David Andolfatto has an interesting take:
The question is this: Would you expect the labor market in the U.S. border states to look more like the Canadian labor market or more like the U.S. labor market?
There is good reason to believe, I think, that the demographics across the two countries are pretty similar (though certainly not identical). If the type of economic activity along the border is roughly similar across the two countries, then one might reasonably expect similar labor market behavior in Canada and the border states. But this is not what we see at all.
One chart shown is of the Employment-Population ratio in the two countries (with the US border states shown separately):
And Andolfatto sums up:
Once again, the border states look more like the U.S. in general, rather than Canada.
A preliminary conclusion is as follows:  If Canada-U.S. demographics are roughly similar; and  if U.S. border states are roughly similar to their Canadian counterparts in terms of sectoral composition; then the differences we observe between the two countries (in terms of labor market activity) are quite possibly driven by policy differences.
Exactly what sort of policy differences we are talking about here remains an open question.
One possible answer comes from “matching shapes”.
Unfortunately navigating in Statistics Canada was “above my pay grade” (why don´t they all emulate FRED?). So what I did was to break up Andolfatto´s chart to show data only after 1992 (that´s the time the inflation targeting scheme began). The bottom chart depicts “monetary policy” (where an “easy stance” means NGDP is above or growing more than trend and vice-versa for “tight stance” – the NGDP Gap).
In (1), monetary policy “tightens” in Canada and “loosens” a bit in the US. In Canada the E-P ratio falls and in the US it rises.
In (2), monetary policy “loosens up” in Canada (NGDP growth higher than trend to take the economy back to trend) and “loosens” more strongly in the US. The E-P ratio goes up in both countries. Note, importantly, that the E-P ratio of the US border states goes up by more than for the rest of the country. One conjecture is that the border states were experiencing a “double-pull” from “monetary easing” in both the US and Canada.
In (3) in both countries NGDP is close to trend (“neutral policy”) so that the E-P ratio in both countries is similar (and there´s no difference in the E-P ratio between the border states and the rest of the US). Towards the end of that interval, monetary policy “tightens” in the US and “loosens” in Canada. That difference is reflected in the behavior of the respective E-P ratios,
In (4) NGDP falls by more in the US. The fall in the E-P ratio is also bigger in the US. Note that in US border states the E-P ratio falls less, reflecting(?) a “combined weaker push” from the smaller drop in Canadian NGDP.
(Noteworthy: In his last year as Head of Bank of Canada Mark Carney didn´t do a good job, “tightening” monetary policy by allowing NGDP to distance itself from trend.)