I found Brad Delong's new The Week column really stimulating. The lesson seems to be that the specifically Friedmanite version of classical liberalism isn't a coherent position. Friedman wanted relatively free markets in everything but money. When it came to money, he made a huge exception and wholly endorsed central planning of the money supply. The problem is, the sources of money are in fact too many and too decentralized for Fed Central planners to control supply. That is, the Fed doesn't really have sufficiently monopolistic control over money creation. So Friedman's big monetary policy exception to laissez faire is undone by the fact that rational monetary central planning isn't really possible. That's a neat argument, and I suspect it's true.
The power of Friedman’s theory was, in part, rhetorical. “Keep the money supply growing smoothly” sounds like it means to keep the presses in the Bureau of Engraving and Printing rolling at a constant pace, printing out a steady flow of pictures of George Washington. But that is not how “money supply” actually works. In economic reality, “money supply” means not just cash money but also credit entries the Federal Reserve has made in commercial banks' accounts at the Fed; plus all the credit entries commercial banks have made in households' and businesses' checking accounts; plus savings account balances; plus (usually) money market mutual-fund balances; plus (sometimes) trade credit and the ceilings between credit card limits and consumers' current balances.
No central banker controls all these vast and varied sluices of the money supply – at least not in economic reality. When banks and businesses and households get scared and cautious and feel poor, they take steps to shrink the economic reality that is the “money supply.” Businesses extend less trade credit. Credit card companies cut off cards and reduce ceilings. Banks call in loans and then take no steps to replace the deposits extinguished by the loan pay-downs. Without a single bureaucrat making a single decision to slow down a single printing press, the money supply shrinks—disastrously in episodes like the Great Depression. Thus in emergencies, to say that all the central bank has to do is to keep the money supply growing smoothly is very like saying that all the captain of the Titanic has to do is to keep the deck of the ship level.
Friedman thought (a) that the central bank could exercise enough influence over the money supply to effectively control it, and (b) that banks and other financial intermediaries would be regulated tightly enough that what is now happening would be impossible. But he never resolved the tension between his view that banks need controls and the Chicago view that business must be unfettered.
There are a couple ways to go from here. One conclusion we might draw is that successful control over the money supply requires heavier regulation of the financial sector–to in effect centralize control over alternative sources of money supply. It's not clear to me whether Brad is calling for that or not. He seems primarily to be pushing the idea that when Monetariasm has to give way to Keynesian fiscal demand priming when monetary responses to recession get tapped out. But I think one could just as easily infer from Brad's argument that since ideal monetary central planning isn't really possible, we ought to give up trying and fully legalize markets in privately-issued money.