Notes on (Capital) Taxation
Some basic principles
So tax is in the discourse again, in particular the way in which capital ought to be taxed. I’m generally sympathetic to taxes on capital income, but this post will not focus on any specific proposal. Rather, I want to outline some simple Public Economics principles which often get missed in the discourse.1
Basic Structure
Let's start with the basic framework. Consider that there are two basic forms of income:
Labour income - from working
Capital income - from owning assets generating a return.
Each tax will impact one or both of these, or a subset of, such as a tax on labour incomes or owners of specific types of assets.
In simplified terms, a tax is said to burden labour income when it changes the relative price of consumption and leisure. A 30% income tax for instance means a worker earning $30 an hour will only take home $21, and thus has lower purchasing power. A tax is said to burden capital income when it changes the relative price of saving/spending income in one period, relative to future periods. A 30% tax in interest income for instance may mean it is less beneficial to save today relative to tomorrow.
1 Exactly Which Capital Income?
The first question to answer is exactly what kind of capital income are we talking about when discussing capital taxation. This is important as the capital income of economic theory (in particular optimal tax theory) is not the same as any return to capital in the real world.
Stiglitz helpfully delineates real existing capital income into 4 components:
Pure rate of interest (riskless return or effectively the bond rate)
The return for risk
Compensation for effort (such as managing investments)
Rents (excess returns based on monopoly power or other rent seeking).
As Stiglitz notes, most economic literature suggesting a low (or even zero) rate of taxation relates to the first of these. The others should be subject to the following rates:
Return for risk - Complicated, but risk sharing between government and private sector can be efficiency enhancing, particularly when risk markets are imperfect. Note this assumes offsets for losses.
Compensation for effort - As this is a form of labour, it should be taxed at the same rate as any other labour.
Rents - As with any form of economic rent, should be taxed at very high rates.2
Of course these components do not come neatly labeled in the real world, or tax policy would be extremely simple, however we should not assume all returns are (1).
2 Sales Taxes are taxes on Labour Income
“We need to tax work less” is a common phrase you hear, usually referring to the income tax. Instead it is often proposed we move towards taxing consumption more, via a higher GST. This however is mistaken, economic theory suggests consumption taxes tax work more than the income tax, not less.
The intuition is simple, a sales tax reduces the purchasing power for every dollar earned, but it does not change the relative payoff between future and present consumption. In economic terms it is intertemporally neutral, taxing equally between periods. Indeed, in theory a universal consumption tax is equivalent to a labour income tax, that is, one that excludes capital income.
3 Individuals can often recategorize their income - this is an argument for equalizing capital/labour taxation
Tax theory often treats whether an income is from capital or labour as a given, immediately apparent and not contingent on legal arrangements. This however is often not the case, particularly for high income individuals, they may have discretion to recategorize their income as either capital or labour, depending on which has a lower tax rate.
As a simple example, suppose I am currently an employee of Company Inc, paid $400,000 a year to provide expert tweeting services. The government then lowers the rate of capital tax relative to labour. I, and the company I work for, may then have an incentive to stop being Michael Chowdhury, loyal employee, and negotiate to be Michael Inc, private trader contracted and delivering a profit to the owners (me again). Of course this example is oversimplified, but similar strategies were adopted in response to the Tax Reform Act of 1986.
Taxing labour and capital income at similar rates then may be efficient and equitable, to the extent it prevents this behaviour and discourages rent seeking.
Conclusion
The economics of taxation invokes strong emotions. This is understandable, they are arguably the strongest tools governments have for redistribution. We however should be careful when invoking economic theory to prove a particular arrangement is "correct", given the strong assumptions implicit in much of the optimal tax literature. A good example of this is Stiglitz and Atkinson themselves, who despite writing the paper often cited as a reason for low to no capital taxation, themselves favour substantial taxes on capital income. As Stiglitz writes:
"Those who [advocated abolishing capital taxes] typically did not understand (or did not want to understand) the limitations of the model. As always, one has to look carefully at the assumptions going into a model to judge whether they provide an appropriate basis for policy."
In terms of sources, I rely primarily on Lectures on Public Economics by Atkinson and Stiglitz, The Origins of Inequality by Stiglitz and Triumph of Inequality by Saez and Zucman.
A notable exception of course being for Schumpeterian competition.