Three worthwhile Canadian puzzles

So, to continue with my recent interest in Canadiana, impelled by, well, being in Canada!

There is an interesting puzzle about Canadian economic development. Well, there are many many interesting puzzles, but what particularly grabs my attention at this moment is why Canadian productivity has persistently lagged the United States.

Below I show Canadian total factor productivity (TFP) as a percentage of American TFP between 1950 and 2014. Save for a few halcyon years in 1973-80, Canada has consistently lagged the United States. Moreover, the gap turned into a chasm in the 1980s and grew even larger after the Great Recession. 

The productivity gap predates WW2. After World War 2, Canada lagged the United States in total factor productivity. That is to say, a Canadian worker in the same industry with the same skills and working on the same machinery produced less than her American equivalent. In the early 20th century, the superficial gap was just as large: the average Canadian worker added only 81% as much value per hour as his counterpart south of the border.* (Table 6 on page 25.**) Canadian workers, however, used rather less machinery than their American counterparts, so TFP closed to about 94% of the U.S. level ... remarkably similar to the gap in 1950-67.

In fact, the gap widens if you go further back. In 1900, GDP per capita (the crudest measure) in Canada ranked only 80% of the United States. In 1870, it was 70%. Canada appears to have had more-or-less that income gap with the United States back to 1820. 

Heck, the gap goes back to even before the existence of either the United States or Canada! Vincent Geloso (Texas Tech) has a paper comparing 18th-century Quebec with New England and finds that living standards were about 31% lower in what eventually became Canada. (Remember, there was no settled European population in modern Ontario back then.) 

So what caused the gap? A seminal 1964 paper by H.E. English pinned the gap on American protectionism: Canadian businesses operated at too a small scale and tried to produce too wide a variety of products. In other words, Canada wasn’t big enough. (You can find a 1980 study making the same argument here.)

This seems plausible. And as theories go, it has the advantage of explaining why the United States has enjoyed such a lead for so long over so many places. Everywhere else is too small! There are just too many damn countries in the world. Those of you who have read my opinions about  secessionism will know that I would find this a very gratifying conclusion.  

But the scale explanation does not seem to fit with the big widening of the gap since free trade.

Or does it? It is true that trade costs between the U.S. and Canada are still pretty high. And it is true that interprovincial trade barriers place even bigger obstacles on Canadian firms than American protectionism.  And it is true that the returns to scale in Canadian businesses are pretty high. (Page 130 of the 2009 World Development Report.) Canada has lots of small firms compared to America (consistent with the scale economy argument) and those small firms tend to be less productive than their American counterparts (sort of maybe could-be consistent with the scale economy argument). 

Statistics Canada ran an interesting thought experiment to test the hypothesis. They found that if you altered the size mix of Canadian firms to look like American ones you would close the productivity gap by 5 percentage points while if you made Canadian small firms as productive as American ones you would close it by nine. (Table 6, counterfactuals one and two.) Fixing both effects would close two-thirds of the gap, call it 16 points.

But that is not the entire gap. In fact, it leaves an unexplained gap bigger than the total gap in the 1950s.

Or the 1870s, for that matter. Kris Inwood (University of Guelph) and Ian Keay (Queen's University) conducted a similar exercise for Canada in 1870. They could get Canadian TFP to rise from 93% of the U.S. level to 97%, which cuts most of the gap. (You can find an outline of their paper here; the main one is paywalled.) Strangely, they downplay their finding, which seems like a very Canadian thing to do. The punchline is not that a gap remained, the punchline is that they explained more than half of it!

So the fact that Canada had a protected small market explains a little more than half the 19th century gap and about two-thirds of the modern one. But that still leaves two puzzles. First, what explains the rest of the gap in both periods? And second, why has the gap grown so much larger recently, despite NAFTA and all the ways modern communications should make service exports easier?

It is a great pair of questions.

But I mentioned three in the title, no? What’s the third puzzle?

I will post on that later, but it’s even weirder than the productivity gap ... it’s that median incomes in Canada may have finally pulled even with the United States. That is to say, for the first time ever it matters not to most Canadians that they live in a fundamentally less-efficient economy than their southern neighbors. And that’s the biggest puzzle of all.  

