Using 1971–90 panel data from a Siberian province, two econometric methods are used side by side to examine technical inefficiency with a suggestion as to how the methods might be used in sequence. Estimates derived from a random effects method reveal that technical inefficiency is both substantial and not time invariant. Results using either a random or fixed effects method suggest that existing estimates of technical inefficiency in centrally planned economies may be biased downward because of the choice of the estimation method. Using either method, the increasing technical inefficiency found is likely to be one cause of the decline in the performance of centrally planned economies and their regions.