The case for taxing basic groceries

Economists frequently argue that taxing basic groceries is a good idea - for example, see these papers/posts making the case for taxing food in the US,  Canada, and New Zealand. 

The equity argument for taxing groceries is straightforward. Suppose everyone spends $500 a month on groceries. If groceries were taxed at 10 percent, everyone would pay about $50 in tax (or slightly less, if people cut back on their food expenditures when the tax is introduced). If part of the revenue raised by taxing groceries was used to give every low income individual a $60 tax credit, the tax on groceries would actually increase the well-being of the worst off members of society. Any additional revenues raised could be used either to decrease other taxes, leading to greater economic efficiency, or to provide needed social or infrastructure programs, further enhancing efficiency and/or equity.

This equity argument is the one emphasized in, for example, Michael Smart's recent paper. Yet it begs the question: why groceries? Why not increase the top rate of income tax instead, or raise capital gains taxes?

The heart of the economic argument for taxing basic groceries is that it increases economic efficiency.

There's a short version of the efficiency argument and a long version. The short version goes something like this: "If groceries aren't taxed, but other goods are, then people's choices will be distorted. They will substitute groceries for other goods, hence the economy will devote too many resources to producing food and too few to producing other goods. Economic efficiency will be compromised."

The short explanation is, on the surface, clear enough, but it begs as many questions as it answers. How do people substitute groceries for other goods? What does "too many" or "too few" resources mean? What is economic efficiency anyways?

The standard undergraduate way of answering this question is with budget constraints and indifference curves. A budget constraint shows all of the different combinations of food and, say, clothing that a person can afford to buy. For example, if a person has an income of $200 a week, meat pies cost $10 each, and t-shirts cost $10 each, the person can afford 20 meat pies, or 20 t-shirts, or any other combination that costs $200, such as 10 meat pies and 10 t-shirts, as shown in the figure below.

No matter how constrained our circumstances, we always have choices. Economists assume that people choose the one point on their budget constraint that makes them happiest. In the picture below, that consumption bundle is shown as the "optimal choice" of 8 t-shirts and 12 meat pies. 

Under Canada's current sales tax regime, t-shirts are taxed, but not food. The tax on t-shirts changes the consumer's budget constraint, as shown in the next diagram. To exaggerate the effect of taxation, I've shown a 100% tax on t-shirts - one that raises the price of t-shirts from $10 each to $20 each. Now, if a person had $200 and spent their entire income on t-shirts, she would only be able to afford 10 t-shirts. Hence the budget constraint rotates as shown.

The consumer changes her consumption decisions in response to the tax. In this example, I've shown the consumer reducing her purchases of both goods, so she buys 5 t-shirts and 10 meat pies after the tax is imposed. The revenue from the tax is the amount of meat pies it takes away from the consumer. Prior to the tax, if the consumer bought 5 t-shirts, she could afford to buy 15 meat pies as well. After the tax, if she buys 5 t-shirts, she can only afford 10 meat pies. The revenue raised by the tax is thus 15 meat pies - 10 meat pies = 5 meat pies (or $50, since 1 meat pie costs $10). I've shown the revenue raised by the tax on the vertical axis, but I could also have shown it as the vertical distance between the new, after-tax choice and the original budget constraint.  (Tax revenue in terms of t-shirts is measured horizontally.)

There are two ways to use this diagram to show that a tax on both meat pies and t-shirts is more efficient than a tax on t-shirts alone. One is to compare the tax on t-shirts with a tax on both t-shirts and meat pies that raises the same amount of tax revenue, and demonstrate that the later makes consumers better off. Revenue-neutrality could be achieved by lowering the tax on t-shirts when the tax on food is introduced, or by increasing refundable tax credits to compensate consumers for having to pay tax on food.

A revenue-neutral switch to a tax on both goods is shown in the figure below. It intersects the "Choice (t-shirt tax)" point because because raises the same amount of revenue as the tax on t-shirts only - taking the equivalent of 5 meat pies away from the consumer.  It is parallel to the original budget constraint because the it reflects the same relative prices. The slope of the budget line shows number of meat pies a person can get in exchange for one t-shirt. It doesn't matter whether t-shirts and meat pies are both $10 each, or if there is a 33.33 percent tax raising the price to $13.33 each, a person can still trade one t-shirt for one meat pie.

The consumer's consumption choice after a revenue-neutral tax change is shown on the diagram below as "Choice (tax on both)." It has to be to the right of the original choice - when food is taxed, it becomes relatively more expensive and so, all else being equal, people will buy less food and spend more on other goods.

The crucial point is that the consumer is better off at the "tax on both" point than the "t-shirt tax" point, even though the revenue raised in the same in both cases. Take a look at the new, tax on both goods, budget constraint. That budget constraint would have allowed him to buy the bundle of goods marked "choice (t-shirt tax)", but he choose the bundle marked "Choice (tax on both)" instead. Therefore it must be a better bundle; it must make him happier than he was when he was consuming the t-shirt tax consumption choice.

Intuitively, the person is happier when both goods are taxed because he no longer has to modify his behaviour, buying t-shirts instead of meat pies, in order to avoid paying so much in taxes. That is the reason why economists think that taxing basic groceries is efficient - a tax on all goods will make consumer better off than a tax on just one good when the two taxes raise the same amount of revenue.

I have made the argument so far in terms of "revealed preference." Another way of making the same point is with indifference curves. Economists assume that everyone has an infinite number of indifference curves; these curves are like contour lines, tracing out a map of people's preferences. All points on a given indifference curve give the consumer the same amount of happiness; indifference curves further away from the origin - ones representing consumption bundles containing more stuff - represent higher levels of happiness. A person's optimal choice is the one that places him on the highest possible indifference curve, given his budget constraint. I have added indifference curves into the diagram here:

The consumer is on a higher indifference curve with the tax on both goods, than with the tax on just t-shirts, even though the revenue raised by the two taxes is the same, therefore he must be better off.

