Image: HikingArtist.com/Flickr

A new report came out from the Fraser Institute this week looking at income mobility. It certainly doesn’t intend to make this conclusion, but a thorough look at their data shows that the rich stay rich as everyone else fights for entrance to this exclusive club.

Plenty has already been written about growing income inequality in Canada from CCPA’s own Armine Yalnizyan and Hugh Mackenzie, among others. Their examinations show how the top 10 per cent and top 1 per cent of Canadians are running away with all income gains. Each year when we calculate who got a raise and by how much, most Canadians’ raises only barely match inflation with the top 10 per cent and the top 1 per cent giving themselves much more substantial raises.

Income mobility is a slightly different issue. Instead of looking who gets the raises in a given year, it tracks individual people to see how they fared over a longer period. So if someone was in the top 10 per cent of earners in 1990, how likely are they to still be there in 2009? Those were the types of questions that the Fraser Institute report examined.

There are some interesting results in it, although not the ones they highlight. (There are also some serious methodological problems that I’ve highlighted at the bottom.)

The report makes hay out of the fact that people who were in the bottom 20 per cent in 1990 ended up seeing a 635 per cent increase in income by 2009 (See Table 5). Wow, if only we could all be poor!  If the poor are making so much why even provide them with social programs?!

Unfortunately, this is a misunderstanding of the data. The bottom 20 per cent — and to some degree the bottom 40 per cent of the distribution — represent people who are unemployed either forcibly (because they were laid off), voluntarily (students or parents home with children), or people working part-time. If you take a look at the average income for the bottom 20 per cent it was $6,000 in 1990. These aren’t people who are working full-time — try living on $6,000 in say Toronto for a year, you’d be homeless by June. Remember seniors are mostly excluded because they likely didn’t “survive” the period (see methodology note below).  It shouldn’t come as a big surprise that someone who was a student in 1990 or who was unemployed is making more, sometimes much more, 20 years later.

If anything, what this examination of the bottom end of the income scale shows us is that a lot of Canadians are poor at some point. Not permanently mind you, but temporarily and while they’re down it makes sense to help them get back on their feet through more generous EI programs, for instance.

What should be very shocking to people is that in the shorter-term panels, 50 per cent of the poorest people in 2002 were still in the poorest category five years later (table 3). That’s a pretty stunning conclusion. For the second poorest category, a shocking 69 per cent either ended up in the second poorest category or the poorest category (making even less). So about 60 per cent of the people in the bottom poorest categories were stuck there between 2002 and 2007. That’s a lot of people living on very little income. They are a bit mobile, but they are mobile between very poor and poor. That’s not really the kind of mobility we should be looking for.

While the study focuses on the low end, there are some very interesting things going on at the upper end. Take a look at table 6 for the top 20 per cent. The table does a “where are they now?” analysis over 19 years from 1990 to 2009.  For people that were in the top 20 per cent in 1990, how many are still in the top 20 per cent in 2009? The answer is a whopping 64 per cent. Six out of 10 people that were in the top 20 per cent in 1990 are still there today, two decades later. Over a 10-year period between 1990 and 2000, almost 80 per cent of the wealthy stayed at the top. This is the opposite of income mobility; if you manage to slip into the exclusive club of wealthy Canadians it looks like you get a 10, if not a 20-year membership. None of the other income groups are anywhere near that level of exclusivity.

It isn’t only that the rich get more raises over time and gain a bigger portion of the economic pie (income inequality), once you get into that top bracket, you’ve got an incredible 79 per cent chance of being there a decade from now.

Although this isn’t the report’s conclusion, my conclusion from the same data is that the rich stay rich and everyone else churns around in the bottom.

Methodological issues for data geeks

There are also some serious methodological problems:

1. When the report looks at “income” it means only wages and salaries, but very importantly excludes capital gains and dividends. Wealthy Canadians actually rely more on wages today than they used to but the income from capital gains and dividends is heavily skewed to the top end. Excluding this portion of income, although it can be somewhat volatile, makes wealthy Canadians look much less wealthy in a relative sense.

2. The shorter-term panels of 5 years are using the Survey of Labour and income dynamics (SLID) dataset. The issue with SLID is that it is well known to under-represent high earners because high earners rarely agree to be part of surveys like SLID. One needs be very careful about concluding anything about the top-end based on this dataset. However, for some reason the longer-term panels of 10 and 19 years rely in the LAD dataset that is actually based on tax returns and doesn’t suffer from the same issues as SLID.

3. There is going to be significant “survivorship bias” in the longer term samples. What this means is that to be included in the 19-year panel from 1990 to 2009, you need to be alive in both 1990 and 2009, i.e. you need to have “survived” the period. So in 1990 you’ll need to be between 18 and about 58 (since life expectancy in 1990 was 77, 77-19 year panel = 58 yrs). You’re basically excluding all people who were say 59 and rich and may well have become much richer as their investments paid off but died in 2008 and so couldn’t be included in the panel.

4. Unfortunately, the study looks at quintiles (groupings of 20 per cent of the population). Previous income inequality studies have focused on the top 10 per cent and sometimes top 1 per cent and found that that’s where all the action is. It’s too bad this study doesn’t allow us to see those more specific breakdowns, even though methodologically it would have been easy to do.