On Baker & DeLong on House Prices Mortgages and Consumption

I commented on Brad DeLong commenting on Dean Baker asking Brad Delong a challenging question.

Brad decided my comment belonged on his front page, so I guess it belongs here too.

Baker wrote

Brad DeLong tells us that he is moving away from the cult of the financial crisis (the weakness of the economy in 2014 is somehow due to Lehman having collapsed in 2008 — economists can believe lots of mystical claims about the world) and to the debt theory of the downturn. Being a big fan of simplicity and a foe of unnecessary complexity in economics, I have always thought that the story was the lost of housing wealth pure and simple. (And yes folks, this was foreseeable before the collapse. Your favorite economists just didn’t want to look.)

Just to be clear on the distinction, the loss of wealth story says it really would not have mattered much if everyone’s housing wealth went from $100k to zero, as opposed to going from plus $50k to minus $50k.


Finally, getting to the question in my headline, the current saving rate out of disposable income is 5 percent. This is lower than we ever saw until the stock wealth effect in the late 1990s pushed it down to 4.4 percent in 1999, it hit 4.2 percent in 2000. The saving rate rose again following the collapse of the stock bubble, but then fell to 3.0 percent in 2007. The question then for our debt fans is what they think the saving rate would be absent another bubble, if we eliminated all the negative equity.

I commented.

Baker is very smart and reality based, but he only discusses two components of aggregate demand, consumption and net exports. Consumption has, indeed, recovered. The anomalies are in government purchases of goods and services and residential investment. As I recall, your version of the debt story is that mortgage finance remains blocked. This is not a story about consumption by households burdened by mortgages and the counterfactual does not involve a lower estimated savings rate.

I do not entirely share your view that changes in mortgage finance are the cause of low residential investment. It could be that lending standards remain extraordinarily tight on the once burnt twice shy principal. But it could also be that the perceived user cost of owner occupied housing is extraordinarily high because the forecast relative appreciation of house prices is extraordinarily low (Shiller argues that it is also accurate, but certainly lower than it was for decades).

Back to Baker: I think he has trouble communicating with many economists exactly because he is reality based. His argument is based on the empirical regularity that the aggregate consumption time series is well fit mostly be disposable income and then by wealth including stock and housing wealth. Most economists consider the correlation of housing wealth and consumption an anomaly. High house prices imply both wealth and a high cost of housing services. The effects should roughly cancel. High house prices should cause high measured consumption (measured because not including consumption of owner occupied housing services) by causing people to substitute out of housing. In fact, they are correlated with high residential investment.

A simple model of rational inter-temporal utility maximization without liquidity constraints doesn’t fit the data at all. The immediate reaction of many economists is to conclude that the key issue must be liquidity constraints. I tend to suspect that Baker has no problem with assuming people are irrational and that high house prices make home owners feel richer even if they have no intention to sell and do not make young people who don’t own a house and will buy one eventually feel poorer. In any case, that’s what I believe. but no one should be surprised that most economists aren’t attracted to a simple story which depends on irrationality.