Poverty could be measured relatively or absolutely. A relative measure of poverty would be some level of consumption such that some specified proportion, say 80 percent, of the population had a higher level. This has a certain reasonableness but such a definition would mean there will always be 20 percent of the population that is in poverty.

In order to measure progress in reducing poverty, poverty has to be measured in reference to an absolute standard. The problem is what should be the poverty level of consumption. The poorest in the United States are relatively well-to-do in comparison to the average consumers in most places in the world or to the advanced world of several centuries ago. In America most poor families have color television sets, something that the wealthiest families of past could not have.

Establishing the absolute level of consumption or income for purposes of defining poverty is a task which is probably not objectively solvable. The definition of poverty-level income that is used for statistical purposes by the Federal Government was formulated by Mollie Orshansky of the Social Security Admininistration in 1961. Orshansky's definition is based upon the proportion of income that families spend on food. She wanted to find the level of income for families of different sizes such that the family was voluntarily purchasing enough food to meet the basic nutrition requirement of the family. The Department of Agriculture calculated for various family sizes the cost of an adequate diet. It was known from surveys that a typical family spent about one-third of its budget on food. Orshansky multiplied the computed cost of the adequate nutritional diet by three to get the poverty line or Official Povery Threshold (OPT). As one can see from the definition this OPT is a rather arbitrary figure and not too much significance should be attached to its value. The poverty rate is the proportion of families or single individuals with incomes below the Official Poverty Threshold. From the arbitrariness of the computation of the OPT one can see that the poverty rate is also rather an arbitray figure. About the only real validity that can be attached to the poverty rate statistics has to do with the trends. The the arbritrarily defined poverty rate is declining then a poverty rate based upon any other definition of the poverty line would also be declining. Despite the arbitrariness of the poverty statistics the news media and political figures treat them as thought they were based upon some scientific standard.

Even aside from the limitations of the official poverty threshold the poverty rate is based just on a head count and doesn't distinguish between people who are almost at the poverty line and people who are far below it. A statistic formulated to remedy this problem is the poverty gap or income deficit, which is the amount of income total that would be needed to bring everyone up to the official poverty threshold.

In 1994 the Official Poverty Threshold for a family of four was $15,141 per year or $1262 per month. Clearly $1262 per month means a lot different in the rural South than it does in the urban North. But based upon that OPT the poverty rate in 1994 was 11.6 percent of the families and 14.5 percent of the population. This poverty rate represents 8.1 million families containing 38.1 million persons. The poverty gap in 1992 was $54.8 billion.

There are many problems with the definition of income used in the statistics. Current income may not be an adequate measure of peoples long term income. The people in the lowest quintile (20%) of income are not necessarily permanently there. A study done by the Treasury in 1979 on taxpayers over a ten-year period found that one third or fewer of taxpayers in the lowest quintile of income in 1979 were still in that category ten years later. More people moved to the highest quintile ten years later than remained in the lowest quintile. For example, someone who took a year or so off from work to get additional training would show up in that lowest quintile while getting the training and then jump back to a high quintile. The income that should be used for distribution of income analysis is long term or what Milton Friedman called permanent income. The best measure of this would be the level of consumer spending.

The income used for determining the poverty rate is cash income, which includes means-tested cash transfer payments but not in-kind food and housing payments. Another problem with the income measure is that it is reported income, whereas for the poor there is a lot of their income that is unreported. It is unreported because of among other reasons, income taxes and payroll taxes.

The official poverty statistics tend to deter people from developing the right statistics. For example, there should be evidence from statistical surveys on what proportion of children are not getting adequate nutrition. Children may not be getting adequate nutrition because of low income or because of unwise spending habits of their families, including drug and alcohol abuse. The crucial problem is not poverty per se but inadequate nutrition and health care of children. Unhealthy excess weight is fairly common in poor families. The official poverty statistics divert attention from the real problems by making people think the government is keeping close tabs on poverty.

With all of the above questions of validity now we can look at the statistics for the poverty rate over the past four decades.

(To be continued.)