new theory: house prices correlate to availability of credit

[sniglet]

Here's a new theory I have been working on: there is a much higher correlation between the availability of credit and housing appreciation than there is between jobs or population growth.

If this theory of mine is true, then a contraction in credit will negatively impact real-estate more than either job losses or population declines. If true, this theory would also imply that the best way to determine how far prices will rise and decline is to track the availability of credit.

This is the big question. How can we track credit availability? I know there are statistics to show the money supply for the US, but that is such an aggregate number that it hard to parse how much is from JUST credit, and mortgages specifically. Worse, I have no idea how one would look at this from a regional level, like a particular city.

If the credit markets are national then maybe this doesn't really matter. I suspect that this hasn't always been the case. Back in the '60s and '70s I think that banks operated much more on a "local" level, and that they were more responsive to local conditions than global investor interest in mortgage securities, thus the national money supply stats might be useful for trending against real-estate prices in recent years, but might not be relevant for past years.

Maybe trend data for the types of mortgages offered (e.g. negative amortization, 100% finance, 30 year fixed, etc) might be a good proxy for the credit availability in particular regions. If my theory holds, then the periods with the greatest house appreciation would also be those with the largest proportion of "easy" mortgage types.

Oh well, proving my theory is likely just a pipe dream. I doubt that such historical data on trends of mortgage types even exists, and certainly not on a local or regional level.

[sniglet]

Here's an idea. How about tracking home prices according to credit spreads on T-bills and junk bonds? This should be a pretty good proxy as to the availability of easy credit. If the credit spreads are really low, then we can reasonably infer that lenders are offering money pretty freely to all-comers.

Does anyone know where we could find a chart showing junk bond vs T-bill credit spreads over time? Then we could just plot that against US housing price apprecation/deprecation rates.

[mike2]

You'd need to know how many people can access credit at certain levels. Not easy to figure out unless you have access to insider financial information. All us serfs can see is that credit has contracted substantially. At this point in the cycle, serf credit is rapidly contracting. Perhaps S-crow has some front line estimates, but how are we to really know where the line is going?

[sniglet]

[deejayoh]

Good luck finding the data, but not sure what it will get you.

I think most agree that the easing of credit/loosening of standards explains the run up since ~2000 - but I don't know (nor have I ever seen it hypothesized) that it explains the run ups in the 70's, 90's or even '95-2000. When I got my first mortgage back in '97, there was no such thing a zero-down, piggy back no PMI loan. At least not that I could find - and I needed it!

I can send you the data on jobs (either creation or unemployment) and population growth vs. home prices. There is almost no correlation between any of them on either a time series basis or across markets.

The best predictor I have found is income/treasury rate (as a proxy for affordability) which correlates to housing prices (OFHEO at a state level) almost perfectly for every state in the country for the period from 1975 to 2001 (r>92%). FWIW, that covers all of the boom periods on Tim's chart and as I said, it's for 50 different states. Not a bad fit considering all that. And you can plainly see the impact of the credit boom in that data as the correlation for the last 5 years is only 60%!

[sniglet]