Taxation and Economic Growth

I had a bit of an intellectual crisis this evening as I pondered the conventional wisdom in economics regarding the choice between reducing consumption taxes or income taxes.  Briefly put, the simple conventional wisdom is that taxes on consumption are preferred to income taxes because they encourage saving and long-term capital formation and economic growth.  Income taxes, on the other hand, distort the labour-leisure choice and can reduce labour supply and therefore reduce economic growth.

Yet, as I thought about the Canadian economy and how well it seems to have weathered the global economic downturn, one thing that stuck out in my mind is that the federal government reduced the GST rate just prior to the Great Recession.  In 2006, the rate was reduced from 7% to 6% while in 2008 it went down to 5%.  Could the reduction in the GST rate have stimulated consumer spending in Canada just enough to help offset some of the effects of the economic downturn?  After all, a reduction in the GST can be interpreted as economic stimulus.  If spending increases are economic stimulus, so are tax decreases.

I’m afraid I don’t have an answer to that specific question for Canada.  However, it should be possible to see if countries with greater reliance on consumption taxation have better growth rates than those that were less reliant.  What I did do was quickly collect data from OECD Statistics for the 34 OECD countries on the annual average growth of per capita GDP in US PPP dollars over the period 2007 to 2009, the share of GDP accounted for by taxes on goods and services in 2008 and the share of GDP accounted for by taxes on income and profits in 2008.  I then proceeded to plot the tax to GDP shares against the growth rates as shown in Figures 1 and 2 with a linear trend.  The results show a positive correlation between growth in per capita GDP and the share of GDP accounted for by consumption taxation and a negative correlation between growth and the share of GDP accounted for by taxes on income and profits. Well, I’m glad I sorted that out.  Good night.

 

Related articles

How OECD governments generate tax revenues

Comments

[Colin Percival]

"taxes on consumption are preferred to income taxes because they encourage saving and long-term capital formation and economic growth" ... "Could the reduction in the GST rate have stimulated consumer spending in Canada just enough to help offset some of the effects of the economic downturn?"

These comments make me wonder if there's an argument to be made for shifting taxation back and forth between income and consumption based on the economic cycle. If you want to reduce the variance of annual GDP growth, shift towards taxing consumption during high-growth periods (to encourage saving), but shift back towards taxing income during low-growth periods, to encourage those savings to be spent.

Can we achieve Keynesian aims by manipulating taxation policy without deficit spending?

[foosion]

Causation and correlation. Check. Adequate sample size. Check.

[Jacques René Giguère]

@Colin Percival : maybe. But one of the preasons we abandonned fiscal and budgetary policies was that the populace didn't like tax rates manipulated. Monetary policy is uncomprehensible to laymen and policymakers can more easily tinker with. Though I am still in favor of boosting expenditures if the need arises.

[Giovanni]

"...conventional wisdom is that taxes on consumption...encourage saving and long-term capital formation and economic growth. Income taxes, on the other hand, distort the labour-leisure choice and can reduce labour supply and therefore reduce economic growth."

This line of thinking is indeed conventional wisdom, but I've never understood why. What matters to a worker - and governs their labour-leisure choices - is the rate at which they can trade their time for purchasing power: for this purpose there is no difference between a 20% income tax deducted from their paycheque and a 20% consumption tax applied at the shopping mall. To a first approximation at least, the distortionary effects of income and consumption taxes on labour-leisure choices are additive and there is no a priori reason to favour one over the other.

[Stephen Gordon]

I once wrote a blog post on that exact question:

"Why the GST is a good idea"

[Wonks Anonymous]

Giovanni, the reason to prefer consumption taxes is investment. Investing allows people to consume more later, but if you had paid the tax when you received the paycheck, you would not have as much to invest.

[Jacques René Giguère]

And, with the withdrawal at the source system, people know their net pay and have barely any idea of the taxes they pay ( it's usually "It's too much" and they confuse taxes with insurance and even their parking fees...). Consumption taxes are added to the posted price ( except at the SAQ) where the pain is felt.
(For non-QC readers: SAQ is the provincial gov't liquor monopoly which includes taxes in the posted price.)

[Bob Smith]

Livio,

It's off topic, but any thoughts on how the government squares the estimated $26 billion 2012-13 deficit, with today's fiscal monitor showing a deficit of $11.8 billion for the first 11 months of 2012-13. Sure, there are always some year-end adjustments, but $14 billion worth? Maybe a topic for a different post.

[Bob Smith]

Jacques: "Consumption taxes are added to the posted price ( except at the SAQ) where the pain is felt."

We could mandate tax-inclusive pricing - the GST/HST legislation already includes provisions for tax-inclusive pricing, they just have to be activated.

Or you could implement a consumption tax using the existing income tax system by providing a deduction for all investments (sort of a super RRSP). Giovani's right that for most people, whose principal (if not only) savings are tax-exempt savings (RRSPs, TFSAs, registered pension plans, etc.) there's no difference, in terms of incentives, between an "income" tax and a "consumption" tax because in substance our income tax is a consumption tax.

[Stephen Gordon]

Bob - apparently the govt is expecting to book a charge for some Atomic Energy of Canada Limited's liabilities. The budget didn't give a number, but it must be pretty big.

[Livio Di Matteo]

Hi Bob:
I've noticed they have been ahead of schedule for the last few issues of the fiscal monitor. I recall it was similar last year (the deficit was 12.6 billion for the same period last year compared to 11.8 this year. There are some additional charges coming and I think Steve nailed them on the head. As well, debt service costs are down by nearly 2 billion dollars so far.

[Giovanni]

As described in Stephen’s earlier post, there is a better theoretical argument to be made for consumption taxes with respect distortionary effects on household saving behaviour. Income taxes - at least as they are applied in Canada - do tend to reduce after-tax rates of return on standard forms of household wealth. But whether this appreciably reduces household saving is an empirical matter. My literature knowledge in this area is woefully out-of-date, but my recollection is that at least circa the early 90s the empirical consensus was that rates-of-return had very little impact on real-world consumption-savings choices. (To be clear, I mean ex ante expected rates of return and not ex post rates. If the stock-market tanks or a housing bubble bursts and people suddenly find their net wealth to be far less than they want it to be this may indeed dramatically raise the household saving rate.)

Moreover, even if household saving did respond strongly to after-tax rates-of-return this wouldn’t lead directly to a general presumption against income taxes. Rather, it would only imply governments should be concerned about how heavily income taxes fall on investment income. As Bob points out, there are all kinds of targeted measures governments can use to reduce the effective tax rates applied to such income.

[Bob Smith]

Giovanni,

Just to be clear, though, if you provide a deduction for investments, you no longer have an income tax, you have a consumption tax. Since consumption = income - investment, a deduction for investments means that you're only taxing consumption.

Stephen,

You know, I have a feeling we've already had this exchange, and I think you're right. On the other hand, that explanation would suggest that their 2013-14 number badly overstates the likely actualy deficit since, presumably, the government isn't planning on taking more AECL charges next year.

