Oskar Morgenstern, “When Is a Problem of Economic Policy Solvable?” (1953)

This file is a single-author methodological essay in economic theory and policy analysis. Morgenstern’s central thesis is that the solvability of an economic policy problem cannot be inferred from the solvability of an economic theory problem, because policy adds requirements of formulation, timing, side-effects, computation, admissible means, and social acceptability.

Even if we could assume that there were no unsolved problems in economic theory, as we understand it today, still it would not follow that every problem of economic policy would likewise be solvable.

The essay first insists that policy problems must be stated before they can be judged solvable. A policy aim must be formulated quantitatively, or at least directionally; it must specify when results are to occur; it must include side-effects and their timing; and it must tolerate only explicit degrees of approximation. Morgenstern notes that ordinary policy language often evades these standards, sometimes deliberately, because vagueness shields both instruments and outcomes from verification.

Only when all these requirements are fulfilled, can we speak of the fact whether or not a policy problem is solvable, has been solved in a concrete instance, and so on.

He then classifies possible theoretical answers: unique solutions, multiple solutions, sets of alternative realizations, no solution, undecidability, or present ignorance. Gödel’s undecidability theorem appears not as a direct economic tool but as a warning that solvability is a deep logical matter, not a mere practical inconvenience. In economics, however, the most common condition is not formal undecidability but incomplete knowledge.

The incomplete nature of economic science expresses itself in the largeness of the area of unsolved and, indeed, incompletely formulated problems.

The essay’s core conceptual move comes with the “problem of application.” Even if a theoretical model yields an optimum price, employment, or allocation system, the policy problem is wider: it must be implemented in a concrete historical and institutional setting. Morgenstern’s decisive question is not merely what result is desired, but what means are allowed in pursuing it.

In other words, no problem of economic policy can be called “solvable” or “unsolvable” unless it is clearly understood which means of obtaining a solution are admitted and which are excluded.

This turns solvability into a constrained concept. Morgenstern illustrates the point with squaring the circle, Columbus’s egg, and the Gordian knot: a problem impossible under one set of permitted operations may become easy if the tools are changed. Conversely, a solvable problem may become unsolvable when society removes instruments on moral, legal, or political grounds.

The statement that a problem has no solution for a given set of means is exactly equivalent to stating that a contradiction prevails.

The economic examples sharpen the argument. A theoretical optimum may require full labor mobility or perfectly flexible resource allocation, but these assumptions may be unacceptable in actual society. Unemployment could be “solved” by slavery, execution, work camps, conscription, wage cuts, or compulsory relocation, but the moral exclusion of many such means changes the policy problem itself. Exchange-rate equilibrium, European integration, and stabilization policy similarly become harder when traditional instruments are forbidden.

It is clear that one of the main gaps between economic theory and its application lies in the fact that the former admits far more than the latter.

The essay concludes with an ambivalent historical diagnosis. Civilization often advances by restricting permissible means; this is ethically desirable, but it raises the difficulty of policy problems beyond the level treated by much economic theory. At the same time, modern war shows the opposite tendency: the accepted means of destruction have expanded, producing vast new economic problems.

The restriction of the means in many ways expresses a progress of civilization.

Morgenstern’s relevance lies in his refusal to separate technical feasibility from institutional and ethical constraints. A policy is not “solvable” in the abstract; it is solvable only relative to a precisely stated aim, a time horizon, allowable side-effects, available computation, existing theory, and a socially admitted set of instruments. The essay therefore anticipates later concerns in mechanism design, implementation theory, welfare economics, and policy evaluation: economic knowledge matters, but so does the admissible grammar of action within which that knowledge is used.