Momentous changes are under way in what central banks are and what they do. We are used to thinking that central banks' main task is to guide the economy by setting interest rates. Central banks' main tools used to be "open-market" operations, i.e. purchasing short-term Treasury debt, and short-term lending to banks.
Since the 2008 financial crisis, however, the Federal Reserve has intervened in a wide variety of markets, including commercial paper, mortgages and long-term Treasury debt. At the height of the crisis, the Fed lent directly to teetering nonbank institutions, such as insurance giant AIG, and participated in several shotgun marriages, most notably between Bank of America and Merrill Lynch.
These "nontraditional" interventions are not going away anytime soon. Many Fed officials, including Fed Chairman Ben Bernanke, see "credit constraints" and "segmented markets" throughout the economy, which the Fed's standard tools don't address. Moreover, interest rates near zero have rendered those tools nearly powerless, so the Fed will naturally search for bigger guns. In his speech Friday in Jackson Hole, Wyo., Mr. Bernanke made it clear that "we should not rule out the further use of such [nontraditional] policies if economic conditions warrant."
But the Fed has crossed a bright line. Open-market operations do not have direct fiscal consequences, or directly allocate credit. That was the price of the Fed's independence, allowing it to do one thing—conduct monetary policy—without short-term political pressure. But an agency that allocates credit to specific markets and institutions, or buys assets that expose taxpayers to risks, cannot stay independent of elected, and accountable, officials.
In addition, the Fed is now a gargantuan financial regulator. Its inspectors examine too-big-to-fail banks, come up with creative "stress tests" for them to pass, and haggle over thousands of pages of regulation. When we think of the Fed 10 years from now, on current trends, we're likely to think of it as financial czar first, with monetary policy the boring backwater.
A revealing example of where we are going emerged last spring, admirably documented on the Fed's website. Using its bank-regulation authority, the Fed declared that the banks that had robo-signed foreclosure documents were guilty of "unsafe and unsound processes and practices"—though robo-signing has nothing to do with the banks taking too much risk.
The Fed then commanded that the banks provide $25 billion in "mortgage relief," a simple transfer from bank shareholders to mortgage borrowers—though none of these borrowers was a victim of robo-signing.
The Fed even commanded that the banks give money to "nonprofit housing counseling organizations, approved by the U.S. Department of Housing and Urban Development." Why? Many at the Fed see mortgage write-downs as an effective tool to stimulate the economy. The Fed simply used its regulatory power to help meet that policy goal.
Even if you think it's a good idea (I don't), a forced transfer from shareholders to borrowers in pursuit of economic policy is the province of the executive branch and Congress, subject to reproof from angry voters if it's a bad idea.
The Fed said candidly that it was acting "in conjunction" with the state attorneys general and the Justice Department. So much for an apolitical, independent Fed.
True, $25 billion is couch change in today's Washington. But you can see where we are going: Hey, nice bank you've got there. It would be a shame if the Consumer Financial Protection Bureau decided your credit cards were "abusive," or if tomorrow's "stress test" didn't look so good for you. You know, we've really hoped you would lend more to support construction in the depressed parts of your home state.
Conversely, when the time comes to raise interest rates, how can the Fed not consider that doing so will hurt the profits of the too-big-to-fail banks now under its protection?
This is not a criticism of personalities. It is the inevitable result of investing vast discretionary power in a single institution, expecting it to guide the economy, determine the price level, regulate banks and direct the financial system. Of course it will use its regulatory power to advance policy goals. Of course, propping up the financial system will affect monetary policy. If we don't like this sort of outcome, we have to break up the Fed into smaller agencies with narrowly defined mandates.
The European Central Bank's political power is, paradoxically, even greater. The ECB was set up to do less—price stability is its only mandate, and it is not a financial regulator. But the ECB holds the key to the euro-zone's central fiscal-policy question. It has bought the debts of Greece, Italy, Spain and Portugal, and it is lending hundreds of billions of euros to banks, which in turn buy more of those sovereign debts.
Eventually, the ECB will have to suck up this volcano of euros, by selling back the bonds it has accumulated. If it can't—if the bonds have defaulted, or if selling them will drive up interest rates more than the ECB wishes to accept—then the ECB will need massive funds from German taxpayers to prevent a large euro inflation. It might ask for a gift of German bonds it can sell, as "recapitalization," or it might ask for a bond swap of salable German bonds for unsalable southern bonds. Either way, German taxes end up soaking up excess euros.
Our views of central banks have changed every generation or so for centuries. The idea that central banks are centrally responsible for inflation and macroeconomic stability only dates from Milton Friedman's work in the 1960s. It's happening again, and it would be better to think clearly about what we want central banks to do ahead of time.
Mr. Cochrane is a professor of finance at the University of Chicago Booth School of Business, a senior fellow at the Hoover Institution, and an adjunct scholar at the Cato Institute.