Federal Taxation of Labour Income in Canada is Regressive (in Terms of Marginal Rates)

Or at least, it is for some ranges of income.  Don't believe me that marginal tax rates are regressive?  Follow me:

For this analysis, I will need to do the following:

• Ignore provincial income taxes since the tax system differs from province to province.
• Assume that the person does not live in Québec (for reasons we will discuss later).
• Ignore industry specific insurance premiums on income, such as WSIB premiums since these vary from industry to industry (and, as it turns out, do not change our story in any fundamental way.

Federal Income Taxes

If we include the Federal Basic Personal Amount as a separate tax bracket (it's the amount of money a person needs to earn before they start paying federal income taxes), Canada has five marginal income tax brackets which are as follows:

• 0%: $0 - $10,382 
• 15%: $10,383 - $40,970
• 22%: $40,971 - $81,941
• 26%: $81,942 - $127,021
• 29%: $127,022 - inf

(all tax rate data from Taxtips.ca) So far this looks pretty progressive, with low-income earners paying no income tax.  Interestingly, the first tax bracket starts up well before Statistics Canada's Low-Income Cut-Off (LICO) which is a widely used (but far from universally accepted) definition of poverty.  In 2004 the LICO for a single person was between $14,000 and $20,337, depending on where the person lived (source: CCSD).

Payroll Taxes on Employees

There are two payroll taxes paid by the majority of working Canadians: Canada Pension Plan (CPP) contributions and Employment Insurance (EI) premiums, both of which are proportional to labour income.  There are some exceptions:

• People can be exempt from either CPP or EI, for a variety of reasons.  For instance, individuals who own 40% or more of the companies they work for are exempt from EI premiums. 
• Québec has their own version of CPP (QPP).  Currently the marginal tax rates are identical, but the plan is administered by the province rather than the federal government.

CPP Employee Contribution Rates

• 0%: $0 - $3500
• 4.95%: $3501 - $47,200

EI Employee Premium Rates

• 1.73%: $0 - $43,200

Note, there is no basic exemption level for EI and both EI and CPP payments kick in well before LICO.

Adding it Together

For a person who is eligible to pay both EI and CPP along with income taxes, there are now eight effective tax brackets on labor income:

• 1.73%: $0 - $3,500
• 6.68%: $3,501 - $10,382
• 21.68%: $10,383 - $40,970
• 28.68%: $40,970-$43,200
• 26.95%: $43,201-$47,200
• 22%: $47,201-$81,941
• 26%: $81,942-$127,021
• 29%: $127,022-inf

Note that marginal tax rates are declining from $40,970 to $127,021, and a person earning $41,000 a year faces almost the same marginal tax rate as someone making $410,000.  And it gets worse from here.

Payroll Taxes on Employers

Employers are required to pay EI premiums and CPP contributions for their employees as well.  Canadian law requires employers to match CPP contributions dollar-for-dollar, and EI contributions 1.4-to-1 (that is, for every dollar employees pay, employers pay $1.40).  Thus the tax rates for employers are as follows:

CPP Employer Contribution Rates

• 0%: $0 - $3,500
• 4.95%: $3,501 - $47,200

EI Employer Premium Rates

• 2.422%: $0 - $43,200

As such, 50% of the legal incidence of CPP Contributions falls on employees along with 41% of the legal incidence for EI premiums.

But Who Really Pays Payroll Taxes

Stephen Gordon examined this question a few weeks ago in The economics of tax incidence: paying the tax is not the same as bearing the burden:

Payroll taxes. These include employer contributions to EI and C/QPP as well as Worker's Compensation premiums. But as a HRCD survey notes, long-run labour demand is more elastic than labour supply, so the ultimate effect of payroll taxes is to reduce wages: "labour's share of the payroll tax burden in the long run is in the range of 87 to 100 percent."

