Rob Grunewald - Associate Economist
Published November 1, 2006 | November 2006 issue
It makes sense that poverty rates are related to the overall health of the economy. As the economy grows, so do opportunities for employment and income growth. Stronger labor markets and higher income levels tend to help those families living in poverty move above the poverty threshold.
According to national data, poverty rates have moved somewhat together with changes in the unemployment rate since 1959, and in opposite directions in relation to changes in inflation-adjusted median income (see chart). That is, lower poverty rates coincide with decreases in unemployment or increases in income.
Some studies have suggested that the relationship between changes in the poverty rate and macroeconomic variables have weakened over time. An October 2005 National Bureau of Economic Research working paper by Hilary Hoynes, Marianne Page and Ann Stevens suggests that while the link has weakened, changes in the unemployment rate and median wages nevertheless predict changes in the poverty rate rather well.
The researchers use the U.S. unemployment rate, median wages, a measure for income inequality and region of the country to predict annual changes in poverty rates from 1967 to 2003. Their results imply that an increase in the unemployment rate of 1 percentage point increases the poverty rate by between 0.4 and 0.7 percentage points, while a 1 percent increase in median wages is associated with about a 0.2 percentage point decrease in the poverty rate. The effects were somewhat larger for both variables from 1967 to 1979 and were somewhat smaller from 1990 to 2003, illustrating the weakened effect of macroeconomic variables on poverty rates.
In light of these results from a national model, a similar model using Ninth District county-level data was used to predict changes in poverty rates from 1997 to 2003, the years in which annual county data were available. In the district model, slightly different measures of poverty rates and median income were used compared with the national model, and the district model doesn't include a measure for income inequality. The district model also includes county population, percentage of population that is minority and median age.
Similar to the national model, a 1 percent increase in median income is associated with about a 0.2 percentage point decrease in the poverty rate. The effect of the unemployment rate is smaller than in the national model, indicating that a 1 percentage point increase in the unemployment rate is associated with a 0.2 percentage point increase in the poverty rate. Furthermore, an increase in population of 10,000 people is associated with a 0.05 percentage point increase in the poverty rate. That is, when median income and the unemployment rate are held constant, poverty rates are relatively higher in more populated counties, but the associated effect is relatively small.
Higher minority populations are associated with higher poverty rates, while a higher median age is associated with a lower poverty rate. In addition, the analysis also indicates that a higher percentage of population with a college degree is associated with lower poverty rates, and the presence of a Native American reservation is associated with higher poverty rates.