Institute for Research on Poverty

We examine the effect of macroeconomic performance on poverty in the United States during the 1980s and 1990s. Our study advances research on this issue in a variety of ways: we utilize variation across the states rather than relying on over-time trends for the country as a whole; we analyze cross-state variation in both levels and change over time; we disentangle the impact of three different aspects of macroeconomic performance: economic output (per capita gross state product), employment, and unemployment; we investigate causal mechanisms more carefully than is often the case in poverty analyses, focusing on work hours and wages; we consider both absolute and relative poverty; we base our poverty measure on pretax-pretransfer income; we use a poverty measure that incorporates both the poverty rate and the poverty gap; and we focus on the working-age population. Our findings highlight the importance of employment for poverty reduction. Employment contributed to lower absolute and relative poverty by boosting hours worked and wages in low-income households. Per capita gross state product similarly contributed to lower absolute poverty by increasing hours worked and low-end wage levels, but it had very little impact on relative poverty because it also was associated with increased wage inequality. Unemployment had little or no effect on poverty. Macroeconomic Performance and Poverty in the 1980s and 1990s: A State-Level Analysis Despite rising living standards, poverty remains a seemingly intractable problem in the United States. Since 1973 the productivity of the American economy has increased by two-thirds, yet the official U.S. poverty rate has not decreased at all. Levels of poverty also vary considerably across states and localities. As of 2002, for example, the poverty rate in Arkansas was more than three times as high as in New Hampshire. Macroeconomic performance is commonly considered to be a key—perhaps the key— determinant of poverty. The notion that a healthy economy benefits those at the low end of the income distribution has been studied extensively (Blank 1997a, 1997b; Blank and Card 1993; Blank and Blinder 1986; Brady 2003b; DeFina 2002, 2004; Freeman 2001; Gundersen and Ziliak 2004; Iceland 2003a, 2003b; Mishel, Bernstein, and Allegretto 2005; Sawhill 1988; Tobin 1994). In this paper we explore the impact of macroeconomic performance on poverty among working-age American households. We use data from the Current Population Survey (CPS) to examine variation across the states in poverty levels as of 2000–2002 and poverty trends during the 1980s and 1990s. Our analysis is guided by the following questions: Does macroeconomic success in fact reduce poverty? If so, which aspect of macroeconomic performance has been most important: economic output, employment, or unemployment? What are the causal mechanisms? How, if at all, does our understanding of the effect of macroeconomic performance change when using alternative measures of poverty? THEORY Macroeconomic Performance and Absolute Poverty Macroeconomic performance refers to aggregate, or average, levels of income and employment. The most commonly used indicator, gross domestic product (GDP) per capita, refers to economic output—the value of goods and services produced in a country in a given year divided by the country’s 2 population. It is a measure, albeit an indirect one, of average income. “Economic growth” refers to increases in per capita GDP. A second indicator of macroeconomic performance is the employment rate, which is measured as the share of the working-age population that is employed. A third is the unemployment rate, which refers to the share of people actively seeking a job who are unable to find one. It has traditionally been assumed that a healthy economy contributes to lower poverty. The notion that “a rising tide lifts all boats” implies that as the economy as a whole improves, so too do the fortunes of those at the bottom of the income distribution. An examination of over-time trends for the U.S. economy as a whole suggests some support for this notion. Figure 1 plots GDP per capita, the employment rate, the unemployment rate, and the U.S. government’s poverty measure from 1960 to 2003. The poverty rate is calculated as the percentage of persons living in households that have incomes below the poverty line (see below). Generally speaking, in periods of economic growth—in which per capita GDP and the employment rate rise and the unemployment rate falls—the poverty rate has declined. And during economic recessions, such as those of 1973–75, 1982–83, 1990–91, and 2001, the poverty rate has increased (Blank 1997b; Blank and Binder 1986; Freeman 2001). Why does a healthy macroeconomy matter? Those most vulnerable to poverty usually have no investment income and receive little or no income in the form of interpersonal transfers from family or friends (Atkinson, Rainwater, and Smeeding 1995; Kenworthy 2004). Along with government benefits, earnings from paid work are thus likely to be the chief income source. Annual earnings are a function of two things: hours worked and wage levels. As GDP per capita rises, employment increases, or unemployment declines, work hours and/or wage levels for those at the bottom of the distribution may increase, thereby reducing poverty. Recessions are commonly defined as periods in which GDP declines for two or more consecutive quarters (three-month periods), though alternative definitions take other factors into account, such as changes in employment, personal income, and industrial production (National Bureau of Economic Research 2003). 3 Figure 1 Trends in GDP per Capita, Employment, Unemployment, and Poverty, 1960–2003