Shift-Share Analysis of Rural and Urban Indiana

Note: This is the latest post from our Rural Indiana study.

Regional economies change composition over time. Some sectors grow, while others shrink. These changes may be linked to the fortunes of the national economy, such as a recession or recovery, while some changes are due to idiosyncratic changes within a particular place. Still other changes are linked to an individual industry that may be over-represented.

Shift-share analysis decomposes these three different effects on a region by comparing changes in national, industrial mix, and local growth. In the following tables we compare the actual employment growth in Indiana’s rural and urban counties coming from these three sources of change. The national growth column illustrates the change that would have occurred within the region (urban or rural) if overall employment had grown at the national rate. For example, if the U.S. economy grew by 2 percent and the region had one million jobs, we would expect the national shift-share effect to be 20,000 jobs, or two percent of one million.

The industrial mix column illustrates the overall change in employment connected to each sector of the economy. This column measures how many jobs would be expected from each region due to differences in the industry mix in that region. This is calculated by subtracting the national growth rate from the industry growth rate to estimate how many jobs should have been created. For example, if there were 100,000 manufacturing jobs and nationally manufacturing grew by 3 percent while the overall economy rose by 2 percent, the industry mix contribution of growth would be 1,000 jobs:

(3%-2%)*100,000 = 1,000.

The industrial share of growth tells us how much more or less a region is growing based upon its particular mix of industries.

Finally, the local effect is often termed the ‘local competitiveness’ effect, and represents how many more or fewer jobs are created in each sector than should be given the national growth rate and industry mix of the region. This calculation is perhaps the most important for the region, because it is a summary of what is happening locally for good or ill to affect employment. Importantly, the shift-share does not provide a causal explanation for local competitiveness, but it does offer insight into how much better or worse a location or industry are performing than is expected, with a focus on idiosyncratic local conditions.

In the following tables we focus on the last decade (2009-2019), which represents the recovery from the Great Recession. We chose this time-period because it illustrates the differential effects of employment growth in rural and urban Indiana during a period of growth. 

During the historically lengthy recovery from the Great Recession, Indiana’s rural counties experienced net employment growth of only 924 jobs.[1] Had Indiana’s rural counties grown at the national rate, we would anticipate growth of 21,709 jobs. The structure of Indiana’s rural economy is somewhat over-represented by nationally growing industries, so should have expected growth of more than 1,700 jobs due to the mix of rural industries.  Finally, the local competitiveness illustrates the number of jobs that were created after accounting for national growth and local industry mix. In Indiana’s rural counties, this accounts for the loss of more than 22,000 jobs.

---

[1] We use employment data from multiple sources (Bureau of Economic Analysis, the Census County Business Patterns and the Census, Current Population Survey).  These elements use different definitions of year to year changes (e.g. March numbers versus an annual average).  Thus, there are slight differences in jobs counts between these sources, but these differences are trivial.  

In contrast to the slow rural growth of the recovery, urban Indiana created more than 111,000 jobs. For the purposes of this study, it is useful to know that more than 126,000 net new jobs were created in the Indianapolis metropolitan region alone, meaning that it is not simply urban places, but specifically the Indianapolis MSA, that led economic growth during this time. Despite the growth, Indiana’s urban counties should have expected 168,000 jobs given the national growth rate (146,000) and the industrial mix (22,000). So a gain of 111,000 jobs actually reflects a loss of more than 57,000 jobs, as highlighted by the Local Competitiveness column above.

It is not the case that economic conditions in Indiana are primarily a rural-urban issue. Much of Indiana’s urban places heavily underperformed the national average. Indeed, the only metropolitan area in total to outperform the national average job growth was the Indianapolis Metropolitan Area. Again, this analysis is not causal. The Shift-share analysis cannot tell us what elements specific to ‘local competitiveness’ result in these outcomes; we explore the question of “why” through other elements of this study.