Meet the Author

Dionissi Aliprantis |

Research Economist

Dionissi Aliprantis

Dionissi Aliprantis is a research economist in the Research Department of the Federal Reserve Bank of Cleveland. He is primarily interested in applied econometrics, labor and urban economics, and education. His current work investigates neighborhood effects on education and labor market outcomes.

Read full bio

Meet the Author

Kyle Fee |

Economic Analyst

Kyle Fee

Kyle Fee is an economic analyst in the Research Department of the Federal Reserve Bank of Cleveland. His research interests include economic development, regional economics and economic geography.

Read full bio

Meet the Author

Nelson Oliver |

Research Analyst

Nelson Oliver

Nelson Oliver is a research analyst in the Research Department of the Federal Reserve Bank of Cleveland. His primary interests include urban revitalization, housing policy, and applied microeconomics.

Read full bio

02.12.14

Is a Neighborhood’s Unemployment Rate Influenced by Its Metro Area?

Dionissi Aliprantis, Kyle Fee, and Nelson Oliver

When people compare employment conditions around the country, they usually think in terms of large regions like the Midwest and the West Coast or cities like Cleveland and Pittsburgh. But employment conditions vary widely within major metropolitan area as well. Even if a metro area experiences rising average levels of employment and income, the changes in specific neighborhoods in that metro area may be well above or below that average.

We looked at unemployment and income data by neighborhood in the100 largest metropolitan statistical areas (MSAs) in the United States to see if we could identify any factors that help explain the differences that are observed across neighborhoods. The MSAs selected were the 100 largest in terms of population in 1980, and the factors we considered come from Census data gathered on neighborhoods, or census tracts, between 1980 and 2008. All dollar measurements are expressed in terms of 2009 real dollars, so they are comparable over time.

In general, high-income neighborhoods have much lower unemployment rates than low-income neighborhoods, as one might. The chart below shows the strength of the relationship in 1980. The typical unemployment rate found in low-income neighborhoods would rarely be found in a high-income neighborhood, while neighborhood unemployment rates of over 50 percent can be found in some low-income neighborhoods.

Given the strong association between a neighborhood’s income and its unemployment rate in 1980, we might expect the effects of negative changes in the labor market to be concentrated in low-income neighborhoods. Contrary to this expectation, we find that unemployment rates increased on average in all neighborhoods between 1980 and 2008, regardless of their income. Neighborhoods in the 25th percentile of the national distribution of household income saw their unemployment rate increase by 2.8 percentage points, while neighborhoods in the 75th percentile of household income saw theirs increase by 1.9 percentage points.

We looked more closely at low-income neighborhoods—those in the bottom quartile, or fourth, of neighborhoods in 1980 in terms of average household income—to see why some did better than others. We find that changes in neighborhood unemployment rates were related to the characteristics of the larger metro area of which the neighborhoods were a part in 1980, such as the metro area’s average household income or its share of residents with a bachelor’s degree (BAs). Being in a metro area that was in the top quarter of all metro areas in terms of average household income decreased a low-income neighborhood’s unemployment growth by about 1 percentage point on average (relative to those in the bottom quarter). Likewise, the share of residents with a BA had the same impact.

Income growth in the metro area over the last three decades is very predictive of unemployment growth in its low-income neighborhoods over the same period. Low-income neighborhoods experienced much larger growth in unemployment rates if they were located in a metro area with low income growth relative to those low-income neighborhoods that were located metro areas with high income growth. High-income neighborhoods, on the other hand, were immune from this effect; they experienced similar unemployment changes regardless of the type of metro area in which they were located.

What might explain the relationship between growth in a low-income neighborhood’s unemployment rate and the income growth of its metro? We speculate that low-income neighborhoods in high-growth MSAs may have experienced an influx of new residents with low unemployment rates, or alternatively, low-income neighborhoods in low-growth metros could have experienced a loss of residents with low-unemployment rates. Another explanation could be that the sectors employing residents in low-income neighborhoods are especially sensitive to the performance of the metro area as a whole.