Resilience and Dynamic Adjustments in Agroecosystems

The paper presents an investigation of agroecosystem dynamics with an application to yield data from England over the period 1885-2012. The analysis relies on a Threshold Quantile Autoregressive (TQAR) model. The model allows for lag effects to vary across quantiles of the distribution as well as with the values taken by the lagged variables. While it includes as special case the quantile autoregressive (QAR) model, the model specification provides a flexible representation of dynamics and asymmetric adjustments. Applied to wheat yield, the analysis documents the dynamics and persistence of yield adjustments to shocks. It finds statistical evidence against QAR models. The estimates indicate the presence of instability in the lower quantile of the distribution. This is consistent with presence of resilience in wheat yield.