« The legal scope of Fed purchases | Main | Existentialism and the non-neutrality of money language »


Feed You can follow this conversation by subscribing to the comment feed for this post.

Interesting graphs, but I'm not sure I see the distinction between keeping the "distribution of net incomes relatively stable" and pandering to the median voter. The former is just a policy rational for the latter.

This sort of exercise is interesting, though, in so far as it challenges some of the usual assumptions about the welfare state, at least in practice (i.e., that it is inherently redistributive or beneficial to the disadvantages). Since those assumptions often provide a moral foundation for the welfare state (or aspects of it), if they're wrong, that raises serious questions about the modern welfare state.

In fact, the apparatus of the welfare state is often used for the benefit of politically powerful groups (a statement, if not the only statement, on which both the "Occupy" movement and the "Tea Party" types would agree, even if they diverge on who they think those groups are (i.e., bankers vs. public sector unions - it's possible they're both right)). Timothy Smith (a historian from Queens - but I won't hold that against him) wrote a book a few years ago looking at the French welfare state, and one of the points he made was that it was, essentially, a reverse Robin Hood state, which to a large extent used its power to redistribute income to politically influential (and reasonably well-to-do) groups (which, in France, means public sector employees and employees of state managed/protected industries) at the expense of the politically weak (immigrants, the young, etc.). At the time that seemed remarkable, though given what we've learned about the economies of other EU members over the past few years, nothing surprises me now.

My memory's a big vague, but I seem to recall that he had some shocking statistics that suggested that, whereas in the big, bad, US (and the other "anglo-sphere" economies), income inequality is reduced when you take into account taxes and transfers, in France it was aggravated. I remember thinking that such a state of a affairs was simply appalling, and something that, whether you're on the right or the left, should be morally untenable.

I was going to say that demographics explains it. Number of kids should affect taxes and transfers, and the echo boom changed the number of kids. And you get different pictures for your unattached vs 2 or more. Or maybe more retired people? But I can't get my head around any *particular* demographic story that could explain your graphs. So it's just a hunch.

It's a terrible idea to destroy the incomes of the median relative to everyone else. The median-income part of a modern society is the part that clawed its way out of feudal and industrial poverty. A small amount of redistribution to the very poor, if its done in a way that increases the space between the wealthy and the median, is not a socially sustainable policy, but such has been the policy of Western societies for a few decades now.

Stephen, I'm sending you a file showing the remarkable decrease in the % of children living in low income after tax/transfers since the mid- to late-1990s. Unfortunately we don't have the most recent #s yet so it's hard to show the impact of the recession.

I suspect part of what is driving your results is, as Nick suggests, demographic changes - together with a deliberate policy decision to focus transfer resources on families with kids.

I was wondering if that might be the case, especially when you see how differently unattached individuals are treated.

Very interesting post.

Seems consistent with how Canadian politicians these days tag "for families" onto almost all policy announcements.

Very interesting post. However, I must confess I still don't quite understand why inequality is a bad thing or something to worry about. Is the purpose of redistribution from the top to the bottom, that of helping the poor or creating more equality? I understand why the former is a good idea. We should help the poor. But I dont understand why inequality in itself is important. Why does it matter what the distribution of income is?

Best wishes!

Besides taxes and transfers, an important component of the Canadian welfare state which reduces market inequality is decent public services. For the most part, good education and health care are available to anyone, no matter how poor they are. Health care spending was 10% of GDP in 2007; the fact that 70% of this was funded through the tax system rather than private insurance should have a significant impact on inequality, compared to a hypothetical system where health care funding is entirely private.

I wonder how important social norms are in terms of political support for the welfare state. Subjectively, it seems to me that Canada is more egalitarian than the US, despite Seymour Lipset's observation in "Continental Divide" that Canada has historically been more unequal than the US. Apparently the gap between the salary of managers and average workers is considerably smaller in Canada than in most countries (an average ratio of 2.3 in Canada, 3.7 in the US).

I think another factor is that Canadians are generally satisfied with public education and health care, and this bolsters support for the welfare state in general. (A 2007 survey found that 86% of Canadians were satisfied with their health care over the last year, and the most recent BC survey on school satisfaction showed 70-75% of parents who responded were satisfied.)

Joe: relative status matters. People's happiness depends a great deal on how they're doing relative to their peers, not just their absolute income. Robert Frank, writing in the American Interest, points out that there's evolutionary reasons for this:

"Since reproductive success has always depended first and foremost on relative resource holdings, it would be astonishing if the evolved brain didn’t care deeply about relative position. Most vertebrate societies, including the vast majority of early human societies, were polygynous, meaning that males claimed more than one mate when they could. It was the high-ranking males in those societies who claimed multiple mates. And given the inexorable logic of musical chairs, it was the low-ranking males who were left with none.

