Economists, especially of the libertarian variety, often discuss how the market increases personal freedom. Milton Friedman (1962) has ably described this relationship in his book Capitalism and Freedom. Paul Samuelson (1963) has recently taken Friedman and others to task for this. He has claimed that "complete freedom is not definable once two wills exist in the same interdependent universe. . . . What is actually called 'freedom' is really a vector of almost infinite components rather than a one-dimensional thing that can be given a simple ordering." While no one claims that complete freedom is possible, it is the intent of this paper to show that freedom can be measured in the same sense that welfare can be measured, and that a discussion of maximizing freedom can be made operational. Before we do that let us consider why an individual in the United States is considered to be freer than one in Russia. There are many reasons, of course, but one element that I believe we would all agree on is that a citizen of the United States can criticize the government, the president, or his employer with relative impunity. In Russia (or any generally considered non-free society) such action will usually have severe repercussions-such as a few years' rest in Siberia. But even in the United States such action can have repercussions. If my employer is a Republican and a great supporter of the president, I may find myself looking for new work. If I work for the government, I may have to try to find employment in private industry. If I happen to be supporting communism I may actually find it difficult to secure a good job. In other words, the exercise of free speech, even in a " free " society, may not be costless. In fact, the essential difference between a citizen here and a citizen in Russia is one of relative cost. In both places I can criticize the government and advocate a different economic system, but the relative costs of this behavior differ significantly. The thesis, then, of this paper is that freedom can be defined in terms of welfare. A change in the cost of action (or non-action) can be considered