Steve Sailer sends in some more thoughts on the issue of cost-of-living adjustments. (See here for background.):
Here’s the link to the latest Accra numbers, which are a little different from my [Sailer's] table from a couple of years ago. My vague impression is that Accra sells its cost of living data to large corporations concerned about transferring employees and adjusting their pay so that they feel they aren’t being disadvantaged by the move.
This link shows Missouri’s presentation of the data, which is self-interested: they want to show how cheap Missouri is.
Is Missouri playing games by converting city data to state data to make a point about Missouri being cheap? Perhaps. I just don’t know that much about the Accra methodology. It looks like they may be taking the average of seven metropolitan areas within Missouri as comprising the state average. And if so, are they weighting by population, or just doing a crude average?
On the other hand, California is so heavily urbanized that simply averaging the cost of living in its major metropolises would be fairly accurate for the whole state.
I suspect Accra is heavily weighting the cost of buying a house rather than renting, since its target market is corporations looking to transfer middle class homeowners. The kind of people who rent are seldom moved around the country by their employers — they’re laid off and new ones are hired in the new location.
As a Los Angeleno, I would defend the notion that California is vastly more expensive than other states. The cost of land in California filters into a lot of other costs. For example, gasoline is expensive in part because the land on which gas stations sit is expensive (there are also environmental costs, too). Car insurance in LA County is much more expensive because there are so many cars on the road to collide with. Lawyers are expensive because you have to pay them enough to be able to afford a million dollar home. And on and on.
Californians respond to the high cost of land ownership in various ways, which raises methodological issues. A much higher percentage of residents rent than in the rest of the country. Another change that’s taking place is that Americans, who don’t like having more than a nuclear family living in one house, are leaving the state, while people from other cultures where it’s normal for an extended family with three to five paychecks to live in one house (over the last few years, West Asians especially) are filling up the state. Another change is that since modern Americans frequently don’t get married until they can afford to buy a house and have kids, they get married later or not at all, and they have fewer children.
So, this raises methodological issues. If more families in California than other states are renting an apartment for $1500 per month because they can’t pay $5,000 per month to buy a home, does that make the cost of living lower than if there was higher rates of home ownership in California? If an American nuclear family of four sells their home (and moves to Utah) to an extended Armenian family of seven, that includes four wage-earners, does that make the cost of living lower? I could see how some methodologies would say that those actions do lower the cost of living in California, and for some purposes they would be valid.
I’m mostly interested in the question from a political perspective: what makes states vote Republican or Democrat?
For my purposes of looking at how the question of “affordable family formation” drives voting behavior (states with cheap housing have earlier marriages and more children and vote more Republican), the cost home-ownership in a district with decent public schools is key. So, for me, the Accra methodology is very helpful: Accra data says that Bush carried the 20 states with the cheapest housing costs, while Kerry won the 9 states with the most expensive.
Also, housing costs have diverged a lot between states over the years. The size of the cost of living difference between California and, say, Oklahoma depends a lot upon whether you bought your house in 1999 or 2006.
So, the more I think about it, the more I realize that calculating the cost of living is a very tricky problem. My impression is that the cost of a fraction of an acre just wasn’t that big a deal in calculating the cost of living until the last two or three decades — the cost of the house tended to be more than the cost of the land. And you can treat the house like a long-lasting consumer durable. But the cost of land has become such a big deal in the cost of living in recent years, and it varies so much from place to place and year to year, that it poses sizable methodological problems.
I agree that this is tricky. If a house in California costs twice as much as an equivalent house in Arizona, should that be adjusted for in cost of living, or should we say that this house in California, since it costs more, is actually worth more than the house in Arizona? I guess it depends on what purpose you’re using the numbers for.