How is the effect of a tax reform calculated?

The tax calculator on this webpage allows you to predict:

  1. how much a reform of (wealth) taxes would cost or benefit you, depending on the tax exempt amount and the tax rate,
  2. how much more or less other people with different levels of wealth would have to pay,
  3. and how much government revenue it would generate.

To explain what this calculator does, it is useful to briefly talk about how economists think about the impact of tax reforms, and about how tax rates should "optimally" be set. When the government changes some tax rate, a number of things can happen at the same time:

  1. Just mechanically, everybody has to pay more or less taxes on the wealth they had before the tax reform, and correspondingly the government receives more or less revenues. ("Mechanical effects")
  2. Individuals might react to the changing taxes, for instance by saving more or less money. "Saving" means that individuals don't spend their money on consumption, but instead invest it in savings accounts, real estate, stocks, funds, etc. They might also react by changing how much they cheat on their taxes. This might affect (a) government revenues, (b) those who change their behavior, and (c) other individuals. ("Behavioral effects")
  3. Changes in behavior might lead to a change in market prices or interest rates, which can make different individuals better or worse off. ("Price effects")

How do these effects enter our calculations?

1. Mechanical Effects

This is the simplest, but also most important part of our calculations. To calculate mechanical effects, we use the information about the distribution of wealth that is provided by the Survey of Consumer Finances (SCF). We probably underestimate mechanical effects on government revenues, as we underestimate the assets of very wealthy households.

2. Behavioral effects:

a) ... On government revenues:

When individuals save less or start to evade taxes more frequently after an increase in wealth taxes, this means that the government might gain less in tax revenues than we would think based on mechanical calculations (since there is less declared wealth to be taxed). How large this behavioral effect on government revenues is, is a difficult empirical question. What we know is that most of the effects are taking place because of increased evasion or avoidance, rather than through decreased savings. Most empirical studies have found little to no effect of returns (net interest rates, rental income from real estate, capital gains from stocks, etc.) on how much people actually save.1

We base our calculations on the assumption that, because of tax evasion, increases in government revenues are about 20% less than suggested by mechanical calculations. This number is probably overestimating the behavioral effect of tax increases. If taxes are well enforced then the behavioral effect might well be smaller.

b)... On those changing their behavior:

When individuals save less or start to evade taxes more frequently after an increase in wealth taxes, this means that they might actually be better off than suggested by the mechanical calculation of how much they have to pay. After all, they could simply have done the same as they did before the tax reform (that is, they could have saved the same and not changed how much they cheat on taxes), but chose not to. Taking this line of reasoning further, it can be shown that in fact behavioral changes can be ignored when calculating the effect of the reform on individuals.

c)... On other individuals:

In a competitive market with property rights, such as the capital markets of the US, the owners of capital (savings, real estate, stocks, etc.) generally gain all the returns to their capital. Contrary to claims often made in public debate, no one else profits from their continuing participation. That means that nobody else is directly affected by any of their behavioral changes.

3. Effects on prices and interest rates:

In principle, a change in wealth taxes might shift savings to such an extent that the available capital in the economy and thereby the real interest rate do change. In practice, these effects are very small. Two reasons for this are that (i) capital is globally mobile and changes in savings in one country will have a negligible effect on the global capital stock, and (ii) the behavioral effects of interest rates on savings (as opposed to tax evasion) are empirically very small. Following standard practice in the academic literature, our calculations therefore assume that there are no price effects.

Based on these different effects, we know how much any given person or household is affected by the tax reform. To get an effect "for the entire economy," we have to somehow sum these effects up. But of course an additional dollar for a poor person is not the same as an additional dollar for a rich person. It makes no sense to say that giving $100 to a person earning $10,000 per year is the same as giving $100 to a person making $100,000 per year. We have to make a trade-off between different people, and choose how much weight we assign to poor versus rich people, when summing up. These weights are a matter of distributive justice that economists do not have more to say about than anyone else; at the end of the day this is a matter of "picking a side" – who do we think should get additional income?
Building on such considerations, economists then discuss the "effect on social welfare" of a change in tax rates. This effect on social welfare is given by the sum of the weighted effects on all the individuals in the economy. Economists say a choice of tax rates is "optimal" if no change is possible that would increase "social welfare."
On this website, we show you how much different people would gain or lose from tax changes that you propose. It is up to you to decide how you evaluate these gains and losses.

1 For further information on the empirical literature concerned with this question see Bernheim, B. D. (2002): Taxation and saving, in: Handbook of public economics, 3 and Attanasio, O. P./Wakefield, M. (2010): The effects on consumption and saving of taxing asset returns, in: Dimensions of tax design: the Mirrlees review, in particular chapter 7.3.3.

Further Literature:

Saez, E. (2001). Using elasticities to derive optimal income tax rates. Review of Economic Studies, 68(1): 205–229

Chetty, R. (2009). Sufficient statistics for welfare analysis: A bridge between structural and reduced-form methods. Annual Review of Economics, 1(1): 451–488.

Mirrlees, J. et al. (2010, Hg.). Dimensions of tax design: the Mirrlees review. Oxford University Press.