Radio commentary, September 15, 2010
[Been a while since I posted one of these. Too many things to do, too little time. Sorry. Much more, including graphs, in the forthcoming LBO.]
Thursday morning brought the release of the Census Bureau’s annual income, poverty, and health insurance numbers. These are drawn from a special edition of the Bureau’s monthly Current Population Survey. The regular survey, which covers about 60,000 households, is what the monthly unemployment figures, among other things, are based on. This special survey, done every March, covers 100,000 households. This is a very large sample, though it’s far from perfect, as I’ll say in a bit.
The headline findings are that incomes were down, poverty was up, and millions lost their health insurance last year. This is just what you’d expect in a recession, of course. A few words about each.
First, income . The median income of all households—the income level at middle of the distribution, meaning that half the population is richer, and half, poorer— fell by almost 1% last year, but the move was deemed not large enough to be statistically significant. That is, the likely error in the estimate is larger than the change itself. But the overall number was helped considerably by a near-6% rise in incomes for households headed by someone 65 or over. (The concept of a head of household—usually the person’s name on the deed or lease—is problematic, but it’s not my terminology.) Under-65 households saw their income fall by 1.3%, and this did pass the test of statistical significance. Especially hard hit were the young, foreign born non-citizens, and blacks. So-called Hispanics held their own. Curiously, residents of central cities did a lot better than suburbanites or country-dwellers.
Taking a longer-term look, median household income in 2009 was just 2% above where it was in 1989, even though real GDP was up 63% over the same period. Of course, population is up about 25% over the period, and not all of GDP takes the form of personal income—some of it goes to corporations, some to government. But real disposable income per capita—the total of income after taxes throughout the entire economy divided by the population—is up 40% over the last twenty years. So, on paper, or its silicon equivalent, were things being distributed equally, the average household would have gained about 20 times more income than it has in reality. The reason for this, of course, is that the rich have gotten most of the benefits of economic growth over the last few decades. The top 5% of the population, according to the Census numbers (which badly understate things, for technical reasons I’ll go into in a moment) has gotten about 25% of the growth in income over the last twenty years—the top 20% altogether has gotten just over half, 54%.
There are several reasons why the Census numbers understate incomes at the top. One is that the very rich can’t be bothered answering questionaires from pesky enumerators. And another is that the Census Bureau treats all incomes over a certain amount—it rises every year, but it’s around a million dollars now—as if they were that amount. The stated reason for this is to protect confidentiality, since the records are available for public use. Maybe. But the effect is seriously to understate how well the very rich have done. For example, the Census shows the top 5% gaining 28% from 1989 to 2009. Work based on tax records done by the economists Thomas Piketty and Emmanuel Saez show that to be roughly true for households at the 95th percentile—that is, those richer than 95% of the population. But by missing the seriously rich, the Census understates things by almost 2/3: Piketty and Saez have the top 5% gaining almost 75% from 1989 to 2007. (Sadly, 2007 is as far as their data goes, but it probably wouldn’t change much if you took it out another couple of years.) The reason for the difference is the action at the high end. The top 1% was up over 100%. The top 0.01%—the 12,000 or so richest households, with incomes averaging $35 million a year—were up 215%.That almost 30 times the increase of the bottom 90% of the population. In other words, an enormous portion of the gains of economic growth have gone to just a few thousand hyper-rich.
All that said, the Census numbers are a good measure of broad economic trends, even if they miss life in the stratosphere.
A few other notes on income. The average black household’s income fell to just below 60% of the average white household’s, down nearly 2 percentage points from last year, and extending a downtrend that began when the boom of the 1990s burst in 2001. At its 2000 peak, the average black household hit 65% of the white average. The average Hispanic household’s income was just below 70% of its white counterpart—down a couple of points from a few years ago, but basically in the neighborhood it’s been in for years. The average Asian household’s income was 120% of the white average last year, also more or less where it’s been for several years.
