A Lake Woebegone Problem

Back in 2001, when we were colleagues in the Economics Department at the University of Montana, Tom Power and I wrote a book - Post Cowboy Economics - in which we tried to pick apart the relationships between environmental protection, natural resource exploitation and family incomes in the Mountain West. One of the troubling questions that we had to confront was why incomes in places like Missoula were so much lower than they were in the rest of the United States. And the answer, at least in part, lay in geographical differences in the cost of living (COL for short).

The simple fact was that when we compared the incomes of Missoula families to the incomes of other  Americans, those other Americans were living predominately in pretty large cities.  And while they typically took home bigger pay checks than Missoulians, they also had higher living costs, particularly when it came to housing. In fact, when Tom and I did the math, to a pretty substantial degree, the earnings disadvantage from living in Missoula and not, say, Los Angeles, was offset by the lower cost of living in the Garden City.

Now flash forward to today’s Missoulian, which reports that the Missoula Economic Partnership has just released a study of the city’s competitiveness in attracting new business, which includes, among its findings, that while Missoula families have incomes below the national average, their living costs, rather than being low, are above the national average. Evidently, what worked for Tom and me a couple of decades ago just isn’t working any more.

Or is it? After all, don’t most American still live in big, expensive cities and pull down wages commensurate with those high living costs?

Well, yes they do, but here’s the problem: when we talk about average earnings and the average cost of living, we’re talking about different kinds of averages. US average family earnings are calculated across all the families in the nation, while average cost of living is calculated across all the cities in the nation, regardless of each city’s population. That means that New York City, with a COL index of 166, and Missoula, with a COL index of 103, count equally in computing the US average COL (100), even though New Yorkers outnumber Missoulians by about 100 to 1. If big, high COL cities carried more weight in computing the national average, that average would be higher, and Missoula would come in below, not above it.

Here’s an entirely hypothetical numerical example of the problem. Suppose we have three cities; call them Big, Medium and Small.  There are 300,000 families in Big that earn an average of $57,150 per year and face a COL index of 120.  There are 30,000 families in Medium, with an average income of $50,000 and a COL index of 105, and 20,000 families in Small, with average incomes of $35,700 and a COL index of 75.

Now the first thing to notice here is that real incomes (that is, income adjusted for the cost of living) are the same in all three of these communities. $35,700 goes just as far at the store in Small as $50,000 does in Medium and $57,150 does in Big. 

The next thing to notice is that the average COL index for these three cities is 100 (that’s the average of 120, 105 and 75). So by that calculation Medium has a cost of living 5 percent above the national average and Big's COL is 20 percent above the average. Viewed in terms of families instead of cities, the situation is a little like Lake Woebegone's. When it comes to cost of living, nearly all families - 330 out of 350 thousand - are above average.

The last thing to notice is that if we average the incomes of all 350,000 families living in the three cities, the result is $55,790, so income in Medium is below the national average.*  And there you have it. Medium, like Missoula, has below average income and above average living costs. That sounds like there’s a significant economic disadvantage to living in Medium. But there’s not. In this example, at any rate, real incomes in Medium are the same as real incomes anywhere else.

To be clear, I am not claiming here that the difference between average income in Missoula and the rest of the country can be entirely accounted for by differences in living costs.  But there is a sort of Lake Woebegone problem with the way the MEP report measures COL. The fact is that for most Americans, who inhabit places like San Francisco or Chicago or Seattle or New York, Missoula would be a relatively cheap place to live, and if they were contemplating coming to Missoula to take a job in some spanking new startup, the cost of living here would be the least of their worries.

*To calculate average family income for the 350,000 families, you need to compute the total income they earn and divide by 350,000. The total income earned by families in any city equals (average income) x (number of families), so for Big, for example, total income is 300,000 x $57,150. For all three cities the total is (300,000 x $57,150) + (30,000 x $50,000) + (20,000 x $$35,700). Divide that number by 350,000 and you get $55,790.

Reefer, Crashes and Boatloads of Cash

Given Montana's relatively bad traffic fatality statistics, it seems to me that we can't promote legal recreational marijuana use without at least understanding what impact it might have on public safety on the highways. And in that regard the statistics cited by Pew in a new report might give us pause. But I have my doubts about how useful they are.

