I suspect that the typical Canadian - and this includes economists - has the impression that the tax system in Canada is progressive. Maybe too progressive, or not progressive enough according to their taste, but progressive. It turns out that this isn't the case. According to this paper (pdf), the most recent study that tries to figure out what proportion of a household's income goes towards taxes is
Frank Vermaeten, W. Irwin Gillespie, and Arndt Vermaeten, “Tax Incidence in Canada,” (1994), Canadian Tax Journal, vol. 42, no. 2, pp 348-416.
Sadly, there appears not to be a version of this paper on the internet, so I had to actually go to our library to take a look. It was well worth the trip, because their results aren't what I was expecting at all.
Here are their estimates for the effective tax rates on broad income (this includes direct transfers) by income level and by type of tax:
'Commodity taxes' include such things as sales taxes.
The most striking feature of that graph is that the average effective tax rate across income groups fluctuates in a fairly narrow range between 30% and 35%. It turns out that the progressive effects of personal income taxes are almost exactly balanced by the regressive effects of commodity taxes.
Of course, that was in 1988, so it's not clear we'd want to make any strong conclusions about the current tax structure:
- The GST will have increased the importance of commodity taxes, but since the GST rebate has also increased the income of lower-income households, the two effects may cancel each other out.
- Personal income tax rates have moved around quite a lot since 1988.
- The authors assume that the corporate income tax is entirely borne by owners of capital, so it's strongly progressive. I disagree, for reasons I discuss elsewhere.
I suspect that an analysis that took into account these points would find that the tax system is even less progressive than we'd think, and may even be generally regressive.
Update: This is indeed the case.
You're not going to get too far with 18 year old data. StatsCan has very current series available, and if you want to do the leg work you can update that table fairly easily. The changes in the tax system over 18 years are very substantial and this old data doesn't tell you anything about what is going on today.
Just one example: as income goes up, the incidence of capital gains and dividends goes up. Each are taxed at lower marginal rates than income from employment. Both rates have been subject to various changes for political and fiscal reasons over the past 18 years - for example, the Liberal Government ended full integration of taxation for non-public corporations and their shareholders sometime in the '90's.
I suspect that if you do the work you will find that a very large precentage of all income taxes are paid by the top decile of payers, but that the average rate payed by them is still under 35%.
Posted by: David Baskin | August 14, 2006 at 05:15 PM
The study is out of date, certainly. But I don't think the changes since then (including the ones you mention) have had the effect of transferring the tax burden to upper income households.
I'll be posting again on this in a day or two using more recent data - the basic pattern is still there.
Posted by: Stephen Gordon | August 14, 2006 at 07:00 PM
Looking at the graph, I can't see how the commodity tax and income tax add up to the total tax. What else is in there--property taxes?
Posted by: Brett | August 19, 2006 at 09:14 AM
While waiting for the next post you may want to check:
http://www.ctf.ca/pdf/04ctjpdf/04ctj3-kessel.pdf
A nice survey with some more recent studies.
The data behind the graph above (from the original article) is presented in Table 11, page 763. In particular, it shows what other taxes are included.
Posted by: Leon | August 21, 2006 at 10:56 AM