Wednesday, March 16, 2011

DeLong and Cowan Debate

The question is should the US Federal Government spend more than was previously optimal because of the Japanese earthquake. Cowan argues that both sides have a point. DeLong says obviously yes.

Here is DeLong's latest

The stylized fact is that the yield on US Treasuries has fallen (that is their price has risen). Brad refers the question to McCulloch's parrot which says "Supply and Demand." If the price of something has gone up, one should make and sell more of it.

Cowan has a moderately interesting counter argument

1) My post and question is about spending, but Brad has shifted the discussion to borrowing. It’s easy enough to borrow more without increasing spending, if that is needed.


With the third (actually fourth) mover advantage, I have phrased the original question so that this argument is weak. I carefully wrote "more than was previously optimal" not "more than was previously planned." I think this is the source of the disagreement. Cowan is convinced that US Federal spending is generally much too high. DeLong thinks this is true only of military spending and agricultural subsidies. This disagreement about optimal spending levels should not spill over to a disagreement about changes in optimal spending levels. But I'm absolutely sure that it does.

With the clever inclusion of "optimal" (which is the question in) Brad's reply is valid

"Unless there is something really weird going on, a government that borrows more should both tax less and spend more."

Cowan has arguments for lower spending.

2) The countervailing forces which might favor lower government spending simply aren’t mentioned. Those include lower wealth and higher tail risk ...


Here I think Brad doesn't understand Cowan's argument. He replies

"Which, if they were of first-order importance, would show themselves as higher interest rates on government debt."

Cowan is talking about the lower level and higher spread of private consumption. If we weren't in a liquidity trap, this would increase the social cost of government spending. Brad's claim is not in general true. An increase in risk can cause higher prices of safe assets and a higher cost of government spending. Basically people want to get out of real estate investments in North Eastern Japan (which real estate might be radioactive soon) and into something else including T-bills. To the extent that US government spending comes at the expense of consumption by those people, the optimal level declines. The interest rate on Treasuries is not relevant to the question of whether this is an important issue. It obviously isn't.

Destruction of wealth will affect interest rates via the marginal product of capital which should increase assuming market clearing, that is no excess capacity, that is ignoring current reality. Here I think Brad is right. There is a story for how total consumption (public plus private) should decline as the world invests in damage repair (with a high return). Again if nothing weird is going on, that means both public and private should decline. Clearly there is no sign of that as huge amounts of wealth are sitting around earning tiny returns right now.

But here I think the word "optimal" which I snuck in creates a huge problem for Brad. His view (which I share) is that the social cost of government spending is very low right now. If the current multiplier is greater than one, it might even be negative. If the multiplier is 1, the cost is the cost of leisure of the unemployed. It is likely this is negative (ask them how happy they are). In any case, the effect of a loss of wealth is that some retired people whose wealth was destroyed will want to return to paid work, so the cost of government spending is lower.

But this means that current spending was not optimal. We are in a liquidity trap with 9% unemployment. The US government should spend more than it is. But that doesn't mean that optimal spending increased.

No comments: