Intergenerational mobility in Canada and the United States

Southern Ontario — the most populous part of Canada lying north of Lake Ontario and Lake Erie — displays a similar bottom to top quintile mobility as adjacent regions in Michigan, Ohio, and New York State, most regions being categorized in the 0.05 to 0.10 range.  This said, some areas of Quebec adjacent to New York State, Vermont, and New Hampshire display a lower probability than their counterparts in Ontario and New England.  Regions with rather high chances of escaping low income and rising to the very top quintile cover the American midwest, where in most Commuting Zones the probability is more than 20 percent, similar probabilities being experienced in the adjacent regions of western Canada.

That is from a new paper by Marie Connolly, Miles Corak, and Catherine Haeck.  I take that to be fairly strong evidence for the “culture matters” view of mobility, rather than the “policy is everything” view.

To the extent the United States has lower mobility than Canada, it is largely because so many people here live in the low-mobility regions of the American South.


The juxtaposition of 'so many people here live in the low-mobility regions of the American South' and “culture matters” is very entertaining.

You’ll never get out of Harlan alive.

I Have No Mouth & I Must Scream? The point of that story was the exact opposite, actually, though the story itself was able to escape from Ellison's typewriter.

Then does it mean that if we favor mobility then we shouldn’t vote for a party rooted in the Sourh?

If you want to use that logic, perhaps you should be more specific: don't favor the party supported by the South's lowest mobility people.

So, don't favor the Republicans. They love the poorly educated.

The lowest mobility voting block in the south would be blacks males.

So Southerners have the dumbest whites, dumbest blacks, and dumbest men. Why should I subsidize all this stupidity? Every other region even ones filled to brim with Mexicans are smarter than this confederation of dunces.

I analyzed Raj Chetty's 2015 database on intergenerational mobility by counties in depth here and came to some unexpected conclusions that haven't yet been incorporated by the economics profession:

The highest generational working class income mobility from the later 1990s to the early 2010s were found in lightly populated rural counties on the Great Plains. I attribute this in part to economic trends such as the rise of Chinese demand for natural resources (which has a similar effect in lightly populated Australia).

In contrast, the Carolinas tended to be at the bottom of Chetty's tables. For example, the county that is home to the vast Myrtle Beach golf resort region on the Carolina border was booming in the 1990s with golf-related construction jobs, but it was dead in the water economically in the early 2010s.

Whether these patterns will remain in place for another generation is difficult to predict.

"many people here live in the low-mobility regions of the American South."

Is this a new euphemism for black people?


Greater Toronto: about 9% black

Harlan County: about 2.5% black

It's the new euphemism for dumbass Southerners of all races. Everyone is represented.

Partly, but possible not totally. Probably some contribution from high status white southerners being better at educating their children than equal status midwesterners, and low status southerners being more interested in following the rhythms of life and remaining within their community than more atomized midwesterners.

Also, much more outmigration in midwest (an effect of low population), probably helps churn in that region. Like the Nords in Europe having the benefit of lowered inequality through a good share of their population with preferences for high risk, high reward work migrating out to the US and more other unequal countries.

'being better at educating their children than equal status midwesterners'

Not even close to accurate, at least a generation ago.

'and low status southerners being more interested in following the rhythms of life and remaining within their community than more atomized midwesterners'

Also not even close, though not as wide of the mark as the first one. Though if one wishes to speculate that the Southerners who remained in the South instead of moving any where else were less successful than the emigrants, it is probably justifiable.

(And good look defining South, by the way - does the Texas oil industry reflect the South? Or how about Florida and its major demographic shifts in terms of retirees from the Midwest and more northern states?)

'Good luck,' though a good look is not exactly off point.

What's the evidence on high status southerners being outperformed by equal status midwesterners? Some evidence on the other point would be good to gather, though. If you feel it's not your role to produce these though, up to you.