This is not, however, the only way of showing the efficiency of taxing basic groceries. Another way of showing the efficiency of taxing groceries is to show that a tax on both food and clothing raises more revenue than a tax on clothing only holding the consumer's level of well-being constant.

When there is a tax on t-shirts only, the consumer makes some consumption choice, shown as point B in the diagram below. This choice puts her on the red indifference curve. To keep the consumer's level of well-being constant, we construct a tax on both food and groceries that leaves her on that same, red, indifference curve. Again, we could do this by introducing a refundable tax credit just sufficient to compensate for the additional sales tax paid, or by lowering the rate of tax on t-shirts when the food tax is introduced.

This new tax-on-both-goods budget constraint is parallel to the original budget constraint, because it represents the same relative prices, and is just tangent to the new indifference curve, because it represents the same level of well-being.

With the tax on both goods, the consumer makes a new consumption choice, say point C. The movement from point B to C is called a "substitution effect"; the change in consumption choices associated with a change in prices holding well-being (or real income) constant. 

The crucial point is that even though the consumer is indifferent between point B and point C, point C raises more tax revenue! The revenue raised by a tax is the vertical distance between the consumer's choice and the original budget constraint. The revenue raised by the tax on t-shirts is shown by distance AB (5 meat pies). The revenue raised by the tax on both goods is distance CD (equivalently, distance AE, or 5.5 meat pies). The difference between the two, distance BE, is the efficiency gains from taxing groceries.

When the case for taxing basic groceries is presented in these simple terms, the assumptions underlying the argument become apparent. The equation of choice with happiness rules out any paternalist arguments for exempting basic groceries from taxation. For example, at present soft drinks are subject to sales tax, but milk is not. A tax on milk would be expected to decrease milk consumption and increase soft drink consumption, all else being equal. Those who would argue for taxing basic groceries would respond in one of two ways: first, that we should respect people's choices whatever they are; second, so many of the basic goods exempted from sales tax at present are teeth rotting, IQ-lowering sugary junk anyways that the paternalist argument has little force.

Note that the advocates of taxing food cannot respond that substitution between milk and soft drinks is unlikely to happen. The efficiency gains from taxing food derive solely from the substitution effects, from people responding to price changes. It is contradictory to argue that "people won't change their shopping patterns when food is taxed" and, at the same time claim that there are efficiency gains from taxing basic groceries.

There are other assumptions underpinning the arguments for taxing groceries. They generally assume, for example, that markets are basically competitive, and goods are priced at marginal cost, so the primary impact of a tax increase is felt in the form of higher prices to consumers. It assumes, as just noted, that people substitute in a non-trivial way between basic groceries and other goods, such as restaurant food. Also, it assumes that the labour supply distortions created by giving people a refundable tax credit to compensate them for paying sales tax on groceries, and then taking that tax credit away again, are relatively unimportant. 

Michael Smart's paper setting out the case for taxing basic groceries in Canada acknowledges these and other difficulties:

It is known that uniform taxation of all commodities is optimal only if all commodities have the same degree of substitutability with labour/leisure in preferences. [see the Ramsey rule] In highly stylized economic models, it can be shown that it is desirable to depart from uniformity by taxing more than average a commodity (like ski equipment) that is complementary with leisure. Nevertheless, a full, formal description of the optimal tax structure has eluded economists. An alternative approach is to study tax reforms that are desirable (though not necessarily optimal). Smart, building on the tax reform model of Ahmad and Stern, showed that a reform moving towards greater uniformity of tax rates increases economic welfare when consumers’ ability to substitute between taxed commodities is large, and ability to substitute between taxed commodities and leisure is small. The intuition is that taxing close substitutes at different rates creates tax avoidance opportunities that reduce government revenue and increase economic distortions of the tax system.

In spite of the theoretical ambiguities, for most commodities, uniform taxation is a reasonable benchmark for an efficient tax system.

Even the latest high-tech models of commodity taxation produce the same basic results as the simple indifference-curve budget constraint analysis, and for the same basic reason: the efficiency cost of taxation arise because consumers substitute one good for another in order to avoid paying so much in taxes, the most efficient tax system is the one that minimizes these substitution effects.

Comments

[nemi]

@ Frances the economist
If there are many persons, instead of one (representative?) person, some will obviously pay more and other less after a revenue neutral change in taxes.

@ Frances the human
Increased “welfare” payments to the poor would be an easier target for the right than to low taxes on groceries.

[Stephen Gordon]

I love these kinds of posts.

[Frances Woolley]

Nemi: "If there are many persons, instead of one (representative?) person, some will obviously pay more and other less after a revenue neutral change in taxes."

Generally expenditure on food is an increasing function of income. Taxing food and introducing a refundable "food tax credit" could be made revenue neutral for the least well off Canadians. Sure, wealthy people who run up vast grocery bills buying organic avocadoes at $2 each and imported stilton at $30/kg will end up paying more in taxes, but from an equity point of view, this is all to the good, isn't it?

Nemi: "Increased “welfare” payments to the poor would be an easier target for the right than to low taxes on groceries."

Yup, that's the classic political economy argument for universal benefits - and there's a lot in it. In the recent debate over cuts to pensions in Canada, for example, how much press did the cuts to Guaranteed Income Supplement (for low income seniors) get, as compared to the cuts to Old Age Security (a close to universal program.)?

Stephen - thanks!

[Ryan]

Just a few random thoughts according to my crazy world-view...

1 - The above discussion assumes that "more tax revenue = better." Why is that a valid assumption?

