The Wealth Effects of House Price Declines

Dean Baker reckons that wealth effects alone would be sufficient to cause a massive recession, even absent any kind of financial crisis:

Homeowners have lost more than $5 trillion in housing wealth. There is a very well established wealth effect whereby $1 of housing wealth is estimated as leading to 5 to 6 cents of annual consumption. This implies that the loss of wealth to date would cause consumption to fall by $250 billion to $300 billion annually (1.7 percent to 2.0 percent of GDP). If you add in the loss of around $6 trillion in stock wealth, with an estimated wealth effect of 3-4 cents on the dollar, then you get an additional decline of $180 billion to $240 billion in annual consumption (1.2 percent to 1.6 percent of GDP).

Alex Tabarrok agrees, and I have quite a lot of time for this position myself. But it's worth pointing out what Baker doesn't mention here: that his $5 trillion number is speculative, and doesn't come from any official statistics.

Baker explained to me that in order to arrive at $5 trillion, he started with the Fed's valuation of America's housing stock as of end-2006: $21.8 trillion. If US residential houses were worth about $20 trillion a couple of years ago, and if they've fallen by 20% in nominal terms since then, it's not hard to come up with a total loss of $5 trillion in real terms. Baker got the 20% figure from the Case-Shiller house price index.

Note that using Baker's real-money math, if you had a $100 bill in your wallet for the past two years, he'd say that you've lost $7 over that time, thanks to inflation. Does that kind of loss show up in wealth effects? I'm not sure. But even if you ignore the inflation adjustment, 20% of $21.8 trillion is still a loss of $4.4 trillion: more than enough to cause a big drop in consumption.

The problem is that the most recent figures from the Fed peg the value of America's housing stock at... $21.8 trillion, exactly the same as it was two years ago. In order to buy Baker's calculation, you have to believe that the Fed has miscalculated to the tune of more than $4 trillion over the past two years.

Is that possible? Yes. As Baker explained to me:

For [house price] movements the Fed uses the OFHEO index which has shown a much lower rate of decline. The main reason is that OFHEO excludes subprime and jumbo loans. In the bubble markets, the median house price is close to the conformable limit, so the OFHEO index is missing much of the price decline in the bubble markets (it also missed much of the increase).
The Fed does rebase using the Census of Housing, but there is a 2-3 year lag for this.
It's certainly possible that the Case-Shiller index overstates the decline somewhat because it overweights the bubble areas, but the Case-Shiller data is likely to be much closer than the OFHEO index, since it doesn't exclude such a large portion of the market with the most movement. The Fed measure also includes new construction and improvements, which would have been around $1 trillion over the last two years.

This is all entirely reasonable, and if you buy Baker's reasoning here then his estimate of housing-related wealth effects could be quite accurate. Indeed, it could be an underestimate: after all, housing wealth is a function not of home values but rather of home equity. And if home values have fallen by 20%, home equity has surely fallen much more than that. If during normal times a marginal drop in home equity has a 5% wealth effect, it's easy to imagine that the wealth effect associated with an outright eradication of home equity could be substantially greater.

Update: Ryan Avent weighs in. "Household consumption was a positive contributor to GDP, right up until the last quarter. Surely that says something about the action of the wealth effect, no?"