The aim of this paper is to analyse the regional productivity differentials on dairy farms in Denmark, Finland and Sweden. Several methods have been suggested for analysing productivity differentials in agriculture between groups of farms or countries. Hayami [5] and Hayami and Ruttan [7] suggested the meta-production function approach. This idea has been further developed by Lau and Yotopoulos [9] and Fulginity and Perrin [13]. Battese and Rao [2] suggested the meta-frontier analysis for these comparisons. One of the advantages of meta-frontiers with respect to metaproduction functions is that they are able to separate technological differences from the differences in technical efficiency. Battese et al. [5] and O’Donnell et al. [16] have extended this idea and developed both parametric and nonparametric approaches. In this paper, we extend the metafrontier analysis to the concave nonparametric least squares estimation of the production function suggested by Kuosmanen [18,19]. In addition, we compare the results with the approach where the estimation of meta-frontier can be avoided. The reference can also be the maximum output providing technology that is the one that yields the maximum estimated output, given inputs [21]. In this case the estimation can be based either on average or frontier production functions. The farm level data is obtained from the EU’s Farm Accountancy Data Network data set for Denmark, Finland and Sweden. They cover 954 dairy farms in 2003. The results suggest that different method provide slightly different results but in all approaches productivity differentials are considerable in favour of Danish farms. In addition, the Danish technology is not only dominating at the mean but also at most of the data points.