[anon]

Colin Percival, wouldn't you want to do the opposite? When growth is low, shift taxes away from capital income and towards consumption, in order to incent investment spending. When growth is high, tax capital income more to reduce investment spending and prevent the economy from overheating or blowing bubbles (Sometimes this is called "macroprudential" policy). Obviously it's a third-best policy; you'd want to use monetary and fiscal policy first. But it might be relevant in regions with fixed currency regimes and balanced-budget constraints, such as European countries and perhaps North American states/provinces.

[Determinant]

Why we worry about the supply of investment funds is beyond me when we quite clearly have a demand problem. We have low investment because there is low demand for it, period.

Secondly, a progressive consumption tax system is much harder to achieve than a progressive income tax system, both policy-wise and especially politically. It is much easier politically to identify and tax high-wealth individuals than to make the "poor" self-identify and sign up for credits or allowances.

Third, as the Canadian income tax system has had deductions for RRSP's and Pensions since the 1950's, Stephen's model does not bear much resemblance to reality. As was pointed out in the GST thread, "consumption = income - investment" is rather like "S=I", sure its accounting but its misleading. Do we tax consumption broadly and hit basic consumption hard like food, housing and transportation, or do we go for a progressive income system that recognizes that money in a bank account isn't much an investment? It comes down to the Keynesian idea that investment is usually the short side of S=I and it's really min (S,I) so taxing excess S does not impact investment.

[Giovanni]

Bob,

If by "provide a deduction for investments" you mean something along the lines of RRSPs, I agree. This in turn is analytically equivalent to exempting returns on investments from taxable income. And this is really my point. To say there is an efficiency argument for shifting from income to consumption taxes is to argue that generating public revenue by taxing returns to personal investment leads to greater efficiency losses than taxing other things such as labour earnings. This may very well be the case, but it is something that needs to be demonstrated empiricially. Yet many economists seem to have adopted a "strong view" on this based on the belief there is powerful apriori case for preferring consumption taxes.

[Stephen Gordon]

Um, there *is* empirical evidence. Quite a lot, all saying the same thing. Here is a recent OECD effort. And here is a recent IMF working paper.

[Bob Smith]

Giovanni,

I think there's actually voluminous empirical evidence on the costs of taxing investments/investment income vis-a-vis other forms of taxation and the efficiency gains from not doing so. Once upon a time I was familiar with the leading literature on the point, but now I'll have to rely on regular posters here who are more familiar with it than I am to provide examples of that literature (is Kevin Milligan in the room?). If economists have adopted strong views on the points it because there's a solid theoretical basis for the proposition supported by an abundance of empirical evidence.

Moreover, from a practical policy perspective, it's telling that successful high-tax jurisdictions (think Sweden) have chosen to raise money with hefty taxes on consumption and labour, but relatively modest taxes on investment income. That's not a coincidence, rather it reflects the conscious application of good tax policy to raise public funds using the most efficient taxes possible.

[Bob Smith]

"Secondly, a progressive consumption tax system is much harder to achieve than a progressive income tax system, both policy-wise and especially politically. It is much easier politically to identify and tax high-wealth individuals than to make the "poor" self-identify and sign up for credits or allowances"

I don't get that. In Canada, our income tax system is, in effect, a progressive consumption tax system for most Canadians (since most don't/can't max out their RRSPs, and so, in theory, can avoid tax on investments). It's no harder to implement the one than the other. We could implement what would be, substantively, a progressive consumption tax sytem by deleteing a few words from the Income Tax Act (by removing the contribution limts from RRSP), and it would be no harder to administer than the current Income Tax Act (in fact, it woulod probably be a heck of a lot easier because no one would worry about taxing investment income - I should watch what I lobby for, it would probably put me out of a job).

[Andrew F]

The only difference is that it is very difficult to invest in small private businesses with before-tax income. This is compensated for by the lifetime capital gains exemption, I suppose...

[Giovanni]

Bob,

In my comments about the effects of taxing returns to investment I was referring to household saving behaviour only, and the difference between income and consumption taxes specifically in that regard. Taxing business income is a completely different story. That replacing high taxes on that sort of "investment income" with hefty taxes on consumption/labour might raise business investment and real growth makes perfectly good sense to me. That replacing income taxes with equivalent consumption taxes would significantly raise household saving, not so much...that's all I was trying to say.

As for the general question of consumption versus income taxes, maybe I just can't see the obvious. Is there some standard "story" that clearly implies cutting income tax rates across the board by one percent while raising the HST by one percent (actually, it really should be a comprehensive, uniformly applied consumption tax...but let's say the HST to be concrete) would cause people to work more or acquire more human capital or change their behaviour in any other way that suggests a lessening of the efficiency burden of taxation?

[Determinant]

"Secondly, a progressive consumption tax system is much harder to achieve than a progressive income tax system, both policy-wise and especially politically. It is much easier politically to identify and tax high-wealth individuals than to make the "poor" self-identify and sign up for credits or allowances"

I don't get that. In Canada, our income tax system is, in effect, a progressive consumption tax system for most Canadians (since most don't/can't max out their RRSPs, and so, in theory, can avoid tax on investments). It's no harder to implement the one than the other. We could implement what would be, substantively, a progressive consumption tax sytem by deleteing a few words from the Income Tax Act (by removing the contribution limts from RRSP), and it would be no harder to administer than the current Income Tax Act (in fact, it woulod probably be a heck of a lot easier because no one would worry about taxing investment income - I should watch what I lobby for, it would probably put me out of a job).

What do you want to be progressive with respect to? We can easily spot somebody who makes more than somebody else. But how do we judge somebody who spends more than somebody else? Power inequalities and situation differences mean some people will have to spend more than others, and in reality lower income earners spend more of their incomes. It's not a recipe for progressive taxation, which says we tax people with the ability to pay, not those who through no fault of their own have to spend more.

It's also a false contention that our income-tax system is a progressive-consumption tax system. Those RRSP contributions are taxable in full on withdrawal, so you can't armwave the investment deduction away. And we're generous to RRSP's so that people will acquire a decent retirement income, which is taxable on withdrawal. That is a conscious policy decision.

Unlimited investment deductions allow for unlimited accumulation of capital in individual hands, which is a recipe for power concentration. No, we don't want that either.

[Jacques René Giguère]

Bob Smith: Couple of short points:
There was a debate when GST was introduced about hiding thet ax in the price tag. But the Conservatives wanted the tax visible so there would be a public demand to lower it. Which it what happened twenty years later. That it is an efficient tax is of no concern to LIVs.

There is a large litterature ( too busy with grading to do the research now) about how tax-exempt savings scheme do not raise savings among the poorer segment as they have no money anyway and do not raise it among the well-off as they would save it anyway.

