Soil Security (Dec 2023)
Testing for complementarity between the use of continuous no-till and cover crops: An application of entropy approach
Abstract
The carbon sequestration and overall crop production sustainability benefits of no-till (NT) farming are fully realized only when NT is used continuously over several years. Research shows that most U.S. farmers do not use the practice continuously. One of the commonly suggested reasons for intermittent tillage is the risk of a significant yield penalty during the first several years after converting to continuous no-till (CNT). Agronomic research suggests that continuous cover crops (CCR) can be both an economical and biological answer to the risk of yield reduction associated with the use of CNT, as CCR accelerates the processes of converting and storing nitrogen in the soil and improves soil structure and water infiltration. However, whether farmers consider the complementary benefits of CNT and CCR is largely unknown. The objective of this study is to test the complementarity between the uses of the two conservation practices in the State of Indiana, U.S. We combine Quadratic Programming and Entropy approaches to estimate 1st-order Markov transition matrices for tillage and cover crops dynamic models, respectively. Then, we apply Bayes' theory to test the complementarity. The data used for the analysis come from the Conservation Tillage Information Center and Indiana Tillage and Cover Crops Transect, 1992–2019. The findings show that there is no evidence supporting complementarity between the use of CNT and CCR in Indiana. The results also show that the use of CCR in the State is growing steadily at a rate of approximately 30% during 2011–2015 and then stabilize during 2016–2019, whereas the share of land allocated to CNT remains flat in the same period. In addition to being the first formal test of the complementarity between the uses of the two sustainable crop production practices, the novelty of our contribution relates to the econometric methodology and the data. We introduce an approach for estimating dynamic models of farmers' yearly choices and demonstrate the possibility of testing for complementarity between the choices with very limited – aggregated and missing – data.
