The official poverty rate is statistically flawed, it doesn’t actually measure poverty in America
Eberstadt, 08 – senior fellow at the American Enterprise Institute (Nicholas, The Poverty of “The Poverty Rate”
http://www.aei.org/docLib/20081117_PovertyofthePovertyRate.pdf OPR = Official Poverty Rate
When we compare trends in the official poverty rate and other statistical indicators bearing upon want and deprivation in modern America, two striking and peculiar findings immediately jump out. First, the official poverty rate today does not at all exhibit what one might describe as a “normal and customary relationship” with any of the other important macroeconomic or demographic “poverty proxies” one would ordinarily wish to examine; to the contrary, against the backdrop of those other indicators, the soundings from the OPR often appear erratic, if not positively perverse. As our analysis will attest, nothing like a “normal” relationship exists today between the OPR and what should ordinarily be the main drivers of poverty reduction: economic growth; employment; education; anti-poverty spending.
Second, the mismatch between reported trends for the official poverty rate and practically all the other indicators bearing on poverty and want has been especially pronounced since 1973—the very decades for which the OPR has been reporting that the prevalence of poverty in the United States has been stagnating, or actually increasing.
Specialists on the poverty question commonly presume that the official poverty rate is strongly influenced by the state of the U.S. economy (that is to say, by macroeconomic conditions). This is not only a commonsensical premise; it is a notion confirmed in the past by empirical research. In a series of influential publications in the mid-1980s, for example, Harvard economists David Ellwood and Lawrence Summers argued that changes in the official poverty rate were mainly determined by “general economic developments.” In “reviewing trends in poverty, poverty spending, and economic performance,” they argued, “. . . it is immediately apparent that economic performance is the dominant determinant of the measured poverty rate over the past two decades.”1 Presenting a persuasive graphic that compared changes in the poverty rate and changes in median family income over time, Ellwood and Summers explained in words what the trend lines demonstrated visually: “Almost all of the variation in the measured poverty rate is tracked by movements in median family income” [emphasis in the original].2
Ellwood and Summers accurately described the correspondence between trends in the official poverty rate and real median family income— at the time of their writing. But that work was published over two decades ago. The time series data they were examining only ran from 1959 through the year 1983. Moreover, even back then they found it worth mentioning that “one does see a slight divergence of the trends in the eighties.”3 By now, that “divergence” amounts to a virtually complete disconnection. Indeed, over the decades the correspondence between the official poverty rate and median family income appears to have broken down altogether. We can see this by comparing trends since 1973 for the official poverty rate of the nonelderly population, on the one hand, and real median family income, on the other, as shown in figure 3-1. These are the same variables for which Ellwood and Summers found such a powerful association over the years 1959–83. For the period 1973–2005, there is no discernable relationship whatever between the measured poverty rate and median family income. In fact, over the past three-plus decades, the “correlation coefficient” for these two trends has been approximately zero. In other words, over the past generation, data on the median income for American families have basically provided no clue about the numbers that would be registered for the official poverty rate.4
If the true prevalence of poverty in the United States is indeed affected by “general economic developments”—as it is commonly taken to be— then this uncoupling of the official poverty rate and the median family income, their pronounced and dramatic transition to a new relationship characterized by utter randomness, should in itself raise serious questions about the reliability of the country’s primary poverty index. But this is hardly the worst of it. A quick review will confirm that the OPR has, for the period since 1973, severed what might be regarded as a commonsensical relationship with all of the other major statistical indices bearing on poverty and deprivation in contemporary America. Even worse: For the period since 1973, the behavior of the official poverty rate looks increasingly aberrant and perverse, registering retrogression when the other indices commonly record progress.
Ext #2 – Poverty Down
The official poverty rate ignores actual spending patterns by the poor, their actual income is substantially higher than what is reported
Eberstadt, 08 – senior fellow at the American Enterprise Institute (Nicholas, The Poverty of “The Poverty Rate”
http://www.aei.org/docLib/20081117_PovertyofthePovertyRate.pdf OPR = Official Poverty Rate
The OPR fails to match up with reality in other, less econometrically exacting ways as well. As initially constructed, and as it still operates to this day, the OPR assumes implicitly that a household’s reported annual pretax money income establishes its spending power for that same year; more than that, it presumes that these two quantities are identical. But reported annual pretax money income is not a reliable predictor of expenditure levels for lower-income households today; in fact, income has become an ever less faithful indicator of actual spending patterns for poorer Americans over time. As we have seen, spending already exceeded income for the poorest fourth of the American public in the early 1960s, when the now official poverty rate was being constructed. Today, however, reported annual expenditures exceed reported income for a third or more of all households in any given year—and, within the lowest fifth of households, reported spending is nearly twice as high as annual reported income.
Some part of this reported disparity may be a statistical artifact. Yet the great (and possibly still growing) discrepancy between reported incomes and reported expenditure patterns among our bottom quintile also appears to reflect genuine and important changes in the American economy, most notably the secular rise in year-to-year volatility in household income since 1973. All other things being equal, greater “transitory variance” in household incomes will mean a greater disproportion between annual spending and annual income for households at the lower end of the annual income spectrum. The material well-being of lower-income households is, of course, directly tied to their spending power—that is to say, to their expenditure patterns—but the OPR, by its very design, is incapable of tracking those same spending patterns.
Share with your friends: |