Additional evidence to this point - The Incidence of Payroll Taxation: Evidence for Chile (Jonathan Gruber, Journal of Labor Economics, July 1997) has a useful literature survey which includes the following:

More recent work in the United States has examined the effects of changes in the costs of government mandated employer benefits within different states over time, controlling for correlated time-series and fixed location effects. These studies have found that the incidence of mandated employer benefits is fully on wages, with little to no disemployment effect.

Meaning that the incidence falls on employees.  Gruber's study of Chile came to the conclusion, but this is not a universal finding in the literature.  Holmlund's 1983 study of Sweden found that only half of the employer's share was paid by employees in reduced wages.

Given all of this evidence, it is difficult to imagine that less than 75% of payroll taxes are ultimately paid by employees.  We will use this figure as a lower bound, though HRDC's figure of 87-100 percent is likely more accurate.

If employees are paying 75% of payroll taxes, then their tax rates on EI and CPP are roughly 3% and 7.5% respectively.  Using these figures along with income tax rates leads to the following marginal tax rate schedule:

• 3%: $0 - $3,500
• 10.5%: $3,501 - $10,382
• 25.5%: $10,383 - $40,970
• 32.5%: $40,970-$43,200
• 29.5%: $43,201-$47,200
• 22%: $47,201-$81,941
• 26%: $81,942-$127,021
• 29%: $127,022-inf

Now the highest marginal tax rates are on people making around $41,000 a year.  Marginal tax rates are regressive from income levels of $40,970 to $81,941, and people under the LICO 'poverty level' making $11,000 face roughly the same marginal tax rate as those earning $110,000 a year.

Complications and Considerations

There are at least three additional factors we may want to consider:

1. Are Payroll Taxes Equivalent to Income Taxes

They potentially could differ in one significant aspect.  If a person pays additional income taxes, they do not receive any additional services from the government.  However, if a worker pays higher EI or CPP contributions, they may receive additional benefits.  From Gruber:

However,Summers (1989) noted that this analysis missed an important of feature of payroll taxation: tax revenues are often used to finance programs which benefit workers only, such as retirement benefits under Social Security or compensation for workplace injuries.This restriction of benefits to workers creates an important tax/benefit linkage...

If workers believe that EI and CPP payments benefit them directly, while income taxes do not, the behavioural response to changes in each will be different.  Specifically, we would expect that higher income taxes would lead to sharper reductions in labour supplied than higher payroll taxes.

The question then is, "Do workers perceive payroll taxes as being fundamentally different than income taxes?"  I am not aware of much research in this area.  It would be very difficult to argue that for EI, there is a close relationship between tax and benefit.  Premiums are not based on expected benefits, as the rates for people with almost no chance of finding themselves unemployed are the same as for higher risk persons.  Secondly, since EI premium revenues were incorporated into general government revenues in the 1990s, there is not necessarily a close link between overall EI revenue received and EI payments made.  This distinction is discussed in great length in Economics versus Politics in Canadian Payroll Tax Policies (Jonathan R. Kesselman, Canadian Public Policy, September 1998).  The link between a person's taxes paid and expected benefit is quite weak in the case of EI.

The case for CPP is much less clear, as there is a closer (but far from one-to-one) link between taxes paid.  Younger workers also have an additional concern - will the system still be in place when they retire?  There is not much data for Canada, but there is a significant body of evidence that suggests that young Americans believe Social Security is important, though they will not be able to collect it.  From the New York Times:

 

Yet here’s the odd part: only a third of the younger respondents, and only a third of all respondents, expressed confidence in the program’s future. How do you square these opinions that Social Security is crucial, that you want it to be available decades hence, with the belief that it probably won’t be?

Think tanks and life insurance companies are recommending to young workers not to count on a young social security system - to the point which the Social Security Administration felt the need to respond.  Whether or not the claims of the demise of Social Security are valid or overblown is somewhat irrelevant - if young workers believe they will receive no personal benefits from their contributions, then the behavioural response is the same as an income tax.

