Agriculture’s inter-industry linkages, aggregation bias and rural policy reforms

As agricultural policy reform and its effects have become increasingly territorialised, analyseswhich attempt to explain or predict impacts need to be more localised but also identify spillovereffects. In addition to the predictions of policy shocks predicted by sectoral partialequilibrium models, local and regional general equilibrium approaches which establish thewider effects of such policy shocks have become popular. However, these neglect a major,underexplored difficulty: agriculture is usually described as a single sector in input-outputaccounts, whereas policy shocks with differential impacts have effects on other industries whichare different to those implied by average input-output coefficients. Regionalisation ofaggregated input-output tables adds further to these difficulties. The objective of this paper is todevelop a relatively simple method for dealing with these problems. It establishes thetheoretical basis for aggregation bias and shows how it can be measured, in two contrastingcase study regions in the United Kingdom and Sweden. Having established that this is asignificant problem, a simple but effective procedure is demonstrated, based on additionalinformation on variable costs, which transforms policy shocks from a direct change inagricultural output to that transmitted to the suppliers of inputs. This method provides animpact close to that which could be calculated if the general equilibrium system had indeedbeen disaggregated, and supports use of this approach in impact studies where the researcherdoes not have the time or funding available for completely disaggregating the agriculturalsector’s regional accounts.