Estimating trade elasticities for South Africa’s agricultural commodities for use in policy modelling

The computable general equilibrium (CGE) model is often used to analyse the effects of policy changes because of its ability to capture multi-sectoral inter-linkages within the economy. The results of a CGE analysis largely depend on the database, policy shock and elasticities. Trade elasticities, such as the Armington elasticities, play a central role in CGE models to determine the demand substitution between commodities from different sources as a result of changes in relative prices. Because of their role, modellers are keen to know the correct elasticities for use in CGE models. Despite their importance, elasticities for South African agricultural commodities are outdated, leaving researchers to rely on value judgements. We address this limitation by estimating the Armington and export supply elasticities for individual and aggregate agricultural commodities using updated time-series data (1980–2016). The results for the two sets of trade elasticities show that estimates for an aggregate agriculture tend to be inelastic compared to estimates for an individual product, indicating a higher sensitivity of products to relative price changes. The Armington estimates were found to be closer to unity for the majority of products, suggesting that agricultural imports are imperfect substitutes for domestic products. The export supply elasticities for grains were found to be more elastic than for fruit and meat, implying that domestic grain production is relatively more responsive to price changes in the export markets. The long-run estimates for the two sets of elasticities were found to be larger than the short-run estimates for all agricultural products.