From 1984 to 1993, in Guarapuava, Paraná (PR), Brazil, four crop rotation systems for wheat were assessed: system I (wheat/soybean); system II (wheat/soybean and common vetch/corn, from 1984 to 1989, and wheat/soybean and white oats/soybean, from 1990 to 1993); system III (wheat/soybean, flax/soybean, and common vetch/corn, from 1984 to 1989, and wheat/soybean, common vetch/corn, and white oats/soybean, from 1990 to 1993); and system IV (wheat/soybean, legume/corn, barley/soybean, and white oats/soybean). A randomized block design having four replications and plots totalizing 60m² was used. Risk analysis over that period is presented in this paper. Two types of analysis were applied on the net return of the systems: mean-variance analysis and risk analysis (safety-first and stochastic dominance). Separation of system II as the best alternative to be offered to the one who decides was feasible by both types studied. The system II showed the highest profit and the lowest risk to be offered to the farmer, as compared to the remaining systems studied.
costs; net return; stochastic dominance; mean-variance
