IDI Methodology Change

IDI Methodology Change

Beginning in the first quarter of 2016, the Regional Economic Research Institute designed the FGCU Industry Diversification Index. The goal of the IDI is to measure industry diversification and, by extension, that of the workforce, in each of Florida’s 24 workforce regions, the state of Florida, and the remaining 49 states. The index measure is computed quarterly, soon after the release of workforce data from the Florida Department of Economic Opportunity’s Quarterly Census of Employment and Wages and it spans from 0 to 10.  The index so far has been calculated using the share of employment from the top 5, 21 two-digit NAICS industry sectors.

Starting with the 2020 Q3 publication, the RERI will begin to publish a revisited Industry Diversification Index. The index will remain spanning from 0 to 10 in which the former indicates a completely specialized economy and the latter indicates an economy that is completely diversified, i.e., all industries have the same employment share. The revision of the IDI aims at improving its accuracy and reliability.  Below we briefly discuss these changes to the index.

Overview and Limitations of Old Methodology

The old IDI was calculated each quarter using 2-digit NAICS employment data collected from the Quarterly Census of Employment and Wages.  The top 5 employment shares from each region and state were used to calculate the FGCU Industry Diversification Index, using the following formula:


Where  a(i)  represented the ith largest industry share for the region/state, N represented the number of industries used to calculate the index (in this case, N=5), and a-bar represents the average of the industry shares.

After a review of the methodology, it was determined that the old IDI imposed two limitations to the index:

  1. The two-digit NAICS employment data comes from the Bureau of Labor Statistics, and in cases in which data can potentially be used to identify firms in a region, the BLS would suppress such information from its publication. This data suppression could lead to potential fluctuations in employment levels from quarter to quarter, with an industry reporting employment in one quarter and data being suppressed the next;

  2. Considering the top five industries in the region allowed us to overcome the data suppression issue, but it created some different concerns. More specifically, we were ignoring the other 15 industries in the region, and the share of employment in the top 5 industries varied across each region.

New Methodology

Therefore, to address these limitations, the research team has made the following changes to the IDI: 

  1. We now use NAICS super sectors instead of the two-digit NAICS codes.  By doing so, we avoid data suppression since this is a higher-level industry classification (12 industries versus 21 industries). 

  2. To help further reduce data suppression issues, we include employment in all level of ownership, private and public.  

  3. We take into account all the super sectors and not only the top 5. This helps solve the issue of omitting employment in other industries and guarantees that we are comparing always the 100 percent of employment in each region.

As a result, the new IDI is now calculated using the following formula, which can be interpreted as a Gini Index, i.e., a measurement of how equal the distribution of employment is among the different sectors in a region.

FGCU IDI Formula

Where n is the number of industries, x_i is the share of employment in industry i, i = 1 to 12.  Industries are ranked by employment share in ascending order.

While no measure of diversification will ever be perfect, we feel the changes made to the methodology help improve the quality and accuracy of determining industry diversification for a geographical region.  For comparability, all trends in the 2020 Q3 report, along with future reports, will contain updated historical data to reflect changes to the methodology.  Furthermore, data found in the IDP Rankings portion of the website will also be updated to account for the methodology change.