Global Insight and "undervalued" housing: methodology (buying a house, buying)

[elflord1973]

Since it proved futile getting the Bulls to come clean about the methodology used in the Global Insight report (as they prefer to accept things on blind faith instead of read them), I dug it up myself.

http://www.globalinsight.com/gcpath/...ethodology.pdf

There are at least two substantial issues I have with this method (which, incidentally, concluded that NY metro is "fairly valued").

(1) property taxes are not taken into account. For areas (such as Westchester and Essex!) where property taxes are a substantial proportion of ones housing costs, this is a glaring omission.

(2) as I understand it, they are essentially doing a multi factor linear model, and using interest rates on an "annuitized basis". I think this means that they are using the monthly payment per $1 borrowed a factor for the model. The problem is, if the market has a tendency to overcompensate for low rates, then they will conclude that the values are "fair" during a bubble caused by low rates.

The numbers I have run on this do support the idea that this actually happens -- housing becomes less affordable when rates go down, because the market overcompensates for the low rates.

[pfwnj]

Actually, that's a pretty decent accounting of how fairly valued houses are at this instant in time. Interest rates are at historic lows so that will make things much more affordable. Omitting property taxes will definitely skew things, but their overall methodology seems pretty fair for an instant in time.

However, since anyone buying a house today will not be selling tomorrow, but rather some years out, it would be a mistake to assume that since the house is fairly valued based on existing conditions, it will appreciate or holds its value as those conditions change. It's difficult to believe that interest rates have anywhere to go but up in the long term, and as they point out, this will drastically reduce the ability of consumers to buy homes.

The funny thing is that since rate hikes tend to be gradual, these guys could be calling everything fairly valued each quarter despite quarter over quarter price drops. Q1 at 4.5%? Fairly valued. Q2 at a rate of 5% and with a 5% drop in average value? Fairly valued! And so forth.

[elflord1973]

Originally Posted by pfwnj

Actually, that's a pretty decent accounting of how fairly valued houses are at this instant in time. Interest rates are at historic lows so that will make things much more affordable. Omitting property taxes will definitely skew things, but their overall methodology seems pretty fair for an instant in time.

It's not so much a "fair value" methodology as it is a data-driven prediction method that doesn't take prior state into account.

I like their method as an elegant prediction technique, and there are advantages to steering clear of judgments about what you think the market "should" do, and just throwing a very simple model at the problem instead.

You make a good point about prices dropping in the face of a rate hike and yet remaining "fair". To use this method as a prediction of future prices, one really needs to enter forecasts of the state variables into the model.