Over at the excellent LAANTA blog, Terence has written a couple of posts discussing issues I have raised here.
The first post covers our discussions on Utilitarianism. Fundamentally I think we have reached a point where our only disagreement stems from value judgments (specifically what we view as the fundamental measure of value, something I plan to write about separately another time) – which implies that his post is a good place for us to leave it.
The first point that Terence raises is that those with high lifetime incomes leave bequests that are more than proportionally larger than the bequests left by those with low lifetime incomes. As a result, a goods and services tax will take a smaller proportion of wealthy peoples income than poor peoples income.
However, this point confuses me a little. The purpose of a bequest is to give income to your children. If you view children as an extension to yourself (overlapping generations type idea – fundamentally saying that you value their satisfaction) then this makes a lot of sense. Now the purpose of giving a bequest is so that they may spend it in the future – and it is then taxed at the rate of GST. As a result, over time the entire stream of income will be taxed at the GST rate for both the low and high lifetime income individual/household.
The only counter argument I can see for this is that people may “value savings” beyond any sort of insurance or future consumption motive. If we believe this impact is sufficient – such that savings is actually like consumption for wealthy people we could justify GST being regressive (as even if the world had a known endpoint the wealthy would keep some savings at the end period). However, I am not convinced.
The next argument states:
Yet, even if the wealthy really did not save more than the poor, GST would still be a regressive tax. Why? In three words: diminishing marginal utility. Which, in this context, is a fancy way of saying that money matters more to those who have less of it.
This of course depends on our definition of a regressive tax. The definition I use of a regressive tax is:
a tax for which the amount of an individual’s taxes falls as a proportion of income as the person’s income increases (*)
In this case we ARE just looking at where the tax liability falls – not on its impact on social welfare. Now I fully agree with Terence that we should care about the impact on welfare, not the impact on the amount of money people have. If we take the argument of diminishing marginal utility at face value, then we have one of the main justifications for having a “progressive tax system” – however this does not change the fact that the GST tax still falls under the “neutral” definition.
Now, as we may know, I prefer targeted assistance to a broad progressive tax system as a way of achieving equity goals and maximising social welfare. Now, if we fully believed that everyone had the same utility function my belief would be silly – as a progressive tax system would be a no-nonsense way of achieving equity goals. However, utility functions do differ substantially – and they differ in such a way that a progressive system becomes inappropriate.
Economists don’t know how to measure utility or welfare, this involves making special types of assumptions that we are unwilling to (cardinal utility) – as a result, we can’t compare individuals utility.
Now, we could have a case where some people gain very little utility from products that they can buy from income, while other people gain a great deal of utility (assume that the cost of working increases for both people, but in the same way). In this case, if people can choose how much to work and therefore how much income to make, the person with the higher income will be the person that gains more satisfaction from the income they receive.
As a result, even if we have diminishing marginal utility we cannot say that we should tax from the rich and give to the poor – insofar as people can choose how much they work.
Furthermore, as we are interested in lifelong utility, a progressive income tax will impact more on people who decide to (or only have the opportunity to) work hard during certain parts of their life then those who are able to spread work (and thereby income opportunities) over their lifecycle – even if they have the same lifetime income. As a result, if we believe that these agents have the same utility function, a “progressive” tax system (in the static sense we have now) would be unfair – and if we do not assume the same utility function we have no idea what will happen.
As a result, I still believe that GST is a neutral tax (in the intertemporal sense). I accept that it may lead to a reduction in welfare for some agents more than others – however I believe that targeted assistance is a better way of solving this than the blunt instrument of messing with the GST system.