Download experimental Business Dynamics Statistics of High Growth (BDS-High Growth) data tables below. Variable definitions can be found in the Definitions dropdown below.

The BDS-High Growth is an experimental product of the U.S. Census Bureau. The BDS-HG was developed by the Center for Economic Studies (CES). The BDS-HG data are compiled from the Longitudinal Business Database (LBD). The LBD is a longitudinal database of business establishments and firms with coverage starting in 1976. Two methods are used to classify firm growth. For details about how firm growth is computed and growth bins are generated see the Methodology dropdown.

The Census Bureau has reviewed this data product to ensure appropriate access, use, and disclosure avoidance protection of the confidential source data used to produce this product (Data Management System (DMS) number: P-7083300, Disclosure Review Board (DRB) approval number: CBDRB-FY24-0446).

The year t release replaces the year t-1 release e.g. the 2022 release (1978-2022) replaces the 2021 release (1978-2021). Please contact us if interested in prior releases.

fempgr_gr – Firm employment growth rate bin. For a description of how firm growth rates are calculated, please see the Methodology dropdown.

The BDS-HG experimental data product classifies firms by their growth rate between t-1 and t as described in Kim et al. (2024). Here we briefly summarize the methodology describe by Kim et al. (2024). Firm-level growth rates are computed as the aggregation of establishment-level changes in employment divided by the sum of average establishment between t-1 and t. Specifically, establishment growth g_it is defined as follows:

For establishment i at time t with employment E. The denominator, X_it is defined as follows:

Firm-level growth, g_ft, is then the sum of employment changes weighted by the sum of average establishment employment, for all establishments i associated with firm f at time t. This is computed as follows:

This growth rate measure, first developed by Tornqvist et al. (1985), has become standard in the firm dynamics literature (Davis et al., 1996). g_ft is bounded from -2 to 2 with firms that transition from positive to zero employment having a firm growth rate of -2 and firms that transition from zero employment to positive employment having a growth rate of 2.

This growth measure has a number of desirable properties. First, it approximates log differences when growth rates are near zero. Second, it can be easily aggregated across cells given information about positive and negative changes and the sum of average employment. Third, it is defined for both entrants and exits. Fourth, it is symmetric for employment increases and decreases. For example, an establishment that changes from 10 to 15 employees will have a g_it growth rate of 0.4 (50% increase) and an establishment that changes from 15 to 10 employees will have a growth rate g_it of -0.4 (-33% decrease).

There are two firm-level growth rate classifications used in the BDS-HG tabulations. The first classifies firms by their firm growth rate g_ft in a given year (fempgr_gr). This classification uses the following nine bins: a) -2; b) (-2 to -0.8]; c) (-0.8 to -0.2]; d) (-0.2 to -0.01]; e) (-0.01 to 0.01); f) [0.01 to 0.2); g) [0.2 to 0.8); h) [0.8 to 2); i) 2. Note that the bins area not of equal size. Firms that enter (2) or exit (-2) are classified separately. Moreover, firms with growth rates near but not exactly zero are grouped together ((-0.01 to 0.01)).

The second firm growth classification identifies where a given firm sits on the within-year average employment weighted firm growth rate distribution (fempgr_grpct). To compute weighted percentiles, we order firms within a given year by their firm growth rates g_ft, breaking ties randomly, and compute the cumulative share of total average employment. The cumulative share of total average employment is then used to assign firms to one of five percentile-based bins as follows: a) [0 to 10], b) (10 to 25], c) (25 to 75], d) (75 to 90], e) (90 to 100]. Again, bins are not equally sized. This method, by construction, involves cut-offs that vary over time. As the firm growth rate distribution changes over time so too will the percentile-based cutoffs. This is illustrated in Figure 5 of Kim et al, shown below. The growth rate associated with the employment-weighted 90^th percentile fell from 0.35 in the late 1970s to 0.23 in the 2010s.