[rsj]

A 2 year sample size? With an R^2 of 0.09? Looking at this, I assumed you had concluded that there was no significant relationship. Then I remembered the topic. What exactly is the bar here? Is R^2 = 0.00001 good enough?

[rsj]

Also, what's important for policy discussions is the tax *rates*, not the tax collected, and certainly not TOPI as measured by NIPA, which don't map to any reliable notion of consumption tax. For example, if I pay $100 in consumption taxes, but receive an offsetting federal tax credit for the same, then TOPI measures this as a consumption tax nonetheless, even though the effect on incentives is very different.

Tax rates are set by policy, tax collections are the response to the policy. Certainly those nations undergoing consumption booms are going to have higher consumption tax collections and higher GDP growth rates, all things equal, than those nations undergoing consumption slumps. But you would not infer from this that there should be higher consumption tax rates because these drive GDP growth.

In any case, you would need to look at something at least across several business cycles and then make an effort to tease apart correlation/causation.

[Dan Kervick]

Briefly put, the simple conventional wisdom is that taxes on consumption are preferred to income taxes because they encourage saving and long-term capital formation and economic growth. Income taxes, on the other hand, distort the labour-leisure choice and can reduce labour supply and therefore reduce economic growth.

This conventional wisdom makes very little sense to me. Consumption is the final activity for which all other economic activities take place. So how could be the case that taxes on consumption are generally preferable to taxes on income or taxes on investment? Certainly discouraging certain kinds of present consumption and encouraging more saving could be the right policy under the right circumstances. But by the same token taxing certain kinds of investments to encourage more consumption could be the right policy under other circumstances. Every prosperous and growing society needs some kind of balance between present consumption and investment in the future, but I don't see any reason for thinking there is a general rule that an imbalance is more likely to be on one side than the other.

I would think that conservatively-minded economists would prefer the tax on income to the tax on consumption, because by taxing income the taxing government leaves the choice over how to allocate the remainder to the receivers of the income, while taxes targeting consumption (or targeting investment) are more prescriptive and paternal, attempting to steer activity one way or the other.

[rsj]

" Certainly discouraging certain kinds of present consumption and encouraging more saving could be the right policy under the right circumstances."

Bingo, The problem being that that the wealthy receive most of their income from capital whereas those in the bottom 99% receive most of their income from wages. Having government put its thumbs on the scales and tax labor income more than capital income is equivalent to taxing the wealthy less and the bottom 99% more. I remember Warren Buffet pointing out that he pays a lower tax rates than his secretary, and that this couldn't be good for the nation. It's odd to see the contortions that some twist themselves into when arguing that Warren should be paying a lower rate.

[rsj]

There is a list of some studies here with a longer term perspective than the 2 year sample offered by Livio and the 20 year sample offered by Stephen Gordon:

http://thinkprogress.org/economy/2012/11/29/1252751/no-low-capital-gains-taxes-dont-boost-the-economy/

Money quote:

"The Congressional Research Service, in a report that Congressional Republicans tried to hide, found no correlation between capital gains taxes and economic growth. Former White House economist Jared Bernstein noted that the capital gains tax has had no effect on investment. The University of Michigan’s Joel Slemrod found that “there is no evidence that links aggregate economic performance to capital gains tax rates.”
As billionaire investor Warren Buffett wrote in a New York Times op-ed this week, “let’s forget about the rich and ultrarich going on strike and stuffing their ample funds under their mattresses if — gasp — capital gains rates and ordinary income rates are increased. The ultrarich, including me, will forever pursue investment opportunities.” Buffett said that the capital gains rate affects investment behavior “only in Grover Norquist’s imagination.”

[Colin Percival]

rsj, the "Buffet pays a lower tax rate than his secretary" claim is only true if you ignore the corporate taxes which he pays by proxy.

I run a business -- a one-person corporation -- and pay myself in dividends rather than salary.
Thanks to the dividend tax credit, my personal tax rate is about 15% lower this way; but that doesn't mean that government is getting any less in taxes; they're collecting the rest of the money from my corporation, and the two cheques add up to just as much as if I had taken money out of the corporation as salary and paid all the taxes personally.

The US tax system is broken in many ways, but comparing personal tax rates while ignoring the taxes paid by corporations does not give you an accurate picture.

[rsj]

"rsj, the "Buffet pays a lower tax rate than his secretary" claim is only true if you ignore the corporate taxes which he pays by proxy."

First, corporations pay very little in terms of corporate taxes in the U.S, primarily because of loopholes. The effective corporate tax rate is about 13%, which is lower than even the capital gains rate and lower than payroll tax rates.

Second as corporations are legal entities with their own books, the equity holders of the firm are not in a unique position to pay the corporate tax rates "via proxy" rather than, say, customers of the corporation. The incidence of taxation, particularly with monopolistic competition, will depend on a lot of factors, but certainly it is not all on shareholders (and remember that Buffet makes a lot of money by lending to firms as well).

I agree that there is a case to be made for cutting corporate taxes, but this does not change the fact that Buffet and the wealthy in general pay taxes at a lower rate than the middle class, who must pay payroll and income taxes.

[anon]

RSJ, a consumption tax is not a payroll tax. With a consumption tax, income from capital gets taxed normally as soon as it's spent. The point is that you should tax present and future consumption at the same rate, instead of overtaxing future consumption. This is also why RRSP contributions are taxed at withdrawal.

[rsj]

Anon,

I'm talking about income taxes, which neither tax present nor future consumption, but market income. Governments levy taxes primarily for social insurance payments. Imagine if your fire insurance was a function of whether your income was primarily from asset sales or from labor sales. It would be grossly inefficient. Similarly, imagine if your insurance payout was less if you decided to spend the insurance proceeds on consumption (renting), but more if you decided to spend it on buying. Your insurance payout should not try to influence your future rent/buy decision after your old house burns down. That consumption/investment decision is your own. The function of social insurance payments are there to reduce nominal income volatility, not to pressure households to save more or less. Really one can argue that they exist -- and the modern tax and spend state exists -- because without nominal income insurance, there would be wasteful savings demands as people excessively self insure in case of idiosyncratic risk.

Similarly when governments levy taxes to redistribute income for social welfare, it makes little sense to give a free pass to the wealthiest members of society who either do not work at all or if they do work, receive a negligible proportion of their income as wages. The whole point of redistribution would be to take more income from those folks and to give that income to those who earn less, creating a minimum level of income for at least some members of society. Again, the total amount that you earn determines the optimal income smoothing policy in this case, not how the income is earned. For the (small) remnant of public spending that consists of direct purchases of goods and services, the largest component here is defense. Again, why should the burden of paying for the national defense disproportionately fall on those who do not have a lot of market wealth? National defense is a service also consumed by retiree. The retirees did not "pre-fund" their national defense contributions buy buying future jet fighters in the past to be delivered to the government when they retire. Everyone who is alive has some burden to contribute to the national defense each year that they benefit from it. Healthcare even moreso. Whether they work or not.

