Introducing medium- and long-term productivity responses in Aglink-Cosimo

This report aims at enhancing the Aglink-Cosimo model by incorporating agricultural productivity growth. It contains a first attempt to develop a measure of the productivity response of agricultural commodities represented in Aglink-Cosimo. In the same spirit as Griliches (1963), this work takes into account the role that specific input changes have on explaining medium- and long-term productivity responses. Aglink-Cosimo model is a partial equilibrium model used to analyse the short to medium term development of annual supply, demand, and prices for the main agricultural commodities produced and traded worldwide. The model improvements described in this technical report aim at capturing endogenously shifts in commodity productivity growth. In the report we first present how productivity growth is considered at present in the model. Subsequently, we review the literature on how technical progress and price elasticities of yield relate to yield growth. From the literature, we extract a way of capturing endogenously productivity growth. We document the Aglink-Cosimo model changes and focus on some specific cases. Finally, we prepare a scenario analysis on the actual baseline and on an economy with a higher GDP growth rate. The scenario, applied to both cases, studies the effect of a 20% labour price increase on endogenous productivity growth. In this report, we present several findings relevant to the Aglink-Cosimo community. First, inelastic Aglink-Cosimo crop yields used for the ten-year period commonly simulated by this model are not consistent with indications in the literature of greater long-run response. The relevant literature suggests that yields will be more responsive to sustained changes in price levels if given time for all reactions of crop supply, certainly including farm input supplier investment decisions as well as on-farm adjustments, than if assessed using responses to annual price variations alone. Second, Aglink-Cosimo users can exploit relationships we develop between the economic development of a country, input composition, and yield elasticity. These relationships can be used to calculate long-run yield elasticities for the model. Moreover, these relationships can be built into a set of equations that can help prepare Aglink-Cosimo for long-run projections. Long-run projections predicated on decades of economic growth imply fundamental changes to the agricultural sector. Thanks to the relationships we define here, the implications of development for crop input price indices and crop yield elasticities can now be taken into account. Our third finding is that long-run crop yield response can be successfully introduced to the model. We adjust existing yield elasticities by adding long-run returns to crop production and long-run responses to these returns. We tie key parameters to the level of economic development. The final, successful simulations of this report use a version of Aglink-Cosimo with these revisions. While we do not at this time project far into the future, our simulations show how the relevant literature and available data can be used as the basis for long-run analysis with Aglink-Cosimo.