In the previous analysis, I have treated CPP contributions as a straight tax.  That may be excessive, though I am not convinced.  However, it may make more sense to treat it as 75% tax and 25% forced saving, or 50% tax and 50% forced saving.  The latter would drop the CPP tax burden by workers down from 7.5% to around 4% - which does not look too much different than the marginal tax rate figures that did not take into account employer contributions, which were still regressive.  The tax/benefit linkage may reduce the regressivity of labour income taxes, but it does not eliminate it.

2. Provincial Income Taxes

We have only considered federal income taxes.  Perhaps once we take into account provincial income taxes, the labour income tax system ceases to be regressive.

The difficulty: Provincial income tax systems are not progressive enough to swing the pendulum back to an overall progressive tax system.  Alberta has a flat tax on income, so it does not make the overall situation more progressive.  Ontario, on the other hand, has a progressive income tax regime which is as follows:

• 0%:  $0-$8,943
• 5.05%:  $8,944-$37,106
• 9.15%:  $37,107-$74,214
• 11.16%: $74,215-inf

If we add this to the marginal tax rate schedule that includes both employer and employee contributions, we get a rather lengthy marginal tax rate schedule that looks as follows:

• 3%: $0 - $3,500
• 10.5%: $3,501 - $8,943
• 15.55%: $8,948-$10,382
• 30.55%: $10,383 - $37,106
• 34.15%: $37,107-$40,970
• 41.65%: $40,971-$43,200
• 38.65%: $43,201-$47,200
• 31.15%: $47,201-$74,214
• 33.16%: $74,215-$81,941
• 37.16%: $81,942-$127,021
• 40.16%: $127,022-inf

Once again, we see the highest marginal labour tax rates paid by individuals earning roughly $41,000 a year, with the system being highly regressive between the income range of $41,000 and $74,000.

3. Tax Incidence of Income Taxes

One assumption we have made through this entire discussion is that the incidence of labour income tax rates fall wholly on those earning the income.  What if, instead, some of the incidence falls on firms or customers?

Remember, though, that the income tax portion is the progressive part of the whole labour income tax system.  The higher the level we assume is paid for by non-workers, the more regressive the system actually is.  The only way this consideration would make the system more progressive is to assume that the tax incidence faced by workers is greater than 100%.  That is theoretically possible, but in reality unlikely.

Labour Income Tax System - Conclusion

When payroll taxes and tax incidence is taken into consideration, the Canadian labour income tax system is highly regressive in the income range of $40,000-$70,000.  Taking into consideration the benefit/tax linkages of payroll taxes mitigates this, but only partly.  Given that both EI premiums and CPP contribution rates are both likely to rise in the near future, expect an even more regressive Canadian tax system soon.

Comments

[Stephen Gordon]

Some of my colleagues went through a similar exercise for Quebec, including the effects of transfers; I blogged about it here (OMG - 4 years ago!!).

They got a similar story, only the effect of withdrawing transfers as income increases makes the effective marginal tax rates much higher. There's even an interval where they exceed 100%.

[Determinant]

Mike, your analysis of the CPP is fundamentally flawed. CPP bears little relation in its current form to US Social Security to the point that comparisons are irrelevant.

CPP and QPP are currently identical in form and implementation: a universal defined-benefit pension plan supported by investment returns. Both CPP and QPP funds are invested in external financial instruments (often out of Canada). They are not solely invested in Government bonds and therefore do not depend on the government's credit and tax guarantee. In this it is unlike Social Security.

With defined-benefit pensions it is customary to carve out both contributions and benefits and let the DB RPP "wrap-around" CPP. Benefits from CPP are tied directly to contributions. Universal social welfare payments to seniors are delivered through Old-Age Security and Guaranteed Income Supplement. Both are funded out of the Consolidated Revenue Fund, that is through tax revenue, just like any other welfare problem. CPP provides benefits to contributors, period. Canada is different from the US in that old-age welfare comes explicitly out of general tax revenues.

