The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
April 15,1995 eCONOMIC GOMMeiMTCIRY Federal Reserve Bank of Cleveland Growth and Poverty Revisited by Elizabeth T. Powers I n the 1960s, economic growth seemed to be the tonic for poverty. From 1960 to 1969, real gross domestic product (GDP) grew at a 4.1 percent average annual rate, while the percentage of all persons in poverty declined at an annualized 5.9 percent. Had the 1960s' relationship between real GDP growth and poverty reduction been maintained, the share of all Americans in poverty, or the "poverty rate," would have been only 8.2 percent by 1992. The actual 1992 poverty rate, however, was nearly twice as large.1 A significantly weaker relationship between poverty relief and aggregate growth in the decade following the 1960s is understandable. In contrast to the long uninterrupted expansion of the 1960s, economist Rebecca Blank notes that "... the thirteen years between 1970 and 1982 contained four business cycles,... five years of negative GNP growth, and a rapid increase in both inflation and unemployment. It is perhaps not surprising that poverty is less responsive to short and sequential upturns and downturns in the economy."2 What is surprising is the apparent lack of progress against poverty throughout the 1980s. In a decade that featured one of the longest peacetime economic expansions in U.S. history, the share of all persons in poverty actually rose moderately. It appears that the benefits of robust economic growth "trickled down" to the poor in declining measure. A more formal economic analysis seems to confirm these simple findings. During the 1960s, a 1 percent increase in the annual growth of real GDP was typically accompanied by a 0.4 percentage-point reduction in the poverty rate (see table 1). In the 1970s, however, the effect of GDP growth on poverty reduction was slightly more than half that amount, and this weaker relationship persisted throughout the 1980s. What caused the formerly strong correlation between output growth and poverty reduction to break down? Economists have offered several plausible explanations, including demographic shifts, policy changes, and increased wage and income inequality. This Economic Commentary explores each of these notions, then considers an alternative view: When poverty is defined based not on annual income but on what is actually consumed in a year, there is no discernible break in the "trickle-down" mechanism. • Why a Breakdown? Demographics A significant demographic shift that might have affected income poverty rates in the 1970s and 1980s was the surge in divorce and out-of-wedlock births. Because female-headed households typically have the highest poverty rates of all family types, growth in this population translates strongly into poverty growth. By 1989, slightly more than half of all poor families were headed by a single woman. The shift toward more female-headed households Official poverty rates remained stubbornly high throughout the 1980s despite a prolonged expansion. Much has been made of this apparent breakdown in the poor's share of economic gains. But when poverty is measured based on consumption rather than on annual income, a different picture emerges. Comparing growth in various macroeconomic indicators with changes in consumption poverty rates reveals no change in the "trickle-down" mechanism in the 1980s. was largely completed by the late 1970s, however, leaving us without an explanation for the persistence of poverty rates in the 1980s. Single women with children may also be less responsive to changes in overall labor market conditions. The "feminization" of poverty would thus tend to imply that the growth solution to poverty is becoming increasingly inappropriate. However, while the share of female-headed households in poverty was increasing during the 1960s and 1970s, the representation of the elderly—a group that is even less responsive to labor-market changes—was in dramatic decline, partially mitigating this effect. Still, the proportion of the poor accounted for by these two groups rose from slightly more than 30 percent to nearly 50 percent from 1959 to 1969, and has hovered around 50 percent since then. This pattern is consistent with the decreased responsiveness of poverty to output growth in the 1970s and 1980s. Even within demographic groups, however, the value of growth as a tonic for poverty has declined over time.3 This suggests that the experience of the 1980s really reflected a fundamental structural change in the way aggregate growth translates into individual well-being. It also implies that aggregate growth would have become a less effective poverty-fighting strategy in the 1980s even in the absence of significant demographic trends.4 Government Transfers Changes in government transfer policies (such as Aid to Families with Dependent Children) appear to be a more promising explanation. Reforms initiated by the Reagan administration in 1981 made welfare participation less attractive, which may have resulted in lower incomes for some of the poor and near-poor. Real transfers grew at an average annual rate of 6 percent in both the 1960s and the 1970s, but at 2 percent in the 1980s. Such policy changes may have offset gains to the poor from overall economic growth, consistent with the persistently high levels of poverty in the 1980s. Even accounting for demographic and policy factors, though, the qualitative differences between the 1960s, 1970s, and 1980s persist. The