Recent Posts

Wednesday, January 6, 2016

Book Review - Taxation: A Very Short Introduction

The book Taxation: AVery Short Introduction by Stephen Smith provides an introduction to mainstream analysis of taxes. It offers an excellent overview of the history of taxation, the forms of taxation, and current issues. The weakness of the analysis is that it is firmly conventional, and ignores the insights of functional finance (described in this primer). The book focusses on minimising the distortions created by taxation, while in fact those distortions are the necessary objective of taxation.

Book Description

The book was published in 2015, by the Oxford University Press, and is part of a series of books ("Very Short Introductions"). The book is 121 pages, plus 9 pages of end matter. Stephen Smith is a Professor of Economics at University College London, and was previously Deputy Director at the Institute for Fiscal Studies.

The chapters of the book are:
  1. Why do we have taxes?
  2. The structure of taxation
  3. Who bears the tax burden?
  4. Taxation and the economy
  5. Tax evasion and enforcement
  6. Issues in tax policy

What Are Taxes?

The author starts with the observation that we need to distinguish taxes from other forms of revenue a government has. A tax is a compulsory payment to a government, which is determined by parameters set by the government. (At least in modern states; historically, taxation could be fairly arbitrary, which is a mark of bad governance.) Taxes are different from various user fees (such as an entrance fee to a public camp ground) which are incurred voluntarily. For governments with extensive state-owned enterprises, such fees can represent a large portion of the governmental revenue base.

In some rare cases, such as petro-states, such revenues can be large enough to make the levying of taxes largely unnecessary. This appears to be a counter-example to the argument that currencies need to be supported by taxation; but we can view the taxes as being paid by foreign consumers of oil. This is politically rather convenient (although at the cost of allowing weaker standards of governance), but that is an option that is not really available to the main developed economies.

Analysis Framework Used

The analysis framework that underlines the text is a "partial equilibrium" framework, in which we look at supply and demand curves in a single market, and see how the imposition of a tax affects the equilibrium state of that market alone. The book is written at an introductory level, and even the discussion of the supply and demand curves is relegated to a text box that could be skipped. It should be noted that mainstream macro texts have generally abandoned the simplistic "partial equilibrium" framework, and instead looked at "general equilibrium" across all markets, but is unclear from the text whether this has spread to the analysis of tax policy. The text is largely based on the logic of the partial equilibrium results, which seems to imply that similar logic is followed in the tax literature.

To summarise the results, we start off at an initial equilibrium between the supply and demand in a market. We then add a tax on the price of the good, and hold "all else equal".

  • For the graphical analysis, we use the after-tax price of the good. This means that the demand curve does not shift; a higher price to consumers just shifts the demanded quantity lower.
  • The original curve was based on the original pre-tax price of the good. If we plot the supply curve based on the after-tax price (which is lower), the supply curve shifts. This means that the new equilibrium is at a lower quantity of goods demanded and supplied. In other words, taxing a good results in a lower amount of it being produced.
  • The "excess burden" created by the tax is proportional to the drop in quantity times the price of the good. If we assume that the supply and demand curves are linear (which would be approximated for very small tax changes), the distortion is proportional to the square of the tax rate. That is, a 1% increase in tax will cause more distortion if the tax starts out at 20% than if it started out at 10%. 
The first pass of this theory led to the theory of "optimal taxation," developed by Frank Ramsey in the 1930s, in which taxes ought to be imposed on goods based on the "elasticity" of their respective supply and demand curves. However, Smith notes that the optimality of these results was based on fragile assumptions, and the consensus has shifted towards broad-based taxes across all goods and services. By broadening the base, we reduce the maximum rate of taxation on any particular good, which reduces the "excess burden" if it increases at the square of the tax rate (as under the linear supply and demand curve assumption).

One of big debates in tax policy revolves around the effect of taxes on the labour supply. Opponents of increased income and payroll taxes will argue that they reduce employment, as employment is being taxed. The book notes that this topic has been studied intensively, and there does seem to be a consensus about the empirical results (Chapter 4, page 66), such as:
  • The evidence is that taxes have very little influence on the labour supply for men and women who work full time. These people have little opportunity to vary the hours worked in response to tax changes.
  • Taxes reduce the incidence of employment amongst those earning little on a part-time basis. But if they are employed, taxes do not effect the hours worked (since schedules are imposed by employers).
  • The effects of taxes are greater on two groups -- women with school-aged children, and those above 50 years of age.

Who Pays?

Chapter 3 revolves around the question of who bears the burden of taxes. The argument is that we need to distinguish between the "formal" and "effective" tax burden. For example, if we tax the output of an industry, it will most likely embed the tax into the price at which it sells the good. If the industry has sufficient market power, the tax may end up being effectively paid by the consumer of the output.

