Our "best places" rankings are based on each region's current and long-term employment growth and its diversification over time. These measurements provide objective, readily comparable evidence of the extent to which area educational and training systems, housing and living costs, taxes, regulatory burdens and quality of life -- factors commonly cited to identify supposedly "hot" economies -- actually stimulate real-world business opportunities. Regions that consistently generate jobs in a broad range of industries rank at the top of the list. Those with poor and worsening growth and increasingly undiversified economies do less well in the rankings.

The rankings are derived from three-month rolling averages of U.S. Bureau of Labor Statistics "state and area" unadjusted employment data that were reported by the Bureau for the period January 1993 to September 2003. The data reflect the new North American Industrial Classification System (NAICS) industrial categories, including total nonfarm employment, manufacturing, financial services, business and professional services, educational and health services, information, retail and wholesale trade, transportation and utilities, leisure and hospitality, and government. All areas for which full datasets and uniform area definitions were available from the BLS for the past 10 years -- 277 regions in total -- were included in the analysis.

This Web archive supplements the rankings data that was published in "The Top 25 Cities for Doing Business in America," from the March 2004 issue of Inc magazine. It provides the following additional information:

  1. Comparative rankings of all regions in the dataset divided into three size groupings according to the September 2003 nonfarm employment of each area (based on three month rolling averages), including:

    (a) "Large" regions with total nonfarm employment of 450,000 or more;

    (b) "Medium" regions with total nonfarm employment of 150,000 to 450,000; and

    (c) "Small" regions with total nonfarm employment of up to 150,000.

  2. A list of the "growth index" results for each area in the dataset. The growth index comprises approximately two-thirds of a region's final ranking score. It is calculated from the normalized, weighted summary of each region's:

    (a) Current year employment growth as of September 2003 (based on three month rolling averages). This element of the index measures the region's most recent performance and is weighted by two points in the growth index;

    (b) The sum of 1998-2003 and 1993-1998 employment growth rates multiplied by the ratio of the 1998-2003 growth rate over the 1993-1998 growth rate (September to September based on three-month rolling averages). This element of the index measures a region's capacity to sustain growth and opportunity over time. The sum of growth in the first and second half of the last decade provides a measure of a region's absolute level of job expansion. The higher the total, the better a region will score. The second component measures the extent to which the growth occurred more recently or earlier in the decade. Regions that had better growth in the last five years score higher than those that had substantial growth early in the decade but slowed thereafter. The long-term performance element is weighted by two points in the growth index.

    (c) The difference between the current year's growth rate and the average 2000-2003 growth rate (September to September based on three-month rolling averages). This element measures each region's current (as of September 2003 based on three-month rolling averages) employment momentum and whether the economy is heating up, or slowing, relative to the most recent trends. This element is weighted by half a point in the growth index.

  3. A list of the "balance index" results for each area in the dataset. The balance index comprises approximately one-third of a region's final score. It is calculated from a normalized, weighted summary of:

    (a) Current sector balance (as of September 2003 based on three-month rolling averages). This element of the index measures how well each region's employment base is currently distributed among the major NAICS industry groups, including manufacturing, financial services, business and professional services, educational and health services, information, retail and wholesale trade, transportation and utilities, leisure and hospitality, and government. Areas with better balance among these sectors score higher than those with an employment base more dependent on only a small number of sectors. It is derived from the standard deviation of the employment totals for each major industry group relative to total nonfarm employment and is weighted by one point in the index;

    (b) Longer term balance. This element is derived from the standard deviation of the percentage contribution of each major NAICS industry group to each region's total 1998-2003 nonfarm employment growth (September to September based on three-month rolling averages). Areas that experienced balanced growth among a wide variety of occupations score better than those that built their growth on a smaller range of industries. This element is weighted by one point in the index; and

    (c) The standard deviation of each sector's recession period growth rate during 2000-2003 (September to September based on three month rolling averages). This measure isolates the extent to which the last recession and "bubble economy" distorted a region's industrial base. It is weighted by half a point in the index.

  4. Summary NAICS sectors for each of the regions in the dataset, including manufacturing, financial services, business and professional services, educational and health services, information technology, retail trade, wholesale trade, and government. The sector data includes:

    (a) Current sector employment as of September 2003 (based on three-month rolling averages);

    (b) Percentage of regional nonfarm employment attributable to each sector (as of September 2003 based on three-month rolling averages); and

    (c) Sector growth from 2000-2003 (September to September based on three month-rolling averages).