THESE ARE ELEMENTS OF THE AMERICAN DISEASE THAT ARE NOT SO EASY TO QUANTIFY.

To my regular readers this will all sound like the set-up to one of those columns where I dig into the guts of a scholarly paper, pointing out surprises and possible pitfalls. To be honest, Milligan and Schirle's article is a little above my pay grade. (My eyes start to wobble out of focus at the sight of the phrase “instrumental variables”.) But I understand what the authors are trying to do, because it's a classy quantitative version of a familiar newspaper columnist trick: they are applying the good old Canadian lens to a foreign issue.

The paper is an effort to examine the heavy use of disability insurance in the U.S. workforce — a “clear outlier... among OECD countries,” as they put it—by using Canada as a sort of statistical control. They have hog-strangling amounts of detailed micro-data on workers from both countries, and — by using a lot of modelling tricks and assumptions I am not qualified to judge — they use that data to guess what would happen to U.S. “disability” rates if the U.S. had our economy and our system of public disability benefits.

It is, as they put it, a question of “push” versus “pull.” Canada's more resource-based economy did well with high commodity prices during parts of the study period (1996-2016), while manufacturing regions of the U.S. suffered: struggling labour markets could have “pushed” more American workers onto disability. But Social Security also offers more generous income replacement than the CPP and the QPP do. Maybe public policy is “pulling” Americans onto disability.

One interesting wrinkle that Milligan and Schirle highlight is not really a result or a finding, but just part of their background research.

Still, it's something I didn't know, and that you probably don't, even if you are an American. The U.S. benefits formula is closely linked to a national average wage, rather than a median. In Canada, the comparable component in the equation is just a flat rate—a number that increases automatically with inflation, blind and deaf to labour market changes.

This means that on the U.S. side, if a small number of high earners are enjoying wage gains while everyone else stagnates, the average wage still goes up, and that makes disability insurance relatively more attractive. The fat cats' and technocrats' insane high-end incomes trickle down, in a surprising way, directly to America's most miserable. Perhaps this is a poorly understood way in which American income inequality feeds on itself. Gains among the one per cent end up causing workers at the bottom to drop out and assume a lifelong sick role, all encouraged by an equation.

The actual result of the study is a bit boring. The “push” and the “pull” turn out to be about equally important, with the “pull” of numerically high benefits dominating in the first half of the study period and the “push” of labour force misery more relevant in the second. Between them, the push and pull factors seem large enough to account for the 1996-2016 trend differences between Canada and the U.S.

But, since their paper is only explaining recent relative trends, it does not account for pre-existing differences in disability-insurance usage, or for the pretty obvious effects of the United States' feebler screening criteria (established long before 1996) and its vast, growing kudzu of opportunistic lawyers and administrative “disability courts.”

These are elements of the American disease that are not so easy to quantify.