Over the past year, I’ve had a friendly disagreement with a few co-workers.
Together, we have written several studies on economic development. Our work focuses on the role quality of life plays in population and employment growth. This work uses statistical models to untangle how families and businesses rank each county in the United States on quality of life. The way it works is quite intuitive, but it benefits from economic thought experiment.
Suppose every house in America is identical, with the same construction, the same color scheme, the same size and the same appliances. In this case, the difference in house prices would tell us how much residents value the community containing the house. Higher prices mean nicer communities and lower prices mean less attractive communities.
Now suppose that all workers in America are identical, with the same education and experience, and that only one profession is available. In this case, any difference in salary would reflect how desirable the location was. Workers would demand a premium wage to live in an unpleasant place and accept a lower wage to settle in a nice place.
These dynamics are familiar to every real estate agent and human resource professional in America.
Of course, we don’t have identical houses or workers, but we can create statistically identical houses and workers by controlling for things that we know affect the price of a house or worker wages. Using this statistical model of house prices and wages, we can predict quite well what house prices and wages will be in almost every county. But, the part we can’t predict — the unexplained real estate price and salary — tells us how American families and businesses value every county in the country.
This small, modest measure isn’t perfect, but it explains much of the population growth and an even greater share of job growth among US counties. In fact, this explains job growth more than differences in tax rates, tax incentives, economic development spending, highway spending or infrastructure combined.
This metric looks a lot like a stock price for each county. This measure alone does not explain why families and businesses choose these locations. To explain why they chose each location, we correlated our measure of quality of life with approximately 500 different measures of local amenities and conditions. This is the interesting part that causes the friendly disagreement within the research team.
Some of the results made sense, like a slightly warmer January temperature or a cooler July, as well as mountains. There is not much to change here for a decision maker. We saw a better quality of life in places that had more arts and recreation establishments, more grocery stores, and more recreational establishments. These are the great things that Mother Nature and the private sector provide. However, none of these things really explained much of the difference in quality of life in the United States.
The most important factor was the share of the county’s GDP spent on K-12 education. The second was a low crime rate and our third most important effect was a healthy environment as measured by average resident health. These findings were not surprising. Economists have known for half a century that school quality alone accounts for about 30% of housing price differences. We also know that crime lowers the value of a home, while locations with more recreational opportunities affect home prices.
What surprised us was the magnitude of these effects, and the best example was school expenses and hills. Our model suggests that a 10% increase in school spending has a greater effect on quality of life than moving Pike’s Peak in your county.
These results are not the source of our friendly disagreement. We agree on mathematics, which remains relatively immune to political disagreement. We also believe that the results do not accurately measure school spending, crime, or local health. Expenses are really a symptom of something more fundamental in local government. The reason is that places with good schools also have lower crime and healthier people. Also, nicer places spend more money on their schools, and better-funded schools are nicer places to live.
Again, that’s not my low opinion; it’s the result of a statistical model that revealed household and employer preferences in more than 3,100 US counties. Of course, some readers will doubt the value of statistical models, and we recognize their imperfections. Looking at the results, the least desirable county in the country, which lost 85% of its population, is in strip-mined West Virginia; the most wanted county in the country includes a Hawaiian beach.
Even more impressively, our model identified several outliers. These are places that have done tremendously better or worse than expected. We did this in order to identify places to visit to find out what our data might be missing. The large outlier we found was initially a big headache. It has mountains, a nice lake and an airport, a local college and decent schools. After some digging, we discovered that it was the national headquarters of the KKK. That’s good work for a statistical model of unexplained variation in house prices and wages.
The friendly disagreement between the co-authors concerns how to interpret these results in a political context. My co-authors are progressive and passionate about progressive ideas. I am a conservative and like limited governments and free markets. Our friendly debate centers on their belief that the quality of life issues we identify translate into progressive policy options. On this they are wrong.
Certainly, higher government spending on schools, hiking trails or eliminating the blight is popular among progressives. But fundamentally, quality of life policies are classically conservative rather than progressive.
Traditional conservatism in the United States has always cherished local governance, local institutions, and the sovereignty of the individual. It is the conservatism of James Madison’s vision of government, Edmund Burke’s description of civic life, and Adam Smith’s description of free markets.
Ideas about quality of life are not just conservative political theory, they are conservative in practice. For example, the state of Indiana calls for “knowledge and learning…essential to the preservation of free government” and asks the legislature to fund and encourage education. This is not the cry of progressivism, and with crime rates ranked as the second most important factor, there are no calls to “defund the police” in our statistical model.
Additionally, about half of Indiana’s counties have higher quality of life measures than San Francisco. This should come as no surprise to anyone familiar with the deep failings of local government in San Francisco. Yet we Hoosier curators have to admit, at least to ourselves, that we focus too few resources on the most important aspects of quality of life. That’s why most of our state bleeds people. One thing my fellow progressives and I agree on is that if we don’t do better, we should expect a smaller, poorer state in the years to come.