The article deals with the relation between assumptions of economic theories and their political implications. Two canons of economic science are being analyzed according to the degree of abstraction. A hypothesis is that the more abstract formal canon is connected with a liberal kind of economic policy whereas the more concrete canon presupposes an active state intervention in economic affairs. Several attempts at integrating both canons are studied separately (Marshall, Schumpeter, Eucken). Historic evidence is more or less consistent with the hypothesis stated above, but there happens to be one important exclusion: the general equilibrium theory is so abstract that it can imply opposite policies.