Certainly some home depreciation in select metro areas is going to happen, but the impact is marginal, as in at the margin, and there is significant doubt as to whether the national average home price will depreciate. Too many so-called “economists” act like everything throws a switch from one binary position to another, and they don’t think marginally, which makes me think that they don’t think. How exactly would one analyze the marginal impact of any overall changes in home prices?

Any future changes in home prices would have to be taken in concert with past changes, i.e. appreciation since purchase, the terms of purchase, affordability of reset, credit-worthiness considering built-up equity for a possible refinance, etc. After all, a homeowner with a zero down interest-only loan that resets today after three years in place, would have had an average appreciation of about 33% and would have 25% equity, which may well allow him to refinance fixed on better terms than his adjustable reset. Since past appreciation is a function of location and time in the home, and equity is a function of loan terms, time in the home, and money down, one needs to know all of this information for all homeowners, or at least have some very definite concrete distributions of who falls into which buckets, before one could put together a histogram of reset impacts. One must also have, in addition to reset impacts, a knowledge of the affordability of the loan resets in some histogram form, in order to estimate what marginal proportion of resets would cause a marginal increase in defaults by the marginal homeowner. Assumptions would obviously be present in abundance.

Following a projection of homeowner defaults, one would then have to take into account ameliorating factors including foreclosure laws by state, the banks’ abilities to hold REO given their various balance sheets (so one knows how motivated they are to re-sell), the real estate prospecting market (with the RE P/E ratio, how much can it rent for vs. how much it costs) so one can estimate how competitive the investment market is, and probably a few other things before one could begin to get a grip on the pain the banks might feel.

Also following a projection of defaults, one would have to examine the impact on consumer spending by the person defaulting (which is minimal IMO, because they’ll just spend the same housing dollar elsewhere, probably renting, without impacting their discretionary spending) and the impact on those who kept their homes and their subsequent spending decreases or savings/investment decreases.

Combine that with the corporate pain of the lenders and how that might marginally impact other spending, and there is your impact calculation.

Then and only then, after amassing all that data and analysis, would one be able to make a reasonably definitive (based on the quality of data and assumptions baked into it) estimate of what the impact MIGHT be. Assuming one can even get the data in sufficient detail.

Of course, one also assumes that everything else baked into their G-D-P model is 100% accurate, too, because inaccuracy with the starting estimate bleeds through.

Does any reader actually think that the “economist bloggers” have done that work, or do you think they just pulled some large number like total dollar value of loan resets (which is immaterial, what is material is the marginal increase in loan servicing costs, not the total amount of the loan) or total count of borrowers (immaterial without a discussion of their financial condition, time in the loan, terms of the loan, etc.) out of their ass so that they can go forth and scare folk?

I’m of the opinion that the so-called “experts” haven’t even thoroughly thought out how they might calculate the impact, much less taken any concrete steps towards doing so. After all, the above process isn’t likely to land you the role of “Dancing Bear” on Kudlow, because it lacks “pizzazz” and can’t be delivered in a sound bite.

For those wanting more information on the housing price series, there are two primary sources, the HPI and the S&P Case-Shiller index.

In its 17-year existence, the HPI has never had negative annual appreciation, i.e., losses on the national level. Standstills, yes, but not losses. Now, in the 20-year history of the S&P Case-Shiller (C-S) index, there have been annual losses of 6% or so, but those should be asterisked, and occurred in 1991.

The C-S used only 10 metro areas until 2000! Today, it uses only 20! One of them is Detroit, which taints their data, making it more levered to the domestic auto industry than home appreciation overall, and turns the index negative for recent months. In 1991, one of the then-only-ten areas was Boston, which had its own, local troubles at that time. Meanwhile, the HPI uses data from all metro areas, hundreds of which are ranked.

The HPI has sales and refinances, the C-S uses only sales.

C-S has practically all transactions for its limited sampling area, whereas the HPI has only Fannie and Freddie data but has it everywhere.

C-S data is updated monthly with a lag, HPI data is updated quarterly with a lag.

Methodology is the same in terms of index construction as proposed by Case and Shiller in the ’80s.