How to Lie about Poverty

I read an interesting column by Nicholas Eberstadt about the measuring of poverty in this country. The major point of his column is that the way in which we measure poverty is deeply flawed. He goes on to explain that, despite the apparent increase in the poverty rate when comparing 1973 to 2005, the poverty of 2005 just ain't the same as the poverty of 1973. He points out that, by many measures, the poor of 2005 are better off than the poor of 1973. I myself have often wondered whether serious abject poverty is truly as prevalent as it used to be.

However, there are serious problems with his analysis. First, we must address his bias. He's from the American Enterprise Institute, a think tank that is devoted to conservative ideals and has supplied the Bush administration with many of their public policy ideas. Certainly, the man has a major interest in showing how poverty did not increase in 2005, though oddly enough his statistics always use 2001 (2001 was a recent high point in American wealth).

Now, let's look at the premises he uses to advance his argument. He writes, "To understand the problems with the official poverty rate, compare the America of 1973 to the America of 2001. In 2001, the country's per capita income was far higher than in 1973 -- according to the Census Bureau, roughly 60 percent higher -- and unemployment rates were lower. In 2001, only one in six adults lacked a high-school diploma; in 1973, two-fifths had not finished high school. And government anti-poverty spending was more than twice as high in 2001 as in 1973."

As any student of statistics and income measures knows, looking at the average income is a horrible way to assess the wealth of a society, particularly if one is interested in the distribution of wealth or income. As income curves skew toward the right (i.e. there are some VERY, VERY wealthy people in this country), the mean is pulled toward the right. The median income (the halfway point in an income distribution where half of people make more and half of people make less) is a far better single measure of income. So, the fact that our per capita income is higher tells us absolutely nothing about the actual poverty rates (which are just a measure of the distribution of people living in poverty).

He also notes that unemployment rates were lower in 2001. Of course, as anyone paying any attention to current politics knows, the minimum wage has not been raised in roughly 10 years and a family of four living on a minimum wage income is most certainly living below the poverty line. Therefore, employment does not equal "not being in poverty". In fact, if one takes into account inflation, it's safe to say that making minimum wage now makes you poorer than making minimum wage in 1973.

In his next point, he explains that more people have finished high school in 2001 than in 1973. However, once again, that does not relate to the poverty rate. In 1973, there were far more jobs that did not require a high school degree and, thanks to unions, these jobs paid a living wage. Now, there are many fewer such jobs and, with the few that remain, the unions have been so punished by the ascendancy of the Republican party that these jobs do not pay nearly as well as they used to (inflation-adjusted, of course). So, in other words, just to keep up with inflation and the job market, it became necessary to achieve more educationally. I don't believe this is a bad thing (I'm always in favor of finishing high school and college), but it certainly doesn't mean fewer people are in poverty.

And, finally, he says that government anti-poverty spending is twice as high in 2001 compared to 1973. If the other two measures (unemployment rates and educational degrees) were unrelated to poverty, then this statement is a complete non sequitur. First off, is it twice as high with or without adjustment for inflation? We don't know because he never informs us. Assuming it is adjusted for inflation, you would think that this fact indicates that either poverty is twice as prevalent or the various governments are now twice as generous. It's hard to know. Frankly, the suggested relationship between anti-poverty spending and the poverty rate is completely opaque to me, though in a logical world anti-poverty spending would correlate positively with the poverty rate, not negatively as he suggests.

Eberstadt goes on to several more arguments that assert that the living conditions of the poor have improved since 1973. He cites better nutrition, more material belongings (such as televisions and phone), and better infant mortality rates and lifespan. Unfortunately, he doesn't always state whether these have improved for the poor or just in general for the country (e.g. it could be just like per-capita income -- more for the haves and nothing more for the have-nots). Moreover, it's extremely easy to argue that some of these are a direct result of government's anti-poverty campaigns (e.g. better nutrition through free breakfasts and lunches at schools). So, in other words, many of these "improvements" that he is suggesting indicate people are no longer poor are really just the government's support of the poor -- i.e. it's not that the poor themselves are no longer "poor".

In fact, ironically, he rarely addresses the basic definition of being poor -- a lack of money. Instead, he talks about quality of life conditions, which are somewhat correlated (but not perfectly correlated -- see Europe as an example of lower average income but better quality of life standards) with income. The closest he comes is when he discusses the purchasing of various material goods. He points out that, in reality, income tells you nothing about people's lives because economists have shown that people spend more in a year than they make. Apparently, poor people support their material belongings through credit cards and other debts. This clearly is worse than just being poor because poor people will now have to pay the usury-like interest rates on their credit cards while still actually producing very little income, but Eberstadt isn't really interested in making that point.

Based on Eberstadt's column, what can we conclude? Virtually nothing about the poverty rate, based on the evidence presented. Certainly, I must assume that, if he had better evidence, he would have produced it. And if this is the best he can produce (a tangle of misleading statistics and non sequiturs), I have to assume that his overall assertion is simply wrong. In particular, he appears to suggest that poor people are less poor than they used to be. Now, however, with such poor evidence to support that suggestion, I am forced to wonder whether the opposite is true. I do not disagree that measuring poverty is difficult or that it is essential to combatting poverty. However, the use of truth and the best available statistics is far preferable to engaging in the kinds of arguments made in Eberstadt's column.