2 - The above diagrams assume perfectly continuous preferences, i.e. it is possible to buy 1.36 t-shirts, 3.89 4-liter containers of milk, etc. This is a simplifying assumption, but one that may not always prove to be realistic. People don't necessarily measure their preferences this way. (Even though I might prefer 1 apple to 2 oranges, that does not necessarily mean I prefer 2 apples to 4 oranges.) The matter gets even more complicated the more goods you take into consideration.

3 - While taxing two goods at the same rate may sometimes be less distorting in terms of relative prices, it need not. There is no reason (that I can think of) to simply assume that a 10% tax on t-shirts impacts me the same way that a 10% tax on milk does. Not only are we talking about different kinds of income (disposable vs. non-disposable), but for reasons outlined in my 2nd point above, there is no guarantee that my price-elasticity of demand is the same for all goods. If it's not, then even though the *tax* is equal, the impact on my preferences will not be.

4 - Is there an economic justification for any tax?

[Frances Woolley]

Ryan: "The above discussion assumes that "more tax revenue = better." Why is that a valid assumption?"

More tax revenue = better *holding the consumer's level of well-being constant*. There is always the possibility that the revenue will be used for something useful.

"The above diagrams assume perfectly continuous preferences"

That's the least heroic of the various assumptions they make!

"While taxing two goods at the same rate may sometimes be less distorting in terms of relative prices, it need not."

Yup, if there's a good with highly inelastic supply or demand, there are efficiency gains to be had from ramping up taxes on these goods. That's why cigarettes, alcohol etc are taxed, and why historically it wasn't uncommon to tax salt in hot countries.

Apart from imposing excise taxes on goods with particularly inelastic demands/supplies, designing a customized tax system where each commodity's tax rate is tailored to its own specific elasticity of demand isn't worth the hassle - the admin costs almost certainly outweigh the efficiency gains.

[DavidN]

Love it. Frances you’re the new Greg Mankiw (before he cut down blogging output).

I agree with Nemi’s second point though, this is an econ. prima facie case for getting rid of exemptions, i.e. ignoring political economy dynamics.

With respect to Nemi’s first point. The tax burden of different individuals will depend on the composition of their consumption bundle. Frances already made the point that you can theoretically raise the welfare of low-income individuals with a tax credit. So if you extend the tax to a good that is a bigger share of consumption for low-income persons than theoretically they can still be better off with a tax credit. On the other hand if you extend the tax to good that is a bigger share of consumption for high-income persons than the tax becomes progressive.

It’s late over here so not thinking straight, but Is there a special case where a low-income/high-income person is worse/better off after extending the tax+tax credit for low-income person if the good taxed is small share of persons consumption bundle?

[Frances Woolley]

Ryan "Is there an economic justification for any tax?"

Equity gains from redistribution (both in kind and in cash) and efficiency gains from provision of public goods and the provision of social insurance, both of which markets fail to provide efficiently due to free riding, moral hazard, adverse selection, etc.

[DavidN]

*Ignore the second part of my last comment. Forgot about elasticity.

[Frances Woolley]

DavidN "It’s late over here so not thinking straight, but Is there a special case where a low-income/high-income person is worse/better off after extending the tax+tax credit for low-income person if the good taxed is small share of persons consumption bundle?"

My guess - and this is where things get interesting - is that young people eat restaurant food and/or food purchased by their parents and old people cook at home. So your typical university student would really benefit from taxing basic groceries plus introducing a tax credit, whereas middle income seniors would lose. I need to crank through the numbers on the Survey of Household Expenditures to confirm this however. This is important because one of the key challenges Canada faces as the baby boomers age is working out how to get older folks to pay tax - this is one possible way of doing it.

[Lord]

When food is exported and clothing imported and currency manipulation is involved, this is not necessarily the case.

[Ryan]

Frances - You said: "More tax revenue = better *holding the consumer's level of well-being constant*. There is always the possibility that the revenue will be used for something useful."

^^ Can this really be true? Jane's total wealth is decreased while her well-being is held constant? Also: Is the possibility that Jane's use of her wealth for something useful less remote than the possibility that the government will do it?

Also not sure I'm convinced of "equity gains from redistribution" or "efficiency gains from the provision of public goods," but probably best if I leave that one alone. ;) Assuming such gains are possible, then okay.

[Bob Smith]

As an aside, there's are some interesting case on some of the incongruities of the "food" exemption from GST/HST. One involved a hunting preserve where the court ruled that 40% of what the hunter paid was for the right to shoot an elk, and was taxable, while the balance was for the elk itself, which was food for human consumption, therefore exempt. Go figure. It also means that there has to be a special exemption from the exemption for food for rabbits (rabbits being the form of "farm livestock" regularly kept as pets)- presumably because otherwise people might go buy their pets on a tax-free basis from rabbit farms.

Now, to some extent the GST/HST system does tax certain "tax-exempt" products like residential rent or financial or medical services, albeit indirectly by denying input tax credit for GST/HST paid by suppliers (in contrast, basic groceries are "zero-rated" so they're nominally subject to tax at a rate pf 0%, allowing their suppliers torecover the GST/HST they pay on inputs). Obviously, taxing the supplier isn't equivalent to taxing the consumer since value-added by the supplier isn't taxed. On the other hand, it avoids the politically unpopular prospect of explicitly taxing "essential" products like housing or medical services - since Joe Q. Voter doesn't see GST/HST on their bill (and nobody cares when landlords or docters complain). Maybe as a part-measure for taxing food without the political hit, we might stop treating it as a "zero-rated" supply and make it a tax-exempt supplies as a way of imposing a partial GST/HST on groceries.

[Frances Woolley]

Ryan: "Can this really be true? Jane's total wealth is decreased while her well-being is held constant?"