None of this has anything to do with taxing consumption twice. None of this has anything to do with consumption at all. Taxation and spending is first and foremost about nominal income smoothing and to a lesser degree about marshalling both factors of production for the government's own use. If, in order to build a lot of aircraft carriers, or hospitals, the government requires 10% of the nation's capital, and 10% of the nation's labor, then why should this be paid for with a 20% tax on labor income only? Why tax labor twice?

[Determinant]

On further thought, if lower-income individuals have a higher marginal propensity to consume their marginal income (last dollar earned), then does not a consumption tax fall much harder on them? Worse, the if the MPC is inversely proportional to income, low-income individuals face very high marginal rates.

Ah, I see now, this is just an attack on a progressive tax system.

[Stephen Gordon]

*sigh*

[Patrick]

Mmmm. BBQ'ed sacred cow.

[Determinant]

And the debate will continue, Stephen, because neither side will budge. Before you cite Sweden again, it has high income taxes* too, so do as proper analysis, you have to analyze income taxes and sales taxes simultaneously, they are not independent which is what Livio assumed above.

And in my lefty opinion, rsj has put forth the better case.

It's also a case of politics. It's ten times easier to argue politically that the rich should pay more rather than giving allowances to the "poor". People don't want allowances because they don't want the government to call them poor. Much easier to send The Man after The Rich.

*Sweden uses the Ghent System, in which unions and union does are one of the main deliverers of social welfare, so you have to include union dues as taxes to get a fair comparison.

[Stephen Gordon]

And yet in Sweden, the heavy lifting of reducing income inequality is done using transfers. The Swedish tax system has about the same redistributive effect as Canada's.

The Canadian Left's intellectual - and moral - failure is its fetish for policy instruments and utter lack of concern for policy outcomes.

Yeah, yeah, yeah, I know: "This goes to eleven."

[Stephen Gordon]

"People don't want allowances because they don't want the government to call them poor."

Any empirical support for that rejection of policies that demonstrably reduce inequality?

[Determinant]

The Canadian Left's intellectual - and moral - failure is its fetish for policy instruments and utter lack of concern for policy outcomes.

I thought we were to stay away from personal attacks on this blog? Moral failure? I am a involved activist and member of the NDP, sir.

"People don't want allowances because they don't want the government to call them poor."

Any empirical support for that rejection of policies that demonstrably reduce inequality?

It's called practical politics. Positioning matters, emotion matters. People matter. I go with what sells in this riding. Leadership involves convincing people that your solution is right. Politics is voluntary and people have to be convinced, which is why politics is a variant of sales.

And your study didn't take into account that the Scandinavian countries, as I said, use the Ghent System, which funds most of the "welfare state" out of insurance dues based on union membership, which is why Swedish union rates are incredibly high. Swedish direct government benefits aren't generous and are comparable to Canadian rates. It's the Ghent system of union membership that takes everything to the next level.

Canada, OTOH, uses the Beveridge model which funds most of the welfare state directly through government taxes and transfers.

The Left in Canada is still committed to the Beveridge model which features higher tax rates and uses government as the delivery agent.

[Patrick]

Stephen's critique seems fair to me. The Canadian left has a history of embracing policy measure with no hope of accomplishing the stated policy goals. The rational is usually that the policy goal is popular, but so is the antithetical policy instrument.

As it is, the NDP is known as the party of high minded hippies who could never be trusted to run the country. Don't you think the irrationality of advocating mutually inconsistent policy goals and instruments reinforces that perception?

Stop trying to win by telling people what you think they want to hear, and start proposing coherent policy. Have the courage of your convictions! Be the alpha dog.

[Eric L]

One difference between consumption and income taxes is that consumption taxes are much more cyclical, dropping much more quickly and sharply in a recession and bouncing back more quickly in a recovery as well. If your data set is comparing countries at different points in their business cycle, that could explain some of the correlation -- a country in a recession will see its consumption taxes drop while income taxes rise relative to gdp, and in a recovery the reverse happens.

[Shangwen]

"People don't want allowances because they don't want the government to call them poor."

That would certainly explain all those cheques for the Universal Child Care Benefit that never got cashed.

[Determinant]

As it is, the NDP is known as the party of high minded hippies who could never be trusted to run the country. Don't you think the irrationality of advocating mutually inconsistent policy goals and instruments reinforces that perception?

I was at the NDP Convention in Montreal, and have our Policy Book to hand. It's online at montreal2013.ndp.ca. It's very realistic and practical, with few of the high-minded hippy resolutions.

If you say the NDP can't be trusted to form the government, then you haven't seen the NDP lately. And I think the people of BC are about to disagree with you, Patrick.

Explanation: There is only one NDP with a federal section and thirteen provincial sections (the Quebec Section only competes federally) and a common membership roll. There aren't 13 different NDP's.

[Giovanni]

Eric L,

Good point. The OECD study Stephen Gordon cites above actually tries to address this issue. Table 5 gives estimates obtained with regressors “purged of all possible correlations with the business cycle, by regressing them on the output gap”. These results suggest the business cycle effect you describe may indeed be important. For example, the principal results [Table 1, Col. (2)] indicate that shifting 1% of tax revenue from property/consumption to personal income taxes knocks 1.13% off annual per capita GDP growth. When the business cycle correction is applied this estimate becomes (maximally) 0.27% [Table 5, Col. (2)]. (Conceptually, this latter value is actually weighted-average of underlying personal and corporate tax-share impact coefficients. Since the results consistently show the latter impact to be the larger of the two, the “true” personal tax-share impact implied by these results is likely somewhat less than 0.27%.)

The authors report on other robustness tests in Table 4. Measures of the rate and volatility of inflation – which one might expect to be business-cycle related - are strongly statistically significant when added as regressors, but seem to have no effect on the estimated tax share impacts. Other (possibly cycle-related) variables are more interesting. The principal results [Table 1, Col. (1)] indicate shifting 1% of tax revenue from property/consumption to personal/corporate income taxes knocks 0.98% off per capita GDP growth. This becomes 0.65% when a measure of openness to international trade (the population-adjusted sum of exports and imports) is included as a regressor [Table 4, Col. (6)], and 0.28% when a measure of national civilian R&D activity is added [Table 4, Col. (8)].

Unfortunately, the authors don’t give results where trade-openness, R&D-intensity and business cycle effects are taken into account simultaneously.

[Jacques René Giguère]

Shangwen: that's why Family alllowances and OAP are not means tested. There is no shame getting them.And the gunmint get it back in the Income Tax Return.
As they say in the movie "La grande séduction" ("Seducing Dr. Lewis" in english Canada) : "They give you money for two weeks and shame for the whole month."

http://fr.wikipedia.org/wiki/La_Grande_S%C3%A9duction
http://en.wikipedia.org/wiki/Seducing_Doctor_Lewis

[Stephen Gordon]

There may be a stigma associated with welfare - but is anyone suggesting eliminating it? I don't see why there's a stigma in receiving supplements to your market income. It's not as though people look through your mail for govt cheques or can see when the govt makes a direct deposit into your account.