I doubt the midwest is particularly mobile relative to race and population density though or particularly good at education relative to race. General Social Survey seems to indicate its not.

'What's the evidence on high status southerners being outperformed by equal status midwesterners?'

Perhaps the more proper comparison is how many high status southerners felt the need to perform in terms of education at all. This is basically the complaint that Franklin had concerning the perils of slavery (a system which was modified after the Civil War, but not exactly replaced in terms of what Franklin wrote in 'Observations Concerning the Increase of Mankind, Peopling of Countries, etc.') - 'Slaves also pejorate the Families that use them; the white Children become proud, disgusted with Labour, and being educated in Idleness, are rendered unfit to get a Living by Industry.'

This is in the realm of stereotype, of course, but generally, Midwesterners were not noted for their servants, whether of the wealthy New England or Southern variety. Individuals remain individuals, but education as a means of advancement has never been a noted part of southern culture, certainly not among those able to trace their ancestry back for several centuries. (Again, that defining the south is not easy - Virginia or Charleston fit this model, Alabama or Florida not so much.)

Something a bit more recent than Ben Franklin might be useful.

To try and test what I've put out there, General Social Survey data:

* Among White Americans, slight differences in highest education between the Midwest (W. Nor Central and E. Nor Central) and the South, but mainly between the least population Midwest and South regions (W. Nor Central vs E. Sou Central). Little difference most populous regions of each, between E. Nor Central and South Atlantic and W. Sou Central.

* For White Americans, education very slightly more correlated between parent and child in the South regions, but not very much (max South 0.56, min Midwest 0.47).

* Comparing average parental and child education status, very small differences in the South vs Midwest regions. Mainly some cases where father has very low education, son has lower education in the South. But no real difference in mid-higher education brackets.

So I'd say I'm wrong that the Southern high education status do transmit education more effectively, but there isn't really a lot of difference other than the South having a slightly stickier floor, not much difference in the mid to upper ranges.

+1 nice and informative post

In Chetty's 2015 database, the most upwardly mobile working class families in the country were found in Sioux County, Iowa, the farming county represented by Steve King. The most downwardly mobile working class families were found on the Sioux Indian reservation in Pine Ridge, South Dakota.

My impression is that some of this is economic cycles: Sioux County has the finest, most expensive farmland in the Midwest and benefited from the China Boom. Another pattern is regression toward the mean: counties with large numbers of American Indians, Eskimos, and blacks regress toward lower mean incomes than do counties with high proportions of Asians and whites, as you might expect.

I'm a bit suspicious of Corak on these topics due him being the popularizer of the debunked Great Gatsby Curve.

See - - no relationship between inequality and IEG mobility within OECD states. Only a relationship when including poor, large and immobile countries.

See OECD -

Within developed countries if anything we see two isoclines where mobility is *positively* correlated with inequality, a Nordic one, and then a general one including all other states.

This is contra the dubious presentations of data by Corak that used carefully cherrypicked countries and dubious data (high income mobility in France and Germany, eh?). For example -

This stuff is as bad as "The Spirit Level" on misleading about the implications of inequality.

Thanks. Useful.

Urgh....The Spirit Level..... *shudder*

Yeah, k means and AI algorithms are so, so manipulable particularly when you can't specifically object to the data or the presentation and can only shout your opinion about his other works.

Present evidence.

It seems one can seldom go wrong these days expecting that any “social science” result that supports the progressive narrative reflects the author’s preconceptions and/or just plain bad faith.

Admissions against interest are much more credible; I recall a case a few years ago concerning law school admissions where the researcher held the non-PC results back for years before reluctantly publishing.