Yes. But it's not obvious, you have to really think about it.

Do you know the idea of the Lafffer curve, that when taxes are high enough it's possible to cut taxes and raise tax revenue at the same time? This is a similar kind of concept - people pay more in taxes, but because they aren't devoting so much effort to tax avoidance activities, i.e. buying t-shirts instead of meat-pies, they're better off.

Lord, yes, in the background there are all sorts of assumptions about the way markets operate.

[Frances Woolley]

Bob: "Maybe as a part-measure for taxing food without the political hit, we might stop treating it as a "zero-rated" supply and make it a tax-exempt supplies as a way of imposing a partial GST/HST on groceries."

These are important points. A lot of the press coverage around the Smart paper centered on the factoid "Canada is one of only three OECD countries that don't tax groceries." In fact, Canada is one of only three countries with *zero rated* groceries - in other countries groceries are tax exempt. But you can imagine that not many column inches were devoted to the tedious task of explaining the difference between zero rated and tax exempt groceries!

[Bob Smith]

4 - Is there an economic justification for any tax?

Yes, unless, at the limit, you don't believe in the existence of public goods (justice, national defense, fire protection). Mind you, that's the minimal econimic justification for taxes, one can go beyond that (externalities, equity, market imperfections, etc.)

[edeast]

What good is a tax credit when you don't make the minimum taxable income? And poor people are too busy hustling to be able to keep track of a year's worth of receipts. And cheap clothing isn't taxed, thrift stores, even new shoes aren't taxed below $29. At least they used to be in Ontario not sure with the HST. Getting a tax receipt is a one time income, doesn't help throughout the year when your food costs go up 5:13%. In my town I guess the gst cheques come quarterly so some people could splurge and buy milk, and gas, at least 4 times a year. I can't imagine what they would do with 60 bucks in the spring.

University students do not need help. They've already won the lottery and I'm assuming their IQ is above average.

[Bob Smith]

"These are important points. A lot of the press coverage around the Smart paper centered on the factoid "Canada is one of only three OECD countries that don't tax groceries." In fact, Canada is one of only three countries with *zero rated* groceries - in other countries groceries are tax exempt. But you can imagine that not many column inches were devoted to the tedious task of explaining the difference between zero rated and tax exempt groceries!"

It would be an interesting political battle if the government switched groceries to tax-exempt status (not that I expect that to happen). On the one hand, you might expect left-wing parties/activists to rant and rave about taxing food. On the other hand, the tax is imposed on the supplier, just like the corporate income tax that they're in favour in increasing. Maybe they can explain why they think that corporations will pass along one set of taxes to their customers, but not the other.

[J.V. Dubois]

First, this is a very good blog. I would love to see more taxation/externalities and other micro blogs here on WCI. Even if it may seem to be obvious for micro expert, judging by the level of debate in Macro I would imagine that there will be many people who did not really think these obvious thinks through back in olden days.

As for the topic itself, I think that you kind of cheated. By assuming that really poor people will be exempted from the tax you fundamentally changed the question. Then it is just matter of course that lowering tax rate and broadening tax base has welfare gains. However to get there, you really have to deal with your assumption first. There are still questions like: is the cost of administration of these tax credits higher or lower than deadweight loss from higher taxation of T-Shirts, or if it is even possible for this policy to work in environment most poor and socially marginalized people live in.

[Frances Woolley]

edeast: "What good is a tax credit when you don't make the minimum taxable income? And poor people are too busy hustling to be able to keep track of a year's worth of receipts. And cheap clothing isn't taxed, thrift stores, even new shoes aren't taxed below $29."

A food tax credit would only make sense if it was based on income, not food purchases, like the GST/HST credit.

On thrift store purchases - purchases made at commercial thrift stores, such as Value Village, are taxable (a pet peeve of mine, that I've blogged about before.) Charitable thrift store purchases may not be taxable, but that would be because the thrift store is run by a charity, not because it's a thrift store.

In Ontario the tax exemption for shoes now only applies to children's sizes (less than a size 6), I think all adult clothing is taxed, regardless of price: see http://www.fin.gov.on.ca/en/taxchange/taxable.html and scroll down the page.

[Shangwen]

Frances, does efficiency thus not improve even more if essentially all goods and services are taxed at the same, but much lower rate? I personally do not understand why the services of lawyers, accountants, health care providers, or architects are not taxed. It might be help to stop electing members of those professions to office. More seriously, if monopolistic regulations eliminate or severely restrict your options (e.g., your mechanic will be sued for pulling your teeth), is taxing those services inefficient?

Note that I think that taxing everything at a lower rate is very different from the left-wing "everyone must pay" rhetoric. This is not about forcing participation, but about removing distortions.

[Ryan]

Bob: In my mind there's a difference between economic rationalization and economic justification. I see it as a pretty tall order to actually and effectively argue that society is better off with a tax, as opposed to other methods of public finance (lotteries, user fees, etc.).

Considering that Frances' example above is a specific tax on a specific market, we can sidestep the question of whether it is possible to fund all of Canada's public expenditures through such means. Nonetheless, it is definitely true that voluntary contributions are more utility-maximizing than compulsory ones. Buying an Ontario lottery ticket and actually getting something for my money (even if it's only entertainment) certainly makes me better off and distorts markets less than levying a tax on food. In this way, it can be shown that virtually any single tax is not justifiable if compared to a user fee or game of some sort.

Am I wrong?

[Frances Woolley]

J.V. Dubois -

Thanks for the kind words! I don't do these posts more often because they're too much like hard work, especially those indifference curve/budget constraint diagrams. This one was primarily for my long-neglected ECON 3405 students.