"Famines were also a frequent survival threat in early environments. But even in the worst famines, there was always some food available. And the question of who got fed was almost always settled by relative income. Then as now, the poorest in every group were most likely to starve."

Russil Wvong,

And do low-income Canadians assess their standing relative to Mike Myers? Or are they more likely to regard their neighbours as peers?

Some of the increase is almost certainly due to increased CPP/OAS. The only really universal transfer that spans all income groups is CPP/OAS (the fact that the poor elderly might not get much CPP, and the rich elderly get OAS clawed back, might explain the increased transfers to the middle group), and given the rapid (and fairly steady) increase in the oldster population over the past 30 years (as a result of both an aging population, and longer life expectancy for olders, result in a bigger chunk of them), you would expect that per capita OAS/CPP spending would increase even if the per retiree spending didn't. I'd love to see those charts broken down by age.

I also wonder if there's a technological explanation for the rise of transfers to the middle class. The big revolution in social policy over the last three decades has been the use of the tax system to provide credits and other transfers to taxpayers - it's essentially a new technology for governments (and was likely made possible by the rapid decline in the cost of computing power allowing governmetns to more readily administer these credits). So accross the western world we see various child tax credits, GST tax credits, family tax credits, earned income tax credits (to say nothing of the fitness, art, student loan, public transit, working, medical caregiver and firefighter tax credits that clutter up my copy of the Income Tax Act).

Because you're using an existing infrastructure to administer these credits, you can come up with all sorts of specialized ways to funnel money to favoured groups (families, firefighters, the median voter, whatever), without having to set up a separate administrative systems, reducing both the economic and political cost of such programs (could you imagine what would happen if you had to have a bureau of art subsidies, or a department of firefighter allowances?). I suspect that some of the increase in the middle groups in Stephen's chart is attributable to things like the child tax credit (which, if memory serves, was first introduced in 1978 and has been steadily expanded over the 80's and 90's into the Child Tax Benefit). That would be consistent with the different patterns between families and individuals.

A corollary of focusing programs on families is that young, single people cannot benefit from these programs and policy initiatives.

Singles in the their twenties suffer foremost from job insecurity and therefore loss of income; if you lose your job you don't have a spouse to support you. Therefore you have to apply for EI and provincial welfare programs. But since the 1970's Canada has curtailed the generosity of benefits available under these programs. That hits singles as EI is the main government benefit available to them, aside from Health Care. For people in their twenties that benefit will vary depending on if they have a chronic disease (diabetes, asthma) or not.

In terms of quintiles, we have done more to save the jobs of established workers in their fifties who were at risk of losing both their employment and their benefits (pensions) than in helping young people get on the same track. GM got saved by government action but so many others will never get that level of pay or benefit.

Ontario has pension default insurance for single-employer defined benefit pensions (Multi-employer pensions like the Teachers plan, the Hospital plan and others like that are excluded) but at the same time we don't offer the opportunity for DB pension accrual to new entrants. In fact large corporations like Canada Post and Air Canada are trying to reduce base wages and pension accrual (shifting new entrants to lower-paying DC plans) to young, new hires.

Sure we are trying to save the middle but we are doing it at the expense of hollowing out the bottom rung. I question if this is sustainable at all.


I take your point that humans may have an innate, perhaps even genetically hardwired dislike of inequality arising from our origins as bald apes living a subsistence existence (after all, isn't envy a fundamental human emotion). But that seems to be little more than saying "we care about inequality, because we care about inequality" and therefore that governments should seek to address it.

Unfortunatwlt, that isn't a particularly satisfying rationale. We probably also have a hardwired predisposition to eat high calorie food that makes us fat and, at least amongst the males of the species, a hard-wired predilection for having indiscriminate sex with the females of the species (both, genetically at least, valuable survival traits). It doesn't follow that those are innate preferences that the government should necessary promote. The case for subsidizing McFatty burgers and prostitution seems pretty weak (and not simply because such subsidies are seemingly wholly unnecessary), and there are compelling, if not wholly persuasive arguments that, to the contrary, they are vices to be condemned and/or controlled.

The question is SHOULD we care about inequality? Is there a moral basis for wanting to address inequality? Your post sets out one moral reason for caring about inequality, namely we believe that it is wrong for people to be "poor", to lack the essentials of life (however we want to define that).