The insurance redistribution provided through CPP is the same as that provided through any other defined-benefit pension plan, aka life annuity.

Second, marketing material from life insurance companies is biased. I have worked in that industry. Never forget that they are salespeople, ever. Such material is intended to motivate people to save more in insurance company products. The truth of such material is open to debate however it is not and will never be unbiased research.

[Mike Moffatt]

Stephen: I recall reading a mid 1980s study that suggested that a single mom living in subsidized housing with 3 kids faced a marginal tax rate around 120-130%.

Determinant: Agreed on the differences between CPP and SS. However, I think you missed the major point: "if young workers believe they will receive no personal benefits from their contributions, then the behavioural response is the same as an income tax."

Whether CPP is stable or a complete mess is largely irrelevant - it's the belief that drives behaviour.

That being said, I wish there was better behavioral research on whether people see (and act as if) CPP contributions and income taxes as being identical. I suspect for my age bracket they do, but there really isn't much evidence either way.

[Mike Moffatt]

RE: "Second, marketing material from life insurance companies is biased. I have worked in that industry. Never forget that they are salespeople, ever."

That was exactly my point.

[Mike Moffatt]

Now, it would be true that looking at the U.S. situation is misleading if the *beliefs* of Canadians differed significantly from that of Americans. I can't say for certain either way - I can't find much of anything on Canadian beliefs RE: CPP (let alone on behavioural responses), which is why I discussed the U.S. situation. I'll continue looking and update that section if I find anything good.

[nonrefundable tax credits]

Employee paid CPP and EI contributions and premiums are non-refundable tax credits. Have a closer look at Schedule 1 of a T1. Please rework your tax brackets.

[Mike Moffatt]

Whoops - good eye! I do have to re-do the math here somewhat. Don't think it will change the results appreciably, but I do want to be as accurate as possible.

[Kevin Milligan]

We typically analyze the redistribution in a tax system according to average, not marginal tax rates. You are right that the Canadian tax system does not feature monotonic marginal tax rates. But this doesn't directly have welfare implications. If you want to talk about regressivity/progressivity of the whole system, you ought to use average burdens. Marginal tax rates only matter insofar as they affect average burdens.

[nonrefundable tax credits]

Don't think it will change the results appreciably

Agreed. There is a kink in the lower part of the second T1 tax bracket if federal income taxes and CPP contributions and EI premiums (payroll) are commingled with or without a percentage of the employer share of payroll being allocated to the employee.

Any employee benefits arising from the commingling are not kinked, however.

[Determinant]

Mike:

Your point is not accurate. What American workers believe and what Canadian workers believe are two independent things and cannot be assumed to be correlated. The comparison you attempted is a straw-man argument. Different nations, different policies, different implementations.

In your analysis, you would have to treat a private defined-benefit RPP as a tax. The arguments you attempt to levy are the same as for a DB RPP, except for the sponsor. In this case sponsor solvency risk is significantly less with CPP. The redistribution/cross-subsidization is the same. In fact the only two differences with CPP are the fact that the pool of contributors includes all working Canadians, the manager is the Government of Canada and premiums are collected through the tax system.

You have also not addressed the fact that wealth redistribution in Canada is conducted through the policy instrument of OAS/GIS, not CPP. The benefits provided under OAS are funded through current tax revenue and are means tested. Neither is true for CPP.

Third, it's financially impossible in Canada (not to mention malpractice) to ignore the effects of CPP in financial planning, both income and benefits. If you do it yourself and most especially if you have an employer or counsellor do it for you, you have to take CPP into account. Just because people gripe about the future doesn't mean they don't take present legal realities into account when planning. There is a fundamental difference between what they say and what they do. I've been there.

[Stephan Wehner]

Put a graph up at http://twitpic.com/2jgkeh.

As always it would be nice to see the evolution of the figures over the last decades

-- Stephan

[Mike Moffatt]

Hi Kevin,

You read my mind - planning on following up with analysis of avg. rates, though I do believe marg. Rates have welfare implications beyond avg. Rates.

When I get back to office will fix tax credit issue and do avg. Rates in a couple of days.

[Kevin Milligan]

Mike: If I have 100K income, my welfare changes if I pay 20K or 40K taxes on that 100K. This is the average tax burden. if I pay 99% or 1% on hypothetical dollar #100,001, it does not change my current welfare given that I am at $100K.

Marginal tax rates matter a lot for efficiency. They also matter because for equity because they affect average tax rates. But on their own they do not affect equity.

[Mike Moffatt]

Agreed. I need to clarify my thinking here - it's one of those things that made sense to me at 6am.

[asp]

CPP is not a tax. It is retirement pension. 100% forced savings, if you want to put it that way. But we all benefit from the resulting automatic stabilization.

EI is not a tax. It is an insurance plan. While not all individuals will collect that insurance, everyone benefits from the automatic stabilization during recessions.

These positive effects function whether or not individuals believe in them.

Attempts to lump these two into the "taxation" category smell of hard core free market ideology. And we have seen where that kind of thinking has gotten us in the last few years.

[Nick Rowe]

Mike: While I digest this, one typo:

"CPP Employer Contribution Rates

* 0%: $0 - $3500
* 4.95%: $3501 - $47,200"

Should be EmployeE, I think?

[Mike Moffatt]

Whoops - thanks Nick - fixed now!

Thanks everyone for your comments - I'm going to revamp this tomorrow to clarify a few areas; was going to tonight, but just got back from a 2 year old's birthday party and am absolutely wrecked. Kids are exhausting.

In short: The point about EI/CPP benefits aren't that the programs are or are not valuable. The importance from a behavioural point of view (as pointed out by Larry Summers) is how the individual believes the benefits they receive from the payments they make. For instance, I believe that gas taxes used to pay for road construction is valuable, but if I (alone) pay an extra $1000 in gas taxes, I don't see any personal benefit (it's not as if the Transport Department uses the money to pave my driveway).

[Mike Moffatt]

That should be: "individual believes the benefits they receive ARE RELATED TO the payments they make."

[Frances Woolley]

Nick, CPP is 4.95 for employee and 4.95 for employer for a total of 9.9. Self-employed pay the full 9.9, but can deduct half of it i.e. take it as a deduction not a credit.

Mike, refundable tax credits, e.g. child tax benefits, GST credit, make a big difference to the progressivity of the income tax system and also significantly influence marginal tax rates.

[Geoff NoNick]

The problem is that you've lumped EI and CPP into the meaninglessly broad category of "taxes" - i.e. any deduction administered by the government, according to the definition you seem to be using.

The CPP is not a tax, it's an investment; you get back returns based on how much you invest. That is not the case with income tax.

EI is also not a tax, it's an insurance premium. You pay an amount to insure something; in this case the insurance stops paying on earnings over $47,200. So yes, people earning more than that pay a smaller percentage of their income than people below that, but they also don't accrue any insurance benefit above that amount.

If you want to include such things in your calculation to bolster your premise, why not also include car insurance and property insurance into the calculation? Surely some might consider these necessities of life: indeed, car insurance is government-mandated. The fact that something is simply collected by the government does not make it a tax.

I also object to your insinuation that the LICO is the definition of poverty and that there should be no tax on earnings below that. I can't think of anything other than the most poorly-research newspaper articles that refer to the LICO as the "poverty line". Why would being in the loosely-defined category of "low income" exempt one from making any financial contribution, however small and relatively less than others, whatsoever to society?

