Sunday, October 31, 2010

A Good Krugman Post

It's here. Here's an excerpt:

"The background to the world economic crisis is that we went through an extended period of rising debt. Now, one person’s liability is another person’s asset, so rising debt made the world as a whole neither richer nor poorer. It did, however, leave the borrowers increasingly leveraged. And then came the Minsky moment; suddenly, investors were no longer willing to roll over, let alone increase, the debts of highly leveraged players. So these players are being forced to pay down debt.

The process of paying down debt, however, must obey two rules:

1. Those who pay down debt must do so by spending less than their income.
2. For the world as a whole, spending equals income.

It follows that

3. Those who are not being forced to pay down debt must spend more than their income.

But here’s the problem: there’s no good mechanism in place to induce those who can spend more to do so. Low interest rates do encourage spending; but given the size of the debt shock, even zero rates are nowhere near low enough.

So since the world economy can’t raise the bridge, it is lowering the water: without sufficient spending from those who can, the only way to make the accounting identities hold is for incomes to decline"

One of the biggest problems for professional economists and laymen alike is to understand how accounting identities are appropriately or inappropriately used. This is the right way to use it. Accounting identities have to hold, simply by definition. That doesn't tell you what will happen of course. What happens depends on what the elements of an accounting identity do because of exogenous forces. Note that here Krugman is not saying "Y=C+I+G so if we increase G it will increase Y". He's saying these identities have to hold, so income is going to be set depending on the decisions people make regarding demand, savings, etc. Income has to equal spending and savings has to equal investment - but what levels these come in at are determined by behavioral responses.

No comments:

Post a Comment

All anonymous comments will be deleted. Consistent pseudonyms are fine.