second column of table 1 reports findings after correcting for the growth in real government transfers and the population of female-headed households. As expected, excluding these factors caused us to somewhat overstate the impact of aggregate growth in the 1960s and to somewhat understate its impact in the 1970s and 1980s. However, the effect of growth on poverty in the 1980s continued to be small relative to the experience of the 1960s. Widening Wage and Income Inequality Several scholars have pointed to a widening disparity in wages, including real declines in the wages of low-skilled workers, as the primary cause of the weakening relationship between growth and poverty reduction.5 According to this theory, as the rewards to those at the TABLE 1 EFFECT OF GDP GROWTH ON INCOME ^ | ^ g POVERTY RATE Percentage-point reduction in poverty for 1 percent increase in GDP growth Adjusted for demographic and policy trends -0.403 -0.220 -0.228 -0.381 -0.252 -0.247 1960-1969 1970-1979 1980-1989 NOTE: Reported coefficients are statistically significant from zero. Coefficients for 1960-1969 are significantly larger than coefficients for 1970-1979 and 1980-1989. SOURCE: Author's calculations using data from the U.S. Census Bureau and Economic Report of the President. FIGURE 1 ALTERNATIVE POVERTY RATES, 1959-1989 Percent of persons in poverty 23 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986 1989 SOURCES: U.S. Census Bureau; and Slesnick (footnote 6). bottom of the earnings ladder rapidly diminished, aggregate growth decreasingly translated into poverty reductions. The conclusion is that beginning in the late 1970s, a structural change in the economy led to increasing income inequality and to a decreasing ability to "grow out" of poverty. income, but also on anticipated income in all future years, accumulated wealth (acquired from past income), and the degree of flexibility to reallocate these resources by borrowing and saving. We can infer consumption levels by making adjustments to available data on household expenditures. • The Mismeasurement of Poverty Figure 1 contrasts the official income poverty rate from 1959 to 1989 with the comparable Jorgenson-Slesnick (JS) consumption poverty rate.7 The JS rate uses a consumption poverty threshold, computed to be logically consistent with the official income poverty threshold. Both the levels and trends of the two rates are quite different. Except for a period in the late 1960s, the official poverty rate is higher than the JS rate. Although both rates decline from 1961 to the early 1970s, they paint starkly different pictures of recent poverty trends. Recent work by economist Daniel Slesnick suggests that flaws in poverty measurement have led to misleading conclusions about poverty trends.6 The problem is that annual income is often an imperfect indicator of economic wellbeing. For example, an elderly family may be income poor but asset rich, and thus able to maintain much the same lifestyle as a well-to-do younger family by spending their retirement savings. This suggests that a much more accurate indicator of the economic well-being of families is their consumption. Consumption depends not only on current annual TABLE 2 EFFECT OF GDP AND PCE GROWTH ON CONSUMPTION POVERTY RATE Percentage-point reduction in poverty for 1 percent increase in GDP growth Percentage-point reduction in poverty for 1 percent increase in PCE growth -0.232 -0.236 -0.190 -0.320 -0.330 -0.306 1960-1969 1970-1979 1980-1989 NOTE: Reported coefficients are statistically significant from zero. Coefficients for all periods are not significantly different from each other. SOURCE: Author's calculations using data from the U.S. Census Bureau and Economic Report of the President. While the official rates indicate a strong resurgence of poverty in the early 1980s and persistently high rates thereafter, the JS rate, after a sharp increase around the recessionary trough in 1980, declined substantially over the 1980s. in consumption poverty of 0.32 in the 1960s, 0.33 in the 1970s, and 0.31 in the 1980s. Again, there is no statistical evidence that the relationship between growth and consumption poverty has changed since the 1960s.8 • Aggregate Growth and Consumption Poverty • Interpreting the Findings The first column of table 2 reports the estimated effects of real GDP growth on the consumption poverty rate. A 1 percentage-point increase in the growth rate of real GDP is associated with a reduction in consumption poverty of about 0.23 percentage point in both the 1960s and 1970s, and about 0.19 percentage point in the 1980s. The estimated effects in each decade are so similar that we can reject the idea of a structural change in the trickle-down mechanism in the 1980s. As stated earlier, scholars have argued that most of the stagnation in progress against income poverty can be attributed to structural change, not to more easily understood causes such as a regime change in policy or an obvious demographic trend. However, taking a fresh look at the way we measure poverty leads to the conclusion that the alleged break in the trickle-down mechanism is quite sensitive to the poverty concept employed. Using a measure that is on a superior conceptual footing, there seems to be practically as much scope for growing out of poverty as ever. Real GDP is one of several alternative macroeconomic growth indicators. In the income poverty context, its use amounts to examining the degree to which the income-poor population partakes in