Hi!

We are three hacks and have written this fine op-ed for the Washington Post to tell you that the U.S. housing market is just as fine as the op-ed itself.

Only four states — Arizona, California, Florida and Nevada — have had declines of more than 4 percent in home prices over the past year, according to the house price index of the Office of Federal Housing Enterprise Oversight.

Is that not great news? Of course that is completely wrong too, we are hacks after all. The latest release (pdf) of OFHEO data is for May 2008. It says:

For the 12 months ending in May, U.S. prices fell 4.8 percent. Since the April 2007 peak, prices have
fallen 4.9 percent.

That recent statistic does not separate out state data. But we are hacks. So we have used the latest State HPI Data from OFHEO’s website. That is based on old Q4 data, which in itself is significantly delayed, from 2007 and fits our purpose much better. Not that we would say such in our column. But we do admit that OFHEO data is quite problematic:

Some worry that OFHEO’s index may be missing the full extent of the crisis because it doesn’t include very high-priced homes with "jumbo" mortgages or homes bought with subprime loans — the ones being hit hardest. While one could argue that the index would be more representative if it included these transactions, the properties it does include represent more than three-quarters of U.S. homes.

By using OFHEO data, you know, we simply exclude those subprime housing bubble homes
that are currently going bust. The "very high priced" "jumbo" limit was until very recently $417,000. So we miss at least 50% of the sales in
last years in California. OFHEO only looks at Freddie and Fanny transactions for their data. Before the bubble went bust, F&F had a market share of less than 50%. We claim that is "representative". It is not, but that is fine with us. We are hacks after
all.

The OFHEO index provides broad coverage of large and small markets across the country, and each home is weighted equally. Furthermore, excluding subprime mortgages has an advantage — doing so makes the index a more representative measure of the homes owned by middle-class families. Fire-sale prices from distressed sales of subprime mortgages exaggerate the declines that patient sellers are likely to experience.

We only want to count "middle-class families." All those low class-families that are "patient sellers" in foreclosure? Why would a hack care about them?

This spring, it was much reported that the Standard & Poor’s/Case-Shiller housing price index recorded a 14.1 percent decline from March 2007 to March 2008, and there is every indication that the index’s June results will also be down significantly.

Now that trick is really nice isn’t it? By writing "This spring, it was
much reported …" we can use the March data and avoid to use the current Case-Schiller Data which shows a year over year  decline of 18.7%. Did we say we are hacks?

But this is a poor measure of what is happening to the value of most homes. The Case-Shiller index includes no data from 13 states (representing 11 percent of the U.S. housing stock) and offers only partial coverage of 29 others (with 79 percent of U.S. housing). Homes in the areas omitted or incompletely covered appreciated at a slower pace during the housing boom, and their values have been more resilient over the past two years, so the data behind the index are biased toward the markets most susceptible to dramatic swings.

Case-Schiller is actually the better index. We know that it includes more relevant areas and all relevant mortgage classes. In total it includes more house sales events than OFHEO. But its data doesn’t say what we want to show, so we will stay with OFHEO data.

Also, the Case-Shiller index weights transactions by value. For example, it gives eight times as much weight to the sale of an $800,000 home as it does to a $100,000 home, meaning it is particularly sensitive to what is happening with high-priced homes in the largest, most expensive markets.

You can imagine how we laughed when we wrote that sentence. Remember that an $800,000 home is one OFHEO does not count at all.

Yes, the Case-Schiller index is the much better measure when one wants to look at economic consequences of house price declines.  But we do not really want to do that. We only want to claim we did so. That is why we hacked this column.

See: When House A goes down from 111 to 100 and House B goes from 250 to 200. OFHEO sees an decrease of 10% on House A and an decrease of 20% on House B. It averages those two decreases to arrive at a 15% general house price decline.

Case-Schiller adds up the house prices in an area so that in their calculation House (A+B) has gone down from 361 to 300. Case-Schiller claims that such a decrease is an economic loss of 16.9% in home value in the area of House (A+B).

They are right of course and it is the economically relevant number, but who cares – we are hacks. That is why we use OFHEO data with our models.

Our models predict that as foreclosures continue to climb in many states, house prices will remain flat or decline in those states — but will not collapse.

One reason for this is that the effect of foreclosure shocks on house prices is small. Furthermore, other fundamental factors (such as employment growth and a slowing of the growth of the housing supply over the past year and a half) will cushion the impact of foreclosures.

No, we will not tell anyone what kind of abstruse model we used to calculate that. No, we do not define when a decline is a "collapse". We just say it is not going to happen. Oh yes, the last seven month employment was down sharply. That is why included older data where employment went up. And yes, we know that housing supply is at an all time record and still increases. But in our model, that did not matter at all, so how cares. We are hacks. We conclude:

[F]ears of a huge loss in home values for most homeowners — and especially for middle-income homeowners — across the United States, and fears of the devastating losses by financial institutions that would accompany them, are greatly overblown.

There, we have said it. Everything is fine. Vote McCain!