[Determinant]

Jacques' nailed it. Most people want to believe that they are self-sufficient, regardless of the facts. Having to fill out a means test for benefits is humiliating. Having a cheque regularly arrive that reminds you of that means test is a constant irritant to a person's pride. William Beveridge understood this back in 1942 which is why he recommended abolishing means tests for social insurance. And that's the view that prevailed in the Commonwealth.

You want to get angry over not understanding basic "facts", Stephen, I get angry at having to repeat basic political and economic conclusions which have been clear and understood since the beginning of the Welfare State in Canada and were clearly stated at its inception.

Tu quoque.

Politicians live in a world of emotion, it just as real as anything else and it has to be accommodated along with every other concern.

[Stephen Gordon]

You're making unsupported assertions to denounce policies that demonstrably increase incomes of low-income households and which reduce inequality. Which side are you on, again?

[rsj]

I would have no problem cashing a government income check. But I would use only a part of it for consumption, and would use the rest to purchase capital.

Income redistribution is just that -- tax one person's income and give it to someone else. Not sure why this needs to be conflated with a policy of subsidizing investment at the expense of consumption, or why a the wealthy, who consume a smaller share of their income, should contribute a smaller proportion of their income to be redistributed than the middle class.

Let's separate out the issue of income redistribution from that of subsidizing investment. If there is some irrationality that creates a need to subsidize investment, let's identify that irrationality and make a case that it can be remedied with an appropriate investment subsidy policy, independent of any income subsidy policies.

[Determinant]

I have made no unsupported assertions, Stephen. I did say I agree with rsj, and I agree with the post above mine. Enough with the ad-hominems, already.

You, sir, have carried on about Sweden while patently ignoring exactly how it finances its welfare programmes (union dues and premiums, the Ghent System is rather like Obamacare). That's like assessing government debt in Canada and comparing us to France or the UK while not taking provincial debt into account.

http://www.guardian.co.uk/money/2008/nov/16/sweden-tax-burden-welfare

http://en.wikipedia.org/wiki/Taxation_in_Sweden - note the "Social Fee" which is not included in Income Tax per se, yet is still sizeable and higher than the VAT rate.

You want support for anti-means testing? Go read the history of Old Age Pensions in Canada, especially the reform in 1954 which abolished the means test of the 1927 system. Or the Beveridge Report.

ISTM you are not interested in a debate. The ad-hominems are already flying, and all I've done is agree with rsj and added some political context.

[Patrick]

Determinant - I would like to like the NDP *federally*. As you well know, several provinces have a long history of electing provincial NDP gov'ts. It hasn't helped much at the national level, and I don't think it's relevant and I don't want to get bogged down.

More importantly, I think my assertion stands. The NDP lacks the courage of it's convictions.

I had a look at the policy book. It's actually a good example of what I'm complaining about. It's the same old same old we get from the other parties.

There are four vague bullet points on taxation, which neglect transfers entirely. On poverty there are four points (with D being expanded with many sub-points), and all but A are are so vague as to be meaningless.

Now, picking on one of my own hobby horses, specifically climate change, I see nothing inspiring. Three vague bullet points about what the NDP believes. I don't care what they 'believe', I want to know what they intend to do. Specifically. Carbon market? Fine ... gonna auction the permits? How's it's going to functions? How's it to be governed and regulated? Why a market and not a carbon tax? Polluted pay you say? What about pricing power - they'll just pass it on. Etc ... The details matter. No mention of replacing coal fired power plants (the single biggest emitters). And no specifics for reducing emissions from cars (#2 after coal fired electricity). It's hopelessly inadequate.

Furthermore, climate change - arguably the single largest threat to continued ease and comfort for modern human civilization - merits three bullet points, while Supporting Canadian Creativity merits ... seven? C'mon.

For me, this is the same old drivel. It's a lot of pandering to various factions within the base, a bunch of stuff to try to keep QC, and lots, and lots of hand waving.

Meet the new boss. Same as the old boss.

[Determinant]

First, thanks for even looking at the Policy Book :)

Second, the Policy Writing Guide issued to Riding Associations said that Resolutions had to be general. It IS all about what we believe; implementation comes later and is the responsibility of the Federal Caucus and the Campaign Committee. The Policy Book stands from election to election, so it has to be general to have a lifespan.

No mention of replacing coal fired power plants (the single biggest emitters).

That's a provincial responsibility. The provinces own/regulate specific plants. It's always been that way.

You'll have to wait for the Election Platform to see how we intend to implement things.

[Bob Smith]

Oh my, I turn off for the weekend and look what I miss.

Determiant: "It's also a false contention that our income-tax system is a progressive-consumption tax system. Those RRSP contributions are taxable in full on withdrawal, so you can't armwave the investment deduction away. And we're generous to RRSP's so that people will acquire a decent retirement income, which is taxable on withdrawal. That is a conscious policy decision."

*shakes head sadly* Right, RRSPs are taxable when they are withdrawn and used to... consume. If people consume their lifetime income (i.e., save when young, dissave when old) all income is taxable, because all income is consumed. But in an income tax system, investments are subject to double taxation, first when they're saved, and again when they generate returns. Or, put differently, an income tax system is a consumption tax system which taxes future consumption more onerously than current consumption.

Determinant: "Before you cite Sweden again, it has high income taxes* too, so do as proper analysis, you have to analyze income taxes and sales taxes simultaneously, they are not independent which is what Livio assumed above."

Good lord, how often do we have to hear this nonsense? Sweden has high income taxes on LABOUR income, investment income is taxed at a substantially lower rate (like, half).

"The Left in Canada is still committed to the Beveridge model which features higher tax rates and uses government as the delivery agent."

Instead of higher tax rates, wouldn't it be better to have higher tax REVENUES? You can't spent tax rates.

See, that's the problem. The Canadian (and American and pre-Blair British) left fixates using inefficient taxes with high notional rates (namely income taxes), which never generate meaningful revenue at the margin, instead of using efficient taxes (VAT or payroll taxes) to generate high revenues. You want high tax revenues and government spending, hey, go for it. All we're saying is that if you want to do that you have to use SMART taxes, not stupid ones. The NDP's rejection of consumption taxes puts it in an odd alliance with conservative small government types (for a Canadian example, witness the baptist-bootlegger coalition of Bill Van der Zamm and the NDP in the BC HST debacle). For conservatives, opposition to consumption taxes has a certain practical and ideological appeal. Consumption taxes are an efficient way to raise vast sums of money, while income taxes are an inefficient way to raise revenue. So long as income tax is seen as the principal tax policy pool available, government revenue (and spending) will remain relatively small (at least compared to EU countries), because they simply cannot raise enough money to fund the spending ambitions of the left (at least not in a sustainable way, i.e., without crippling the economy). That's a position which, while cynical, at least is rational, if you're a believer in small government. But it is inexplicable why the NDP shares that view. Indeed, the NDP is complicit in maintaining that conservative position because it presents high, inefficient, income taxes as the only way to fund its spending ambitions - feeding into the conservative message.

