Economically distressed or "at-risk" U.S. regions/counties have a limited set of policy options when it comes to economic development. Most at-risk regions are too small or lack the resources and human capital to implement any of the standard set of economic development strategies (EDS), such as expanding the creative class, industry cluster-based growth strategies or industry diversification.
The regions that succeed--the survivors--are often held up as role models, but those who struggled and didn't make it to the top may provide clues as to what is needed for success as well. For this analysis, we examined the worst-performing U.S. counties to ascertain their similarities and differences. Using county employment growth from 2001 to 2016 as the measure of performance, we assessed a set of county characteristics and conducted several quintile regressions--and found few robust and insight-enhancing results. We posit that the methods and even the variable selection process itself may suffer from survivorship bias. Different approaches are needed because each at-risk region (defined here as a U.S. county) is, in its own way, an outlier. If each at-risk region is, in some characteristic, an outlier, designing a development strategy that is congruent with a region's unique characteristics is a challenge.
Designing a new, feasible set of economic development strategies for at-risk regions is not within the limited scope of this article. Our mission, however, is to light the first lantern to find the way to feasible and appropriate development strategies for at-risk counties. We first provide an overview of four definitional categories of a region's characteristics typically applied for statistical analysis. Second, we visually present a few dimensions or characteristics of the more highly performing regions for which the EDS may be appropriate-approximately 620 top-performing counties. Third, we present visualizations of several characteristics that define the most distressed counties, but they are too few to describe the unique make-up and experience of each region. Finally, in the discussion and conclusion, we suggest guideposts for aligning regional needs/characteristics with research and policy options.
There are at least four categories of data that are of interest when determining the drivers of regional economic performance and dynamism, measured here as growth/change in employment. While GDP growth is the common national measure of economic performance, it is difficult to measure in smaller geographies, so it is not highlighted in this analysis.
* The first, and arguably most important analytical category, is industry structure and how that structure changes over time. Industry structure as it relates to performance and competitiveness is frequently divided into "traded" goods and services that are produced in a region for export beyond the regional geographic boundary of analysis, and "local" production for consumption within a regional boundary. Industry structure is the primary focus of Michael Porter's cluster-based strategies (Porter, 1998).
* Occupational structure, on the other hand, relates to "what we do" rather than "what we make" (Thompson and Thompson, 1993; Feser, 2003). Occupations may or may not define a region given that computer coders and programmers are not physically tied to any place and can perform their work off-site.
* Innovation, the third category, is a very broad spectrum, accounting for everything from educational attainment to foreign investment flows to proprietorship rates in a region.
* Finally, social capital attempts to find measures that provide a signal to the degree to which a region's population has the capacity to solve its own problems, be they social or economic.
We use Quarterly Census of Employment and Wages (QCEW) employment data from the U.S. Bureau of Labor Statistics (BLS)--with suppressed values estimated in-house-to determine the industry structure/profile for a region. The BLS program for Occupational Employment Statistics (OES) is the source for regional occupation structure. Innovation-related and social capital data-the majority of which is sourced from the ACS--come from a wide array of sources. (1)
Given the distribution of how counties performed in terms of employment growth from 2001 to 2016, as shown in Figure 1, we concluded that the two tails provided the more interesting analytical features.
Two common industry components contributed to employment growth for both the top 620 (high performers) and the bottom 620 (at-risk). A sufficient share of business-related service industries appears to be a critical element in driving overall county economic growth. Traditional manufacturing industries on the other hand--industries such as upstream metal, plastics, automotive and heavy machinery--exert a headwind on employment growth. Interestingly, but not...