A few weeks ago, Simon marked the anniversary of the controversial Canada-FIT decision, here. His helpful post – which recalled an interesting discussion we had on this at the 2014 BIICL WTO conference – prompts me now to reflect a little further on why I have always found a lot more to like in that decision than many others.
I should state at the start that what I am interested in is the benchmark analysis under Article 1.1(b) of the SCM Agreement. There is much more in the decision to discuss, of course, but this is what I am focussing on. Remember that for a subsidy to exist under the SCM Agreement, a ‘benefit’ needs to be shown. And a benefit is provided by a measure if the measure makes the recipient better off in comparison with the position it would have been in ‘the market’.
Of course everything hangs on how the benchmark market is defined. Here, I want to distinguish between two issues. First, there is the question of defining the product scope of the market. This was probably the most controversial aspect of the FIT decision, but I want to leave it to one side. Second, there is the question of defining the institutional conditions of the benchmark market. This is what I want to focus on.
It is easiest to explain the term ‘institutional conditions’ with some examples. What are the relevant ‘institutional conditions’ in the context of energy markets? Obviously, there are a rather huge number, but some are very easy to point to. One is whether or not there are secondary carbon markets: does carbon have a price in the imagined benchmark market, or not? Another is the structure of competition: do consumers have a choice between energy suppliers, or is competition only at the wholesale level? A third might be the market structure: have the functions of production, distribution, management etc between unbundled, or are some of these functions performed by vertically integrated firms? A fourth might be the broader policy environment defining goals as regards the sustainability of the energy supply over the long term (eg, a government-defined supply mix), or the ability of suppliers to disconnect non-paying consumers.
There are many many more, of course. The point is that all of these factors can, in the right circumstances, dramatically affect the energy price which ‘the market’ produces. So the choices one makes in defining the institutional conditions of the benchmark makes a huge difference to the analysis. The question for panels and the AB is on what basis does one make this choice?
One of the main difficulties that I have with much of the criticism of the AB in the Canada-FIT decision, especially in the context of informal conversations, is that such criticism disagrees with the AB about the choices it makes in this respect, but never justifies its own (different) choices. Very often, one sees in such criticism a whole raft of assumptions being made about the institutional conditions of the benchmark energy market, without any explicit justification for those assumptions. (For example, the argument that the FIT programme was self-evidently a subsidy, because the whole point of the programme was that ‘the market’ would not provide an economic return to renewable energy clearly smuggles a whole load of unstated assumptions about that market’s institutional conditions – the most obvious one being that in this imagined benchmark market there is no price for carbon.) For lawyers, it is clearly not good enough to rely on an intuitive, commonsense understanding of what a competitive electricity market looks like.
In fact, the problem of how to choose the institutional conditions of the energy market used in the benchmark analysis is a very difficult question to resolve satisfactorily. Why is this? First, because the text of the SCM Agreement does not provide a great deal of assistance. (There is some help in Article 14, but hardly definitive – the subject for another post). Second, because it is just a fact that there are many many different kinds of (energy) markets in place around the world, in different WTO Members, characterized by different institutional conditions. There is no chance that WTO Members can agree on a single set of ‘proper’ institutional conditions for energy markets, even if that were desirable. How then is the AB to make this kind of choice? Third, because this is an inescapably normative question, since the choice of institutional conditions for the benchmark market determines whether or not a measure is subject to justification under the SCM agreement.
What I really like about this decision is that the AB seems to get this problem in a pretty profound way, and in a way what many of its critics don’t give it credit for. And while the AB has been criticized for the famous distinction it draws between market-distorting and market-creating measures, I read that distinction in light of this problem. The AB seems to me to be saying: ‘We cannot make a choice about the ‘proper’ institutional foundations of energy markets. It is not our role, nor the purpose of the SCM Agreement, to restructure domestic energy markets in fundamental ways. The institutional conditions for energy markets reflect in many cases core public policy objectives of sovereign governments, and even sometimes are the historical product of domestic constitutional arrangements around state-markets arrangements. We must leave them in place as far as possible and appropriate within the confines of the SCM Agreement. And we certainly cannot reshape them by arbitrarily re-imagining some or other aspect of these institutional conditions as a ‘subsidy’.’
There are some ways we might want to modify this approach, and balance it against other priorities, but it seems to me a good starting point. And while it is true that the distinctions between market-creating, market-correcting and market-distorting measures may need some further elaboration and rethinking, the alternative – of not drawing any such distinction at all – can’t be right. Normatively, it must matter eg whether or not a subsidy is correcting an existing market distortion.
Now, the most obvious pragmatic ‘solution’ to the problem of defining the market benchmark is to define it simply as: ‘the market as it is, absent solely the impugned measure’. This is implicitly the benchmark which the AB has adopted in many other cases, and in fact in most of the easy cases it will actually perform pretty well, by a variety of different measures. But in the hard cases (of which Canada – FIT was one), this sort of benchmark will not be inappropriate. Most of the reasons for this have already been elaborated extremely well by others (I am looking at you, Alan and Luca). Eg, the market might in fact change in many unpredictable ways in the absence of the impugned measure. Or the market might be otherwise distorted in myriad ways (does it matter, eg, whether or not existing energy prices are the result of subsidies to fossil fuel producers?). And so on. What the AB merely recognized in Canada-FIT was that this ‘market as is, minus the measure in question’ benchmark is inappropriate in another specific case – namely, where the impugned measure constitutes part of the fundamental institutional conditions in which the market operates, as opposed to a ‘distorting intervention’ into it.
All this will explain my reaction to the solution proposed by Simon in his thoughtful post. To some extent I guess we disagree on the question of whether or not it is useful to talk about ‘creating a market’. What I have taken from people as diverse as Polanyi, North, Callon, Harcourt, Pistor and others is that in practice there is a great deal of work that has to happen in order to construct modern functioning markets, and that includes a lot of legal work by the state. But actually for the purposes of this conversation, it may not make a huge difference if we imagine a market always already existing, as per Simon’s thought experiment. What matters more is the use of actually existing markets as the benchmark (in the particular context of renewable energy). Simon, the way I read your post I think this is the benchmark you are using to define a subsidy – and for the reasons given above, I am not convinced that it works in the hard cases. I guess the main reason in this context is that wholesale energy markets are so incredibly distorted right now, that it is hard to see why they should be taken as the benchmark for determining benefit. It’s also true that there are so many different ways of establishing competitive energy markets, and we see so many different kinds in different parts of the world, that it is hard to know which of the many existing models to use as the benchmark.