Keywords

JEL Classifications

One can detect in the literature of economics two important lines of thinking on the subject of neutral taxation. One emphasizes economic efficiency (i.e. the elimination of deadweight loss) as the objective in terms of which the neutrality of taxation is defined. The other emphasizes the generality of a tax as itself imparting the quality of neutrality. Two examples, each with a long history in economic thinking, illustrate the main lines of the distinction.

On the one hand we have the taxation of land rents or land values. It builds on the notion (not precisely true in fact) that each piece or plot of land is totally fixed in supply, with the consequence that any tax levied upon it will ultimately be paid out of its pure economic rent.

On the other hand we have the relatively modern idea of a general tax on value added, the tax being applied at a uniform rate on all activities in the economy. Here there is no thought that the underlying resources are fixed in each activity; quite to the contrary, mobility among the various taxed activities is taken for granted for most of the resources on whose product the tax will fall.

It is easy enough by making artful assumptions to bring these two notions very close together. For example we can assume that no manmade improvements to the soil are possible, or alternatively that the tax assessors can always distinguish between ‘the intrinsic and immutable qualities of the soil’, on which tax is then duly assessed, and the manmade improvements thereon or accretions thereto, on which (under our convenient assumption) no tax is either assessed or paid. Similarly, we can assume for the value added tax that there are just three basic resources in the economy – land, labour and capital – and that each of them is fixed in supply. Therefore a uniform tax on the marginal product of any one of them will be neutral, striking the factor equally regardless of the end use to which it is applied, and leaving the factor (because of the assumed zero-elasticity of its supply) no untaxed haven (not even leisure) to which it might choose to escape.

The above assumptions make it easy to define neutral taxation for a Dictionary. (Neutral taxation is taxation falling on something that is in completely inelastic supply, with the tax being so designed as not to affect resource allocation either within or among the affected categories or between them and the other activities not subject to the tax.) But it would probably not add much to the usefulness of the Dictionary.

To be truly useful, I believe, a definition of neutral taxation should be able to throw away such artificial crutches as the two assumptions presented above. It should be able to live in the real world, where we know that the relevant supply elasticities are rarely zero, but where we do not feel at all sure about their magnitudes nor how they vary as between the short, middle and long run. It should be able to cope with reality that, for tax policy at least, the objects of tax do not have an independent essence as commodities; rather, a commodity subject to tax is whatever the tax law (including the regulations and practices followed in enforcing that law) defines it to be. And finally it should come to grips with the serious claims that can be made for considering equality (among the affected activities) in the applicable tax rate to be an attribute whose presence connotes neutrality and whose absence creates a presumption of non-neutrality.

Economics has come the farthest in responding to the first of the desiderata expressed above. Deadweight loss is a concept completely familiar to the discipline, as is the idea of minimizing the deadweight loss of raising a certain amount of tax revenue subject to given constraints. A clear line of thinking runs from Ramsey in the 1920s through Hotelling in the 1930s, Meade in the 1940s, Corlett and Hague and Lipsey and Lancaster in the 1950s, Harberger in the 1960s, to the modern writers on optimal taxation of whom Atkinson, Diamond, Dixit, Mirrlees, and Stiglitz are a representative few. Flowing through this strand of thought are the related ideas (a) that uniform taxation is not always neutral; (b) that the special condition under which uniform taxation of a subset of commodities or activities minimizes the deadweight loss of raising a given amount of revenue from that subset is met when the equilibrium quantity (or activity level) of each member of the taxed subset would respond in the same proportion to a (hypothetical) uniform tax on all goods or activities that are not in the taxed subset; and (c) that whenever the condition stated in (b) is not met then instead of uniform taxation the minimization of deadweight loss requires higher-than-average taxation on goods whose quantities would fall as a result of a (hypothetical) uniform tax on the uncovered group and lower-than-average taxation on those whose equilibrium quantities would rise most sharply.

The analysis underlying the above statements is straightforward, and one can even call economic intuition into play to explain the conclusion. If the tax authorities are denied the possibility of taxing certain goods or activities, then it can to some degree ‘get around’ the ban by putting higher taxes on those items within the taxable subset which are complements of those that cannot be taxed. In a similar vein, since one way of thinking of the resource misallocation that occurs when only a subset of activities is allowed to be taxed is that resources are ‘artificially’ shunted from the taxed to the untaxed subset, it seems quite plausible that the optimal patterning of tax rates within the taxed subset should entail taxing at somewhat lower-than-average rates those particular activities in which a percentage point increment of tax would lead to notably greater-than-average ‘shunting’ of resources to untaxed activities.

The line of reasoning just presented is persuasive – sufficiently so that some economists have been tempted to write off uniformity altogether as a plausible objective of tax policy. There remain many, however, who adhere to uniformity as a goal. Given the ease with which propositions (a) through (c) above can be derived, one should hope that most of those who hold to uniformity base their adherence on considerations extraneous to the derivation, say, of the Ramsey rule and other similar propositions in the literature on optimal taxation. The discussion that follows assumes so.

