by Calculated Risk on 6/04/2008 12:04:00 PM

First, the outlook for housing is negative. Just because I mention a few minor shreds of good news for housing, doesn't mean my view has changed. It hasn't, especially for existing home sales and prices.

OK, for those hoping to buy at lower prices, the outlook is rosy. Now that we’ve gotten the semantics out of the way, the overall outlook remains negative for house sales and prices. With tighter lending standards, demand will remain weak, and supply is already at record levels and still rising - especially the supply of distressed homes. This record supply, combined with continuing weak demand, will put pressure on house prices, probably for several years, in real terms, in many areas.

Now let’s return to cancellations. The cancellation issue can be confusing.

When looking at new home sales, we are interested in net sales, but the Census Bureau reports gross new sales. A simple equation would be:

Sales (net) = Sales (gross) – Cancellations + Sales of earlier cancellations.

In the long run, the cancellation terms balance out, and the Census Bureau numbers are what we want. In other words, Sales(net) = sales(gross). But in the short run, with cancellations increasing, the Census Bureau probably overestimates sales; and with cancellations decreasing, the Census Bureau underestimates sales.

We don’t have the raw data for cancellations and sales of earlier cancellations. However the public builders typically report net sales and cancellation rates. Using the public data, we can estimate net vs. gross sales for the industry, and adjust the Census Bureau estimates accordingly. Luckily the analysis isn’t too difficult: when cancellations rates are rising, net sales are typically below gross sales, and when the cancellation rates are falling, net sales are usually above gross sales. Right now cancellation rates are falling and the builders are reporting they are reducing their inventory of “unintended spec homes”, and net sales are above gross sales.

Currently cancellation rates are still significantly above normal for the home builders. As an example, Toll Brothers just announced a cancellation rate of 24.9%, far above their historical rate of 7%. But the key for adjusting the Census Bureau numbers is that the cancellation rate has declined from 38.9% two quarters ago. What matters for this calculation is the change in cancellation rate over the previous six months since that is the time it typically takes to build a home.

The same is true for other builders. Another example is Hovnanian: they reported a cancellation rate of 29%, down from 40% two quarters ago. Hovanian averaged 23% cancellation rate in 2004 and 2005 (cancellation rates are builder specific because of their downpayment and pre-qualification requirements).

Since cancellations rates are now falling, this suggests that the Census Bureau is currently underestimating sales for new homes. This is not a positive comment about these individual builders, but this helps analyze the entire market.