Krueger (Texas Law Review, 2012) have a nice description of the settlement process that yields this result. For now I will take it as given.) In my first post I questioned whether drug patents reduce total welfare. In this post I question whether extending drug patents raise producer welfare at the expense of consumer welfare. I will argue that the profits pioneer drug companies make under patents overstate producer surplus. Producer surplus depends on not competition in the drug market but rather on how competitive the market for research and development for the drug was. But we have little evidence on how competitive that market is. It is possible that that market is perfectly competitive, in which case, in expectation, drug companies are making no supra-competitive profits. No such profits would mean no excessive producer surplus, and no antitrust concern, even with its consumer surplus focus.
At the risk of being repetitive (and thereby pedantic), let me restate the conventional tradeoff when setting patent duration, but from the perspective of producer versus consumer surplus as conventional antitrust analysis sees it. An innovator – in drugs or another product – gets a patent if they come up with a valuable innovation. This patent allows the innovator to charge a high (monopoly) price and thereby earn supra-competitive profits. These profits are treated as producer surplus (though I will question that). The high producer surplus comes at the cost of low consumer surplus. This is partly because surplus is a zero sum game: total surplus is either consumer surplus or producer surplus. This partly because the high prices that generate high producer surplus reduce total surplus by pricing consumers out of the market (ignoring my first post on reverse settlements). When a patent ends, competition starts and the market price of the previously patented drug falls. This increases consumer surplus, at the expense of producer surplus. If total sales also rise, total surplus also rises, which also favors consumers. Thus the duration of a patent determines how long producers enjoy high producer surplus and when high consumer begins.
Given this background, it is possible to see why antitrust law cares about the duration of patents. Antitrust law and antitrust authorities – for distributional reasons it appears to me – favor consumer surplus over producer surplus; I will take this preference as given. The more quickly a patent ends, the sooner consumers start earning higher surplus. For this reason, antitrust law is opposed to reverse settlements if they increase the expected duration of patents.
The problem with this logic is that the producer surplus created by patents is not fully producer surplus. The purpose of this producer surplus is to encourage innovation. In the absence of innovation, consumers would be worse off because they would not have the innovation required to generate high consumer surplus once patents expire. Thus, antitrust law should not judge producer surplus in the patent setting the same as it is in the non-patent setting. It should not be judged against a baseline of zero-producer surplus. Instead, it should be judged against a baseline of innovation with shorter patents. If patent duration is shortened, consumers will obtain less innovation and less consumer surplus. That reduced consumer surplus should be subtracted from producer surplus observed due to patents and credited as consumer surplus. (If this were not the case, antitrust law would want to eliminate all patents!).
A defense of the traditional antitrust view of patents would argue that surely not all producer surplus that innovators earn from patents are really consumer surplus. Doesn’t it depend on how much innovators spend on research and development? Correct! But even those expenses should be subtracted from producer surplus observed while a drug is under patent. Yet it is not. The reason is that R&D expenses are (typically) incurred before a patent is granted. Thus, profits under a patent do not net out all costs.
How do we determine how much companies spend on R&D? And do those costs offset all the supra-competitive profits? To answer that question, we have to look at how competitive the market for R&D for a drug is. This is not the same as the market for the drug itself. There is an R&D race for each drug. The question is how competitive that race is. As Dick Posner pointed out long ago: when the government offers a monopoly such as a patent, firms will expend resources equal to expected monopoly rents to obtain that monopoly [1]. In the context of patents, the resources expended are for R&D. Moreover, across all the firms that race, R&D expenses will total expected rents from the patent. This implies one cannot even look simply at the profits that the firm that wins the patent earns and subtract just that firm’s R&D expenses. Some other firms may have incurred R&D costs that should be subtracted because they count towards expected profits – expected producer surplus. Moreover, even if one did look at the profits of drug firms, F.M. Scherer from the Kennedy School argues – rightly – that one must normalize for risk, especially scientific and regulatory risk [2]. Once that is done, it is not obvious that drug companies make supracompetitive returns.
