Monday, October 26, 2015

The Housing Bubble involved banks

Two economists from my alma mater have an article on 538 about why the Housing Bubble tanked the economy when it burst in 2008, but the Internet Bubble of 2000 did not.
What explains these different outcomes? In our forthcoming book, “House of Debt,” we argue that it was the distribution of losses that made the housing crash so much more severe than the dot-com crash. The sharp decline in home prices starting in 2007 concentrated losses on people with the least capacity to bear them, disproportionately affecting poor homeowners who then stopped spending. What about the tech crash? In 2001, stocks were held almost exclusively by the rich. The tech crash concentrated losses on the rich, but the rich had almost no debt and didn’t need to cut back their spending.
It pains me to say this, because I love economics and I love Chicago, but Sufi and Mian get this wrong because of the core gap in monetary economics -- they miss the finance system.

In general, economics treats money as an "illusion" in that it facilitates the trade and exchange of real goods and services, but fundamentally does not impact or distort that exchange (at least to no great degree). A rose is a rose is a rose, and therefore a good is a good is a good regardless of whether it's prices in dollars or shekels. Therefore, money in general and banks in particular do not play a central role in macro monetary models, which instead focus on things like time preference, consumer expectations, etc.

In reality, money, or more particularly credit, plays a central role in the economy because of how bank lending works. When banks lend, they lever up their balance sheet, and therefore create money out of "thin air", constrained only by capital requirements on the supply side, and the number of qualified borrowers on the demand side.

Saif and Mian are unaware of this dynamic, as they show in their opening paragraphs:
In 2000, the dot-com bubble burst, destroying $6.2 trillion in household wealth over the next two years. 
Five years later, the housing market crashed, and from 2007 to 2009, the value of real estate owned by U.S. households fell by nearly the same amount — $6 trillion
These two $6T are not comparable. In the dot-com bubble, the loss wiped out venture accounts and household wealth in brokerage accounts, but neither was enabling additional lending (and therefore money supply). In the housing bust, $6T of bank capital (which collateralized the loans) was propping up an additional $120T or so (at a 5% capital requirements ratio) of money supply, so the impact on the economy was over an order of magnitude greater.

Without understanding credit, and the role banks play in the economy, economics will continue to struggle to explain the business cycle, and fall for fads like distributional spending effects.

7 Comments:

Blogger Greg said...

Good find and good commentary too.

I do think though that their analysis is more incomplete than wrong. They do miss the full impact of credit market by simply saying that the cost of the housing bubble was 6 trillion but I think their point about the dot com bubble and the housing bubble affecting different distributions is correct.... and a huge reason for the lingering poor performance of our economy. They are simply underestimating the magnitude of losses which means they will underestimate the level (and type) of response needed.

If you think the entire problem is "just" about getting housing prices back up you won't make the necessary changes to the financial system.

7:36 AM  
Blogger Ramanan said...

I am not sure what your point is Winterspeak.

This "thin air" sometimes get overemphasized. It's possible for economic dynamics to be sustainable with creating money out of "thin air", so you cannot blame it for unsustainable processes. The argument needs more details.

I am also not sure of your number $120 trillion. Are you saying banks' outstanding loans was that much. Not possible.

Roughly banks' balance sheet was of the size $10 trillion and an additional $10 trillion (or less) was off-balance sheet: loans which were securitized.

Btw question for you about the article, since you've read it: This graph doesn't say $6 trillion destroyed in/around 2000.

https://research.stlouisfed.org/fred2/series/HNONWRA027N

10:10 AM  
Blogger winterspeak said...

Hi Greg

I think the distributional impact is very small, and frankly I can tell a story about it going the other way. The rich have far more discretion about how much they save therefore they represent the margin of aggregate demand moreso than the poor.

Irregardless, residential real estate is the business cycle, and mortgages and bank lending are at the very heart of that. They got that wrong by an order of magnitude precisely because capital constrained lending isn't in models. GIGO.

RAMANAN: Never blamed "thin air" for anything! I have no problem with the process, just the analysis.

10:30 AM  
Blogger Greg said...

Im not sure what you mean when you say the rich...."represent the margin of aggregate demand moreso than the poor."

Are you saying that their affects are at the margins while the poor are the bulk of the affects? If so I agree and I thought that is what your excerpt of the Sufi and Mian paper was claiming as well.
If I understand what they are saying its that the wealth/spending power lost in the dot com was where it could be easier absorbed (in the wealthy margins) while in the housing bubble it was not, which I agree with. I just think that their number of 6 trillion misses much of the magnitude, as I understand you to be saying as well.

I concur completely about residential real estate BEING the business cycle.

11:03 AM  
Blogger winterspeak said...

Hi Greg

I apologize for not being clear, when I said "at the margin" I mean "right at the line between spending happening or not happening" or "amount of spending that is discretionary".

Aggregate demand falls when the number of purchases falls, and discretionary purchases are (by definition) more likely to fall than non-discretionary purchases. And, in general, the rich make more non-discretionary purchases than the poor, and have more ability to scale their purchasing up or down.

But again, not relevant to the discussion at hand. If you want to know why the dot com bust and the housing bust had such different impacts on the economy, you need to look at banks and bank lending, and understand how money works.

1:06 PM  
Blogger NeilW said...

"constrained only by capital requirements on the supply side"

There's no constraint on the supply side. Loans create deposits which creates demand for 'capital bonds/equity' at a price.

When the banking system has created extra deposits into the dynamic flow it has effectively reduced its cost of capital since there is now more money chasing that capital.

The only constraint on lending is borrowers willing to pay the current price of money. The current price is a dynamic feedback loop that depends upon the amount of loans issued.

Capital requirements change the price of money, but they don't stop anything in themselves. There is a dynamic amount of capital out there.

12:52 AM  
Blogger JD said...

That's a good point Neil. I'd never thought about that before. You're sort of feeding the increased capital requirements with your new money. So the ECB set banks a requirement to hit Common Equity Tier 1 ratios of between and 9 and 12 percent... and that impacts the cost of money higher than it would be if it were 3% ... but does not limit the quantum per se?

6:28 PM  

Post a Comment

Subscribe to Post Comments [Atom]

<< Home