The Haves and Have-Nots -- Why the Wealthy Tend to Keep Getting Richer at Our Expense

Economist Branko Milanovic explains why the middle class inevitably fails to reap the benefits of fiscal redistribution.

The following is an excerpt from Branko Milanovic's new book,The Haves and the Have-Nots: A Brief and Idiosyncratic History of Global Inequality(Basic Books, 2010).

Who Gains from Fiscal Redistribution?

When we speak of income distribution, we almost invariably have in mind the distribution of what in economics is called “disposable income.” As the name says, this is income that remains at the disposition of households to save or spend, after they have paid direct taxes to the government and received from the government cash benefits like social security and unemployment compensation. But there is another income concept that is, at times, useful. This is “market income,” which is income received from wages, profits, interest, rent, and so forth, before fisc (that is, before either taxation or receipt of government transfers).

Obviously, people with very low market income will be people who either are unable (or unwilling) to sell their labor services and lack property from which they could derive some income. Oftentimes, in developed countries, they are the unemployed. Incidentally, state-funded pensions (like social security) or private pensions are considered equivalent to wage payments (only delayed) and thus included in market income.

Having thus set the stage, this question is often asked: What income groups benefit the most from government redistribution (taxes and cash benefits)? There is a theory that says that in democracies, where people vote on redistributive policies, the main beneficiaries should be people in the middle of (market) income distribution. The rationale is as follows. Suppose there are three people with their market incomes being, respectively, low, medium, and high. (Market income is important here because you decide on tax-and-spend policies you prefer based on the market income you have earned.) The poor person will tend to prefer high tax rates and a lot of government spending because he is likely to benefit from it.

For exactly the opposite reason, the rich person will prefer low tax rates. Then the pivotal vote is held by the person in the middle. With whomever he aligns, that side wins two to one. The median voter, or what we may call “the middle-class voter,” becomes the decisive voter. We would in principle expect him to gain through the process of redistribution. He may not gain more than the poor because tax rates are generally progressive (that is, they increase as market income goes up), but, for sure, we would expect that the middleclass voter would gain in the process of redistribution since, after all, he chooses the tax rate and the benefits that go with it. That means that he would be better off when measured in terms of disposable than in terms of market income. Is this the case?

As it turns out, not exactly, and not unambiguously. First, we find that it is the poor who gain the most from redistribution. Their income share, which is generally tiny in terms of market income, increases, after government redistribution, to still small but more reasonable. For example, in major democracies studied over the period 1980–2000, the share of the poorest decile (10 percent of the population) in total market income is very small, only 1.2 percent. After government redistribution the income share of these same people rises, on average across countries, to 4.1 percent. The poorest decile thus gains almost 3 percentage points. The equivalent numbers for the second-poorest decile are 3.6 and 5 percent. The gain here is 1.4 percent. For each successive (richer) decile, the gains diminish, until they turn to be slightly negative for the fifth and sixth deciles, and then, of course, continue to be even more negative for the higher deciles. Gains are not the same in all countries.

Two conclusions are immediately apparent. First, the biggest beneficiaries of the redistribution process in developed countries are those who “start” the poorest, that is, those with the lowest market income. This is not surprising. What is surprising is that the middle-class deciles (fifth and sixth), whom we expected to gain through redistribution, do not. Their share even decreases slightly. Moreover, this result does not hold only for the developed countries as a whole; it holds for every country individually:

The extent of the share loss of the middle class varies from country to country, and from year to year, but it is overwhelmingly present. This opens up a question to which we do not have a good answer: Why, presumably, do the decisive (middle-class) voters vote for a process that, in the end, leaves them with a smaller share of the national income pie than before? There are two possibilities, neither of which we can decisively prove or reject. The first is that the middle-class voter votes for some redistribution policies as potential insurance. Although at the time when we take our snapshot picture of redistribution, he is not receiving anything from such redistribution, he votes for it “just in case.” For example, the middle class may accept higher taxation of current incomes in order to fund unemployment benefits in the expectation that, were they to be in need of such benefits, they would be eligible.

It is a reasonable hypothesis, but, short of having longitudinal data that follow the same groups of people year after year (so that we can indeed establish whether they eventually gain from these transfers), it is hard to find out if it is true. Moreover, suppose that we find that, even after following these middle-class families throughout their lives, we find that they still lose through redistribution. We can explain that, too: Perhaps voting for unemployment benefits, even if they never receive a penny from them, is still valuable because it is an insurance policy. When we buy car insurance, we do not hope to be able to make any money from it; if we do, we are likely to be worse off than if we did not. But we buy “peace of mind.” The same may be the case here.

The second possibility is that there are transfers received mostly by the middle class that our concept of disposable income fails to catch. This is particularly true for the European welfare states with socialized health systems and public education. Benefits that are derived from both are not included in disposable income. (Disposable income is just the cash income that is ours to spend or save; if we receive free medical care or free education, these benefits are received in kind and are not part of disposable income.) Yet free health care and education are being paid out of people’s direct taxes.

It is likely that while we estimate middleclass taxes correctly, we underestimate middle-class benefits because some are received in kind. Thus, were we somehow to account for the cash value of these benefits, it might turn out that the middle class does reap a net benefit from government redistributive policies. Both of these explanations are possible, but, unfortunately, we lack the data to clearly prove either of them. But we are left with one additional interesting question. Let us go back to the very poorest, who, as we have seen, are the biggest beneficiaries of government redistribution.

Suppose now that the situation of this poorest group becomes even more dire: Their market income share, already very exiguous, diminishes even further. What will happen? Will the redistribution process kick in for the drop in the share or not? It turns out that the answer is an unambiguous yes. In advanced democracies the decline in the market-income share of the poorest will be compensated, one for one, by more redistributive government tax-and-transfer policies. This is, overall, a comforting message, particularly at a time of crisis.

If people who are at the bottom of the income ladder, since they cannot sell their labor or have to sell it very cheaply and also lack any assets, lose even more, the tax-and-transfer system will fully compensate for this fresh shortfall.

Thus, we see that despite the large increases in disposable-income inequality in the past 25 to 30 years in developed countries, and despite the differences in national tax-and-transfer systems, these systems can be shown to operate probably in the way they were intended: to help the most those who start from the lowest position and to reduce the income share of the richest. What remains somewhat puzzling is why the middle class— the class that plays a pivotal role in deciding the degree of redistribution—cannot be shown to clearly benefit from such systems. This result might also show the limits of economic analysis since our voting behavior is influenced, at times even more strongly, by ideology, principles, or values. We do not live by bread alone.

Branko Milanovic, lead economist at the World Bank’s research division in Washington, DC, and professor at University of Maryland, is author of "Worlds Apart." He lives in Washington, DC.