This is not to say drug companies do not make supracompetitive expected producer surplus in the process of innovating.[3] The point is that that we don’t have great evidence it is true. And unless we do, perhaps we ought not punish drug companies for making profits when those may not amount to expected producer surplus, even if antitrust law prioritizes consumer surplus over producer surplus.
Before I wrap up this post, let me make two points. First, I want address a potential objection. The objection is that extending patent life increases R&D expenses (or patent-race expenses beyond R&D) and that these expenditures are a social waste. Perhaps. But I don’t how we could know that with the evidence we now have. For every article that complains about technology driving cost growth or flat of the curve medicine, there are articles about the returns to R&D. Moreover, the complaints about cost growth and flat of the curve medicine is equally a complaint about products with valid patents as invalid ones. Finally, perhaps a bigger driver of that R&D expenditure than reverse settlements is the $1.4 trillion or so the Affordable Care Act invested in expanding insurance coverage over the next 10 years. While that will have significant social benefits for the un- and udner-insured, its effect on R&D expenditure will swamp any minor correction in reverse settlements.[4]
Second, I want to offer a different perspective on the producer surplus that providers make. I think one could interpret producer surplus as future consumer surplus, in which case patent duration would really be a tradeoff between current consumer surplus and future consumer surplus. Recall that part of the ex post producer surplus that drug companies make is actually consumer surplus because it raises the probability of innovation that benefits consumers. Moreover, if there is a competitive race for the patent, the rest of the producer surplus was eaten up by R&D expenditures, which is a legitimate cost of production. It is possible then that there is no supracompetitive ex ante producer surplus.
To this let me add one other fact about the pharmaceutical industry. Drug companies tend to use profits from one patented drugs to finance R&D for future drugs. It appears to me that they do not appear to rely very much on external financing,[5] perhaps because it is hard to convince banks and outside investors that a first future drugs will be successful. Because drug patents have such variable profits, drug companies also typically hold a portfolio of drugs to ensure a relatively stable flow of revenue to finance R&D for future products.
This reliance of drug companies on retained profits for financing means that, at any point in time, drug company profits are really R&D expenditures for future consumers. From that perspective, attempts to reduce patent duration – by barring reverse settlements, for example – are really attempts to shift consumer surplus from future consumers to current consumers. I do not dispute that there is an optimal tradeoff between current and future consumption, or that lower patent duration may improve social welfare. I just don’t think we have adequate evidence about this. More importantly, once patent duration is cast as a current v. future consumer surplus tradeoff, it is not at all obvious that antitrust laws preference towards consumer v. producer surplus would tend argue for shorter patent duration.
In my next post I will take up the question of what purpose drug patents serve and how that impacts whether a particular drug patent should be judged valid.
[1] Posner, Richard A. “The Social Costs of Monopoly and Regulation.” Journal of Political Economy 83, no. 4 (1975): 807-27.
[2] Scherer, F.M. “The Link between Gross Profitability and Pharmaceutical R&D Spending.” Health Affairs 20, no. 5 (September 1, 2001 2001): 216-20. [Note: there is a better cite than this. I just have to dig it up in my digital library.]
[3] Indeed, there may be reasons to suspect they do. There are barriers to entry into R&D. There are a limited number of capable scientists. Large fixed cost investments in lab facilities may be required. And equally important are the big regulatory hurdles – the process of FDA approval – that favor large companies that have specialized knowledge of how to overcome those obstacles. See Thomas, L.G. 1990. “Regulaton and Firm Size: FDA Impacts on Innovation.” RAND Journal of Economics 21(4): 497-517.
[4] A related criticism is that drug companies do not invest profits in R&D but in marketing – and that marketing is a social waste. First, see my first post. That marketing may offset some of the static welfare loss from patent-enabled monopoly pricing of branded drugs. Second, this is mainly a complaint that drug companies produce R&D inefficiently. However, as far as I understand it, antitrust law is not concerned with inefficient production. It is concerned with market power. Those are not the same thing; indeed, they may be negatively correlated.
[5] This is not to say that drug company stocks do not have value. The question is: how much do drug companies use new share issues to raise capital as compared to retained earnings.