Now, poverty . Before proceeding, it’s important to say that the U.S. poverty line is an extremely stingy thing. It’s based on research done in the 1950s that showed that the average household spent a third of its income on food. So, the Johnson administration, eager for metrics in its war on poverty, decided that a poverty line would be three times a minimal food budget computed by the Agriculture Department. Never mind that that food budget was considered something of an emergency measure, not something to live on. They needed something in a hurry and went with that. And they’ve just adjusted that line for inflation ever since, with no notice paid to rising GDP or average incomes, or the changing nature of household budgets (like medical inflation or the need for child care). So, conceptually, a poverty income today is exactly the same as it was almost 50 years ago, even though average incomes have more than tripled. A more honest poverty line would produce numbers probably twice what officialdom reports.
Yet despite that undemanding standard, 14.3% of Americans were officially poor in 2009, up from 13.2% in 2008, the highest level since the early 1990s. It’s almost certainly up this year, since it tracks the unemployment rate pretty closely. The rate among white households was just over 9%; for black and Hispanic households, over 25%, or nearly three times the white rate. It was almost that high for households of any race with children under 6.
And the percentage of people without health insurance for the entire year—and not just a spell, so these are lowball numbers—rose more than a percentage point, to 16.7%—or 51 million people. Since people over 65 are covered by Medicare, giving them near-100% coverage, it’s worth looking at the under-65 crowd: almost 19% of them were without insurance for the whole year. The share rises to 30% for young adults, those under 35.
So, a bad year for the mass of the American population, and it’s likely that 2010 was equally bad or worse. And this news will be reported for a day or two, and then forgotten, as the TV moans about the sufferings of the rich, who desperately need their tax breaks renewed.
The use of household income by the Census is a bit problematic as well. There are significant differences in household sizes among our made-up ethnic groups in the U.S.
And in general, the Census’ use of racial categories is increasingly concealing rather than illuminating. For instance: lumping together Indians (most highly educated in the case of US immigration) with Cambodians (a generation removed from refugee status) is almost deliberately a way to conceal class issues – in this case, under the Asian banner.
The Financial Times says Poverty among those aged 18 to 64 it’s a 50-year high.
To what do you attribute the greater income jump for principal city residents?
Dunno. NYC has done pretty well in the recession. Don’t know about other cities.
Just curious, what do you make of the Census reporting data for non-Hispanic whites as well as for all whites? Not surprisingly, non-Hispanic whites on average earn higher incomes and have lower poverty rates. I realize that race and ethnicity intersect in some pretty profound ways, but it seems Census data reporting has created a purer form of whites by distinguishing between Hispanics and Nots. And you won’t find data on non-Hispanic Blacks – odd if the goal was to help keep track of social outcomes for the so-called Hispanic population.
It all gets as complicated as medieval theology, but about 13% of (self-identified) white households are also (self-identified) Hispanic. (Who knows what the composition of those households is? Could be the “head of household” identifies as Hispanic, yet has a “white” spouse. Dunno.) Median household income for non-Hispanic whites is about 5% higher than the broader category. There are some Hispanics that also identify as black, but I don’t think it’s all that many. There’s also the whole multiracial issue, which complicates everything.
I hadn’t even thought about establishing the race/ethnicity of households. That’s gotta be problematic.
Since I have the 2009 P60 income and poverty report in front of me, I can do some quick population numbers from their tables.
Total pop for whites is 190,940 million non-Hispanic whites and 43,608 million white Hispanic or some other ethnic identifier (that’s not European). I would assume the bulk of those 43 million are Hispanic. The base number for the CPS is 47.398 million Hispanics so assuming most of the 43,608 million are white leaves 3,790 of some other race (presumably Black).
Now after this futile effort at calculating today’s residue of the one-drop rule, what to make of this non-Hispanic white category? The thing is, I don’t think the Census is doing this slicing and dicing to key us in how the current premium whiteness gives whites who just happened to be not Hispanic.