"As for the topic itself, I think that you kind of cheated"

Fair enough. There is still a lot of value, I think, in making assumptions explicit, deconstructing and demystifying the economic arguments that people make. What does it really mean to say something is "more efficient", for example? "More equitable"?

[Frances Woolley]

Shangwen: "Frances, does efficiency thus not improve even more if essentially all goods and services are taxed at the same, but much lower rate?"

Absolutely, this is precisely the argument that's made in Smart's paper. B.t.w. I thought of you last Saturday when I went out to dinner at a Chinese restaurant, and wasn't allowed to order from the Chinese menu.

[K]

Frances: A tax on basic foods, the demand for which is relatively inelastic would likely be far better than taxes on stupid t-shirts. But even better than that would be a tax on land value and hydrocarbons, the supply of which is (at least in the long run) totally inelastic, and which make up a very large fraction of food production costs. A carbon tax would help too. Hand those revenues out as a citizen's dividend and you would have a system that is not only highly efficient buy also extremely fair.

Bob: How do externalities (e.g. stealing from the commons) not fall under the category of "justice?"

[edeast]

I've since calmed down and read the linked papers, and it makes sense. It sure isn't poor people I see at farmer's markets.
Thanks, for responding, on the type of tax.
I'm not a even a leftwinger, so I can imagine how hard it would be politically to tax food. I was speaking from experience working at a gas station growing up. The line about Ontario being able to reduce it's tax rate from 13% to 7% convinced me.

[Frances Woolley]

edeast - personally I don't think the case for taxing basic groceries is a slam-dunk, though I would like to see the definition of what counts as basic groceries narrowed to exclude more empty-calorie food-like substances. Nemi's initial point, that a food tax credit, once in place, is an easy political target, is one worth remembering.

[JC]

"so many of the basic goods exempted from sales tax at present are teeth rotting, IQ-lowering sugary junk anyways"

Frances, my question is tangential, but this caught my eye. Do you know of any studies showing that sugary junk lowers IQ? I did a quick google scholar search and didn't see any. I'm hoping this wasn't just a figure of speech - I'm constantly nagging my husband about buying so many slurpees, and I'm sure this argument would be more important to him than cavities!

[Bob Smith]

Shangwen: "I personally do not understand why the services of lawyers, accountants, health care providers, or architects are not taxed."

Under the GST/HST/QST the services of lawyers, accoutants and architects ARE taxed. The services of health care providers are generally exempt (although, since in practice, many of them bill the government, query whether that makes much of a difference).

K: "How do externalities (e.g. stealing from the commons) not fall under the category of "justice?"".

Sure, that's one way of characterizing it. But, it's only a "justice" issue, if it's "stealing", whereas taxes on externalities could still be justified if one has a right to "take" from the commons

[edeast]

Income based vs income tax is very freakin confusing, just fyi to any politician.

[Bob Smith]

"I see it as a pretty tall order to actually and effectively argue that society is better off with a tax, as opposed to other methods of public finance (lotteries, user fees, etc.)."

It's hardly a tall order at all to make that argument, given that taxes are the means of public finance that civilizations have generally chosen to finance themselves - it reflects the revealed preference for societies for taxes rather than other means of public finance. From a practical perspective, voluntary user fees aren't workable for true "public goods", because everyone uses them whether they pay the fee or not, that's what makes them public goods (and what's an involuntary user fee if not a tax) and lotteries are seldom effective means of raising funds.

And of course, this totally ignores the arguments that public spending other than on public goods can be beneficial.

[Ryan]

Bob, that's not really an economic argument. It's more like saying, "This has happened so much, that it must be what society wants." You could make a similar - and similarly incorrect, in my opinion - argument about war. Just because it has happened frequently through history doesn't make it a revealed preference.

[Frances Woolley]

Ryan "Just because it has happened frequently through history doesn't make it a revealed preference."

Any good economic argument begins with the premise that people aren't stupid. Which creates two challenges: first, explaining why people buy lottery tickets when the expected return is negative; second, explaining why so many non-stupid people throughout history have opted to finance government through taxation rather than lottery tickets or user fees.

[Ryan]

Frances - People buy lottery tickets because there is entertainment value in gambling. That's not stupidity, it's utility.

As to why governments have opted to tax rather than use public finance methods that require the consent of the governed, it's perfectly obvious, and it has nothing to do with utility maximization. Yes, governments can forcibly expropriate more money than they can persuade people to contribute voluntarily. That's all well and good, but to suggest that society is actually better off because governments have throughout history preferred to force people to do things rather than allow them a choice is kind of a strange suggestion to me.

From the sounds of it, the rationale seems to be "Given that the government always does things correctly, efficiently, and fairly, then people will always be better off if everything they buy is taxed." Sure, okay. If we assume that the government is always flawless and benevolent, then sure every tax maximizes utility for all of the state's hapless wards.

But...

[Bob Smith]

I'm not sure your counter-example is all that compelling, people have often been quite eager for war,at least ex ante - that's why it happens so much (and if you think otherwise, look up some of the old newsreel footage from the summer of 1914 or read some of the rhetoric leading up to the US civil war).

In any event, you ignored the real economic argument that I made - i.e., user fees inherently don't work for public goods (otherwise, they wouldn't be public goods), and lotteries aren't particularly effective means of raising public funds.

And as I noted, that the's minimal argument in favour of taxation. On fairly reasonable assumptions about utility, one can make plausible arguments that redistributive taxation is socially welfare improving (i.e., a millionaire won't notice a $10,000 tax, while $10,000 will have an order of magnitude impact on the life of someone living in poverty. There are good arguments for taxation to fund public education (both on redistributive grounds and on the grounds that inefficient capital markets would otherwise result in a suboptimal levelof education - i.e., poor people can't borrow to pay for education even where there are huge private and public returns to that education). I could go on.