But in that case, we really don't care about inequality, per se, we care about poverty, and inequality is seen as a proxy for poverty. Moreover, inequality can be a pretty poor measure of poverty. If we doubled everyone's real income, I'd wager that most Canadians would accept that poverty had been reduced - even though inequality is unchanged. On the otherhand, a thermonuclear war would produce a rapid improvement in income inequality, albeit at the price of reducing the handful of survivors to subsistence income levels.

Now, we clearly have other moral reasons to objecting to inequality, where such inequality arises as a result of "unearned" wincome (although, as Rawls points out, the distinction between "earned" and "unearned" is often morally arbitrary). So the "trust fund" kid, ot the banker who gets a bonus from bail-out money drives us up the wall. But, that strikes me as less a concern about equality, per se, than about perceived fairness. After all, the same guy who complains about bailing out the banker can turn around and complain about "welfare bums" (hence the appeal of the Tea Party). As with poverty, inequality is just an imperfect proxy for what really bothers us. And the proper policy for dealing with what really bothers us is likely to be very different for dealing with poverty (in my example, taxing the rich and giving it to the poor might address inequality, but not bailing out banks (or doing so on strict conditions) and making able-bodied welfare recipients work for welfare is likely to satisfy the underlying fairness concern).

Bob: I remember that when I filed my first tax return, back in 1986, Revenue Canada included a table so that people could do percentage calculations using table lookups. Presumably pocket calculators weren't widespread yet!

"I take your point that humans may have an innate, perhaps even genetically hardwired dislike of inequality arising from our origins as bald apes living a subsistence existence (after all, isn't envy a fundamental human emotion). But that seems to be little more than saying 'we care about inequality, because we care about inequality' and therefore that governments should seek to address it."

It's not really a moral argument, it's more about utility (*): in a more equal society, people are happier (within limits, of course). Government comes into it because income redistribution requires collective action. (In other circumstances, alternate institutions might provide a similar function, e.g. the medieval Church.)

(*) My apologies if I'm using "utility" incorrectly.

I'm not arguing that inequality is more important than poverty (in particular, I'm not arguing that a poorer but more equal society is always preferable to a richer and less equal one). I'm arguing that other things being equal, a more equal society is better than a less equal one. If you had to sacrifice a significant amount of aggregate income in order to achieve greater equality, you might well conclude that it wasn't worth it.

However, comparing Canada and the US, it's not obvious that there is such a tradeoff. Canadian public spending on health care represents a significant redistribution of income, and yet it's much more efficient than the US: the US spends 16% of GDP on health care, Canada spends only 10%.

Besides the utility argument, there are moral arguments in favour of greater equality (it's unfair for children growing up in low-income households to have vastly different life chances than children growing up in high-income households--Richard Freeman's "apartheid economy" argument), but I'm not sure getting into them makes sense. There's also moral arguments in the other direction (the "taxation is theft" argument); how do you choose between them?

W. Peden: that's an interesting question. In the media, at least, high-income people are much more visible than low-income people, which may skew people's expectations upward. The BC Securities Commission released a survey recently showing that young people expected that in 10 years they'd be making an average of $70,000/year. The actual number is more like $31,000/year.

Russil Wvong,

I'm not sure, in the absence of any evidence whatsoever, that visibility is what determines people's impression of their peer group. I regard my friends as my peers, at least economically, so anecdotally the media doesn't have any effect.

Isn't the phrase "Keeping up with the Joneses", i.e. the neighbours, rather than "Keeping up with the Rothschilds"?

It's an excellent question: what's the reference group that people compare themselves against? Friends? Family (e.g. parents at the same age)? Co-workers? Neighbours?

A quick Google search on "reference group relative income" turns up a recent paper:
"Life Satisfaction and Relative Income--Perceptions and Evidence" (October 2010), by Mayraz, Wagner, and Schupp.

The background section of the paper makes it clear that this is still an open question, but they do report some interesting findings. They asked people in the 2008 German Socio Economic Panel Study to compare their income against a number of different reference groups, and ran regressions against reported life satisfaction. What they found, using this particular data set (about 1000 respondents):

"For women the small effect combined with the small sample size means that none of the comparisons is statistically significant at the 5% level. It is therefore not really possible to rank the difference income comparisons by importance. For men the effect size is much larger, and there is consequently also better statistical power. The results in Table 3 indicate that the important comparisons are work related comparisons (same profession and with co-workers), and even more so comparisons with other men in general. Comparisons with friends and with other individuals of the same age are less important. Finally, comparisons with neighbours are almost completely unimportant."

It's not surprising that co-workers would be an important reference group, but it's surprising that men in general would be. The gender difference is also interesting (and consistent with Robert Frank's comments about the evolutionary connection between relative status and being able to find a mate).

