Lars Christensen's blog The Market Monetarist, which I make sure to read regularly, used the recent centenary of Milton Friedman's birth to discuss Friedman's views on exchange rates. The standard view of Friedman was that he was an advocate of flexible exchange rates, pure and simple. I think this view is based on an incomplete reading of Friedman, as Steve Hanke argued several years ago. (Anna Schwartz told Hanke she disagreed strongly with his interpretation of Friedman, but I think it's the only way to view Friedman's pronouncements on exchange rates as consistent over the years.)
Friedman wrote his most frequently cited essay on exchange rates, "The Case for Flexible Exchange Rates," as a proposal for a quick way for Western European countries to eliminate the exchange controls that they had established before World War II and that persisted in the early 1950s. Exchange controls hindered trade. Flexible exchange rates could allow countries to remove their exchange controls quickly, Friedman thought, thereby improving opportunities for international trade and the wealth created by the international division of labor.
In the same essay, though, Friedman also discussed the sterling area, a zone of rigid exchange rates with the pound sterling. He wrote, "In principle there is no objection to a mixed system of fixed exchange rates within the sterling area and freely flexible rates between sterling and other countries, provided that the fixed rates within the sterling area can be maintained without trade restrictions." At the time, the sterling area included most of Britain's colonies and protectorates as well as India, Pakistan, Australia, and New Zealand. It therefore extended over a considerable portion of the globe. Friedman likewise showed no objection to, or even praised, fixed exchange rates on a number of other occasions, cited in Hanke's article.
The key to reconciling Friedman's apparently contradictory positions is to understand that clean fixed and clean floating exchange rates, though differing in their degree of nominal rigidity, are similar in that both give market forces free rein. Under a clean fixed exchange rate, the nominal exchange rate is fixed and market forces determine the nominal monetary base. Under a clean floating exchange rate, the nominal monetary base is in the short term fixed (or perhaps a better word would be "set") and market forces determine the nominal exchange rate. Under intermediate arrangements where the monetary authority intervenes in the foreign exchange market to influence the nominal exchange rate and the nominal monetary base simultaneously, market forces do not have free rein and market adjustment is to some extent frustrated.
The overall impression Friedman's statements on exchange rates leave is that he considered flexible exchange rates to be the system most desirable and most politically sustainable for large and medium-size economies that were politically independent and able to keep inflation relatively low. In his policy advice, which took account of the particular circumstances of various countries, he did not, however, advocate flexible exchange rates across the board, nor were the cases where he thought fixed exchange rates would work acceptably mere unimportant exceptions.