Front Page Fiasco!

NYT's wacky "poverty line" story.

home

archive

masthead

contact

Posted Wednesday, October 20, 1999

        [Note: You can sign up for the free kausfiles.com e-mail service at the end of this column.]

        A strange, confusing story, pushing what looks like a not-so-hidden liberal agenda, appeared on the New York Times' front page Monday. Let's feed the piece -- "DEVISING NEW MATH TO DEFINE POVERTY" -- through the Snipe-O-Meter and see what turns up. As usual, boldface text quoting the story is followed by context and perspective from the kausfiles.com staff:

        "The Census Bureau has begun to revise its definition of what constitutes poverty in the United States, experimenting with a formula that would drop millions more families below the poverty line. The bureau's new approach would in effect raise the income threshold for living above poverty to $19,500 for a family of four ... Suddenly, 46 million Americans, or 17 percent of the population, would be recognized as officially below the line, not the 12.7 percent announced last month ...."

        These lead paragraphs from NYT reporter Louis Uchitelle are misleading, or worse. The Census Bureau has been putting out experimental poverty lines for 15 years. Uchitelle is apparently referring to a July report, authored by Kathleen Short, that outlined 12 different possible definitions of poverty along lines recommended in 1995 by a National Academy of Sciences (NAS) panel. There's no evidence, and Uchitelle provides none, that the bureau has actually endorsed any one of these experimental definitions. (The bureau responded to Uchitelle's story with a same-day press release stating that "none [of the experimental definitions] has been selected to replace the current definition," and noting that the study of alternative definitions "began some years ago and will not be completed for several more years.")

        Uchitelle does highlight a bland quote from a Census official, Edward Welniak, who says "It is certainly our opinion . . . that something should be done to update the poverty measure." But here Uchitelle confuses two distinct issues: (1) whether what is counted as income should be changed to make it more accurate in reflecting the resources poor people actually have; and (2) whether the amount of income that's considered too little (the poverty line) should be raised. Was Welniak talking about issue (1) or issue (2)? There's no reason --except the way Uchitelle has strategically placed the quote -- to think he was talking about issue (2), and he almost certainly wasn't, since that's not the Census position. And when it comes to issue (1) -- accuracy -- you can "update the poverty measure" so there are fewer people counted as poor (e.g. if you count as income government benefits such as food stamps) or more poor people (e.g. if you deduct the Social Security payroll tax). Welniak's quote doesn't support one sort of update over the other.

        Actually, it's not the Census Bureau's job to revise the poverty line anyway. That duty officially falls to the Office of Management and Budget, part of the White House. Is there a move within OMB to raise the poverty line? Quite the opposite -- OMB is widely seen as dragging its feet, and crack OMB analyst Richard Bavier has written a memo criticizing parts of the NAS report. Uchitelle himself quotes an "administration official" as saying "we have at least a couple of years more work to do" before the poverty measure is revised. Translation: The poverty line ain't gonna be raised; the entire premise for this story is bogus.

        And which of the 12 experimental measures did Uchitelle anoint when he came up with his magic $19,500 figure? He didn't return my phone call, so I don't know. His calculations don't seem to match any of the experimental measures.

        Uchitelle here implies that a) setting the line at which Americans should be deemed "poor" is a scientific question for "sociologists and economists," as opposed to a value judgment to be made democratically; and b) there is a unanimous scholarly consensus that this scientifically determined point should be "between $21,000 and $28,000." Both implications are unsupported by the piece, and are indeed rubbish. Uchitelle doesn't manage to name a single one of the presumably numerous "sociologists and economists" who share in this "new thinking." (I'll name one who told me he doesn't: John Cogan of Stanford, an economist on the NAS panel.)

        "A strong economy has undoubtedly lifted many families, but not nearly as many as the official statistics suggest."

