Paul Krugman uses a graphical analysis to explain what the Fed is doing with the TAF:
What’s Ben doing? (Very wonkish), by Paul Krugman: The financial crisis seems to have entered its third wave. Panic in August, then partial recovery thanks to lots of money thrown at the system by the Fed. Renewed panic late fall, then partial recovery thanks to even more money thrown in, especially the Temporary Auction Facility. And panic has set in yet again...
So the Fed is throwing another wave of money in, via the TAF and also additional loans to banks. All this lending is backed by collateral: the banks are setting aside various stuff, but probably mainly mortgage-backed securities.
How do we think about all this? I was trying to straighten out my own thoughts, and realized — after reading Steve Waldman — that the old framework I learned back in grad school for thinking about sterilized intervention in the foreign exchange market applies pretty well. [...continue reading...]
The bottom line?:
OK, this is just like the way you analyze sterilized intervention in currencies. And the usual problem with such intervention applies: the financial markets are so huge that even big interventions tend to look like a drop in the bucket. If foreign exchange intervention works, it’s usually because of the “slap in the face” effect: the markets are getting hysterical, and intervention gives them a chance to come to their senses.
And the problem now becomes obvious. This is now the third time Ben & co. have tried slapping the market in the face — and panic keeps coming back. So maybe the markets aren’t hysterical — maybe they’re just facing reality. And in that case the markets don’t need a slap in the face, they need more fundamental treatment — and maybe triage.
Update: Here's how I see the policy choice.
Suppose there are twenty people in distress due to financial market turmoil. One person in the group caused their own problems and therefore there is no reason to bail them out from facing the consequences of their choices. But the other nineteen did nothing wrong, their problems are the by-product of the choices of others, and they deserve help.
Also assume that, by changing the rules of the game, you can prevent people from repeating the behavior that caused the problems. Here's the catch. You can't help anyone without helping everyone, i.e. if you choose to help the nineteen people who deserve help, you must also help the undeserving member of the group.
Do you (a) help all twenty people and change the rules so that the same problem doesn't happen again, essentially forgiving and forgetting, letting bygones be bygones in the interest of stabilizing the economy and preventing harm to the rest of the group, or do you (b) help nobody to make sure that we don't bail out the undeserving member of the group and create moral hazard problems going forward.
The uncertainty here is whether or not regulatory changes, or changes instituted by the private sector to protect itself, will actually prevent the problem from emerging again in the future. But so long as you can effectively prevent the behavior from reemerging, then I think the choice of (a) is the correct choice. If you cannot be sure that the behavior can be prevented in the future (a) might still be the correct choice, but the cost-benefit calculation becomes a bit more complicated because the moral hazard cost must be factored into the analysis (with a change in the rules that is effective moral hazard doesn't matter because the behavior cannot be repeated). I would still choose (a) in the present situation, but the choice isn't as obvious.
One additional note about the TAF. There is a lot of concern about the Fed taking in assets as collateral that may deteriorate in value. This is not a big worry of mine, but suppose some of the assets did deteriorate in value, what would happen? Then the money created when the Fed traded for these assets turns into the equivalent of helicopter money. That is, normally the Fed trades one asset, a T-Bill, for another asset, money it just created, to increase the money supply. But this is not the only way the Fed can increase the money supply, it can also give it away - drop it from helicopters, send it in the mail, toss bags out of cars, etc.
Now, suppose the Fed trades newly printed money for a mortgage backed security rather than a T-Bill. So long as the mortgage backed security retains it's value, this is just like trading for T-Bills. But what if the value of the security the Fed is holding falls to zero? Then it is just like the Fed flew a helicopter over over a particular bank and dropped a bag of cash equal to the value of the mortgage backed security at the time of the initial trade. Essentially, the bank got the money without having to trade anything of value - it is just a money drop. We can worry about the incentive effects on those banks lucky enough to get a helicopter flyover, but, though I need to think about this more [and see this comment on what happens if margin calls are missed], from a macroeconomic perspective the defaults would not have dire consequences, at least not as dire as if the banks themselves had been holding the assets when they lost value. This is a shift of risk to the Fed with the advantage that the downside is less severe if the Fed rather than a bank is holding the assets when their value deteriorates.