Monday, February 4, 2008

Housing Affordability in Oregon

Apropos of nothing. As a part of some other research we are doing, a colleague and I have collected county by county average housing values and average incomes from 1960 to 1999 for most US counties. I was curious about Oregon. It turns out that Oregon and Washington have hade two of the biggest increases in home values over incomes during this period. Here are the figures:



This first graph, above, shows all of the counties for which average home vlaues over average incomes have increased in the time period and by how much. There are also a few which decreased (below).


Is this meaningful in isolation? I don't know but I thought it was interesting. Note that some of the biggest increases are probably due to property being bought by people who earn incomes in another county, another state, or are retired and relocating. In other words, mega-commuters, second home buyers and retirees (e.g. Hood River).

3 comments:

Stevelle said...

Data only through 99? That's too bad. Could be really interesting to examine through '06 or '07. In fact I tried to put this together myself a little while ago but I haven't been able to find *enough* good income data from the BLS to calculate it myself.

Also, what source are you using for the housing costs?

Patrick Emerson said...

These are both from census data. I should mention there are only 5 data points for each: '59, '69, '79, '89, and '99. In 2010 I can update for 2009 income and house values.

Fred Thompson said...

Very interesting. How much of this variation can be explained by population growth do you think? It is too bad that the data are not more finely grained. It would be interesting to see if Measures 5 & 47 had an effect and how large, especially to the extent that their effects varied from county to county. Indeed, I'd like too be able to throw some kind of interaction term into the analysis of the panel data if it existed. One of the interesting puzzles about Measure 5 is that at both the precinct and county levels (the only categories for which we have data) population change was the most powerful predictor of vote change when compared with previous (unsuccessful) statewide property tax limitation measures.