Friday, July 21, 2006

Income Inequality in The United States

A recent Kevin Drum post posed the question of the vanishing middle class in this country:

Over at The Corner, even John Derbyshire thinks there's some evidence that the middle class isn't doing too well these days:

If the rich get richer while the middle class thrives, and some decent provision is made for the poor, I'm a happy man, living in a society I consider healthy and am proud of. If, however, the rich get richer while the middle class is struggling, or actually declining, I am not a happy man. There are some reasons to think that is happening, and you don't have to be a socialist to worry about this.

It is, perhaps, telling that Derbyshire's post sparked not a single response from his fellow conservatives. Even the neo-Lafferians at NRO seem a little too embarrassed by the whole thing to go through their usual exercise of digging up a few pseudo-statistics to demonstrate that, really, the middle class is going great guns under today's Republican leadership.

Kevin Drum then quotes another piece of news which doesn't really have anything to do with the middle class but is interesting anyway:

....No amount of chaff can hide the failure of our remarkable productivity surge (and the accompanying robust growth of the overall economy) to meaningfully boost average wages, which have barely grown with inflation. Separated by income level, the picture is more dismal. From 2000 to 2005, for example, average weekly wages for the bottom 10% dropped by 2.7% (after adjustment for inflation), while those of the top 10% rose by 5.3%.

In short, the rich are getting richer, as would be expected when the Party of the Rich is in power. But what is happening to the rest of the income distribution? And is it ok that the very poor are getting poorer if the middle class can somehow hang on? But is the middle class hanging on or not?

To answer these and similar questions requires delving into statistics. But before I do that I want to say a few short words on the reasons why we should be concerned about increasing income inequality, even in a capitalistic society.

The traditional arguments for allowing incomes to be unequal are two, one ethical and one efficiency-related. The ethical argument simply states that those who work harder or smarter should get to keep the fruits of their labors and should have the right to will those fruits to their heirs, even if this creates a society where money concentrates in few hands. The efficiency argument states that what drives economic progress and innovations is the chance to make money out of it, to make more money than the rest of the pack, and that economic progress and innovations will ultimately make everybody better off. It's a choice between having low equal incomes and higher average incomes in a system which has some people earning more than others do.

The equally traditional arguments against greater income inequality are also both ethical and efficiency-related. The ethical argument points out that greater incomes are not only a consequence of greater effort or smarts. They can also accrue from illegal activities or from gaming the system, and even when they don't they are helped on by the society and its government, because it is the government which makes markets possible and which guarantees the infrastructure that the money-makers can use. Thus, the society deserves a cut in the greater incomes of some. The ethical argument also stresses the horror of the no-safety-nets form of capitalism where the ones who don't thrive are allowed to suffer and die. The efficiency-related arguments point out that a society which is very unequal in income becomes a dangerous place to live in. Just think of life in the so-called banana republics to get an idea of the problems that income inequality causes. And even a less extreme income inequality causes a society where it will be very difficult to arrive at agreements about public policy, because the income differences will make different outcomes desirable.

In short, extreme inequality in incomes may destroy a country. It kills the markets for the products that the rich built their riches from, and it creates an angry and violent underclass. The vanishing middle class is something that speeds up this development, as the middle class is usually the part of the society which cares about the political choices its government makes and which tends to invest in stability. It is also the source of most of the educated labor force.

Given all this, where is the United States going in terms of income inequality? The general consensus is that income inequality is growing, and that this is true whatever the measures we select to use for it.

Now comes the part where I go all economist. But trust me, you will learn a lot from it. For example, you will meet the Lorentz curve and the Gini coefficient. They are really not that horrible creatures. The Lorentz curve is a picture of income inequality, like this one:

The horizontal axis (direction to the right) is all the families of a country standing side by side, arranged so that the poorest family is closest to the left edge and then the next poorest family and so on, until the richest family is the last one on the right. The axis is then standardized so that when all the families are lined up the total length is 100% or 1.