"You, sir, have carried on about Sweden while patently ignoring exactly how it finances its welfare programmes"

It finances them principally with taxes on labour (payroll taxes and labour income) and VAT - http://www.skatteverket.se/download/18.616b78ca12d1247a4b2800025728/10411.pdf
And according to the Swedish tax authorities - who might know a think or two about the subject - 60% of the "social contributions" are properly regarded as taxes.

[Bob Smith]

"Not sure why this needs to be conflated with a policy of subsidizing investment at the expense of consumption, or why a the wealthy, who consume a smaller share of their income, should contribute a smaller proportion of their income to be redistributed than the middle class."

Whoa, let's be careful there. We're not talking about "subsidizing" investment, we're talking about not double-taxing it. That's a huge difference. We're talking about treating future consumption (i.e., today's savings) the same as today's consumption. To characterize that as a subisdy for future consumption does a great deal of violence to the concept of a "subsidy".

And let's also separate the tax instruments we use with a discusstion of redistribution. The wealthy may well consume a smaller share of their income, but that doesn't mean you can't design a consumption tax that is sharply progressive (using different tax rates, higher exemptions, etc.). Again, one could consider Canada's income tax system (at least at incomes below ~$130,000 to be a defacto consumption tax regime, and yet there's not doubt that, even over that range, it is sharply progressive. You want the rich to contribute to public goods and transfers, great, all I'm suggesting is that we should use an efficient instrument to achieve that goal.

[Bob Smith]

Livio, I think the spam filter ate one of my comments.

[Bob Smith]

Determinant: "Explanation: There is only one NDP with a federal section and thirteen provincial sections (the Quebec Section only competes federally) and a common membership roll. There aren't 13 different NDP's."

Formally, no, but there's a world of difference between the NDP in Manitoba and Saskatchewan, who regularly govern, and their federal cousins who have never seen the inside of a Cabinet room (BC political parties, of all political stripes, are in a world all their own).

Consider Kevin Milligan's timely piece in today's Macleans highlighting the decision of Manitoba's NDP government to buck the trend of raising taxes on the "rich" (which taxes raise little or no revenue) in favour of raising the PST by 1 percentage point - http://www2.macleans.ca/2013/04/29/manitobas-choice/#.UX59_XeW-9o.twitter. The PST is a sub-optimal tax (Manitoba should convert to the HST), but it's probably less inefficient than the alternative income tax increase that would generate the same revenue.

That's how responsible social democratic governments operate, and it's a world of difference from the rhetoric that has formed the basis of NDP tax policy for the last 40+ years.

So there's a NDP government that's

[genauer]

Livio,

nearly totally off topic, but your old post about NBER w18315 paper from Robert Gordon “Is U.S. Economic Growth Over? is closed, and I think, you might be interested in this:

I wondered at that time, whether I believe the factor of 2 drop, from 1450 to 1650.

And I stumbled yesterday upon the Herengracht index, 350 years of house / rent prices in Amsterdam,

e.g.http://www.personal.psu.edu/bwa10/House%20Prices%20and%20Fundamentals%20On-line%20Appendices.pdf

a) growth rates just dominated by inflation index

There are some folks, who say we gained a factor of 100 in GDP over the last 200 - 250 years (2 % growth, as a stylized fact)

The paper from Robert Gordon looks more like a factor of 13,

and the Herengracht data say: basically constant

b) long term expectations
Rent price ratio is historically at 1:16, pretty much the same as the PE ratio for the S&P500

And if we now have services at 80% of the GDP, we are basically consuming just each others time, and this can not grow at all.

[Jordan]

I ran a slightly different version of this analysis, comparing proportion of sales revenue to sales + income revenue as it correlates with growth and found some slightly different results.

http://www.introspections.org/wp-content/uploads/2013/05/sales_tax_v_income_tax.pdf

2008 is roughly in line with what you found. But 2009, when the recovery was really starting, shows a significant, in fact the only significant result in the lot, negative correlation of having a high sales tax:income tax ratio and economic growth.

Between 2005-2009, where the tax data was available, this method of predicting growth doesn't seem to useful. 2009 really stood out to me though because it was the first test I ran and seemed like the most obvious choice.

[rsj]

That's a huge difference. We're talking about treating future consumption (i.e., today's savings) the same as today's consumption. To characterize that as a subisdy for future consumption does a great deal of violence to the concept of a "subsidy".

"Future consumption" is not taxed. Only present activities occuring in the time period in question are taxed. You cannot say that because A is necessary for B, then a tax on A is a tax on future B. I mean, you can say it, but that type of argument is capricious and meaningless. I need to eat today in order to have strength to supply labor tomorrow. Does that mean that a tax on consumption is a tax on "future labor supply"? It is absurd to be making these types of claims. And if you look at what most people spend their consumption basket on -- transportation, education, shelter, food, health care -- that is also critical for the future production of goods and services, nevertheless a tax on these expenditures is not a tax on future production anymore than a tax on investment.

In terms of "double taxation" -- that is an equally meaningless statement. When a worker pays a portion of their wages for FICA taxes, then another portion for income taxes is that "double taxation", or is the total tax burden just applied in two levies? What is important is not the number of levies but the total levy. There is nothing "unfair" about applying X levies instead of a single levy.

[Determinant]

I was also going to say that the assertion that today's savings is tomorrow's consumption is meaningless. Today's savings MAY (and often is) tomorrow's income, a percentage of which may be saved and the rest consumed; there is a fresh choice each time period. So there is no "double-taxation".

What a VAT does is allow untaxed compounding on a very broad basis. A VAT is an excise tax and whenever a country relies on excise taxes it faces Pitt's Dilemma if they are relied upon too much. William Pitt the Younger tried to finance the Napoleonic Wars based on excise taxes; he taxed hair powder (which was de rigeur at the time) and powdered hair and wigs promptly fell out of fashion. So Pitt instituted the First Income Tax in 1798.

Any time a tax allows a choice to escape it (investment, in the case of VATs) it risks narrowing the tax base. An income tax is not so easy to escape, that's why it's such a "burden", all taxes are a burden if they are to be actually paid.

Formally, no, but there's a world of difference between the NDP in Manitoba and Saskatchewan, who regularly govern, and their federal cousins who have never seen the inside of a Cabinet room (BC political parties, of all political stripes, are in a world all their own).

No, there isn't. Been to Convention, seen it all up close.

