This paper examines several elements of the body of conventional wisdom surrounding the likely effects of trade liberalization on agriculture. The first point bears on the likelihood of agricultural output expanding when across-the-board liberalization begins from a position whereby the farm sector is relatively lightly protected (or taxed). Several recent empirical studies incorporating imperfect competition in the nonfarm sector have thrown doubt on this presumption. This paper provides some additional insights into why this comes about, and the relative importance of the socalled "pro-competitive" effects of tariff reform on the subsequent pattern of output changes. The next point has to do with the relative strength of two competing effects on the change in output of unsupported farm commodities when agriculture alone is liberalized. The first is the transformation effect. This reflects the switching of farm resources from supported to unsupported commodities when subsidies are removed. The second effect is the "contraction" effect which reflects the exit of farmers and resources from agriculture when commodity support is removed. This tends to lower all farm output, not just that of the previously supported commodities. The conventional wisdom is that the transformation effect dominates even in the medium run. Recent econometric evidence does not support this. Thus many studies of agricultural trade liberalization may be predicting the wrong direction of change in the output of unsupported commodities. The third issue examined has to do with the anticipated, short-run factor market incidence associated with trade liberalization. Here too, the conventional wisdom is challenged. Finally, the paper closes with a discussion of the critical role of non-agricultural interest groups in reforming agricultural policy.