Fertilizer applications should be based on relevant yield response models and be economically justified. In this study, we defined the yield response models of cereals to phosphorus (P) fertilization on the major Finnish soil types by the means of a research synthesis and meta-analysis. We also calculated economically optimum P rates under different price combinations of P fertilizer (1-3 € kg -1 ) and cereal yields (100-300 € tn -1 ), for 1-year decision interval of P applications. Our material consisted of data on P fertilizer experiments conducted in Finland during the last 60 years on clay, coarse-textured mineral and organic soils, with variable soil test P (STP) status at the start of the experiments. The cereals cultivated were spring barley, oats, spring and winter wheat, and winter rye. The applied P rates ranged between 6 and 100 kg ha -1 . For low STP classes, Mitscherlich-type exponential models were appropriate for all soil groups, predicting 17-27% higher maximum yields when compared to the controls without added P. In contrast, for medium and high STP classes, the yield responses to increasing P rates were scattered around zero in most soils.Phosphorus fertilization had also negligible effect on 1,000-seed and test weights. On Finnish cereal farms, when P fertilizer is purchased, the present P rates allowed by the Agri-Environmental Programme are uneconomically high. It appears that P fertilization can be substantially reduced on majority of Finnish fields, or even omitted for years, without economic loss under current (2 € kg -1 ) or higher P fertilizer prices.