The effects of decoupling on land allocation

Abstract

The purpose of this article is to study the impact of agricultural policy decoupling on land allocation decisions. Our analysis contributes to the literature by formally assessing the effects of decoupling on farms’ crop mix and on the decision to set land aside. The analysis is undertaken within the framework of the model of production under uncertainty developed by Just and Zilberman (1986). Our empirical application focuses on a sample of Kansas farms observed from 1998 to 2001. Results show that US agricultural policy decoupling has resulted in a shift in land use away from program crops towards nonprogram commodities offering higher expected profits and idle land.

Acknowledgements

Authors gratefully acknowledge financial support from the Spanish Ministerio de Educación y Ciencia (Project Reference Number AGL2006-00949/AGR).

Notes

¹The role of risk has also been considered when assessing farmer decisions other than strict profit maximization decisions. Key et al. (2006), for example, analyse the impacts of risk on farm operators’ off-farm labour supply. In another line of inquiry, Abdulkadri et al. (2006) study the impacts of excluding risk and risk preferences on economies of scope estimation.

²In the sample of Kansas farms utilized in the empirical application, decoupled government payments represent less than 2% of farmers’ wealth (see the results section for more detail).

³An anonymous Journal reviewer raised this issue.

⁴Our empirical application is necessarily constrained by data availability. The collection of information on how land quality varies across an agent's land holdings would allow to assess the influence of land quality on land allocation decisions. Such information could be easily incorporated in our model by respecifying the production function: y_ij  = y_ij (x_ij , γ _ij )  + ε _ij , where γ _ij would represent land quality in plot j of firm i. Allowing for changes in land quality across an agent's plots would probably confirm the well known practice that farmers follow to set aside the land with the lowest quality and productively use the more fertile plots.

⁵Because, for our sample of farms, crop land remained almost constant during the period of analysis, A is assumed to be fixed.

⁶The Conservation Reserve Program (CRP) does not factor into the land idling consideration of this article. Because it consists of long-term contracts that promote the establishment of conserving covers on environmentally sensitive cropland, it precludes adapting land use to changing market conditions or policy regulations.

⁷As Just and Zilberman note, the assumption of risk neutrality is very common in models with stochastic production and is necessary for the dual cost and production functions to be independent of risk preferences. This assumption allows to derive a theoretical framework that is more tractable at the empirical level.

⁸Note that this simplification is economically reasonable as it represents two possible corner solutions that can apply to our problem, i.e. that farmers decide not to set land aside or diversify the crop mix.

⁹This was suggested by an anonymous Journal reviewer.

¹⁰To be able to do so, a complete panel of data is built out of our sample.

¹¹We should note here that another allocation mechanism based on profit maximization was also used, but yielded inconsistent results. This is not surprising in light of Just et al. (1990) findings that the behavioural method is superior.

¹²Halloran (2006), on the other hand, finds that for potato cropping systems, rotating program with nonprogram crops can improve economic viability and reduce risk.

¹³The problem applies to crop i  = 2.

¹⁴It is important to note here that other alternatives were also considered, including the use of farm-level values whenever possible (and averages otherwise) or the use of the Kansas Farm Management Association crop budgets (http://www.agmanager.info/crops/). However, these alternatives yielded results in contrast to widely accepted previous research results and thus were discarded.

¹⁵This estimate is compared to actual government payments received by each farm. If estimated PFC payments exceed actual payments, the first measure is replaced by the second.

¹⁶Differences in the variance of profits might partly reflect the fact that y ₁ is a composite output and y ₂ is a single crop.

¹⁷High idle land elasticities are partly because of the low initial values of this variable.