aggregate income growth. Analogously, we can compare real personal consumption expenditure (PCE) growth with consumption poverty. The question is, were the consumption gains of the 1960s, 1970s, and 1980s decreasingly reflected in the consumption of the poor over time? Why do these two measures yield such different interpretations of the past 30 years? One conjecture is that the trend toward female-headed families drove up income, but not consumption, poverty rates over the 1970s, and accounted for some of the failure of aggregate income gains to redress income poverty.9 However, this cannot explain the different conclusions about poverty and growth employing these two measures in the 1980s. The second column of table 2 presents the estimated effects of real PCE growth on the JS poverty rate. Aggregate PCE growth of 1 percentage point was associated with percentage-point reductions Policy changes may also have differentially affected income and consumption poverty rates. For example, Medicaid has come to represent an increasingly important share of welfare benefits. While the value of Medicaid insurance and actual benefits paid are not reflected directly in either the standard (income) or JS (consumption) poverty rate, such transfers may indirectly affect the consumption poverty rate. The increasing availability and generosity of Medicaid, particularly in the 1980s, may have caused people to worry less about potential medical costs and to spend funds formerly set aside against such contingencies. These types of policy changes would show up in lower consumption poverty rates (through increased consumption), but would have no effect on (official) income poverty rates. Thus, it is possible that a breakdown in the growth effect on consumption poverty was obscured by mitigating changes in policy. Finally, recent research suggests that up to half of the greater earnings inequality observed from the late 1970s through the 1980s was unrelated to fundamental increases in inequality.10 Measured annual and permanent income trends seem to have diverged. Because theory suggests that consumption is more closely related to permanent income, this phenomenon could result in increased income, but not consumption, poverty rates. This would help to explain why the two poverty series are so divergent over the 1980s. A first look at the consumption poverty data suggests at the very least that the picture of increasing inequality and the failure of the 1980s' "rising tide" to lift all boats may be exaggerated. The poor appear to benefit from a vigorously expanding economy now as much as before. After considering this new evidence, it is not at all clear that a policy to promote overall growth (and hence to expand aggregate consumption opportunities) is of declining interest to Americans below the poverty line. • Footnotes 1. The official government poverty rate, computed for all years since 1959, is the proportion of the population with annual pre-tax, post-transfer income below specified levels, called "poverty thresholds" or "poverty lines." 2. See Rebecca Blank, "Why Were Poverty Rates So High in the 1980s?" in D.B. Papadimitriou and E. Wolff, eds., Poverty and Prosperity in the USA in the Late Twentieth Century, chapter 2. New York: St. Martin's Press, 1993, p. 27. 3. Ibid. 4. Even so, promoting growth might still be a more effective long-run poverty reliever than other policies. 5. See, for example, Blank (footnote 2) and David Cutler and Lawrence Katz, "Macroeconomic Performance and the Disadvantaged," Brookings Papers on Economic Activity, 1991 No. 2, pp. 1-61. 6. See Daniel Slesnick, "Gaining Ground: Poverty in the Postwar United States," Journal of Political Economy, vol. 101, no. 1 (February 1993), pp. 1-38. Federal Reserve Bank of Cleveland Research Department P.O. Box 6387 Cleveland, OH 44101 Address Correction Requested: Please send corrected mailing label to the above address. Material may be reprinted provided that the source is credited. Please send copies of reprinted materials to the editor. 7. Daniel Slesnick developed the basis for his poverty computations with Dale Jorgenson. For further details, see Slesnick (ibid.). 8. The careful reader might be curious about the estimated relationship between PCE growth and income poverty. A 1 percent increase in real PCE growth is associated with poverty reductions of-0.44, -0.24, and -0.27 percentage point in the 1960s, 1970s, and 1980s, respectively. There is no statistical difference between the 1970s and 1980s estimates, while the latter periods are significantly different from the 1960s. These findings are strikingly similar to those obtained using real GDP growth. Elizabeth T. Powers is an economist at the Federal Reserve Bank of Cleveland. The author thanks David Altig for helpful comments and Kristin Roberts for research assistance. The views stated herein are those of the author and not necessarily those of the Federal Reserve Bank of Cleveland or of the Board of Governors of the Federal Reserve System. 9. Female-headed households are less likely to be classified as poor under the consumption standard, because the average age of children is lower in female-headed households than in other households with children, and younger children tend to consume less. 10. See Peter Gottschalk and Robert Moffitt, "The Growth of Earnings Instability in the U.S. Labor Market," Brookings Papers on Economic Activity, 1994 No. 2, pp. 217-54. BULK RATE U.S. Postage Paid Cleveland, OH Permit No. 385