* The gender shift is deliberate.

** John Baldwin and Alan Green, “The Productivity Differential Between the Canadian and U.S. Manufacturing Sectors: A Perspective Drawn from the Early 20th Century,” The Canadian Productivity Review (Statistics Canada: December 2008).

Comments

[Steven]

The immorally low pay of slaves, sharecroppers, prisoners, and members of a dead-end underclass can make people very "productive." People whiling away in school or staying home to raise children have no "productivity" at all. Preventive health care is also anti-productive, as it lets people save their money or work less rather than scrambling for low-wage jobs and burning their income on medicine.

Of course these cases give the lie to the whole idea of productivity, because raising healthy children, learning things and letting people develop a bit of personal wealth are the whole point of having an advanced civilization in the first place.

Is there any way to control for these three factors? (Lack of slavery/underclass in Canada, failure to measure domestic labor/education, better health care.) Because having gone back and forth between the two countries all my life, I assert that for an average non-rich person, life in Canada feels more privileged, relaxed, and successful than it does in the States.

[Noel Maurer]

The easiest thing to do is compare Canada with the Northeast or Upper Midwest. (The authors of the paper about 1871 are compare Ontario and Michigan; Vincent Goloso explicitly makes regional comparisons for his 18th-century analysis.) People forget that the Northeast scores at Scandinavian levels on most measures of well-being or educational attainment, despite being not-at-all-Scandinavian.

Note that productivity is measured by input of work-time, so the numbers are not dependent on leisure. For example, the U.S. only regained its TFP lead over Belgium and France around 2004; that's not because French workers put in more hours. The same applies to Canada. U.S. health care, meanwhile, is a productivity sink; it drags down the American number rather than push it up.

Canadian median incomes only recently caught up with American ones. (And there is still a little wiggle room in the numbers, so the lead is not -- yet? -- clear.) I'll have a little more to say about this when I get around to polishing up the next post. But at the risk of giving away one of the punchlines, individual wages in Canada show disappointing trends not terribly unlike the United States: see http://www.csls.ca/reports/csls2016-15.pdf.

[Dave K]

I suspect that some portion of the US regression to Canadian median wages is due to anti-trust action. From 1950-1976, monopolies were less common in most US sectors than in Canada. Over the last couple generations the US has caught up to Canada in terms of uncompetitive sectors. Canada had the excuse of protectionism and small market size but non-competitive markets was a policy choice in the US.

If Canadian median wages are equivalent to US median wages while TFP is significantly lower that suggests capital gets less of a return in Canada than in the US. Is that the case currently? It certainly wasn't the case historically.

Canada has a higher female labor force participation rate than the US. There are several ways that could statistically influence median incomes if it is collated by household (more dual income families, fewer unemployed single mothers).

[J.H.]

Hmmm...could linguistic barriers (French/English) be a factor in addition to the fact that historically inter-provincial barriers to trade in Canada were greater than inter-state barriers to trade in the US?

[Randy McDonald]

They wouldn't explain income differences before the American Revolution, when there were no Anglophones in the Great Lakes basin and settlement in the Maritimes was a patchwork linked to New England.

It might explain some divergences now, Québec being poorer on per capita measures than Ontario and most provinces west and a less significant source of migrants to richer provinces than income disparities would suggest. But then, it's not that much poorer, and by at least some metrics that has been convergence as productivity has increased, in keeping with post-Quiet Revolution urbanization and investment in human capital.

My understanding, as a non-economist, is that the big internal division in Canada is not between the two major linguistic communities, but rather between an oil-rich Alberta that has American levels of income and the rest of Canada.

