Civic Health and Unemployment II: The Case Builds

In 2011, we found that states and large metropolitan areas with high levels of civic engagement prior to the Great Recession suffered less unemployment between 2006 and 2010. The relationship between civic health and economic resilience held even when we adjusted for the economic factors that are usually thought to influence unemployment, such as demographics and changes in housing prices.

To be sure, civic engagement is not the only factor that matters. Las Vegas lost jobs because of the collapse of the housing market; Detroit, because of changes in the auto market. But, given two states with similar economic conditions, the one with more civic engagement would weather the recession better.

Since 2011, in partnership with the NCoC, and with support from the John S. and James L. Knight Foundation, CIRCLE has continued to investigate this topic. For the 2012 report, we investigated the relationship between civic health and unemployment in all 50 states, 942 metro areas, and more than 3,100 counties. We added new statistical controls (alternative explanations of unemployment change) to the model, analyzed a Census Current Population survey that follows individuals over time, and incorporated the results of the Knight Foundation’s Soul of the Community Survey, which investigated a wider range of opinions and attitudes than are measured in federal surveys.

The basic pattern found in the 2011 report held up: communities with more civic engagement in 2006 suffered less from unemployment in the Great Recession, even when other possible explanations are factored in.

The new analysis also directed attention to two particular aspects of civic engagement: the role of nonprofit organizations and the effects of social cohesion.

Nonprofits play important economic roles in communities:

• The counties with the most nonprofits lost fewer jobs than the counties with the least. Counties that ranked in the top 10 percent for nonprofits per capita in 2006 experienced an increase of two percentage points in their unemployment rates between 2006 and 2009, compared to an increase of 5.1 percentage points for counties in the bottom 10 percent during the same period.
• If a county had one extra nonprofit for every 1,000 residents in 2005, and everything else were held constant, it would have half a percentage point less unemployment by 2009.
• An employed individual in 2008 was twice as likely to become unemployed if he or she lived in a community with few nonprofit organizations (the bottom five percent in nonprofit density) rather than one with in the top five percent for nonprofit density, even if the two communities were otherwise similar.

A closer look revealed that not all nonprofits mattered equally. The ones that provided concrete opportunities to local members or clients had the biggest impact on employment rates. For example, fraternal organizations and unions that convene their members for local meetings, sports organizations that hold athletics events, and service-providers that directly assist local people all seemed to help, whereas “mailing-list” organizations whose members just contribute checks did not seem to matter for unemployment.

The second aspect of civic engagement that affects employment is “social cohesion.” We analyzed several surveys for this project, each with somewhat different measures. Thus the precise components of social cohesion varied, but the general definition was the degree to which residents socialize, communicate, and collaborate with one another. Some examples of the impact of social cohesion:

• At the state level, stronger social cohesion within a community strongly predicted a smaller increase in the unemployment rate from 2006-2010.

• In 2006, the states with the highest social cohesion and those with the lowest had virtually identical unemployment rates of around 4.5 percent. But by 2010, their unemployment rates were significantly different: states with high social cohesion had an unemployment rate of 8 percent and states with low social cohesion had an unemployment rate of 10 percent.

Social cohesion and nonprofit density each independently mattered for unemployment. Changing either the number of nonprofits that serve and work with citizens, or the degree to which people interact, would benefit a community’s economic resilience.

The statistical models used in this analysis cannot completely explain why these patterns exist, but the Knight Foundation’s Soul of the Community provides suggestive evidence. It measures subjective factors, such as people’s feelings about their own communities. By combining Knight survey data with federal data, we find that nonprofit density and social cohesion both predict people’s pride in, and attachment to, the places they live. In turn, pride and attachment strongly predict positive perceptions of the local economy. Given the links we found between nonprofit density and social cohesion (on one hand) and employment (on the other), we hypothesize that when residents are proud of their communities, they are more likely to promote local businesses and local initiatives. As a result, local business owners prosper and entrepreneurial activities flourish.

For example, businesses and investors have three choices during a recession: (1) keep their own capital and productive assets (such as factories) on the sidelines until economic conditions improve; (2) invest anywhere in the world where the returns seem most promising; or (3) invest in job-creating enterprises near where they live. They may be most likely to choose the last option if they are optimistic about local opportunities, if they are connected to and trust local people, and if they care about where they live. They may assess local opportunities more optimistically, or simply decide to invest locally even if they anticipate somewhat lower returns.

Consider too consumers who have retained some assets in a recession. Like executives and investors, consumers have three options: (1) hold onto their savings until economic conditions improve; (2) purchase goods or services from far away, or (3) pay local people to provide goods or services. Again, consumers may be most likely to choose the third option if they know and trust local service-providers and care about their communities.

Although more research is necessary, the case is building. Civic engagement may not merely be good in itself and good for politics; it may also benefit the economy.