A common theme throughout the day has been the declining number of seed companies–increasing concentration–and its effect. Except no one has talked about the effect. Other than pointing to the structural change itself, no one seems to have any evidence relating to the effect of the change. One farmer at the open mic session (coincidentally one who had been sued by Monsanto) asserted that the move from 70 seed companies to 4 represented a relevant decline in competition. But he didn’t talk about any relevant effect; he had nothing to offer on declining return on investment–no evidence that the change actually affected his bottom line.
Unfortunately, Diana Moss is the lone antitrust expert on the seed industry concentration panel (also known as the “is Monsanto an antitrust problem?” panel), and it falls to her to put meat on these bones. But she fails in the effort, and really just repeats the same mantra as the farmer, with exactly the same amount of evidence (zero, in case I wasn’t clear on this point). (Moss’s AAI paper on biotech seeds is available here; our ICLE paper partially addressing Moss’s is here).
Asked, “how to think about structure and is this a highly-concentrated industry?” Moss answers that we should view biotech seeds as a “platform”–with upstream markets and downstream markets. This move is important for her argument–separating input competition from output competition–because there’s lots of competition among seed distributors, and the overall effect of seed trait development has been pretty much unambiguously positive. I’m not sure why she calls this a “platform,” other than it lets her analogize to Microsoft. But it doesn’t help analytically. For the rest of us, she is talking about competition and concentration in an input market–where the relevant input is patented seed traits. Of course it matters in assessing the effect of concentration in the input market if price and value are improving in the output market, even though a key input may be available from only a small number of firms.
But whatever we call them–“seed platforms” or “seeds” or “seed traits,”–the argument founders on the absence of evidence, and the claim that “there is concentration in the industry” does not (claims to the contrary) mean that there is actually inefficiency in the industry.
Moss starts by asking “what are the alternatives?” She wants to know where Monsanto’s competition is (and of course she forgets that few competitors does not equal no competitors), on the assumption that more competition is better. Well, yes–but she is looking at the wrong dynamic. She asks what alternatives there are to Monsanto’s traits–where are Monsanto’s competitors? But that’s a pure SCP question (structure, conduct, performance–the discredited assumption from the 1960s that competitive outcomes correlated necessarily with industry structure–with the degree of concentration in an industry). The relevant question is not “how can we get more competition?,” it’s “would competitive outcomes improve if we intervened to engineer a market with more competitors?” Moss’s question assumes the only way to improve outcomes is to have more competitors, but she has no evidence to support the claim; it is merely assumed.
Moreover, again reflecting her “platform” claim, she assumes that every farmer and thus every seed manufacturer will need to plant seeds with biotech traits–in other words, if you’re a consumer in the output market, you are necessarily an indirect purchaser in the input market. But that isn’t true.
Keeping with the theme, she analogizes to Microsoft. She analogizes to Intel. She thinks this makes the point. But while a computer can’t function without an O/S or without a processor, soybeans can indeed be grown without biotech traits (I think I also heard Jim Tobin, on the panel from Monsanto, make this point). It may be that almost all farmers prefer seeds with Roundup Ready genes, even at a higher price. But there will always be competition for these seeds from non-GMO seeds. The mix of price and quality may be different, but at some price, for all farmers, the non-GMO seeds would offer a better alternative; non-GMO seeds will always exert competitive pressure on GMO seeds. And of course “buy nothing” is always an available alternative unless demand is infinitely inelastic (and here it is not).
So Moss claims there is “less choice.” But less than what? She implicitly compares the market to some ideal, theoretical world, but she has no measure of the cost of getting to that world, and no support (again) for the claim that the “competitive ideal” is preferable.
Moss says we would get “more innovation from competition, instead of having one firm dominate.” This is an unfortunate and misleading confusing of ex ante and ex post. As Mike noted in a tweet earlier today, “the market for traits is not monopolized. Just turns out the traits farmers like best were created by one firm.” As it happens, the only constraint that I know of on innovation is the willingness and ability to incur the risk and bear the cost of R&D. This is not a “barrier to entry,” unless any positive cost is a barrier to entry, in which case there are no contestible markets in the world. One firm “dominates” here only because its innovation has proved extremely successful. Ex post, there is only one firm that invented the most valuable herbicide resistance trait. Ex ante, any number of firms could have done so. The returns to that investment (and its patenting) are not evidence of diminished innovation, but rather precisely the opposite: The return on that investment is evidence that the innovation has been enormously beneficial.
On patents, by the way, Moss maintains a clever line that, even though it’s true that post-patent-expiration (in 2014 for Monsanto’s Roundup Ready I, FYI) there will be lots of competition, the only way to have real competition on the day the patent expires is for “the pipeline” to be started long before patent expiration. She wants a sort of Hatch-Waxman Act for seed traits (although she thinks H-W is too cumbersome and she wants something better). Again, the claim sounds nice, but she has no analysis of the effect on innovation incentives. It is indeed probably true that some H-W kind of legislation would bring “generic” seed traits to market more quickly. And I have no reason to think that the precise term of patent protection in place today for this precise innovation is optimal, so of course it’s possible that shortening the term could improve efficiency. But the burden should be on the proponent. It’s hard to have confidence in claims that, essentially, patent terms should be shortened when the proponents offer no evidence to show that overall outcomes would be improved, taking account of reduced incentives for innovation.
As I noted earlier, there are always rents to be split, and, not coincidentally, it’s hard not to see this whole enterprise as one big rent-seeking effort. I do hope that the DOJ and USDA get something of value out of this process, rather than just offering (inadvertently, I assume) encouragement for rent-seeking through politics. But until we see evidence that intervening in this market would increase overall efficiency rather than just the wealth of certain farmers relative to that of certain shareholders, it’s hard to take the claims seriously.