This paper examines the relationship between farm size and productivity in China's agriculture. In developing agriculture where there is a broad range of farm sizes, farm size and productivity or output per unit of land are often found to be inversely related. In China, where average farm size is small and the distribution of farm sizes is relatively compact, farm size and productivity are weakly inversely related. However, when we utilize the egalitarian principle during land allocation in China and use imputed homogenous land area rather than actual land area in the regression, the inverse relationship between farm size and productivity disappears. Hence, the strong inverse relationship that some studies have found are undoubtly due to a number of methodological problems, including the failure to account properly for land quality differences and the method of land distribution. Applying the principal agency theory, we also discuss the possibility that market inefficiency may contribute to the inverse relationship. We examine the necessity and validity (Hahn & Hausman 2002) of the instrumental estimation applied in the paper. The corresponding variance estimates are adjusted as Murphy & Topel (1985) suggested.