But whom do they mean by "men in general"? And to what extent are perceptions of "men in general" determined by close environmental (friends, family & neighbours) comparisons?

There's also the problem that people are systematically bad at estimating the distribution incomes of society as a whole. Most Americans, (in)famously, think that they live in a very egalitarian country in terms of income inequality.

(That's assuming, of course, that survey data is suitable here i.e. dismissing the problem that comparisons with one's neighbours are less socially acceptable than comparisons with co-workers and "men in general".)

Still, it is notable that the rich seem not to be the comparators for normal people's assessment of how well they are doing economically.

W.Peden: at some point, the rich are so far away, we have no idea , and for some can't even conceive,of what they have. And in caste societies, like the the U.S. South, comparing yourself with your betters is doubleplus ungood crimethought.

Jacques René Giguère,

I quite agree, though I will suspend judgement here on whether or not a "caste society" is an appropriate description of the US South or a sophomoric rhetorical hyperbole until the term is further defined in its use here.

So, if we assume that no accurate comparison with the rich takes place in most (if any) modern societies, then how seriously should we take psychological-medical explanations of social ills based on supposed accurate comparisons between the rich and the non-rich on the part of populations?

Let's assume that survey data on income perceptions in the US is accurate. Shouldn't the "social stress" of inequality in the US be very low? And yet the empirical arguments for inequality being the cause of many social problems via evolutionary psychology comparisons (such as those in The Spirit Level) on the part of populations seem to get most of their plausibility out of the correlction of the two facts that (1) the US is a very inequal society & (2) the US has a lot of very bad social problems. However, given the survey data, (2) should NOT be true if the explanations of social problems in books such as The Spirit Level are correct.

All of which might suggest that it is indeed poverty that is the great moral tragedy of our age (and, to a much greater degree, all other ages) not income inequality. The claim that income inequality is wrong in itself requires further argument than "social stress" as a result of income or wealth comparisons within populations.

(As it happens, I've been thinking about a Malthusian-esque argument in favour of wealth equality on the basis that 1. we live on a finite planet & 2. it may well be that we will never colonise the rest of the universe. If that is true, then Malthus rather than Smith was right and a lot of the egalitarian arguments that only work on a "zero-sum" basis become sound.)

W. Peden: my comments are getting rejected (perhaps they're getting too long), so I'll try to keep it short. The Easterlin paradox (as societies get richer, average life satisfaction doesn't increase, and may even drop) strongly suggests that poverty isn't the only problem. Branko Milanovic cites a startling example: China.

"Carol Graham, a senior fellow at the Brookings Institution, and Stefano Pettinato from United Nations Development Program, have shown that it is not only the global poor who increasingly feel left behind, but also what she termed 'the frustrated achievers'--those who have done well in real terms but feel deprived because others have done even better. According to Graham, the more the frustrated achievers knew about those who did better, the worse they felt about themselves.

"China is perhaps the best example of this phenomenon today. While China’s overall living standards have improved massively in recent decades, World Values Survey data show an equally large decline in life satisfaction. Researchers credit this unhappiness to ballooning income differences, especially as ostentatious consumption has became more visible."

Milanovic suggests that the problem may be related to perceived inequity, as suggested earlier in this thread by Bob Smith.

We've strayed a long way from the original topic, so I'll let you have the last word!

Thank you for the last word. I think that rhetorical inequality might cause some global discomfort, since I'm rather embarassed at having such a position!

Poverty may not be the only thing that determines happiness (though the Easterlin paradox is in a general state of empirical decay- the early strong results are becoming harder to repeat, while contradictory results are accumulating), but that doesn't mean that inequality is the explanation of the non-correlation. There are many factors involved in economic development that could also affect happiness (urbanisation, pollution, global awareness, industrial working habits etc.) and other particular factors at work in China (the One Child policy & associated demographic inbalances, long working hours, corruption etc.).

With complex multi-variate systems that you can't manipulate, you need a lot of data before probability principles become meaningful and statistical work can detect causal relations & the relevance of particular variables. This is one of the reasons why I am sceptical about "happiness economics" in general.

One place where I do think that relative income might be important is in consumption. Here's a paper you might enjoy-


It has to do with power, unjust capacity for wielding thereof. Too much inequality -> the actual road to serfdom. I dunno if that counts as "utility" or "morality", but it's just wrong to me. As for why inequality might be immoral in se, I think the question is being asked in the wrong direction. I don't see why, in any moral sense, the person who vacuums the office should be paid less than the person sitting at the desk. That's probably a utility thing, I guess.

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad