There are so many misconceptions about tax policy and the size of government that it's almost impossible to sort through them all. But I'm going to do the best I can.
I will deal first with claims that high ratios of government spending to GDP are necessarily deleterious to economic growth. These claims are wrong. To my knowledge, there isn't a compelling theoretical foundation for this conclusion. And as far as I can tell, neither does the available empirical evidence support it. In this survey, Xavier Sala-i-Martin (who speaks with a certain amount of authority in this field) observes that "[t]he size of government does not appear to matter much. What is important is the 'quality of government'." Obviously, it is a very bad idea to let Robert Mugabe allocate a significant share of GDP. But governments in the Nordic countries offer an important counter-example: these countries have large public sectors, and are also rich, democratic and - by all accounts - pleasant places to live. Simply put, there is no apparent trade-off between government size and national income.
So a stated preference for a small(er) government sector can only be justified on ideological grounds. A political party may campaign for a larger or smaller public sector, but the justification cannot be that this choice will have a material effect on national income or on economic growth rates.
What does matter is how those government revenues are generated. The subject of the optimal 'tax mix' of taxes on labour income, capital income and consumption has been the subject of an enormous amount of theoretical and empirical inquiry. The theoretical literature has reached certain broad conclusions:
- Taxes on capital income generate the most distortions. In a small open economy, the elasticity of supply is very large (in principle, infinite), so small changes in the rate of return on capital can have large effects on investment.
- Taxes on consumption generate the least distortions. Since they don't affect post-tax rates of return, they don't affect savings and investment decisions.
The available empirical literature supports this consensus. An interesting recent OECD study on the optimal tax mix finds that a strategy of shifting away from income taxes - and especially corporate income taxes - in favour of consumption and property taxes has a positive effect on national income. (Shifting away from consumption taxes to income taxes has the opposite effect.) Since property taxes are generally left to the municipalities in Canada, this leaves consumption taxes as the preferred instrument for Canadian federal and provincial governments.
This lesson has been learned by all successful social democracies. As Peter Lindert explains in his book Growing Public, the secret to the Nordic countries' success is to make sure that their extensive social programs are financed by a tax mix that doesn't penalise economic growth. In particular, they make sure to offer low tax rates on capital income, and high tax rates on consumption. (See this post for more detail on this point, as well as some nifty graphs illustrating it.)
The broader point is that regardless of what you think the appropriate size of government should be, you should prefer a policy mix that penalises economic growth the least. And that means a recipe of (relatively) low taxes on capital income and (relatively) high taxes on consumption.
Sadly, Canadian political discourse is contaminated by what can only be described as tax policy fetishism: an irrational obsession for - or aversion to - certain tax policy instruments for their own sakes, without reference to their effects on what really matters. We indulge in these fetishes to our material cost. But as David Dodge pointed out over the weekend, we can't afford to do so much longer. It's long past time that we had an adult discussion about tax policy.
Stephen I'd really like you to make a quick comment on the Quebec government's just released budget. I'm thinking particularly about this new health care fee.
Posted by: Guillaume | March 30, 2010 at 11:08 PM
Actually it'd be very interesting to compare the Ontario and Quebec budgets, as they are taking radically different paths.
Posted by: Guillaume | March 30, 2010 at 11:09 PM
A nice post Stephen; it's good to reminded that while that the level of taxation is an ideological viewpoint, but the method of taxation is a matter of good governance.
Posted by: Kosta | March 30, 2010 at 11:50 PM
Good post.
I'm not sure I would agree, though, that the method of taxation is not an ideological viewpoint. True, the method of taxation should only be a matter of good governnance, but to explain the preference of various political parties for their respective tax policies, one can only look at their respective ideological fixations.
For example, consider the NDP. While they are forever looking to emulate the (admitedly unsustainable) spending policies of their social democratic (if not outright socialist) cousins in Europe, they've failed to grasp that European countries raise a far higher portion of their revenue from "efficient" taxes (notably VATs) than does Canada, the US or the UK. Instead, they commit to financing their proposed expansion of the state by way of inefficient taxes (such as income tax on the "rich" and corporations) which, in the long-term won't be able to raise the revenue they want to raise (at least not without a prohibitive impact on the broader economy). Their commitment to inefficient taxes can only be driven by class envy and their idealogical desire to punish their class "enemies".
Conversely, for conservative parties, the opposition to "efficient" taxes such as the GST make a certain amount of (admitedly cynical) sense. Efficient taxes can raise lots of money to fund an expansive state, inefficient taxes can't (at least not without bankrupting the country). A commitment to inefficient taxes is one way to control the expansion of the state budget (albeit at a high cost). To the extent that conservative parties are driven by an ideological end goal of restraining the size of government, inefficient taxes are preferable to efficient ones.
I will say this about the Canadian tax system, though, we're slowly getting it right. We're getting rid of the hopelessly inefficient sales taxes in all but Manitoba, Sask. and PEI (which three provinces will probably follow the rest into the HST system at some point in the near future). With the advent of TFSAs as a more flexible compliment to RRSPs, our income tax system is now (or will be within the next decade or so, as TFSA room builds up) a progressive consumption tax system for the vast majority of Canadians (though not for the Canadians at the top who earn a significant portion of the total income). The dividend tax credit has been adjusted to remove, more or less, the distortion caused by the corporate tax for domestic taxable shareholders (though it's still a problem for non-residents and Canadian non-taxables, including RRSP and TFSA holders - at some point the feds are going to have to introduce a refundable dividend tax credit for non-taxables). Our witholding tax regime has been amended to stop taxing interest income (which previously drove up the cost of capital for Canadian borrowers). Indeed, while it's largely gone under the radar screen, the last decade or so has seen some profound changes (and improvements) to the Canadian tax system.
Posted by: Bob Smith | March 31, 2010 at 12:35 AM
"The broader point is that regardless of what you think the appropriate size of government should be, you should prefer a policy mix that penalises economic growth the least. And that means a recipe of (relatively) low taxes on capital income and (relatively) high taxes on consumption."
What about a wealth/income inequality "place" that is demand constrained instead of supply constrained?
Posted by: Too Much Fed | March 31, 2010 at 01:29 AM
Interesting post. It will take me a while to digest some of your supporting blogs/presentations as I have been wondering/asking about the optimal CIT rate and how does one know when a particular country is at it, or still above it. I'm not sure that's been answered yet.
The Economist made a comment (embedded in a link in your nifty graphs link) about Norway being an outlier, stating "Throwing out Norway, where high GDP is due to fossil fuel reserves that cannot be achieved through any policy decision..." which caught my attention. In addition, Norway's O&G assets are largely state owned and controlled - so CIT rates on O&G investments are of reduced relevance - profits before tax end up in gov't hands either way (high or low CIT rates).
Now, even though "Nordic countries ...[are] by all accounts - pleasant places to live", what is the appropriate objective measure? How about Best Country to Live: List of Countries on 2009 UN HDI ?
http://tinyurl.com/yzc89my
Top 30 on Human Development Index Ranking
1. Norway
2. Australia
3. Iceland
4. Canada
5. Ireland
6. Netherlands
7. Sweden
8. France
9. Switzerland
10. Japan
11. Luxembourg
12. Finland
13. United States
14. Austria
15. Spain
16. Denmark
....
If you were to also throw out Norway as an outlier, Canada fares pretty well relative to the Nordic countries on this measure.
The other issue - what does Norway do with its huge O&G heritage fund? I believe it is mandated to invest outside of Norway. So, does it invest disproportionately in other Nordic countries due to proximity or cultural similarities? May be a factor in investment decisions. How do these countries fare in terms of FDI from outside the Nordic block?
Posted by: Just visiting from macleans | March 31, 2010 at 02:20 AM
"I will deal first with claims that high ratios of government spending to GDP are necessarily deleterious to economic growth. These claims are wrong. To my knowledge, there isn't a compelling theoretical foundation for this conclusion."
I guess that depends on what you find compelling. On the theoretical side, I'm convinced by Hayek and Public Choice theory. On the empirical side I'm convinced by what I see of government in North America. I have no experience of Nordic countries, but there are so many factors that confound comparison, that I think it's meaningless to try.
Posted by: ari | March 31, 2010 at 10:38 AM
What if the politically feasible options don't include what the left considers the ideal overall rate of taxation (size of government) accompanied by what economists agree is the optimal mix of taxes across consumption, capital, etc.? The left is worried that we won't get lower corporate taxes and higher consumption taxes, but just lower taxes. I think the following post and accompanying article provide some evidence for this political hypothesis in the U.S. context:
http://www.angrybearblog.com/2010/03/declining-progressivity-in-us-taxes.html?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+blogspot%2FHzoh+%28Angry+Bear%29&utm_content=Google+International
Posted by: Andrew Lister | March 31, 2010 at 03:47 PM
One reason why higher consumption taxes might not be politically feasible is that it pleases the Canadian Left to demonize them. If they would stop doing that, then it might become more politically feasible.
Posted by: Stephen Gordon | March 31, 2010 at 04:31 PM
That might be optimistic, Stephen. I think the left (in fact, opposition parties in general) demonize VATs because they are very easy to demonize. I live in Ontario, and am from BC, so I have quite a few friends who have an opinion on harmonization, and it's always bad. I do have some professional background in the merits of VATs, and I try to explain, but it is a HARD sell. Nobody buys it.
People like their taxes to be hidden, even if they are very destructive.
Posted by: Darren | March 31, 2010 at 06:10 PM
I do not think it will be feasable in near future. Left is always against it, and they will stay like that.
Nobody likes taxes, its al about how you package it. Pure Selling... but as always they aren't good at that.
Posted by: Don Antonius @ Currency Day Trading | March 31, 2010 at 06:15 PM
"One reason why higher consumption taxes might not be politically feasible is that it pleases the Canadian Left to demonize them."
Although it was the Canadian Right that actually reduced them, federally.
But partisanship aside, clearly neither side is willing to make the hard sell. Sadly.
Posted by: myron | March 31, 2010 at 07:32 PM
A question, Stephen. Do you know to what extent the two 'broad conclusions' in the literature would apply in a closed economy?
i.e. Are we just seeing the effect of a 'race-to-the-bottom' with respect to having lower taxes on mobile sources of income, or is the difference related to greater economic efficiency, regardless of competition with other jurisdictions?
Posted by: Declan | March 31, 2010 at 09:02 PM
When governments cannot control their spending and run up enormous deficits its the taxpayers that have to get hurt in the long run. They don't cut the salaries of high paid civil servants or cancel bonuses....that would be too difficult.
Posted by: Jeff | March 31, 2010 at 09:12 PM
Do you know to what extent the two 'broad conclusions' in the literature would apply in a closed economy?
The same results hold, but are less strong. The supply of capital would be less elastic in a closed economy, so the effects of capital taxes on investment would be smaller - although still negative. But the consumption tax story isn't affected. So the broad policy implications are the same.
Posted by: Stephen Gordon | March 31, 2010 at 09:32 PM
Jeff, do you have any data to support the claim that cutting the pay of senior civil servants would make the slightest bit of difference to the deficit? I have a hard time believing that the civil service is overpaid to the tune of $50 billion.
If you compare Canada to say a $1.5 trillion dollar company, how doea senior managers pay compare to senior civil servants? And keep in mind that Canada was very 'profitable' in the past decade or so, and the lack of a serious financial crisis here seems to indicate that at least some of the management was doing a pretty good job. What d'ya figure the bonus for that kind of performance in the private sector would be? $4 maybe $5 million? More?
Posted by: Patrick | April 01, 2010 at 12:58 AM
Even if everything you say in this post is correct, Steve, there's lots of room for ideological debate in terms of taxation. For example, even if we decide we're going to tax consumption do we tax:
- people when they spend their money (e.g. a GST)
- people on what they don't save (e.g. a tax system quite similar to what we have now but with unlimited RRSP contributions).
Both are consumption taxes. But there's a big ideological difference between those two options because
- direct consumption taxes like the GST create scope for all sorts of ideologically-motivated (or politically-motivated) exemptions (see debate over what is/isn't taxed under the HSTs being implemented in Ontario and BC).
- indirect consumption taxes - the unlimited RRSP contributions approach - allow for a more progressive tax structure. How progressive the tax structure should be is one of the key ideological differences across parties. Indirect consumption taxes also provide all sorts of scope for preferring some family types over others (joint v. individual taxation, exemptions for kids, for example) which is quite ideological.
What I really don't understand - and perhaps someone can enlighten me - is why the Canadian government hasn't been able to require firms to include GST in the price of goods, as it is in Europe (where it's much less unpopular). The manufacturers sales tax, which the GST replaced, was included in the price. And I don't know if anyone else is old enough and wonky enough to remember this, but when the GST was first announced, the government of the day had all of these mocked-up signs "GST included in price" that they used to sell the tax. I wish I'd kept one for my tax policy course but sadly didn't. A couple of retailers - I think Eaton's if I remember rightly - originally included GST in the price, but they lost out.
Posted by: Frances Woolley | April 01, 2010 at 08:25 AM
is why the Canadian government hasn't been able to require firms to include GST in the price of goods
From wiki:
Much of the reason for the notoriety of the GST in Canada is for reasons of an obscure Constitutional provision. Other countries with a Value Added Tax legislate that posted prices include the tax; thus, consumers are vaguely aware of it but "what they see is what they pay". Canada cannot do this because jurisdiction over most advertising and price-posting is in the domain of the provinces under the Constitution Act, 1867. The provinces have chosen not to require prices to include the GST, similar to their provincial sales taxes. As a result, virtually all prices (except for gas pump prices, taxi meters and a few other things) are shown "pre-GST", at the merchant's choice.
There was quite a large political battle over the introduction of the GST as I recall. Mulroney had to stack the Senate with PCs to get it past the Liberal majority there. Chretien ran on scrapping it in the 1991 election - which he won (PCs reduced to two seats). Sheila Copps had to resign her seat and run in a by-election when they didn't (part of her campaign promise).At the time, there was also concern by opponents that if it was hidden, the gov't would slowly raise the rate over time - which would go unnoticed by the general public.
Posted by: Just visiting from macleans | April 01, 2010 at 11:12 AM
I believe it was Mulroney who leaned toward excluding the GST in marked retail prices.
I would definitely approve of legislation requiring retailer to show the price with taxes included, while given them the option to also disclose the tax embedded. It would make changes in the rate a little less politically explosive, I think, since it might be perceived (incorrectly) as coming out of the retailer's cut rather than an extra charge on the consumer.
Posted by: Andrew F | April 01, 2010 at 02:09 PM
Thanks Stephen, that makes sense.
I agree that things are generally moving in the right direction, with some backsliding (GST reduction), I guess we just need to get moving on the Pigovian taxes (cue Mike...) and pushing for global floors for taxes on the more mobile sources of income.
Posted by: Declan | April 01, 2010 at 09:08 PM
Stephen, does this mean your ideal tax mix would be zero income taxes? And much much larger consumption+wealth taxes? Much larger than the Nordic countries? Or can a little bit of income taxes be a good thing?
Also, I'm only a layman, so I hope you'll forgive my likely naiveté, but if a country followed your logic, and moved towards consumption taxes, wouldn't that make people overly-likely to save rather than consume?
Posted by: Alex | April 02, 2010 at 09:17 PM
"The capital gains tax rate has wandered all over the place in the past generation. Private savings have not shifted significantly in response. Moreover, any increase in private savings must be balanced against the increased government borrowing due to the tax cut. Unfortunately, research shows that each dollar of tax cuts increase private savings by less than a dollar; the net effect is to lower national savings. Continued belief that lowering tax rates will boost American savings is the triumph of faith over experience."
-- Brad DeLong and David Levine, at:
http://www.j-bradford-delong.net/oped/acapitalgainstaxcutoped.html
Also, with regard to Lindert's book, I don't think he means to say (or shows) government spending doesn't matter much, only that higher and lower transfer payments within a conventional reasonable range don't seem to make much difference. Certainly if education, infrastructure, and basic science spending is cut by 95% we shouldn't expect there to be little difference in long run growth compared to if spending on those things were doubled.
Posted by: Richard H. Serlin | April 25, 2010 at 10:06 PM