Where's the Poverty Line?

An Economic View in the New York Times talks about problems with poverty statistics:

A Poverty Line That's Out of Date and Out of Favor, by Anna Bernasek, Economic View, NY Times: ...[T]he poverty line ... has remained essentially untouched since 1963, when Mollie Orshansky, an economist at the Social Security Administration, first came up with it. Today, there is a consensus ... that it is no longer on the mark. "Everyone agrees we need a better measure," said Douglas J. Besharov, a resident scholar at the American Enterprise Institute and a professor ... at the University of Maryland. To him and many others, the public lacks an accurate test of who is poor, making it far harder to justify actions, whether by government or by individuals, intended to alleviate poverty.

No one blames Ms. Orshansky. She devised a simple and elegant measure of poverty based on the data available at the time. ... Her method was to use the cost of a nutritionally adequate diet ... and to adjust for families of various sizes and compositions. Using data from the late 1950's, she estimated that families spent about one-third of their income on food... To come up with a poverty threshold, she multiplied the cost of the nutritionally adequate diet by three. ...

The thresholds have risen over the years to reflect broad price changes, but ... much has changed since the 1950's. Today, families spend something close to 12 percent of income on food, for example, not one-third. And while some of the remaining 88 percent may go to nonessentials, items such as housing, transportation and health care are significant, and expensive, factors.

Much is at stake.  ... So why hasn't such an important statistic been updated...? The answer is politics. ... the poverty indicator, unlike many other economic statistics, is not under the jurisdiction of an authoritative statistical agency like the Bureau of Economic Analysis or the Bureau of Labor Statistics. Instead, it resides in perhaps the most political place of all: the office of the president. And during the last four decades, no president of either party has wanted to draw attention to [the] statistic ..., especially if updating it might cause the number of people regarded as living in poverty to increase. ...

But why not insist on an official yardstick that works...? Rebecca M. Blank, ... at the University of Michigan and a professor of economics, ... thinks ... some straightforward steps could be taken to improve the official indicator. ... Ms. Blank says three changes would provide much help. First, refine what's meant by income. The current gauge uses a family's pretax income. Instead, ... taxes should be subtracted and near-cash income, such as food stamps, should be added in. ... Her second step would be to update existing thresholds to take into account not only food but also housing, clothing and out-of-pocket medical costs. And finally, she said, the scales that are used to adjust for differences in family size and composition should be more precise. These steps are largely consistent with recommendations of the National Academy of Sciences report.

The Census Bureau has been tracking alternative measures ... The bureau found that if such measures had been used in 2003, ... the poverty rate would have risen about two percentage points above the official rate of 12.5 percent. That's not a big adjustment in percentage terms, but ... for some Americans living under economic stress who are not now considered poor, updating the gauge could make a tangible difference.

Perhaps the first step in resetting the safety net's altimeter is to take the poverty measure out of the White House and put it in the hands of professional statisticians.

I have no problem with the call to put this in the hands of an independent agency, nor with developing improved measures. But it would be wrong to conclude from this that existing poverty statistics should be discounted heavily or ignored because of their high degree of unreliability. On this point, here's Brad DeLong:

Brad DeLong: ...Yesterday in the Washington Post Jonathan Weisman told his readers that they shouldn't believe the Census Bureau:

Measuring the Economy May Not Be as Simple as 1, 2, 3: The Census Bureau tomorrow will release the latest statistics on poverty in the United States, the income level of an average household and the number of Americans still lacking health insurance. Don't believe the numbers.

What reasons does Weisman give to support the lead of his article--to support his warning that we should not believe the numbers in today's report?

1. David Malpass wishes that the Commerce Bureau of Economic Analysis had a change-in-wealth rather than a financing-of-investment definition of saving.
2. A number of those without health insurance could afford it, but prefer to spend their money on other things.
3. Marilyn Bryant, making $20,000 a year in Washington D.C., "earns too much money to qualify for any federal assistance" but has a very low standard of living because she lives in an expensive area and pays $5,000 a year in child support to her ex-husband.
4. Because successive White Houses have stuck with the Orshansky measure of poverty, "officially, the poverty rate has drifted upward since 2000, from 11.3 percent to 12.5 percent in 2003. But a more sophisticated measurement... shows the official rate has consistently understated poverty... the percentage of Americans below the poverty line has risen from 12.8 percent in 2000 to 14.2 percent in 2003."
5. We aren't spending enough money collecting economic statistics, and Rahm Emmanuel thinks we should set up a commission to overhaul our statistical agencies.
6. Successive White Houses have stuck with the simple to calculate and interpret Orshansky poverty measure rather than move to a more accurate but less transparent measure.
7. The Consumer Price Index is a fixed-weight index, rather than a chained-weight cost of living index--and if the government used a cost of living index to index Social Security and tax brackets, we would collect more in taxes and pay out less in benefits.
8. The 45-million count of those without health insurance is overstated by a relatively constant 5 to 10 million.
9. "Since poverty levels are not adjusted for regional costs of living, the working poor in expensive urban centers like Washington are routinely excluded from federal programs because their income lifts them above the official poverty line."

Weisman's reason (1) is simply irrelevant: that David Malpass wishes that Simon Kuznets had set the National Income and Product Accounts in a different way doesn't make the BEA's numbers unbelievable. Reason (2) is irrelevant as well: that some people who lack health insurance could afford to buy it does not make estimates of the number who lack health insurance unbelievable. Reason (3) is also irrelevant: that Congress bases program eligibility on income rather than income minus child support does not make the Census Bureau's numbers wrong. Weisman's reason (4) is simply wrong: there are more sophisticated alternative poverty measures that show higher rates of poverty and measures that show lower rates. All measures, however, do show very similar trends and patterns. Reason (5) is bizarre: we should spend more money collecting economic statistics, but that our estimates are not as detailed and informative as we would wish does not mean mean that we should not believe the numbers we have.

Weisman's reason (6) is a reason to report and consider a range of different poverty-level estimates and concepts, and to be cautious in interpreting reported levels of poverty. But it is not a reason to dismiss the trends and movements in poverty over time that the Census Bureau reports. Reason (7) is a reason to adjust real earnings estimates upward: on the most appropriate cost-of-living index basis, full-time year-round median male real earnings probably fell by only 1.5% rather than 2.3% in 2004. Reason (8) is, once again, a reason to adjust the level of uninsured--but not to doubt the trends in health insurance gaps and in the erosion of the employer-sponsored health-insurance system that we have seen over time.

Only reason (9) is truly cogent in the way that Weisman claims: differences in regional and local costs of living do make comparisions of regional and local poverty estimates unreliable, and do channel less federal poverty-support money to people living in high-cost areas.

The conclusion we should draw? Don't believe Jonathan Weisman. When Weisman says, "Don't believe the numbers" the Census Bureau released today on "poverty... income... and the number of Americans still lacking health insurance," he is not playing it straight. The numbers are quite good. The biases in their levels are relatively small. And they are accurate indicators of changes, trends, and patterns.