[Ryan]

Bob - Because I feel I've hijacked this discussion too much already, I'm going to make this my last reply on the topic, although I will definitely read and consider any reply of yours. :)

1 - If you're suggesting that propaganda and pro-war rhetoric reflects a genuine widespread interest in war, I'm not sure what to say other than I completely disagree. As Ludwig von Mises put it: "Wars of aggression are popular nowadays with those nations which are convinced that only victory and conquest could improve their material well-being. On the other hand the citizens of the na­tions assaulted know very well that they must fight for their own survival. Thus every individual in both camps has a burning inter­est in the outcome of the battles."

2 - User fees can and do work for all sorts of things. It's not as if they are a hypothetical method of public finance. One really great example I can give you is that a hospital I used to live near in Southern Alberta has an annual raffle to raise money for the local emergency helicopter. It's a big success.

3 - Your plausible argument about the millionaire assumes you can step into the mind of every millionaire and make those determinations for each of those individuals for certain. I don't consider that a "fairly reasonable assumption about utility." I believe utility is subjective. For this reason, all of these redistributive arguments fall a little flat for me. Even your public education rationale assumes that every poor person is definitely best served by being placed into the public education system. True of many, yes, but true for all? Can you *really* make that judgement call for every single human being?

Food for thought, anyway. Good discussion.

[Chris S]

(Re-Reposted - first time was reported as OK, but didn't show up. Left it an hour before reposting, then another two and half before rereposting.)

It *is* interesting, but it makes me wonder if we would be better off if we continued to call this "political economy" to recognize the scope of what we should be considering. We are missing out on many factors that strike me as outside what this model considers, but which are absolutely necessary to consider.

I see two rather fundamental problems. The first wasn't really the topic of this post, but it was clearly mentioned at the top - that the extra tax can be redistributed to the lowest income members. This treats the redistribution problem as solved (or at least as an externality). That to me is a stretch - it's not clear that we know how to do the redistribution well. Moreover, even if we did know that, that redistribution system will have costs, which must be deducted from the tax revenue. It's not enought to say "same amount of tax with improved efficiency" if you are simultaneously ignoring costs which must come out of that tax. If the redistribution costs too much, it wipes out your gains, and the original "no tax on food" would be a more efficient solution.

The second problem goes to the nature of food. It's handily noted that the tax difference between "tshirt tax" and "both tax" can be described in terms of meat pies. So - taxing both reduced the number of meat pies purchased.

But an insufficient number of meat pies can be an existential risk to an individual. Ten might pies might be enough to survive, while 9.5 meat pies might be malnutrition. The idea that it is ok to reduce the number of items purchased by introducing a tax is an untested assumption. The UofC paper makes what I consider this same mistake -- it states that the ideal VAT is uniform, and apparently expects that this assumption should go unchallenged. The meat pie example is doubly useful, as it provides a key way to point out that there are economic costs associated with NOT purchasing meat pies.

If you want to try and capture this in the model, examine the opportunity cost of the marginal meat pie. It might be an equal trade between one meat pie and one tshirt, but the opportunity costs may be radically different. Maybe I have to stitch a seam in an older tshirt - a cost that changes only slightly the value of the shirt I did not buy. But I can't patch an old meat-pie. If that marginal meat pie is the energy I personally need to do my job and stay employed, then the opportunity cost of that meat pie is enormous. It's so large, that the preference curves would all have a lower limit -- no fewer than ten meat pies, no matter how many tshirts you have to give up.

What would happen if the preference curves all indicated that any solution must involve at least ten meat pies? As I understand this, the value of the tax on meat pies comes off the purchase of t-shirts. In the system we have described, I wouldn't be surprised if the equilibrium point for purchases ended up at the same point as "tshirt tax".

But at that point, we now have a different problem. We are taxing two things, and somehow, we are paying for two tax systems, even though in this case we could get exactly the same outcome and tax revenue by taxing only the shirts. Tax systems aren't free, no matter what our model contains.

So - in some tshirt-meatpie systems, it will be more efficient to tax only tshirts.

The core problem I see to both of the above items is that there is a suggestion we base a real-world decision on a strictly theoretical model. The model is assuming that the external factors will be small and will not influence the outcome. My core argument is that you don't get a free pass on that. I can agree that theoretically this works, but now you need to go improve the model to account for the factors and costs we find in the real world, and then see if the proposed change is still an efficiency improvement.

[Patrick]

OT nit-pick: 'begs the question' vs. 'raises the question'.

[Patrick]

I wonder what would end-up being more of an administrative burden: tax all consumption and do rebates for (e.g.) those with low incomes, or the current system of trying to figure out what goods, and how. My gut says writing the checks would be cheaper and simpler.

[nemi]

Frances: “but from an equity point of view, this is all to the good, isn't it”.
Sure, for a “human” (or at least for someone with the characteristics we usually associate with human nature), that’s why I addressed it to the “economist”.

PS 1: I also really liked the post. Why isn´t there more blogs about micro economics (freakonomics style blogs do not cont).

PS 2: Do you know what Nicks argument is in the last posts about endogenous money? If you do, maybe it could clarify things if someone else described the same thing with their own words.

[Bob Smith]

As Ludwig von Mises put it: "Wars of aggression are popular nowadays with those nations which are convinced that only victory and conquest could improve their material well-being. On the other hand the citizens of the na­tions assaulted know very well that they must fight for their own survival. Thus every individual in both camps has a burning inter­est in the outcome of the battles."

That assumes that perceptions of war have to be symetrical - i.e. that you can't have two sides believing that war will be welfare improving for them. That's demonstrably wrong (and WWI and the US civil war are examples of that sentiment - both sides of those wars thought each would be short and that they would win), as were old-school European (and to be fair, everywhere else) wars for territory, power and prestige.