To build a case for uniformity in taxation in the face of the foregoing logic, one should (appropriately, I think) postulate that one is not dealing with two quite arbitrary categories of goods and/or activities, viz., the taxed subset and the untaxed subset. Instead, one should assume that the taxed subset, rather than being ‘any arbitrary bundle’, is so selected as to contain all the goods and activities that can plausibly and without unusual administrative or regulatory effort be brought into the tax net. One then proceeds to view the problem not as a simple analytical puzzle but as one of guiding or governing the interaction between the society’s fiscal authorities and its members.

With this objective in mind, an advocate of uniform taxation might set up a quite different problem from that posed earlier. He might consider the ‘disturbance’ with which he is dealing to be a consumer changing his mind about how to spend his money or a worker changing his preference about where or for whom to work. A uniform-tax advocate would likely place a considerable value on the authorities’ simply not caring about these various changes of mind.

When one solves the Ramsey problem one takes as given the tastes and preferences of economic agents and maximizes government revenue for a given aggregate level of the agents’ welfare. Under the differentiated set of tax rates that emerges from this exercise, the maximizer is not indifferent to changes in tastes of the agents. The maximizer likes it when agents shift their tastes from low-taxed to high-taxed activities, and is disappointed by shifts in the other direction.

Something of the same thing occurs when uniform taxation is implemented. Here the ‘good’ event would be a shift in tastes that caused untaxed activities to contract and taxed activities to expand; the ‘bad’ event would be the opposite. But there would be a wide range of changes of tastes that would be neutral–these would cover shifts among commodities or activities within the sector subject to the uniform tax, and also shifts among activities in the untaxed sector. To the degree that the authorities are successful in extending the tax net over quite a wide range, it may turn out to be true that most changes in tastes simply lead to shifts in the composition of goods within the taxed group. This is the sort of scenario that would best fit the vision of an advocate of broad-based, uniform taxation and at the same time would (at least if changes in tastes within the taxed sector were frequent and important) create problems for proponents of Ramsey rule taxation.

Subtle overtones of a less technical nature also arise when Ramsey-rule taxation is compared to a broad-based, uniform levy. In Ramsey-rule taxation individuals are genuinely presented with incentives to shift their demand from high-taxed to low- taxed products, and workers are likewise motivated to shift their labour efforts from high-taxed to low-taxed activities. Both these incentives are counterproductive from the social point of view. Subtly hidden in the way the problem is framed is the assumption that people’s tastes are given. The reality of the world is that tax laws change only rarely; once enacted, they stay in effect for long periods of time, over which economists can be certain that there will be important changes in the parameters of tastes and technology. The goal of having a tax system that is robust against these unknown future shifts in demand and supply is not capricious; it deserves to be taken seriously.

In a quite different vein, there arises the question of to what degree we want our choice of tax patterns to depend on parameters like elasticities of supply and demand about which our knowledge is very spotty and imperfect. Proponents of uniform taxation can fairly argue that their choice of such a form does not depend seriously on knowledge about the parameters of demand and supply. Economic theory assures us that the dominant force is substitution (in the sense that a tax on an activity will, other things equal, cause that activity to contract). There is thus a very strong presumption that broadening the coverage and lowering the rate of a uniform tax will reduce the deadweight loss associated with it (for given revenue yield). One can build policy on this basis without having any detailed knowledge of the parameters of supply and demand, without any particular hope of gaining anything more than a very patchy knowledge about them in the future, and indeed with an almost absolute assurance that whatever the relevant parameters might be now, they will undergo substantial changes in the future. If one believes that these conditions come close to describing our present and likely future state of knowledge about the relevant parameters, he will likely be predisposed toward uniform as against Ramsey-rule taxation.

The last line of argument favouring uniform taxation has to do with the interplay between equity and efficiency considerations in governing tax policy. The motivations that fall under the umbrella of ‘equity’ are too numerous and too varied to try to recount here. But nowhere among them can one find that it is fairer to tax more heavily factors of production that cannot flee to other activities or that it is more just to tax heavily those items whose demand happens to be less elastic. To tax salt more heavily than sugar simply and solely because it has a lower elasticity of demand is at least as capricious (from the standpoint of equity) as taxing people differently according to the colour of their eyes.

Ultimately, I believe, the issue of uniform versus Ramsey-rule taxation may turn out to be just one facet of much broader philosophical differences. Consider the philosophy of government that assigns to government the role of creating a framework of laws and regulations within which the private sector then is encouraged to operate freely. Under this philosophy a positive value is placed on the authorities’ not caring about what private agents do (so long as they abide by the rules). It is a position desideratum to create a tax system that is robust against changes in tastes and technology.

On the other side of the coin we have a philosophy of social engineering, in which the detailed tastes and technology of the society enter as data into a process by which the policy makers choose parameters such as tax rates and coverages so as to maximize some measure of social net benefit.

Each of these philosophies has had its own long trajectory within the profession of economics. Each has its representatives today. Each will surely be reflected in the literature of future decades. In my opinion, the future debate as to how the concept of neutrality in taxation should be reflected in real-world policy decisions will swirl around the subtle differences between the ways in which holders of these two philosophies view the world, between the roles they envision for government, and between the ways they see the science of economics interacting with government in the formation of policy.

See Also