Irrigation farmers in the Lower Orange (Kakamas and Boegoeberg) and Lower Crocodile rivers (between Nelspruit and Komatipoort) areas in South Africa were surveyed during October 2003 in order to study whether water marketing has promoted efficiency in water use. This study is a follow-up on research undertaken by Armitage (1999) in the Lower Orange River area and Bate et al. (1999) in the Lower Crocodile River area. Factors associated with future investment in irrigation farming were also studied in the Lower Orange River Irrigation Scheme. Econometric procedures used included principal component analysis, and logit and ridge regression. Results from the two areas will be discussed separately. Econometric results for the Lower Orange River farmers indicate that purchasers of water rights produce lucrative export grapes and horticultural crops with relatively less raisin, wine or juice grapes and less field crops; are more specialised in production (table grapes); have more livestock (probably liquidity factor) and have a less negative view of the five-year water license review period. The water market has facilitated a transfer of water use from relatively lower value crops to relatively higher value crops, and also promoted the use of more advanced irrigation technology. An investment model using Ridge Regression indicates that the following variables are associated with increased future investment in irrigation farming; higher expected profitability and lower levels of risk perception and risk aversion (Arrow/Pratt). Results confirm that farmers who are more risk averse are likely to invest less in the future as can be expected from theory. Policies that increase risk in agriculture will have a significant negative effect on future investment in irrigation. What is significant from the results is that irrigation farmers in the Lower Orange River area are highly risk averse (down-side). Results also show that farmers who feel that water licenses are not secure expect to invest less in the future. The latter effect is thus amplified, as farmers appear to be highly risk averse. This has important policy implications, and measures should be taken to improve the perceived security of water licenses. This could be achieved by keeping farmers more informed about the practical implications of the New Water Act (NWA) (Act 36 of1998) and, specifically, water licenses. In the Lower Crocodile River area, almost all the water trades (permanent and rentals) observed in this study were from farmers above the gorge to farmers below the gorge. It is concluded that in the transfer of water some attributes in the purchasing area such as lower production risk (sugar cane) and lower financial risk and better cash flow (bananas and sugar cane) were more important than the expected income per cubic meter of water. Water supply in this area is highly irregular, while sampled farmers were again found to be extremely risk averse especially as far as down-side risk is concerned. The average water price in this area in recent years (2002 to 2003) was between R2000 and R3000 per ha (l ha = 8000 cubic meters). Buyers have large farms and are progressive farmers that purchase (and rent) from many sellers (or lessors). It is concluded that information on water transfers (sale prices and rents) is asymmetrical. Few permanent transfers have taken place in the Crocodile River in recent years. It is concluded that there are reasons why transfers at present are not processed, such as excess demand for water (due to the irregular flow of the Crocodile River, and role players should discuss these reasons and possible solutions before further action is taken. / Thesis (M.Sc.Agric.)- University of KwaZulu-Natal, Pietermaritzburg, 2004.