        This is the third paragraph of the piece. What's it doing there? You get the impression some editor stuck it in to make the piece look like a smart response to the positive poverty report issued three weeks ago. But there is nothing in the subsequent 3-4,000 words that supports this sentence -- no comparisons of how many have been lifted out of poverty under the "old" poverty line versus under the "new" higher, experimental poverty line, for example. In fact, the economy has almost certainly lifted more people out of poverty if you use a higher poverty line, since that line hits nearer to the fat center the income distribution "bell curve." All other things being equal, a general gain is likely to push a whole lot more people over a $28,000 line than a $16,000 line.

        "But a higher threshold means government spending would rise to pay for benefits tied to the poverty level . . . That would require an incursion into the budget surplus ...."

        Another bogus paragraph. Spending on benefit programs is tied to poverty thresholds set by the Department of Health and Human Services, not to the Census' thresholds. It would be entirely possible -- though it might require some careful draftsmanship -- to raise the Census poverty line without raising benefits. Nor does it appear to be the intention of those lobbying for an updated poverty measure -- such as Gary Burtless of the Brookings Institution -- to trigger an automatic benefit hike.

        "For Crystal McConnell and Austin Johnson, an Indianapolis couple, the poverty line, in their view is definitely north of the $22,250 they expect to earn this year ...."

        Uchitelle's point, it seems, is that we should think of many people who are above the current poverty line as "poor." Hence he devotes a few paragraphs describing the real-world plight of the McConnell/Johnson family. Theirs is not a gripping story. They "do not seem poor," Uchitelle admits. They live in a well-kept home in a well-kept neighborhood, with toys in a living room "which is furnished comfortably." Their son attends "a good public school." But they occasionally eat lunch at Mr. Johnson's mother's house to stretch their food budget. Sometimes dinner too! Mr. Johnson, who is 24, won't get health insurance until his provisional job as an airline load planner becomes permanent; Ms. McConnell is insured, but "last year" she wasn't. She owes a hospital $650. The couple also hopes they can delay paying their utility bill for half a month without getting cut off (something kausfiles.com does too on occasion). That's it!

        Does this striving family have so little money that it is not living in dignity, or not participating fully in American life? It doesn't seem that way. Uchitelle tosses in some laughably banal stats -- "The poor are twice as likely as the nonpoor to rent rather than own their own homes." (Duh! And so what!) He also notes the finding of sociologist Kathryn Edin that when the poor single mothers she interviewed didn't have enough money for "the occasional trip to the Dairy Queen, or a pair of stylish new sneakers" they borrowed money from family members, etc. and made those purchases anyway. But does that argue for raising the poverty line, or for keeping it where it is?

        There is a strong case that the Census definition of income should be updated to include government benefits -- such as food stamps -- that have become society's prime weapons for fighting poverty. That might benefit liberals (e.g. when fewer eligible people get food stamps, as is happening) or conservatives (when counting benefits makes people seem richer.) There is much less support within the government for the cause Uchitelle outlines -- simply raising the poverty line -- though somebody could, unlike Uchitelle, make a good argument for that too, on the grounds that "poverty" is relative. ("Not to be able to afford a movie or a glass of beer," Dwight Macdonald once wrote, "is a kind of starvation --if everybody else can.")

        Yet by hopelessly conflating these two causes, Uchitelle's article has hurt both of them. He has made it look as if updating the poverty measure for accuracy's sake is really part of a liberal/media plot to raise benefits and reclassify a vast new swath of America as "poor" and needy. ('Want to increase poverty? Give a liberal a calculator!') Journalists and policy analysts who study what newspapers must do to stay credible say they can hardly remember a shoddier front-page effort.

        New E-mail service: Sign up, using the ListBot gizmo below, and you will be notified by e-mail whenever there's a new item on kausfiles.com. [Note: this service is free. You'll be asked a couple of demographic questions; if you find them annoying just leave them unanswered.]

        Recently archived:

        Just a Few Tweaks Beatty's still building his bomb.