[Bob Smith]

"Future consumption" is not taxed. Only present activities occuring in the time period in question are taxed"

In an income tax system, any future consumption financed with savings is taxed when the money to pay for it is earned, not when its consumed. Think otherwise? Try this thought experiement. You earn $100 and put 40$ under your bed for next year. When do you pay the tax on the $40 you saved (i) when you consume it next year, or (ii) when you earn it now? Hint, unless you want a visit from the tax authorities, I wouldn't answer(i). It's a simplified example, but it illustrates the point, in an income tax system, future consumption is taxed when income is earned, not when it is consumed.

"You cannot say that because A is necessary for B, then a tax on A is a tax on future B."

Why not? It's true.

Let's start with a couple of simple propositions:

(1) Todays savings becomes tommorow's consumption (i.e., people don't save and bury money in the ground foreever): S(today) = C(tommorow)/(1+r), where r is the real return on savings.

(2) Income is the sum of todays consumption and today's savings: Y = S(today) + C(today)

So a tax at rate t on income is the same as a tax on consumption and savings: (1-t)*Y = (1-t)*S(today) + (1-t)*C(today)

As for double taxation, again, that's easy to show. With an income tax, C(tommorow) = (1-t)*(S(today)(1+r*(1-t)).
See how the tax shows up twice in that equation? Once when the money is saved, and again when it generates a return (i.e., investment income). You can make the double taxation explicit by rewriting the equation to that C(tommorow) = (1-t)*s(today) + (1-t)*(1-t)*r.

The brilliance of RRSPs and TFSAs is that they prevent this double taxation by ensure that savings are only taxed once, either when they are earned (in the case of TFSAs) or when they are consumed (in the case of RRSPs). In effect, they remove one of the layers of tax from r.

[Bob Smith]

"Today's savings MAY (and often is) tomorrow's income, a percentage of which may be saved and the rest consumed; there is a fresh choice each time period. So there is no "double-taxation"."

See my prior post. In any event, last I heard, people don't live foreover (I'd like to be proven wrong on that point) and when they die, everything is taxed. Outside of RRSPs and TFSAs there is always double-taxation.

"What a VAT does is allow untaxed compounding on a very broad basis. A VAT is an excise tax and whenever a country relies on excise taxes it faces Pitt's Dilemma if they are relied upon too much"

Except modern VATs, unlike Pitt's excise taxes, are imposed on pretty much everything. You can't avoid the tax by shifting your consumption patterns. You want to buy wigs? Taxable. Hair powder? Taxable. Ribbons? Taxable. Have the barber shave it off, or give you a mohawk? Taxable. You want to buy brandy? Taxable. Gin? Taxable. Beer? Also taxable. Granted, we have some exceptions in Canada's GST/HST, notably basic groceries, federally regulated drugs, catheters, etc. although they tend to be for goods/services that people aren't likely to shift consumption to (there's only so much milk you can drink in a year and not many people voluntarily consume catheters - I hope).

True, you can avoid taxes today by saving for tommorow, but you can't avoid the tax because the wigs, brandy, and everything else, will be taxable tommorow. Unlike the income tax, which distorts the choice between consumption today and consumption tommorow (see above), the VAT is neutral, it taxes you at the same rate regardless of when you choose to consume your income.

"No, there isn't. Been to Convention, seen it all up close."

And yet, when faced with tough choices the NDP government in Manitoba raises the PST, which will actually raise money, while the federal NDP runs on BS policies of taxing corporations and "the rich" - which will generate sweet f-all in revenue - and their Ontario breathren ask for tax increases that - plausibly - could cost the province revenue (See Kevin Milligan's calculations: https://docs.google.com/spreadsheet/ccc?key=0App-Y0SS83SndDlTYlRKY01qVTBONE9sSlB5UFp0OUE#gid=0. As a proposal, it at least has the merits of creating work for tax lawyers). Same party, different realities.

[Bob Smith]

"Any time a tax allows a choice to escape it (investment, in the case of VATs) it risks narrowing the tax base. An income tax is not so easy to escape, that's why it's such a "burden", all taxes are a burden if they are to be actually paid."

As I noted above, a VAT doesn't allow people to "escape" taxes by saving (investing), since those savings (and any income they earn) will be taxed when they are ultimately consumed. A VAT does prevent such savings from being taxed twice (as does an RRSP). The income tax is a burden not because it it taxes investment, but because it double-taxes investment (or, I suppose, more accurately, the returns on investment income). If you're worried about people having a vast store of untaxed wealth that never gets taxed, hey, adopt my consumption tax proposal of having an unlimited RRSP - when you die, your RRSP gets taxed. Once.

[Frank Restly]

Bob,

Most tax codes allow for capital losses to be realized (deducted from taxable income). Hence your r is not always positive and the double taxation becomes a redistribution means - from those who incur capital gains to those who incur capital losses.

[rsj]

Sure, today's consumption is tomorrow's investment, and tomorrow's consumption is today's investment. You can draw these analogies if you want. If I tax consumption, that will discourage the purchase of consumption, which will discourage the production of consumption, which will discourage investment.

At least we see now that you need a general equilibrium analysis, instead of picking two random points in the circle of cause and effect and marking them special.

The advantage of taxing capital over consumption is that while no one will increase their demand for consumption if sales taxes go up, the demand for savings may well go up if capital income is taxed -- this depends on whether the wealth effect outweighs the income effect. But for consumption, you have no correspondance to the wealth effect. So this is really an empirical question -- how sensitive is investment to broad capital income taxes, which is equivalent to asking how sensitive are savings demands to the interest rate. The answer is not very much. But if savings demands are less responsive to changes in the interest rate than consumption demands to an increase in the price of consumption, you would want to tax capital income more than consumption.

In terms of "escaping" paying capital income taxes, investment occurs as a result of arbitrage. There is no escaping the desire for free money. If the CB sets the policy rate at 1%, and you see an investment opportunity that yields 2%, you will invest -- borrowing from the CB and purchasing the asset yielding 2%, and you will make this investment regardless of whether 50% or 90% of your capital income is taxed. 50% of free money is still better than nothing. So investment will still occur -- the only question is whether the CB will need to hike rates as a result of the capital income tax in order to avoid inflation due to reduced *overall* savings demands at the after-tax yield. And again, that depends very much on whether wealth effects outweigh the income effects, as well as depending on things like distribution of wealth, the social safety net, etc. Again, it's an empirical question, not something to be reasoned out from pure philosophy.

In terms of your fixation on the number "1", I am assuming that if I levy a tax of $5 on you in the morning and $5 in the evening, then you will be obsessed about being taxed twice and will therefore prefer to pay a $10 tax in the morning. I see nothing special about two levies rather than a single levy, and am genuinely amused that you seem to think the latter is always better than the former.

[Giovanni]

Bob Smith writes:

"The brilliance of RRSPs and TFSAs is that they prevent this double taxation by ensure that savings are only taxed once, either when they are earned (in the case of TFSAs) or when they are consumed (in the case of RRSPs). In effect, they remove one of the layers of tax from r."