[Frances Woolley]

Mike, if you're interested in this area, check out Kevin Milligan's web page and in particular his Canadian Tax and Credit Simulator. A gift to economist-kind - I can't say enough nice things about Kevin's work. He also has a forthcoming paper in Canadian Public Policy about the progressivity of the Canadian tax/transfer system.

You might also want take a look at some of the research in this area, e.g. the papers by Jon Kesselman (2004) and by Vermaeten, Gillespie and Vermaeten (mid 1980s) in the Canadian Tax Journal. References:

http://www.sfu.ca/mpp-old/04research/pdfs/CPPR_tax_inequality.pdf

http://qed.econ.queensu.ca/pub/cpp/Sept1995/Vermaeten.pdf

[Mike Moffatt]

"If you want to include such things in your calculation to bolster your premise, why not also include car insurance and property insurance into the calculation?"

Because the premiums you pay are related to the benefits your expected benefits, which is clearly not the case here, for several reasons:

- People who are ineligible to collect EI are still required to pay it
- People who have no realistic chance of losing their job are still required to pay it
- Revenues go into general revenues and are not necessarily related to payments (see Kesselman 1998).

"The fact that something is simply collected by the government does not make it a tax."

Agreed. And the fact that something is called an insurance does not make it the equivalent of car insurance.

Frances: Thanks for the links! I'll have to check out Prof. Milligan's piece. I'm a big fan of Prof. Kesselman's work, as shown above.

[Mike Moffatt]

"Mike, refundable tax credits, e.g. child tax benefits, GST credit, make a big difference to the progressivity of the income tax system and also significantly influence marginal tax rates."

I should have included the GST credit - mea culpa. There was an implicit assumption in all this that I was looking at a single childless income earner. I should have made that explicit. Will add that to the list of assumptions.

[Determinant]

"People who have no realistic chance of losing their job are still required to pay it".

Give me an example. I've seen laid off teachers, public servants, even church ministers.

What you're describing is classic insurance thinking. It's the customer's thinking in the problem of Selection Against the Insurer. The rational economic response is to include as many people in the base as possible. This is no different from things like group health and disability insurance.

"People who are ineligible to collect EI are still required to pay."

This is the only criticism that is fair. However Service Canada would respond that they can't say that you'll move into insured employment in a years time and be collecting in three years. So there are two sides to that argument.

"Revenues go into general revenues and are not necessarily related to payments."

If EI was a private disability plan (DI is the closest relative of EI) you wouldn't make this argument. When a private insurer makes a profit it's OK, when a government does so it's a nasty tax. This also ignores the fact that the Government of Canada is obligated to make up losses in the EI Fund and has done so repeatedly in the last 50 years. The reforms of the early 1990's threw the plan in to consistent surplus.

EI is a hybrid creature between welfare and pure insurance, but that doesn't mean it's a tax for most people.

Insurance depends on cross-subsidization of losses between premium payers. It's an age old struggle between having many members with low claims and therefore a large number of members who think the value of their coverage is marginal and low, and having a small number of members with a high probability of claims and premiums that nobody can afford. It's gone on forever.

Just because you think the risk is low and want out doesn't mean you can get out and the risk isn't there.

[Frances Woolley]

"There was an implicit assumption in all this that I was looking at a single childless income earner."

Over 18 and under 65. Not self-employed.

For this group you also need to bear in mind the new working income tax benefit (don't remember it's name) that's available for singles.

[Mike Moffatt]

Determinant - all I can suggest is that you examine the literature. In either case, it's a tax - the key is, again, as Summers points out, is to determine the level of benefit linkage. We may disagree on the degree, but the benefit linkage is clearly not 100%, as the premiums are not based on any actuarial calculation of the risk at all.

And please refrain from the 'big bad government' stuff. It's hurting my head to be too be criticized from both the right (on my implication that perhaps people under LICO shouldn't be paying income taxes) and on the left (you said government is bad!). It'd be much easier if y'all push from the same side. :)

[Mike Moffatt]

"Over 18 and under 65. Not self-employed."