"User fees can and do work for all sorts of things. It's not as if they are a hypothetical method of public finance."

They work for all sorts of things, except public goods (http://en.wikipedia.org/wiki/Public_good). And note, I didn't suggest that lotteries can't be used to raise public funds (they clearly can), they're just not a particularly effective way of doing so.

"Even your public education rationale assumes that every poor person is definitely best served by being placed into the public education system. True of many, yes, but true for all? Can you *really* make that judgement call for every single human being?"
That's a strawman,I make no such assumption. Public subsidies for education don't have to be welfare improving for everyone in order to be socially beneficial, they just have to be welfare improving for enough people for the gains to outweigh the losses. Frankly, I have no trouble making that assertion.

[edeast]

Peace isn't inherently good. My paraphrase of Richard Feynman, somewhere or other.

On topic. If you made the gst rebate monthly/biweekly, it might work, keep the same name don't call it welfare. But it is an easy target.

[Bob Smith]

"I wonder what would end-up being more of an administrative burden: tax all consumption and do rebates for (e.g.) those with low incomes, or the current system of trying to figure out what goods, and how. My gut says writing the checks would be cheaper and simpler."

In some sense, that's where we've gone with our "income" tax system. For maybe 90-95% of us (i.e., wage earners whose investment income can be sheltered in TFSAs/RRSPs), it is effectively a graduated consumption tax (or, put your way, a 46% consumption tax (in Ontario) with rebates of varrying degrees for people making less than ~$130K). At least in terms of consumption it's relatively non-distorting (although there are some preferences for consumption of health care, child care, children's fitness, arts and crafts, etc.).

[Frances Woolley]

edeast - "if you made the gst rebate monthly/biweekly" - this issue of infrequent big v. frequent small payments is a fascinating one, it would be a good topic for another blog post.

nemi: "PS 2: Do you know what Nicks argument is in the last posts about endogenous money?"

Sorry, I haven't read it yet. Did you see that Paul Krugman has been reading the comments again?

[Bob Smith]

edeast: "On topic. If you made the gst rebate monthly/biweekly, it might work, keep the same name don't call it welfare. But it is an easy target."

Frances: "this issue of infrequent big v. frequent small payments is a fascinating one, it would be a good topic for another blog post"

Interestingly, the Ontario government recently changes an Ontario tax credit that, until recently, was paid out annually (with a tax return) to a quarterly payment. People who qualify for that credit are apparently quite miffed because they preferred a big lump sum of money once a year rather than quarterly, often on the quite reasonable ground they have big "lumpy" expenses - dentists, car repairs, home repairs, etc.

I'm not sure anyone ever got a straight answer out of the Liberals what the thinking was. There was some suggestion that it was a scheme to save cashflow for the province, though if so that would be pretty rinky-dink. But the only alternative rationale is a paternalistic one - that the McGuinty Liberals think recipients will blow it all on lottery tickets and smokes if they pay them all at once (is it cynical to think that when Ontario's running a $16 billion deficit the government should want people spending their money on smokes and lottery tickets?).

[Seamus Hogan]

Francis,

Your link for New Zealand posts draws a blank. You can try

http://offsettingbehaviour.blogspot.ca/2011/08/case-against-dirty-gst-illusory.html

Or, to link to a number of our GST-related posts, many making reference to WCI, try

http://offsettingbehaviour.blogspot.ca/search?q=gst.

[K]

Frances: "Any good economic argument begins with the premise that people aren't stupid."

So the only valid model is a ratex model? Why do we have to play this charade?

People do stuff that's not in their own self-interest all the time. Yes, if you take revealed preference to its logical extreme, then some people prefer to be drug addicts. If you define it not to be stupid to go down a road despite clear and present evidence that it's likely to lead to ruin, then fine, we now have a new definition of "not stupid." And some economists choose to go that route. But does that honestly get us anywhere? What if some people, despite all effort to determine a rational cause of their behaviour, really do turn out to be stupid (conventionally defined)? Why do you insist on ascribing to them rational motives which, if asked, they can't even begin to articulate?

There's a whole continuum of intelligence from the amoeba to the most brilliant human, and there's no sharp cut-off below the dumbest human. At what threshold do we suddenly stop assuming infinite computing ability? In reality, I prefer to think of us all as just different grades of stupid. I do all kinds of stuff that is not in the least utility optimizing. So does everyone else I know. Most of them even know they do it. Why pretend otherwise?

[K]

Bob: "But, it's only a "justice" issue, if it's "stealing", whereas taxes on externalities could still be justified if one has a right to "take" from the commons"

What I meant was that you can't have a "right" to take from the commons without compensation (unless it's a public good, i.e. non-rival). If you "take" from a common limited resource without compensation, then that is morally equivalent to stealing. All external *costs* are therefore an issue of justice.

[Eric Crampton]

Hi Frances,

The search you're linking for Offsetting on New Zealand won't pull up much. Try instead our GST tag, here.

Thanks for the link!

[Frances Woolley]

Eric, Seamus, thanks - I knew I'd read something on Offsetting about this, and that was the best link I could find.

K "Frances: "Any good economic argument begins with the premise that people aren't stupid." So the only valid model is a ratex model? Why do we have to play this charade?

No, there's a world of difference between assuming that on average, most of the time, people avoid doing really stupid things, and assuming rational expectations.

Let's take, e.g., the example that you use, drug addiction. One explanation of drug addiction is that people get addicted to drugs because they're stupid. That's a really boring uninteresting explanation, and yields no useful policy implications. Another explanation - the one that I find much more compelling - is that a lot of drug addiction is a form of self-medication, a way of masking or hiding the pain associated with mental illness, and escaping the hardship and drabness of everyday life. I find this explanation, which starts with the premise that people aren't entirely totally stupid, much more interesting.