Disagree. The brilliance of RRSPs and TFSAs, if such exists, resides in the behavioural and distributional effects they produce by sheltering returns to household financial investment from taxation. It is not a matter of "double taxation" as such. Imagine a world in which we taxed only consumption and household investment income, but not labour income. One might still be able to make a sound efficiency/equity case for removing the investment tax, even though the investments in question had all been made out of previously untaxed labour income. On the other hand, if it turned out TFSAs did nothing to alter savings behaviour while merely providing a windfall to high-income households, then one might be able to make a case for a little more "double-taxation" by zapping TFSAs.

[rsj]

And by the above argument, if we are in a depressed economy, and there is consensus that the CB needs to credibly promise more inflation, then the best situation would be eliminate all consumption taxes and tax only capital. As we know from arbitrage that every investment earning a return in excess of the policy return will continue to be funded, whatever the capital income tax rate happens to be, we also know that if we raise taxes high enough on capital income, at some point savings demand will decrease and the (post-tax) natural rate will drift above zero.

Instead of trying to discourage aggregate demand by taxing consumption or labor, we should be encouraging aggregate demand by successively hiking capital income taxes while the CB keeps the policy rate fixed at zero.

[Frank Restly]

Giovani and Bob,

I presume that capital losses on holdings of TFSA's are not tax deductible? Meaning that selling at a loss generates no reduction in tax burden. This works well at near 0% interest rates, not so well with higher nominal interest rates.

Part of the reason for double taxation is that it allows capital losses to be deducted from taxable income. In essence it is a form of income smoothing and reduces market volatility.

[Determinant]

I presume that capital losses on holdings of TFSA's are not tax deductible? Meaning that selling at a loss generates no reduction in tax burden. This works well at near 0% interest rates, not so well with higher nominal interest rates.

Correct, capital losses in both RRSP's and TFSA's in non-deductible.

[Bob Smith]

Frank:Most tax codes allow for capital losses to be realized (deducted from taxable income). Hence your r is not always positive and the double taxation becomes a redistribution means - from those who incur capital gains to those who incur capital losses.

Two points. First, and its a secondary point, capital gains is only one form that r can take (dividends, interest, etc. being other forms). So even if I accept your proposition, it doesn't apply generally.

Second, and more importantly, ACTUAL r can be positive or negative. But when we're talking about capital gains people don't know ACTUAL r when they're making their investment decision (outside of the odd insider trader). Rather, people are making their investments on the basis of EXPECTED r. And since people alway have the option of "investing" in a sock under their mattress, they're not likely to invest in a risky asset where EXPECTED r is less than zero. So, directly at least, the deductibility of capital losses in the event you lose money doesn't directly affect your investment decision.

Now, there is some literature (and its been a decade, so I can't remember by whom, but they were prominent - diamond, saez?) which suggested that loss deductibility reduces the expected volatility of r, since the government is taking a cut of both the upside and of any loss. So to the extent that people are risk adverse, loss deductibility might increase investment. I think the caveat to that is that its unlikely (but I could be proven wrong) that that effect dominates the reduction in EXPECTED r caused by the imposition of a tax. Moreover, in practice, loss deductibility of capital losses are often sharply constrained (in Canada, for example, you can only deduct them against other capital gains, meaning that capital losses are often of limited use). So I think my general point holds.

[Bob Smith]

Giovanni: "On the other hand, if it turned out TFSAs did nothing to alter savings behaviour while merely providing a windfall to high-income households, then one might be able to make a case for a little more "double-taxation" by zapping TFSAs."

Sure, if the investment/consumption decision is entirely insentive to price (i.e., after-tax return on investment) then I don't disagree that taxing investment would be great tax policy (for the same reason that taxing other price-insensitive commodities - cigarettes, for example, is a relatively efficient way to raise money). Taxing a perfectly inelastic good or commodity is restritributive, but not inefficient.

But that's one hell of a proposition, and to be accepted, I think you'd have to put some evidence on the table. For a somewhat older summary of the abundant evidence to the contrary see http://www.unescap.org/drpad/publication/journal_8_1/PETER.PDF

[Bob Smith]

"Correct, capital losses in both RRSP's and TFSA's in non-deductible"

That's not right. RRSP losses aren't deductible against your non-registered losses (i.e., losses ouside your RRSP - the RRSP trust is a separate taxpayer), but they are certainly deductible against gains in your RRSP. Think about it. If you realize gains and losses inside your RRSP, do you (i) pay taxes only on the gains when you collapse your RRSP, or (ii) pay taxes on the amount actually in your RRSP (the gains net of the losses)? Losses in your RRSP are deductible against gains in your RRSP, because you only pay tax on the net amount in your RRSP when you take it out.

Neither gains or losses are taxable/deductible in your TFSA, because you pre-paid your tax on the funds in your TFSA (i.e., it is paid out of after-tax funds). But the after-tax return on the TFSA is, all else being equal, the same as that for the RRSP.

Frank, note that the deductibility of losses isn't a "benefit", it simply involves the government treating gains and losses consistently (i.e., taxing net gain). Income smoothing and reducued volatility is one of the results of taxing gains/deducting losses (although only if the losses can be applied against other gains and/or are refundable by the government, which as discussed above, they often aren't). But you don't need double taxation to see those effects - the same effects would be present if investments were only taxed once (as the RRSP example makes clear). Go back to my original equation and take out the second level of taxation, think the results you identify cease to occur? Those are arguments for taxing investment income, not for double-taxing investment income.

[Frank Restly]

Bob,

"But you don't need double taxation to see those effects - the same effects would be present if investments were only taxed once (as the RRSP example makes clear)."

If you invest in an RRSP and that RRSP is sold at a loss, you never paid any taxes on the income that you used to fund that RRSP. Hence you go from double taxation to zero taxation. That creates a perverse incentive to pour money into enterprises that have no hope of ever offering a positive real return on investment. See offshore shell companies.

[Bob Smith]

Frank,

First, avoiding tax by losing money isn't much of a tax-avoidance strategy. And offshore shell companies typically have significant returns - they don't pay taxes after-all - people just don't report it.

Second, the example you give is no different from paying tax on the original income, then claiming an offsetting deduction for the loss. And intuitively, in the scenario, the person SHOULDN'T pay any income tax because his net income is nil.

[Frank Restly]

Bob,

"First, avoiding tax by losing money isn't much of a tax-avoidance strategy."

That would depend on how progressive / flat the tax rate structure is and how much control or foresight you have on the amount of money you lose.

[Bob Smith]

Frank, think about that for a second. At any tax rate less than 100% the cost of the loss exceeds the tax saved. Typically, the aim of tax-avoidance strategies is to save you money. At a 99% tax rate, does it make sense to lose $100 to save $99 in tax? Really? At tax rates less than 100%, losing money is NEVER a good tax-avoidance strategy. And, I'm willing to bet if investors had any control/foresight over the amount of money they lose, they wouldn't lose it, losing money being generally a bad investment strategy.