On “Economic Fairness”

A few minutes ago, I made a comment in this thread at Tacitus.org, and I decided to refine it a bit and make it a topic for discussion here.
Say that a society has two different possible means of organizing itself as far as economics go, and that it will know in advance approximately the effects that each strategy will have on the relative wealth of individual citizens. The first option–a more or less free market capitalist economy–will create income distribution as follows (by decile):
10X/5X. . .2X/X
The second option will redistribute all wealth produced by the society and leave a wealth distribution as follows (again by decile):
(X-1)/(X-1). . .(X-1)/(X-1)
Choosing the first one should be obvious, but I submit that any true believing socialists out there will point to the second one as more just, assuming they don’t refuse to answer on the grounds that the situation is artificial and unrealistic. In one sense, they are correct–no state can predict *exactly* what the effects of their economic policies will be. However, there’s enough data in from over the years and from over the globe to say this conclusively: command economies are miserable failures at producing wealth compared to free ones. Any society which chooses to adopt a command economy is doing so with the knowledge that doing so will make their nation poorer than it would have been otherwise, whether for egalitarian purposes or for other ones (propping up a dictator, waging a war, etc.). And, to put it bluntly–collectivists are more concerned with making sure those who would otherwise be rich are poor and miserable too than they are with improving the status of the poor. Am I overstating the case? Look at the scenario above and ask which society the collectivist will praise for its egalitarianism and justice–the fact that everyone in it is poorer than the poor of the free economic state that would have been its alternative does not register with them.



  1. Matt Clarke

    The reason Socialism does not work is human nature.
    If there is no benifit to inovate then why do it?
    If you take away the ability to make yourself more prosperous though working smarter or harder then the only real benifit of rising above anyone else is to gain power.
    Once you get that power you are never going to let it go lest you fall back to the pig farm you crawled out of.
    This then leads to corruption and then onto a Socialist state where some people are more equal than others.

  2. Dave in LA

    I’m guessing that a real economist would say that you haven’t provided enough information. Let’s make it more interesting. Before choosing, you must further satisfy a condition that no one is allowed to starve. To meet this condition, your system must be constructed such that the lowest decile’s income equals or exceeds this ‘starvation’ minimum.
    The only interesting case, as it turns out, is where that minimum is at least X-1, but less than X. That’s the only range where either system will satisfy the ‘no one starves’ constraint.
    The classical approach to redistribution kicks off with the notion of diminishing marginal utility. That is, what we are trying to maximize is the welfare (utility) of the society, and not the total income (subject to that ‘no starvation’ constraint.) Under diminishing marginal utility, each additional dollar over the starvation minimum is less valuable (has less utility) than the last dollar. Thus, the welfare of our top decile individual, which under society A has 10 times the wealth of an individual in the bottom decile, is less than 10 times as great. This utility concept is one of the key arguments about the fairness of redistribution.
    Two other basic concepts which let us get totally wrapped around the axle are:
    – the friction introduced by the income distribution mechanism (costs of actually doing the redistribution, disincentives to work and invest, misallocation of resources, etc.)
    – a notion that income inequality itself imposes costs on the less fortunate. Call it an envy coefficient. A variation of this notion extended to systems over time gets to inequalities of opportunity for one’s children, creation of immobility among deciles (to put it blandly), and a host of social ills.
    and we haven’t even talked about negative externalities, the implications on individuals’ migrating among deciles through time, etc.
    If this isn’t already way off the mark, this exercise isn’t an argument for a command vs. a market economy per se. In your example, individuals could still have freely chosen the activities in which to engage. The society simply redistributes the fruits of their labor at the end of each period. This gets to your definitional issue about socialism vs. communism.

  3. Dave in LA

    After an exchange of e-mails, Scott suggested I elaborate on his hypothetical here. Since I posted my comment above, Bruce Bartlett has written an article in NRO that summarizes some of the economic literature regarding tax rates and economic output. What follows here poses Scott’s original question slightly differently. We are going to let a society choose between two alternatives, but in my case, the choice is independent of the kinds of activities its people choose to engage in. Changes in the distribution of incomes are accomplished through the tax system.
    Assume the same society as in Scott’s case, and the same perfect foresight as to the income implications of its policy choice. The society does not wish to collectivize the means of production, but is willing to let the members of the society engage in the same varieties of activities under either of its policy choices. Its first option will create an income distribution as Scott describes, namely that the median after tax income of the lowest income decile is X, the median income of the second is 2X, and so on, up to the top decile with a median income of 10X, or ten times the median income of the first decile. The overall median for the society under this choice is 5.5X. Note that this distribution denotes the distribution of after-tax incomes. If it helps simplify things, assume that the after tax results are net of a flat tax.
    The second option will redistribute incomes through a steeply progressive tax system, such that after-tax incomes (the median incomes of all deciles) are equal. Using Z to represent the median income of the lowest income decile under choice 2, median incomes by decile are Z, Z, …, Z.
    However, the society realizes that imposing this redistribution of incomes is not costless. High taxes on high-earning individuals will induce some to prefer additional leisure (which is not taxed) to income (which is highly taxed.) Further, the tax system will affect the kinds of work individuals choose to do. Difficult, risky and highly skilled occupations will produce the same after-tax incomes as safer, less demanding and less skilled work. Incentives to invest, especially in risky but potentially profitable ventures, will be drastically curtailed. Finally, as Scott implies, the government will at some point be inclined to interfere in ways other than the collection and redistribution of income.
    In this simple hypothetical, let’s assume that the society knows that its income will be lower, and it growth in annual incomes will be lower, if it elects to equalize incomes than if it allows income inequality. One can insert various values for initial incomes and for growth in incomes, to see what the actual implications might be. To insert a real world example, estimates cited in Bartlett’s article suggest that incomes in the U.S. would be 20% higher if we abandoned our progressive tax system in favor of a flat tax system. Another article Bartlett cites suggests that a large percentage of the differences between U.S. and French per capita GDP differences are attributable to differences in our tax codes.
    Depending on the income and growth differentials one assumes, it will not take many periods until the median income of the egalitarian choice (Z) is far below the median income of the “free market” choice (5.5X). Given a bit more time, the lowest decile median income under choice 1 (X) will exceed the overall median income under choice 2 (Z).
    One modification to the choices above would be to ensure that the truly needy had a level of income sufficient to provide for basic needs, and to modify the tax system to redistribute income such that the lowest incomes were brought up to that basic level. Doing so in choice 1 of our simple example would not meaningfully change the outcome of the exercise. The new choice one would still have significant income inequality, higher total income and higher income growth than would choice 2.
    Getting back to Scott’s original question, what assumptions about the utility of income and the societal costs of income inequality would one need to make in order to justify choice 2? In what sense is choice 2 more “just,” even though it produces an overall level of societal income significantly lower than choice 1?
    One final note: Scott’s original formulation assumed that to achieve an equal distribution of incomes, a command economy was required. The problem formulation above does not make that assumption. In practice, however, I agree that it’s unlikely that a free economy coupled with a massively redistributive tax system could long co-exist. Businesses that required highly skilled or otherwise scarce labor would find it impossible to attract employees. I’ve asked acquaintances who are in favor of a government-mandated “living wage,” for example, that if $12/hour is good, why wouldn’t $50/hour be better? That way, everyone could buy a home here in L.A. It’s only when one thinks about this extreme kind of social engineering that the implications for the health of the overall economy become obvious.