In some cases, it is difficult to see who pays a tax on business -- will it come out of profits, or else will it effectively damage the industry's competitiveness, and wages have to be reduced? I have my doubts that there are easy answers to such questions. But one principle is emphasised throughout the book -- taxes are best imposed on the side of the transaction which is easiest to monitor. For example, it would be near impossible to impose a tax on consumers based on the luxuries imported into the country, but most governments are able to impose a tax on importers at the point of border controls.

The relative ease of imposition has meant that Value-Added Taxes (VAT) and Income Taxes have been "successful" (or a disaster, if you are a libertarian). Since firms are refunded the VAT on their inputs, they have an incentive to report the tax paid on those inputs. This creates a paper trail which forces the firms selling the goods to report their sales. Meanwhile, it is difficult for employees to avoid income taxes, since they are imposed as a source deduction. It is possible to drop out of sight of the tax authorities, but this is only feasible for individuals who are probably earning so little that they would pay little in taxes in the first place (or else they are running a successful criminal operation that obviously has to be hidden from the authorities). A business that attempts to stay completely out of sight of the tax authorities requires the cooperation of employees; it is only one malcontent away from being exposed.

History Of Taxes

The book has a short history of the forms of taxation. Taxation used to be fairly arbitrary, and probably led to the economic dysfunction of earlier societies (such as the Roman Empire). Taxes have moved to being driven by various objective parameters, such as nominal income or sales. Modern developed states have developed tax systems that are more effective in raising revenue, as they are more intrusive. Before the twentieth century, import duties were a lot more common, as border controls were the only area where most governments had at least a chance of monitoring transactions. Currently, some lesser developed countries with weaker governments are still in that situation.

Tax Evasion

The book argues that there is considerable survey evidence that indicates that societies can be in one of two equilibriums with regards to taxation.
  1. The "good equilibrium" in which it is believed that tax laws are enforced in a reasonably fair manner, and so most people believe that most other people pay their taxes. This makes them more willing to be honest with regards to paying their own taxes. 
  2. A "bad equilibrium" in which most people believe that nobody else is "paying their fair share" of taxes, and so the feeling is that only chumps are honest with their taxes.
Surveys are a relatively squishy measurement, but this characterisation seems to hold up fairly well when compared to how tax systems have evolved in practice. 

With regards to combating tax evasion, the book notes that there are two strategies open to authorities: increasing the number of audits and keeping penalties mild, or increasing the penalties. Smith's view is that increasing the penalties has the lower resource cost, but it is probably more damaging in the long run, as these high penalties create an overly adversarial relationship between the tax agency and the private sector, and helps embed a bad equilibrium for attitudes towards taxation.

Why Do We Have Taxes?

Although the book has a chapter entitled "Why Do We Have Taxes?", it is a question that is not delved into very deeply. The book follows the common view that taxes pay for government services, which is a view that has been rejected by Functional Finance. (This primer discusses Functional Finance.) Modern Monetary Theory (MMT) is a now popular school of thought within post-Keynesian economics which has embraced Functional Finance. In summary, Functional Finance argues that taxes are needed to prevent the economy from overheating, and not for raising revenue.

To summarise why I believe that taxes are not needed for revenue for a central government with a free floating currency, there are three key points.
  1. Free-floating sovereigns such as Canada, the United States, and Japan have monetary systems in which there is no realistic chance of sovereign default; government spending refluxes via the banking system to the bond market. Yes, a default is possible, but such an event requires a spectacular amount of incompetence. The possibility of government default is the underlying theme of my ebook, Understanding Government Finance.) Such governments cannot simply "run out of money."
  2. Government liabilities (money and debt) grow in line with nominal GDP, which implies a need for persistent deficits. In turn, that means that there is government spending that is never matched to corresponding taxes, and so the best we can say is that "taxes pay for some government spending," which is a very imprecise statement.
  3. In any given year, there can be little correspondence between spending and taxes. The fiscal balance can swing from a surplus of 2% of GDP to a deficit of 10% of GDP without a whole lot of changes to the parameters defining fiscal policy (such as tax rates).
However, if we have a large increase in government programme spending (from 20% of GDP in a steady state to 30% say), taxes would almost certainly have to rise in the new steady state condition. That is, there is a linkage between taxes and spending on average, so the "taxes fund spending" is not totally baseless, even from the point of view of Functional Finance. (Moreover, sub-sovereigns such as provinces need to worry about default, and so they cannot ignore financing risk. My comments here are reserved for central governments.)

