Salmon farming is a young industry which has experienced many important changes in
the production technology during its short lifetime. A high level of government-sponsored
innovative activity together with on-farm experimentation and learning-by-doing are probably
important factors behind the industry’s productivity growth. The estimated models
will tell us about the nature of productivity growth, both in terms of shifts in mean output
and in terms of shifts in the level of output risk. First-order stochastic dominance (FSD)
analysis based on the estimated models will tell us if all producers experienced technical
progress, regardless of risk preferences. Overall, the results in this paper support the view
that for risky production technologies the traditional deterministic approach to productivity
measurement provides an incomplete picture of productivity growth.