How decoupled is the Single Farm Payment?

Three of my Irish colleagues at the Teagasc Rural Economy Research Centre have conducted an interesting simulation to estimate the extent to which farmers treat the Single Farm Payment (SFP) as coupled or decoupled. Using the EU-wide partial equilibrium simulation model AGMEMOD, Peter Howley, Kevin Hanrahan and Trevor Donnellan project Irish production in the cattle and cereals sectors (these were the sectors with the most important payments in the pre-SFP era before 2005) under two assumptions: first, that farmers treat the SFP as fully coupled, and second, that they treat the payment as fully decoupled.

They then compare the levels of production that are projected under the alternative assumptions of full and zero coupling with actual observed output values in Ireland over the period 2005-08. Based on this experiment, they conclude that farm operators to a large extent treat these payments as fully coupled, but that the supply-inducing effect is smaller than for the previously coupled payments.

There is considerable interest in the question of whether the EU Single Farm Payment is fully decoupled or not, not least because of the controversy over whether such Green Box payments under WTO disciplines are actually not or only minimally trade-distorting or not.

Until now, most of the research on the supply-inducing effects of decoupled payments has used US data, where decoupled payments were introduced in the 1996 Farm Bill, and researchers have arrived at very varied conclusions. The EU SFP differs in important respects from the US decoupled payments. There was a greater likelihood of base updating in the US while in the EU the GAEC requirement to keep land receiving payments in good agricultural and environmental condition probably inevitably implies some degree of coupling.

However, the SFP has only been in force since 2005 at the earliest. To date, we can only make inferences about the production effects of the SFP based on survey data, where farmers were asked about their production plans post-decoupling. These surveys (see Hennessy and Thorne, 2005; Colman and Harvey, 2004) showed that a significant number of farmers planned to use their decoupled payments to continue or expand non-viable production.

The results of the model simulations are shown in the table, where the projected changes in area (cereals) and suckler cows (numbers) under the zero coupling and full coupling scenarios are compared with the actual outcome over the period 2005-08.

It is evident that the level of grain area harvested and suckler cow numbers observed over this period is considerably above that projected when payments are assumed to have no effect on farmers’ production decisions, although somewhat below the levels projected when payments are assumed to be fully coupled. Their conclusion is that the Single Farm Payment maintains a strong effect on farm behaviour, even if lower than previous coupled direct payments.

The research design is a clever way of trying to identify the impact of decoupling payments while taking into account all the other influences (changing relative prices, changes in technology) which might affect production trends. The credibility of the results depends, of course, on the forecasting ability of the model and unfortunately we are not given any information on this.

The larger the forecast error of the model, the harder it becomes to assess the significance of the differences between observed and projected outcomes shown in the table. It would be interesting to run the model for the period up to 2001, for example, and then compare the projected results for the period 2001-04 with the production levels actually observed to give an indication of the likely size of this forecast error.

The authors stress that they are looking at short-run responses and point out that it is conceivable farmers might treat the payments as more decoupled over time. However, they also refer to the considerable body of literature which suggests that farmers do not necessarily behave rationally in the profit-maximising stereotype usually assumed in economic modelling, and that there can be a range of other factors that may influence farmer behaviour.

These potential influences include utility derived from being self-employed, prestige associated with land ownership, family circumstances, benefits accruing from social interaction with other farmers and aversion to change. In other words, farmers engage in agricultural production for non-economic as well as economic motivations, and economic modelling which fails to take this into account may well overestimate the extent of change to adverse economic shocks.