[Vincent Geloso]

Dear Noel (if I may),

I would disagree with many elements of your post while also being in agreement. I say this because the argument is axiomatically true. As such, its probably true. I would question its "amplitude". I say this because Ontario and Western Canada are actually equal or richer than the United States (historically closer - see Altman 2003 in AEHR). The laggards are the Atlantic Provinces and Quebec. There I think free trade matters, but it woudnt explain why provinces like Quebec were 75% as rich as Ontario while Prince Edward Island was 40% as rich (1920s). Without them, Canada looks much richer. Given free trade within Canada, I doubt that free trade is sufficient to explain te gap driven by five provinces account for less than 30% of Canada's population. I think protectionism fits well to the Atlantic provinces being poor. But Quebec is the big chunk. Why? I think this is where language acts as a barrier on mobility and thus, with limited mobility, fewer constraints on politicians (i.e. More captive taxpayers). As such, I think the history is largely institutional outside of protectionism. I would point to the historic role of seigneurial tenure in Quebec as a deterrent to industrialization.

Best

Vincent Geloso

[Tony Zbaraschuk]

Forgive me for asking a silly and obvious question, but has anyone considered the effect of longer Canadian winters and worse weather? Supply chain interruptions, etc., etc.

[Noel Maurer]

Tony: That's a good question. I'd be surprised if the effect was particularly large. The 1871 study, for example, compared manufacturing firms in Ontario and Michigan, which have pretty similar climates in their urbanized parts.

There is a bigger literature on productivity and climate, but its results generally run in the direction of hotter and wetter weather leading to less output per work-hour. I don't think it has much to say about the direct impact of snow, for example.

Things can run in weird directions, however. Consider that winters around D.C. are quite milder than in Boston. But D.C. is also resolutely unprepared for the bad days that do come, leading to more snow days and (I imagine) lost output.

That said, it's worth looking into, and there may be a literature of which I am unaware!

[Noel Maurer]

Hi, Vincent! Looking forward to seeing you in Philly.

My discussion above pulled a bit of a fast one. From 1871 onwards, we have productivity data. The 1871 data compare Ontario and Michigan, and are thus unaffected by low productivity in Quebec or the Maritimes. The post-1926 are national, but they're output-weighted.

Quebec produced about 30% of manufacturing value-add and about the same share of GDP in 1926. If low productivity in Quebec had been driving the national results, then you would have expected postwar Canada to converge on the U.S. as postwar Quebec converged on the rest of Canada. But you only see that pattern briefly in the late-1960s.

Provincial productivity data exist, so we could end the debate! But they're harder to find than you'd expect, at least in a form suitable for international comparisons. But let me state that I'll be gobsmacked if the 1926-67 and post-1982 productivity gaps turn out to be driven by Quebec. (And weknow that that the Maritimes are near irrelevant, given that the nat'l figures are output weighted.)

And the fast one? Well, before 1871 we have either GDP per capita estimates or direct measures of income. We don't have productivity. I assumed that income is more-or-less proportional to productivity before then. And that's normally a reasonable assumption, except for the fact ...

... that in the last ten years Canadian median incomes have passed American ones! And that despite a collapse in relative productivity. Me, I take this as a clear indication of the superiority of socialism ... nah, sorry, I'm poking you.

Me, I have no idea what's caused that, which won't stop me from wildly speculating in a post that I've already written but haven't yet pushed "publish" on. Meanwhile, if there's a hole in the above logic, please pile on!

[Noel Maurer]

Fun fact: output per worker in Quebec manufacturing fell from 82% of the national level in 1926 to 67% by 1959.

[Randy McDonald]

Ah, the period of _la grande noirceur._

https://en.wikipedia.org/wiki/Grande_Noirceur

[Randy McDonald]

Going by the metric of GDP (PPP) per capita, meanwhile, Michigan is still a bit less than 10% richer than Ontario. New York State, probably a better comparison in terms of history and climate, is 70% richer, but I think you really need to figure out a way to subtract New York City for like figures to be relevant. (Toronto is so not New York City. Maybe it is Queens? Maybe.)

Alberta is far and away the richest Canadian province, well ahead of not just British Columbia and Ontario but almost on par with New York State. That can be explained substantially, but not entirely, by its very high per capita exports of natural resources, the same thing that helps Western Australia be so rich.

This provides a convenient overview of GDP per capita in Canadian provinces.

http://www.conferenceboard.ca/hcp/provincial/economy/income-per-capita.aspx