In discussions of monetary policy, I’ve noticed that some discussants fail to note the distinction between giving and buying. A young colleague asked me recently:

If you’re trying to use monetary policy to increase aggregate demand, isn’t it better to give the same amount of money to each person rather than to give it to wealthy people by buying bonds?

Similarly, in the comments on Bryan’s recent post, Steven Roth writes:

Their reasoning? If we’re printing money, it’s better to give it to the elite few who have the knowledge and wisdom to allocate it in everyone’s best interests (yeah: right), rather than distributing it widely and letting the wisdom of the crowds do its magic.

In both cases, my economist colleague and Steven Roth have made the same error: confusing giving and buying. In the case where the money is distributed widely, say, with a check per household, the government is literally giving people money. But if the Federal Reserve decides to buy bonds, it’s not giving people money. Of course, the sellers of the bonds are better off; we know that because they sold. But the Fed didn’t give them money. It bought bonds from them. And their gain is probably pennies on the dollar. In other words, if the Fed had offered to pay them, say, 97 cents per market value of the bond, they would likely sell close to zero of the bonds. In the case where the government distributes money to the household, the gain to a household from a dollar of money distributed is approximately one dollar. That’s a pretty big difference. It’s the difference between giving and buying.

Update: In the comment section of his blog post, Scott Sumner makes the point that I make above. See comment at 16. November 2010 at 17:59.