When fertilizer and crop prices fluctuate, the economically optimum rate of application varies as well. The extent of adjustment that is advisable depends on the availability of data describing the relationships between soil test levels, rates of application, and crop responses in terms of yield and quality. These relationships may also be specific to other soil characteristics as well, such as soil texture and tillage management. This study examined responses of alfalfa, corn and soybeans to phosphorus and potassium in soils of the province of Ontario in Canada. The analysis showed that the optimal rate for profitability in the long term differs from that of the short term. Over the past 15 years, variations in potash: crop price ratios had considerable impact on optimum rates calculated from single-year crop response functions. However, when rate decisions are based on response frequencies as a function of soil test level, the impact of price ratio is dampened by the longer-term necessity of replenishing crop removal. The answer to the question of how far to cut back on application rates becomes dependent on the soil test level and the expectations regarding price ratios in the future. As in many soil test calibration databases, soil test levels explained only a small proportion of the total variability in measured crop responses and single-year optimum rates. The impact of this uncertainty on potential financial losses encountered with various management approaches will be further analyzed and compared.