The vertical axis measures the earnings of the same people, added into a pile of earnings. So the vertical dimension initially measures zero because there are no people or earnings yet. Then the poorest family takes its place and its earnings are measured on the vertical axis at that point. Next the second poorest takes its place and its earnings are added to the poorest family's earnings. And so on. When the richest family enters the lineup its earnings are dumped on top of the vertical dimension. The vertical axis is then standardized so that incomes are shown as fractions of the total. So the highest point is 100% or 1.

The green mountain shape shows the way total earnings have been piled up in this process of adding one family's share at a time. The mountain rises slowly at first, because the first families are poor ones, then more rapidly as wealthier families are added to the lineup. The shape of the green area's edge can give us a good visual grasp of income inequality. The more convex* that edge looks the more unequal incomes are.

If you think about this a little it becomes clear that the straight diagonal line in the picture would show how the earnings pile rises in a society where every single person has the same income. Deviations from that line show increasing inequality. For example, think about a country where one guy owns everything. Then the cumulative earnings line would hug the horizontal axis until the richest guy enters the lineup. At that point, the earnings pile would suddenly appear as a vertical line.

So the two extremes the picture shows are full equality: the diagonal line, and total inequality: a reverse L-shape. To show something inbetween the two we get the kinds of curves that are shown in the graph above. The more convex the curve is the more income inequality the country has. Also, a curve that becomes more convex over time shows that income inequality has increased.

The Lorentz curve is a good way of analyzing income inequality. It can be shown for both pre-tax incomes and for after-tax incomes, for example. But it's not one number. If you want one number to describe income inequality the most common candidate is the Gini coefficient. Luckily, it is derived from the Lorentz curve! Just take the area between the diagonal line and the curve (the pink area) and divide it by the total area under the diagonal line! If there is no income inequality at all, the Gini coefficient is zero. If the income inequality is total, so that one person has all the earnings in the society, the value of the Gini coefficient is 100 (we've moved back to percentages here).

Now we are ready to look at the United States income inequality. The Gini coefficient gives the U.S. a value of 46 according to one source, a value around 40 according to another source. The values for West European countries tend to be in the thirties or lower, the values for South American countries tend to be in the fifties or higher. The U.S. value places it slightly apart from other post-industrial countries and closer to the South American countries.

What about changes in the United States over time? Look at the statistics on this site (scroll down). The increase in inequality is very evident in the growth of the Gini coefficient year after year. That is not good news.

I selected the Gini coefficient and the Lorentz curve for discussion not only for their simplicity (yeah!) but also because the Lorentz curve actually stresses changes in the middle of the distribution, in the area where the middle class has hunkered down, and so the changes we observe are quite likely to affect the middle class position. But if this is not sufficient for you, you could also look at income changes by deciles (groups consisting of ten percent of income earners, arranged in increasing order) or quintiles (groups consisting of twenty percent of income earners, arranged in increasing order). That's what we are talking about when we compare the top ten percent of income earners to the bottom ten percent and so on. Here is one summary of what has happened to the quintiles:

So, the top quintile is making more. That means by definition, the bottom three [sic] quintiles are making less. Again according to the Census bureau for the years 1969 - 1999 the fourth highest quintile saw its percentage of national compensation income drop from 23.6% in 1968 to 22.8% in 1999. This is a 3% drop. The third quintile saw its percentage drop from 17.3% to 15.3% or a 11.56% decrease. The second lowest quintile saw its percentage drop from 12.1% to 9.8% or a drop of 19%. The bottom quintile saw its percentage decrease from 5.7% to 4.1% or a drop of 28%.

In other words, the lower you are, the larger share of national income you have lost over the time period.

What has caused this increased income inequality? Have some people become smarter and harder working while others have just decided to rely on the teats of the welfare state sow? That would be the wingnut interpretation, I guess. A more realistic one would look at the changes over time in outsourcing and international competition in general, the failure of education to respond to the changed needs in the labor force after those changes, and the way taxes have been "simplified" to fall less on the very wealthy and more on the middle class. And we don't yet know the impact of the repeal of estate taxes on income inequality.

Time to end this post. You might do worse than by re-reading the bit about why increasing income inequality is not exactly good news, unless you like the idea of gated communities surrounded by acres of misery.
*Convexity: Imagine yourself standing on the horizontal axis and looking up towards the curve. The more it bulges towards you the more convex it is.