The low-mobility South keeps attracting lots of migrants from other regions. Why, if it has low mobility? Well, Boeing went South to escape unions and union-scale wages. What Boeing got was a work-force that produces inferior aircraft. (No, not the 737 Max, the defectively-designed aircraft, it's produced elsewhere). One might ask why the American mid-west is a relatively high-mobility region: is it cultural? With its history of unions and relatively-high paying industrial jobs, one might conclude it's cultural, a culture with a tradition of shared prosperity. In the South, shared prosperity is anathema to the culture: those at the top were born to be there, while those at the bottom are resigned to be there. Alas, at this blog I suspect "culture" has an altogether different meaning.

"those at the top were born to be there, while those at the bottom are resigned to be there."

Lulz. Perhaps less TV and more travel would be good for a person with that view.

both the south and the north have been overtaken by California's shade of pride. meaning has overtaken being. in the Midwest, the culture of being thrives in impossibility or infiniteness, ie romanticism.

One other thing to note about this paper: Check out Fig 5 and 6 and data Table 4.

In the raw two cluster solution, mobility is equally low in the East North Central and South Atlantic region to the South.

The South only breaks out from the East North Central and West North Central in the four cluster solution, but the difference in mobility between Cluster 3 (including the Midwest) and Cluster 4 (the South), is relatively small. (Difference between Cluster 3 and 4 in mobility rank is about 0.03 while Cluster 1 to 3 is about 0.085!).

High mobility seems more related to low population, with the high mobility cluster 1 beginning *west* of Minnesota, just where population seems to be low, and extending quite far south....

Seems to be a story where "the frontier" is related to high mobility, while dense regions are more stratified, with "culture" forming a much more secondary element. Oddly there's no mention of the word "density" in this preprint...

Actually, the version I have doesn't have the tables, but the text claims the evidence shows that Canadian cities adjacent to US cities had higher mobility, and that the mobility in the west was due to oil patch.

Right, in recent decades the Great Plains/ Great Basin have done very well economically due to the China Boom, while the Carolinas were crushed by the endts of 2008. I'm not sure that long term trend will continue forever, however.

With whites, Hispanics, and blacks having average IQs of 100, 90, and 85, and IQ increasingly rewarded, there are limits to the mobility that a meritocracy will produce.

You have no biological evidence if you are unable to control for social economic status, support networks, etc.

No biological evidence. In this day of cell data, FMRI, etc., no biological evidence.

Here is my evidence:

"You have no biological evidence if you are unable to control for social economic status, support networks, etc."

There's a very large amount of research that IQ leads to higher economic status. Therefore controlling for economic status wipes out any useful information. Since, this is pretty obvious, I'm assuming that piece is more political than scientific.

I don't understand Tyler's comment: "I take that to be fairly strong evidence for the “culture matters” view of mobility, rather than the “policy is everything” view."

If it is culture,

And not government policy (healthcare, education support),

Then we should all be

Watching Canadian television and eating poutine.

Hmmm I wonder how Nunavut does.

For Québec in Canada: francophones still face discrimination for the highest paying jobs. As The Economist once wrote: »Canada is the only country where speaking French doesn’t mark you as a member of the upper class. »

Quebec employers: 67% francophone 23% anglophone/allophone 10% foreign owned

Quebec poverty rate: anglophones 13% francophones 10%

Francophone population of Canada: 21%

Percentage of francophones in management in Federal civil service: 30%

percentage of anglophones in Quebec civil service: 1%

percentage of francophones in Ontario: 4.5%
percentage of francophones in Ontario civil service: 7.7%

Meant as a reply to M. Giguére.

While we're at it here's the abstract from a 2009 study by Serge Nadeau at Université d'Ottawa:

Using a variant of the Blinder-Oaxaca decomposition method, I find no evidence that, outside Quebec, there was at any point in time between 1970 and 2000, a labour market advantage for Anglophones that cannot be explained by a higher relative demand for English skills, whether in the public sector or the private sector. However, I find that in Quebec’s public sector, between 1970 and 2000, Francophones enjoyed a wage premium that may have gone beyond language skill considerations and that I cannot explain. Such a premium also appears to have been present in Quebec’s private sector in 2000.

Comments for this post are closed