How Much Will Refinancing Help the Economy

I see that I rather glossed over the housing portion of the American Jobs Act last night, so it's worth considering the matter now.  Obama wants to help homeowners refinance their mortgages at today's low mortgage rates. It's worth thinking through the effects of this, because it's yet another variation on a proposal that has been bandied about for a very long time.

The details are rather thin but as best I can tell, the government is supposed to "help" refinance a bunch of mortgages at current rates--somewhere between 4-4.5% on a 30-year.  There will be no principal reductions, and it's not clear what percentage of the mortgage market this applies to, who pays the transaction costs, and so forth.

What are the economic effects of this?

You get an immediate burst of economic activity as people process mortgage claims.  Does this create jobs, or more work for existing processors?  I tend to assume that they must be swamped with existing refis, so will probably have to hire more people.

It also puts more money into consumer pockets.  A working paper from the Congressional Budget Office estimates that about 2.9 million people would take advantage of a program like this.  This is expected to reduce defaults by about 110,000, while saving consumers about $7.4 billion in the first year of the program.  Some of that is recurring annual revenue--however, the CBO expects that number to decline in real value as the loans are paid down or the houses sold.

Call it $8 billion in stimulus.  However, against that, you have to set the costs.

What are those?  The net cost to the GSEs is expected to be about $600 million: they lose $4.5 billion in the Fair Market Value of their mortgage bonds, but that is offset with a savings of about $3.9 billion on the loan guarantees they don't have to make good, because lowering the payments staved off 110,000 defaults.

However, there's also the private banking sector.  According to the CBO, they lose $13-$15 billion in the fair market value of their portfolios.

The general response has been to dismiss this loss; the prepayments are perfectly legal and the banksters are vile, so to hell with 'em.  And hey, couldn't happen to a nicer bunch of folks.
Felix Salmon argues that the CBO in fact is overestimating costs, and understimating the benefits, because it's only looking at the effects next year.  From the point of view of the mortgage market as a whole, perhaps this is right (though given the historically unprecedented situation in the housing market, I expect they simply found it too uncertain to forecast home sales, foreclosures, and so forth).  But from the point of view of stimulus it's exactly wrong.  What matters most is what happens immediately.

As Tyler Cowen pointed out the other day, the Keynesians--especially those in the press--are paying way too much attention to flows, and almost none at all to stocks.  But stocks matter too:

Just as there is a spending multiplier, there is also a multiplier from changes in wealth. For instance declines in perceived wealth will cause people to spend less. The Keynesian AD gains from a broken window have to stem from the difference between the spending multiplier and the wealth multiplier. Under the permanent income hypothesis, there's not a lot of daylight here. Furthermore the perceived wealth decline, even if it doesn't lead to immediate one-to-one reductions in spending, can persist over several periods. Granted, PIH is not exactly correct, but still the net impact of stimulus on current employment and income won't be that large because of the negative wealth effects. Fiscal policy remains a weak pill. The declines in housing prices in recent years really have taken their toll on AD so the wealth multiplier is not to be ignored. The notion that a stable and sustainable restoration of AD actually requires some increases in perceived wealth is one of the most underrated ideas among today's Keynesians.
We're getting about $8 billion in new spending stimulus over the first year after we enact the refinance program.  But to get that stimulus, we're decreasing wealth by $13-$15 billion.  The spending multiplier needs to be more than half again as big as the wealth multiplier for this to provide net stimulus.

It's possible that this is so.  After all, one of the explanations for the recession is that money is getting stuck in the banks--it repairs their balance sheets, but it doesn't get put back into the productive economy.  In the long term, this means we make a more robust recovery, but in the short term, people are suffering.  If you think that money is getting stuck inside investment funds, then handing those people a large immediate loss in order to give other, poorer people a smaller immediate gain may on net produce more economic growth.

But the delta between spending and stimulus needn't be that big.  Luxury goods, presumably purchased by people like the wives of hedge fund managers, have actually been doing well during the recession.  This may be morally odious, but does not suggest that you can take money out of their hands without any impact on the real economy.

Heck, the delta needn't even be positive.  We know from behavioral economics that people respond more strongly to losses than to gains.  It's possible that the spending contraction from losing $15 billion would be stronger than the spending boost from $8 billion saved on mortgage payments.  The people who take a bath on their bonds might cut their spending faster than people who get a break on their mortgage increase it.

Or maybe not--the CBO thinks that it will provide a sizeable stimulus relative to the amount of federal spending, though a relatively small one relative to GDP.

Overall, the program probably won't hurt much--the bondholders probably disagree, but in the end I have to agree with Felix that they knew prepayment was a risk when they bought the bonds.  However, it probably also won't help much.  The biggest bang may be whatever controversy it stirs up on Capital Hill.