Essentially, what Colorado, Washington and other states that have legalized recreational use are reporting is a pretty big increase in the number of drivers who were using (or had recently used) marijuana when they got killed in an accident. That's suggestive, but it doesn't prove much. What we really want to know is whether, when you smoke a joint or pop an edible and get behind the wheel, you’re more likely than you otherwise would be to end up killing yourself or somebody else. And the way to answer that question is to compare the fatal accident rate for all the drivers who are under the influence of marijuana with that of all the drivers who aren't under the influence of anything. Even that comparison could be misleading, of course, because the universes of using and non-using drivers may differ from one another in some relevant respect, such as age. That's a typical confounding variable problem, and it needs to be accounted for, but in any case, we don't have the data needed to make the comparison in the first place. 

That doesn't mean that the question can be ignored. Pew reports that in 2016, of all the fatally injured drivers tested, 38 percent tested positive for marijuana use alone (the issue gets a lot more complicated when you have to take into account cases where the driver was using marijuana and drinking, or using other drugs, at the same time). On the assumption that drivers are usually tested when there is some reason to suspect they were impaired, the rate of marijuana use among all fatally injured drivers is probably less than 38 percent. That may yet sound pretty high, but whatever it is, we still don't know how it compares to the rate of use among drivers who don't have fatal accidents. It would surprise me if 38 percent of all the drivers out on the road tested positive, but what do I know? And as Pew points out, coming up with the right data is further complicated by the fact that drivers can "test positive" for marijuana long after any effects that would affect driving have worn off.

I doubt that in what remains of my political career (one more session in the Montana Senate) I will ever have the opportunity to vote on legalization, so my opinion doesn't count for much. But for what it's worth, it seems to me that decriminalization is a no-brainer. It's a monumental waste of time and resources to pursue, arrest, convict and incarcerate people for smoking a little reefer. When it comes to full on, Colorado style legalization, however, and particularly the promotion of recreational use, I have some some concerns - about public safety, for one thing, and about the health impacts of the very powerful marijuana available in the shops, for another. And the last thing in the world I think we should do is legalize because we can then collect a boatload of tax dollars. We shouldn't make ourselves financially dependent on the sale of a product that may turn out to be a stone around our necks. We did that with coal, and you can see how that worked out.

Ships in the Night

Talk about ships passing in the night!

The Missoulian today reports on the Republican US Senate candidates’ positions on the perennial question of the choice between economic health and environmental protection.  And in the same issue, there’s a piece out of the Flathead explaining just how important access to a high quality environment and outdoor recreation is to the health of Montana’s rural  economies.

As politicians of all stripes usually do, the Republican candidates (Rosendale, Fagg, Downing and Olszewski) want you to know they can give you whatever you want. Reconciling the competing demands of the economy and the environment is simply a matter of striking the right balance between protecting natural areas and outdoor recreation, on the one hand, and providing jobs in extractive natural resource industries on the other. To see it any other way, they tell you, is to embrace a false dichotomy.

Now there is a false dichotomy here, but it’s not what they think it is. It resides, rather, in comparing the benefits created by protecting the environment with the labor costs of exploiting it for commodity production. Those benefits and costs are apples and oranges.

If politicians want to fret about the jobs that will be created by using natural resources in different ways - and it seems that that’s about all politicians want to fret about - they can have at it.  Maximizing job creation and labor costs is not really a very sound basis for managing resources - it’s a far cry from finding their highest and best use - but if you are going to do it, you ought to look at the jobs created by every use, not just resource extraction. And as the Flathead article makes clear, protecting the environment and the opportunity to recreate creates lots of jobs. Not because of commercial activity associated with recreation - guiding, fishing gear sales, snowmobile rentals and so forth - but because people want to live and work and do business in nice places where they can have fun outdoors. It simply isn’t true that protecting the environment means people will have to go without jobs. On the contrary, they gravitate toward high quality environments and bring their jobs with them.

Another way of comparing alternative natural resource uses is by weighing the benefit each use provides. In the case of extractive uses, that’s pretty straight forward: the benefit is the market value  - or better yet, the market value net of production cost - of the commodities extracted (coal, for example, or timber). In the case of conservation, the benefits - again, best measured net of production costs - are the streams of environmental services (clean water, wildlife habitat, recreational opportunities, and so forth) that are preserved. Those services may not have a market price, but they do have substantial economic value. 

This comparison of net benefits, if policy makers were willing to engage in it, would drive resources to their highest and best use. But as long as they insist on comparing the benefits of doing one thing with the costs of doing another, we are going to get resource management decisions that are all wrong.