However, there is a large operational difference in the view of taxes if we drop the worry about "funding" government spending. Within the book, there are a lot of different descriptions of the effect of taxes. From the perspective of Functional Finance, we can give a cleaner characterisation:

A well designed tax will target one of the two objectives.
  1. Microeconomic Management. Taxes and levies that are imposed in a manner to influence behaviour. For example, these would include "sin taxes" on cigarettes and alcohol or carbon taxes, and even fines for breaking speed limits.
  2. Macroeconomic Management. Taxes that are aimed to control aggregate demand. They will rise as the economy "heats up," and drop during a recession. Income taxes and Value-Added Taxes are the workhorses of these taxes, and they are now the primary sources of revenue for central governments. (Taxes on land and property are not sensitive to the cycle, and so they do really not fit within this category; they are a form of micro tax.)
Many of the problems in tax policy outlined in the book are the result of attempting macro management with taxes that are really micro-focussed. The classic example was the English "Window Tax", introduced in 1696 (Chapter 4). The advantage of the tax when compared to modern property taxes was that it was easy to administer -- all the tax collector needed to do was to stand outside the house and count the number of windows. Since larger houses generally had more windows, the tax was theoretically fair. Unfortunately, it led people to brick up their windows, and so a good portion of the English housing stock was stuck in the dark for centuries.

Nevertheless, the objective of both types of taxes is to influence behaviour. Micro taxes are "nanny state" policies, which are explicitly aimed to shape behaviour, but revenues should not be the objective. This was eventually discovered in Canada, where excessive tobacco taxes created a thriving cigarette smuggling business (which grew to a point so that the smuggling had to be incorporated into the National Accounts). However, even macro taxes shape behaviour -- they are designed to lower economic activity, so as to prevent a breakout of inflation.

This insight makes some of the analysis within the book rather beside the point. All taxes are distortionary and "non-neutral"; they have to be. Income taxes certainly lower the number of jobs -- if they did not, the labour market would be under too much pressure, and wage inflation would take off. So we do not need any "partial equilibrium" analysis to tell us what the effects of taxes on labour are. The only things we need to worry about are whether the level of aggregate demand is appropriate, as well as concerns about the distribution of income (which rapidly turns into a political debate). If aggregate demand is weak, we either cut taxes or increase spending (or rely on the monetary policy, which has generated the sub-par growth seen since the Financial Crisis).

The insights provides by a shift in analytical assumptions spill over into other topics. For example, if you want to look at the trends in the fiscal balance, you need to take into account the mix of taxes.  Developed governments have shifted tax burdens towards the lower income cohorts (via VAT and social insurance contributions), and these taxes are extremely efficient at lowering aggregate demand per dollar of revenue raised. This means that in order to remain at the same level of capacity utilisation, government deficits have to be higher than would otherwise be the case. If your objective is to reduce government debt, you need to impose a greater tax burden on high income cohorts, as you need to raise more revenue in order to get the same reduction of aggregate demand. As a historical example, note the fiscal surpluses under President Bill Clinton, which were fuelled by an influx of capital gains taxes (mainly paid by the rich). As a result, we cannot look at projected revenues in isolation and conclude anything about the effect on the fiscal balance.

Concluding Remarks

The book is an interesting introduction to taxation, with a lot of excellent data summarising trends over recent decades. It tells us what tax experts are looking at. Unfortunately, it also gives a good indication of what they are not looking at.
Finally, the book is available at Taxation: A Very Short Introduction (affiliate link).

(c) Brian Romanchuk 2015


  1. "The relative ease of imposition has meant that Value-Added Taxes (VAT) and Income Taxes have been "successful" (or a disaster, if you are a libertarian)"

    I've said for a while that the taxation system should tax the transaction stream at a highish rate, and the tax authorities should actually be in the business of giving refunds. Firstly that gets around the loss aversion issue (since it is the banks doing the actual taking money away, which they already do with 'fees' and 'interest charges'), and it means there is an incentive to fill in your tax return.

    1. The Canadian system is somewhat like that; for people earning mainly wage income, you often get a refund (if you use the retirement tax shelter like an RRSP, do any charitable deductions, or have deductions for children). As a result, a good number of people get refunds when they fill out their taxes, which is a constant topic by personal finance writers ("getting a refund is bad because you are giving the government a tax-free loan!"). If you start needing to pay taxes (self-employment, investment income), they make you pre-pay quarterly, which has the effect that you will end up getting a small refund if your situation is the same.

      However, they do not deduct investment and interest income at source, which is what they did when I was in the UK (early 1990s), and which I think is still the case. The U.K. tax system was shockingly simple to me when compared to filling out the Canadian and Quebec forms. So even though the tax system is ultimately similar, they "felt" quite different.


Note: Posts are manually moderated, with a varying delay. Some disappear.

The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.

Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.