Implications of water policy reforms for agricultural productivity in South Africa: Scenario analysis based on the Olifants river basin

This paper uses the water-reallocation scheme created within the National Water Act (1998) to analyze the impacts of water policy on farm livelihoods in South Africa. Based on one of the most water stressed catchments in the country, the Olifants river basin, we provide an integrated modeling approach combining water and agricultural modules to investigate the impacts of compulsory licensing and water market on crop production and investment made to improve water use efficiency. The model maximizes net farm profits and takes into account the characteristics of the agricultural sector in the region in classifying farmers between large-scale (LSFs) and emerging (EFs) groups, according to their land acreage, irrigation efficiency and historical heritage. Compulsory licensing is analyzed through curtailment of water-use rights from large-scale to emerging farmers. The water market is investigated to provide conditions under which farms trade water to complete their irrigation schedules. Our results show that, though compulsory licensing might promote a rise in emerging farmers and a re-balance of past riparian-based water allocation schemes, care should be given to the level of that curtailment rate in order to balance equity measures with efficiency objectives. Indeed, we found that the losses associated with water curtailment for LSFs are not entirely captured by the EFs. Therefore, beyond water policy, there are other factors, which also influence farms' profits and water use efficiency. The results also demonstrate that a water market provides a good opportunity to increase water use efficiency. The introduction of water market induces LSFs with good water storage facilities a possibility to trade their remaining water-use rights. It also offers EFs an alternative to diversify their water supply sources when they encounter shortfalls in amount of water allocated. (Résumé d'auteur