We know the cost of living varies considerably across the country. It can even vary significantly just across town. But what really drives these differences? In a word, housing. Sure, the price of going out to eat can vary by a few bucks per plate here and there, plus we know haircuts and accounting services differ as well. But as seen in this edition of the Graph of the Week, prices across the country for most goods and services fall within a range that varies by 20 percent or so. Of course that’s nothing to sneeze, especially if incomes don’t keep pace. However the variation in costs for most goods and services pales in comparison to the differences seen in housing costs which can vary by 200 percent.
Housing represents a big component of consumer spending and inflation overall of course. In Detroit and St. Louis housing accounts for around 40% of the local CPI basket, while it is nearly 50% in New York and San Francisco. Before BLS murdered the Portland-Salem CPI, housing was 45% of the basket, on par with Denver and Seattle.
Part of the reason housing costs vary so much is demand, a lot of people want to live in the large urban areas with strong economies and a high quality of life. However we also know a key reason is the low levels of supply or lack of new construction. As Tim Duy and I wrote a few years back, expensive cities don’t built much housing, for a variety of macro reasons in addition to local policies.
That said, we need to keep in mind that housing isn’t the only thing impacting household budgets. Non-housing items account for 50-60% of the CPI basket and more like 80% of consumer spending overall. So even as we spend most of our money elsewhere, it is clear that housing differences are the key driver behind the overall cost of living.
Finally, as an aside, product availability and not just prices matter when it comes to quality of life as opposed to cost of living. If you are craving Peruvian food or northern Indian food or the like, those types of restaurants may not be in every city. This is true even if the number of restaurants or number of restaurants in a broad category like Asian is the same. This added wrinkle matters at the personal level and may affect how you feel about a certain city, but will definitely be missed at the macro level when we look across the country using standard economic measures.
Mark McMullen
Josh Lehner
Oregon Office of Economic Analysis 
[…] Source: What Drives Cost of Living Differences? (Graph of the Week) | Oregon Office of Economic Analysis […]
By: What Drives Cost of Living Differences? (Graph of the Week) | Oregon Office of Economic Analysis | eClips on January 24, 2020
at 9:30 AM
In the current election cycle, local discussions about housing costs and availability are key topics. However, too many of these discussions seem to focus on artificial means of reducing housing costs, e.g., rent control, removing zoning and permit restrictions, and/or government mandates to require more “affordable” housing. Some also believe we can build our way out of this problem, but it seems unlikely that builders would wish to construct more homes that they would then have to sell for less and less money. The link to your 2015 and 2017 articles was helpful, but an additional factor to consider would be the growth in incomes relative to housing prices, which seems to be a major contributor to the decline in housing affordability over time.
By: Scott Boone on January 24, 2020
at 9:51 AM
Thanks Scott. You’re right. We also need to consider interest rates for the ownership side. Factoring that in housing costs as share of Portland median family income is relatively tame right now. Challenges are more about down payment or credit scores or whatnot. And of course on the rental side where lower interest rates don’t help renter households.
By: Josh Lehner on January 24, 2020
at 9:57 AM
“Before BLS murdered the Portland-Salem CPI, housing was 45% of the basket…”
Sorry if I missed an earlier post, but to what action by BLS are you referring?
Thanks!
By: Herb Holloway on January 24, 2020
at 10:27 AM
In 2018 the U.S. Bureau of Labor Statistics (BLS) revamped which metro areas they publish local inflation data for. They cut some and added a couple, basically. And Portland lost out. We are the largest metro for which there is no local CPI measure. BLS is still collecting data but no long publishing a Portland specific measure. The State of Oregon and our forecast too has switched over to West Region CPI instead.
By: Josh Lehner on January 25, 2020
at 7:41 AM
Thanks for the interesting blog. Does cost of living take tax rates into account at all or does it just look at goods, services, and rents?
By: WILLIAMS Michael A * ODOE on January 24, 2020
at 11:12 AM
Something like this only looks at the prices of things relative to the US average. It doesn’t take into account incomes or disposable incomes, which one would do if we are doing a full analysis of whether a household can or cannot make ends meet or the like.
By: Josh Lehner on January 25, 2020
at 7:42 AM
I would also add the cost of commuting from “cheaper” locations. So just because rents are cheaper the farther one lives from the more expensive urban areas, the cost of getting to and from those areas needs to be factored into the cost of living.
By: Leo Koen on January 24, 2020
at 11:16 AM
You’re right. This measure only looks at prices. So the cost of gas or transit is in there but not how much of it is used. Housing plus transportation costs is an important measure! But we don’t have great transportation cost numbers in terms of how much someone drives from this neighborhood or the like.
By: Josh Lehner on January 25, 2020
at 7:44 AM
Josh, for the uninformed, what is the BLS you refer to that murdered our CPI?
Nancy DeSouza Sent from my iPhone
>
By: Nancy DeSouza on January 24, 2020
at 1:44 PM
In 2018 the U.S. Bureau of Labor Statistics (BLS) revamped which metro areas they publish local inflation data for. They cut some and added a couple, basically. And Portland lost out. We are the largest metro for which there is no local CPI measure. BLS is still collecting data but no long publishing a Portland specific measure. The State of Oregon and our forecast too has switched over to West Region CPI instead.
By: Josh Lehner on January 25, 2020
at 7:36 AM
Josh, Your post illustrates one of the reasons that economic analysts — such as yourself — need to to replace the “30% of household income” as the metric for “housing cost burdened.” The real measure is what a household has left over of their income after basic costs in relation to their real costs of decent housing. Both values in this comparison are location-dependent, as well as household constitution, including number of adults and children.
By: pconte on January 24, 2020
at 4:36 PM
30% is an imperfect measure, no doubt. See our previous look at estimating residual income by zip code. One issue is we lack good data to do better measures easily. https://oregoneconomicanalysis.com/2015/10/05/housing-affordability-and-residual-income/
By: Josh Lehner on January 25, 2020
at 7:33 AM