by Calculated Risk on 5/01/2007 10:00:00 PM

Bear Stearns chief economist David Malpass writes in Barron's: Two Fed Hikes Loom in the Second Half

The first-quarter GDP data show residential investment just now getting back to its normal share of GDP (4.5% in the first quarter versus the 4.4% average in the 1990s).

These numbers are incorrect and apparently led Bear Stearns to believe that the housing market will stabilize soon. In fact, Residential Investment (RI) as a percent of GDP is currently 5.04% (not 4.5%) and the average in the '90s was 4.08% (not 4.4%).

How is this ratio calculated? The following table shows the ratio using both nominal and chained dollars. It appears Bear Stearns incorrectly used the chained dollars. This is incorrect because the price indexes are different for both series (RI and GDP). The correct calculation is to use nominal dollars for both series for each period (we are calculating RI as a percent of GDP for each period, so RI is normalized by GDP).

Calculating Residential Investment (RI) as a Percent of GDP
	Q1 GDP (billions)	Q1 RI (billions)	RI as Pct GDP
Nominal	$13,632.60	$687.2	5.04% correct
Chained Dollars	$11,549.10	$515.1	4.46% wrong

This graph shows RI as a percent of GDP with both actual (Blue) and the current trajectory into the future (Red). Note that the red line is just an extension of the current rate of decline, and is not a forecast.

For those that believe RI as a % of GDP will bottom out at around 4.5%, the bottom would come in Q3 2007 (at the current rate of decline). This would be the mildest housing slump on record.

For the more pessimistic, they probably believe the current bust will be the worst on record, and the bottom would then come in Q3 2008 or later (assuming current trends).

If Bear Stearns is basing their housing forecast on RI getting back to level of the '90s, they just missed by about a year.