The crux of the debate over the proposed new statistics is tbe purpose of measuring poverty. As originally conceived, the poverty statistics were meant to be diagnostic. They emerged in the mid-1960s as a benchmark for President Johnson‘s “war on poverty.” What Americans wanted to know then—what they should still want to know today—is whether they’re reducing tbe number of families struggling to obtain the basic necessities of life.
The answer is yes. A recent Heritage Foundation study examines the incidence of the bedrock problems of poverty—malnutrition, crowded housing and lack of access to medical care. It concludes that 8.7 million Americans, or just 3.7% of the population, make up the nation’s “hardship population”—the truly poor.
In 1993, University of Texas economist Daniel Slesnick recalculated the poverty rate based on spending rather than income. To remove the vagaries of inflation, he established the poverty threshold at three times the cost of a nutritionally adequate diet for all members of a household. Mr. Slesnick’s results show that the proportion of poor in the U.S., measured by consumption, has fallen steadily, from 31% in 1949 to 13% in 1965 to 2% at the end of the 1980s.
It’s not hard to discern the political agenda of those who want to conjure up another 12 million poor people. Having more poor families enlarges the constituency for programs that dole out money to the poor. But if it’s simply a matter of deciding which families are eligible for government programs, then the issue really comes down to how much American’s are willing to sacrifice to the insatiable god of equality.”
My (closing) comments:
(1) Poverty is always relative, and its definition is always a search for what the better-off regard as acceptable rather than a search for objective truth. Opponents of a reduced welfare state, like Cox and Alm, should rather accept that relative standard, rather than confuse the debate with some absolute arguments. For example, a Dutch poverty debate in the early 1900’s was about whether a table would be part of household necessities or not. Defining poverty as three times the grocery bill would surely answer that question. But it is more likely that society’s standard would start including air-conditioners too (by some regarded as the most important invention this century).
(2) One of my main arguments is that society even tends to update poverty with the general level of welfare. That the US has been using only the CPI would counter that argument. But that the CPI has been overstated, that all kinds of provisions like Medicare have been added for purchasing power, and that one is experimenting with a serious update, is supportive again. Similarly, Cox and Alm p201 even state “What were once luxuries are now viewed as necessities”. It would be better to make welfare indexation the official line, and stick to it.
(3) The political argument given by Cox and Alm is doubtful. The few votes of the new beneficiaries may well lose out against a huge majority that could be against the proposals, including the current beneficiaries. Why start the whole discussion about democracy again ?
(4) Poverty definitions, though relative, nevertheless should be as sound as possible. If wealth is not properly accounted for, as Cox and Alm point out, then the debate gets noisy, and popular support for the poor indeed suffers. (Even though the 302,000 families with expensive homes are only a fraction of the 13 million real poor.) Similary, implementation of anti-poverty policies will often be very murky. (‘Did you really try to get a job - and shouldn’t we not take you from the programme ?’) There is no alternative but to accept this murkiness, and try to instill operations managers with the spirit that they should try for a good performance anyhow.
(5) To clarify the argument, to get rid of some of the murkiness, I myself take a stylized approach. Then we don’t bother with the question whether air-conditioners are part of household necessities. We assume some historic subsistence and exemption level, and then work through the arguments of indexation etc. This thus eliminates much of the need of statistical measurement.