I did mention that there were exceptions (see discussion RE: 40% ownership stake. Maybe useful to list the major ones?

"For this group you also need to bear in mind the new working income tax benefit (don't remember it's name) that's available for singles."

I didn't realize how many of these I missed. Again, I'd like to thank everyone for their helpful suggestions. I'm looking forward to polishing this up this week.

[Determinant]

I disagree with the literature on a fundamental level. I have a well-grounded view on what constitutes insurance and what constitutes a tax. This difference here is political at root. I look at objective facts and have my opinion. I don't see the fact that some revenue may flow to the government as a tax.

The key question for me is "If this plan were operated by the private sector, would the same structure apply?" Since EI resembles pension plans where employers bear sponsorship risk, and the fact that the Government has LOST money on EI, that is it has had to make up deficits in the EI Fund from the Consolidated Revenue Fund (i.e. income taxes) in the past, there is a good argument to be made that EI premiums aren't a tax.

Consider the case of community-rated health insurance. The benefits linkage isn't 100% either. In fact you're stumbling over the fact that benefits for EI are community-rated based on historic unemployment rates. This is a valid insurance stance, it's just that community rating isn't common anymore since it has mostly passed out of use in Canada.

EI is an insurance plan, not a savings account. There is always a significant risk you won't claim anything.

As for the left and the right, I critique inconsistencies wherever they may be. ;)

[Geoff NoNick]

Mike -

I won't bother re-iterating the good points Determinant makes, except to also add that benefit linkage doesn't correlate 100% in the case of car insurance either. Here, again, the insurance is government-mandated but the benefit accrues disproportionately to people who have accidents. I have never had a car accident and might prefer to just be a safe driver than to buy insurance, but I recognize that I may some day benefit from it. So too, I do not have a job that allows me to draw EI should I become unemployed, but I recognize that I may get one some day and my past contributions will allow me to draw the benefit.

The EI system could be re-jigged to exempt premiums for people who are not in a situation to immediately enjoy a benefit, at the expense of charging a higher premium to people who do. Would that make the system more "progressive" and address your concerns? Indeed, most of the people you are concerned about in the $37-47K range are actually EI's principle clientele as salaried employees (it's no mistake that they stop insuring benefits above that amount), and a great many high-income self-employed (lawyers, doctors, government employees, university professors, etc) cannot benefit and are actually those ones you identify as unlikely to need it. Doesn't that mean that the higher-income are actually subsidizing the lower-income employed, making the whole system more progressive than your model portrays?

[Andrew F]

I agree that EI is a tax more than insurance. Premiums are completely insensitive to risk. I don't care where the revenues go--any reasonable insurance would vary premiums according to risk.

[Kevin Milligan]

I'm with Mike on EI and CPP. Benefit-tax linkages are tricky in practice. Exchanges between individuals and government range the spectrum from buying a bottle at the LCBO that confers a direct benefit to the individual to tax assessments on income to which no direct benefit is attached.

The right way to think about this is to ask--if the EI or CPP premium is raised, does that change my CPP or EI entitlement? The answer in both cases is no. CPP benefits depend on your lifetime pattern of earnings, not how much tax you paid. It so happens that CPP taxes also depend on earnings, but that is beside the point. Same with EI--your benefits depend on your earnings, not how much EI premiums you have paid.

It is true that on a macro scale, budgets do need to balance. (Note that this is enforced much more for the CPP than EI. Anyone who looks at total EI revenue vs EI benefits paid over the last 30 years would be very hard pressed to argue for any tax-benefit linkage.)

However, on the individual scale that matters for Mike's analysis, one's own benefits depend not at all on one's own marginal contribution. They are taxes.

[Chris S]

"EI is a hybrid creature between welfare and pure insurance, but that doesn't mean it's a tax for most people."