[Bob Smith]

"What I meant was that you can't have a "right" to take from the commons without compensation (unless it's a public good, i.e. non-rival)."

Sure you can. The "tragedy of the commons" refers to the historical rights of peasants to use common lands for grazing, and the alleged adverse impact it had on productivity. Three centuries ago the notion of taking away such rights was seen (with considerable fairness) as a significant injustice (and a source of a great deal of early marxist and neo-marxist writing). Maybe one "shouldn't" have right to do that, but there's no shortage of either current or historical examples of people having such rights. The justice of it all depends who has the "rights" to the commons.

[edeast]

All I have is anecdotes. And I was about to relate them when I realized holy shit, these online merchants and ad networks must know a ton about people, if a corner store kid knows personal stuff.

Anyway I don't know where you'd get data for the frequent/infrequent difference or if there is one, some people didn't have bank accounts.

I suppose you could argue it's like the patient kids in kindergarten. The more patient end up more successful. But I don't know.

[Bob Smith]

"I find this explanation, which starts with the premise that people aren't entirely totally stupid, much more interesting."

Not only is it more interesting, it's also closer to reality. Drug addicts, to use K's example, aren't stupid when it comes to drugs. Quite the contrary, they know a heck of a lot more about the consequences than the rest of us (Often even before they become addicts they will have a significant knowledge of the realities of drug use. Anyone under the illusion that your average 9-year old in a US ghetto or on one of our aboriginal reserves knows a lot more about drugs and the consequences of taking them than (i)any of us, or (ii) by any rights, they should?).

And you're right, the "not stupid" assumption is much weaker than the "rational expectation" assumption. Frankly, it's hard to take serious any model that's premised on the assumption that people are stupid, because at that point it ceases to be model and become a projection of the modellers ideological beliefs (i.e., if people weren't so stupid (premise), they wouldn't do drugs(belief), ergo, we should ban drugs (policy based solely on belief)).

[edeast]

With the kindergarten comment, I don't want to argue paternalism. It would be hard for anyone to make a lump sum last a year, when your cost of groceries increases 10%. It should show up in a model, just like not having a bank account is rational due to the fees.

[DavidN]

edeast,

But that’s the beauty of it, you can use extension of GST as a bargaining chip to cut income tax rates for low-income earners or increase income tax-free threshold. But of course, you can’t guarantee that tax cuts will be permanent or just bought forward, but then you can argue that for any policy.

[Frances Woolley]

DavidN: " you can use extension of GST as a bargaining chip to cut income tax rates for low-income earners or increase income tax-free threshold."

That really wouldn't do a lot to help many low income households, because they don't pay income tax anyways.

An increase in the income tax-free threshold of, say, $1,000 federally would provide a $150 benefit to everyone who pays a positive amount in income tax - so it's somewhat progressive, but it's very far from being the most progressive change that could be made in the tax/benefit system.

[Bob Smith]

DavidN,

I also wonder about the merits of (effectively) increasing the GST (by taxing food) and using the revenue to either cut a tax that is, at best, a consumption tax.

Think of it this way, if you cut the bottom tax rate (or increase the threshold) it is a marginal rate cut for people who save very little (i.e., people in the bottom tax bracket) or a lump sum tax rebate for the wealthier (I.e., people in a higher tax bracket). In the first case, you're just increasing one consumption tax (since for the poor income = consumption) and reducing another and in the second case you're just giving a lump sum tax rate (which is probably the worst form of tax policy - akin to a permanent commitment to Klein bucks).

That's the ironic point about the (very fair) Liberal criticism of the Tory GST cut - they forget that they were running in 2006 on the (equally silly) tax policy of cutting the bottom tax rate (very expensive in terms of foregone revenue, little upshot in terms on improved efficiency). It's almost impressive that two mainstream political parties managed to fight an election on such a useless (from both an equity and efficiency perspective) concoction of tax policies.

[DavidN]

Frances,

That’s true.

Bob,

I think it depends (on things I don’t know about e.g. elasticity of labour demand, opportunity cost of leisure etc.). If you reduce income tax because you extended the GST, labour increases (income increases), leisure decreases (which directly reduces utility e.g. through reducing family time etc.), also price of goods increase. So are you better off? But as Frances just pointed out, income tax cut or increase tax-free threshold for low-income earners is progressive but not as progressive as direct hand outs (but then as edeast noted direct handouts can lead to other issues...).

[edeast]

I don't think I'm communicating, well. I think direct handouts are good, I'm arguing against getting a onetime tax refund. At least if its a onetime payout it should be at the beginning of winter so people could fill their oil tanks.
These are the anecdotes. One couple did cash the gst cheques,(our town doesn't have a bank) and filled up their truck and jerry cans. (I'm inferring that that was their gas budget for the 4 months)

The milk comes in, as another couple couldn't afford milk for their family, but did buy smokes.

Another couple would walk to the store, and buy groceries, they may have had a car but it wouldn't have been street legal, the wife would come in with a welfare cheque, and pick up groceries, canned goods. It didn't buy much. The husband would stay outside.

No social spending would help our town, but money would.

Also last anecdote, wrt drugs, I've met an ex? heroin addict, his wife cheated on him while he was out at sea, gave him aids. Heroin would look pretty good to me too. Plus there was some racial stuff.

Each of these are a lot more human, involved, and emotional than I'm letting on. Cause I didn't want to engage in poor-people voyeurism.

But being a treeplanter I have met seasonal workers and heard stories of people who take advantage of ei. So that is everything I know, and I don't know what is correct. I was just reacting to the income tax credit vs the income based tax credit.