ECONOMIC statistics seldom rise to the newspaper's front page, and the statistical methodology of economics almost never does. Both did last week, in spades. Carolyn Bessette-Kennedy must have felt slighted.
The occasion was a report on how to improve the Consumer Price Index. The dialogue was about regression analyses and sampling patterns. (Here's Katherine G. Abraham, the Bureau of Labor Statistics' chief bean counter: "The BLS has not yet reached a conclusion about whether full or partial adoption of geometric mean aggregation to construct lower level subindexes of the CPI is warranted.")
The nut of the matter was, will they cut our Social Security payments?
Social Security is tied to the CPI, the best-known measure of price inflation. And if the CPI is inflating inflation, as most economists think it is, then the government is paying retirees too much money.
But economists have been worrying for years about how and how well the government measures business and income statistics. The digits seem as interesting as a pile of mud, but they paint a picture of the common weal and serve as trail posts for policy-makers.
Even the most basic measures of welfare are in dispute. How many poor people live in the United States? The government can tell you, and it does. President Clinton took great credit for the fact that the "poverty rate" declined from 14.5 percent of the population to 13.8 percent last year.
"More and more of our people are sharing in that prosperity," Clinton said last September.
Are they really? The method of computing the poverty rate, solemnly announced by the Census Bureau each year, is unchanged from when the rate was first published 36 years ago.
Census officials at that time figured that the typical family spent a third of its income on food. Analysts then created a "basket" of essential groceries needed to feed a family for a year. And they multiplied the cost of the basket by three to get the poverty "threshold."
Is your income below the groceries-times-three sum, which was $12,278 last year for a family of three? You're poor. Is it more? Maybe $15,000? You're not.
In 1960, this reckoning of the poverty was innovative and meaningful. In 1996, it is primitive and anachronistic.
The federal poverty accounts make no allowance for the fact that rent, utility and transportation prices have risen faster in three decades than the price of food. Typical families now spend only about a fifth of their income on food. The grocery basket should be multiplied by five to obtain a subsistence income, not three.
And even an income of five times food costs might not be enough to live on. Poor people pay taxes, but the census poverty formula doesn't account for payroll taxes. If payroll tax costs were added to the minimum income threshold, the 1995 poverty rate would have been 14.6 percent, not 13.8 percent, said Leatha Lamison-White, a statistician with the Census Bureau's poverty and health branch.
But it's even more complicated than that. The poverty accounts ignore Medicaid, Medicare, welfare payments and other subsidies that effectively raise the incomes of recipients and might push some above the poverty line. "People might live in public housing or subsidized housing," said Lamison-White. "They might get free or reduced-price school lunches, food stamps, all the noncash benefits that the government might give them."
The poverty ledger makes no regional distinctions, either. The government assumes that $12,278 goes just as far in Manhattan, N.Y., as it does in Manhattan, Kan. And the Census Bureau's poverty branch, too, relies on the Consumer Price Index, using it to update the 1960s grocery basket each year. And, now you know, the CPI itself has problems.
The National Research Council has worked with a study panel and published a book, "Measuring Poverty: A New Approach," which discusses the problems. The work hasn't provoked the same attention as the CPI commission's, perhaps because the lobbying and PR budget of the Coalition on Human Needs isn't quite the same as that of the American Association of Retired Persons.
The Census Bureau is still figuring out what to do. But many economists believe that the current gauges give American poverty short shrift, even when welfare and Medicaid benefits are factored in.
"I think we're under-measuring it," said Cecilia Conrad, associate professor at Pomona College in California and a specialist in the economics of poverty. "I would argue it's about 2 to 3 percentage points off," putting the real poverty rate for 1995 as high as 17 percent.
A family of three earning $12,278 a year might tell you that seems conservative.
Pub Date: 12/09/96