You have to be careful about the other side - the benefit. I've tried reading all the comments, and I don't think it's been pointed out that for a high-income earner, EI payments won't help all that much. If your EI benefits won't even cover your base expendituures, like mortgage, car insurance, gas, and groceries, then it likely affects your view of the value of the premiums.

However, on a different note...

I would like to see the last charts (the ones including the provincial details) with a column added that shows the population in each of those brackets. I think we would suddenly see why government tax policy is actually **really** hard. You can play with the rates on the upper end all you want - running them either up or down, but to get enough money to run the government, you have to tax people in the middle. This isn't ideology - it's just demographics.

[Roland]

WRT federal taxes, it seems that the regressivity caused by payroll taxes is just an artefact of the maximum insurable or pensionable earnings level for the EI/CPP programs.

Two ways to solve that: raise the insurable or pensionable income limits, or levy the premiums and contributions on higher earners while still capping benefits. Either way, there's problems, e.g. potential six-digit EI bums, or unapologetic redistributionism. Political hell either way.

Determinant btw did well to point out the OAS/GIS factor which is fair comment if CPP was going to be treated as tax rather than as savings.

Andrew: as for making EI premiums vary with risk, wouldn't that compound the problem? i.e. some regions and sectors would have astronomical premiums, employment levels in those regions and sectors would drop, and the risk being insured against would made more likely to happen. Besides, the politics would blow apart the country--nothing like another unity crisis to remind us we're at home!

Chris: do you tax where the people are, or where the money is? If Canada had a fairly equal income distribution, that would be the same thing. But of course there is a demographic effect on CPP balances.

I don't know if the state-run universal pensions are going to fare much worse over the long run than all the other saved claims to future production, represented by other retirement assets. Society as a whole is to a certain degree always a "pay as you go" phenomenon.

[Determinant]

With EI, premiums are capped at $747.36. The rich individual making six figures may collect EI, but as his premiums and benefits are capped at the Maximum Insurable Amount. I fail to see how this translates to a "tax". It is implicit that these individuals should have enough savings and assets to see them through, but that doesn't keep them from claiming EI by itself.

It is also apparent that many are missing the fact on how EI is rated. Please see http://en.wikipedia.org/wiki/Community_rating. Not all insurance in individually risk-rated. It doesn't have to be. That doesn't mean its not insurance.

EI is Community Rated across Canada; the premiums are supposed to balance expenses "over the economic cycle". Benefits vary by region; regions with higher unemployment have longer benefits. This may seem counterintuitive until you reason that if the purpose of EI is to get people off EI, regions with lower unemployment rates should have lower unemployment periods. This is the same sort of reasoning that employers use with Disability Insurance. Most employer group policies will have 1-2 years of "Regular Occupation" Disability and the balance of the period will be the more stringent "Any Occupation" definition. In both cases access to benefits is restricted motivate people to get off of claim.

[Richard D]

Hey. Long-time reader first-time commenter. I wanted to get a better picture of individual tax burden and incentives, so I took the values here and wrote a little program to compute retained income (in Ontario) for a given level of income. Of course to do that I had to put in the flat-lining CPP and EI limit. If I missed something I'm open to correction. This excludes the effects on wage from employer contributions.

I graphed the results at http://i.imgur.com/GHSeE.png

You can see there are six pivots there, where going up in income lowers your after-deduction income itself. Those (rounding up to the nearest thousand) are at 11k, 38k, 41k, 75k, 82k, and 128k.

[T. K.]

Mike,

Great stuff, fascinating reading -- but unless I'm missing something, your provincial tax rates are too low.
Ontario's top marginal tax rate isn't 40.16 per cent. It's 46.41 per cent.
I think you left out the surtaxes (which are not insignificant) on higher income earners.
See: http://www.rev.gov.on.ca/en/tax/pit/rates.html
Other provinces have similar ratese, whether structured as surtaxes or as additional (and higher) tax brackets.