The proportion of small, medium, and micro enterprises in all enterprises in the United States has been gradually decreasing since the 1980s.
Executive summary
The claim that the proportion of micro, small and medium enterprises (MSMEs) in all U.S. enterprises has been “gradually decreasing since the 1980s” cannot be accepted or rejected as a simple fact on the current record: different datasets point in different directions and no single source in the reporting set provides a continuous, definitive share-of-enterprises series from the 1980s to today [1]. Some official and advocacy reports flag long-run declines in employer firms or low points (for example a 2020 count that was the lowest since 1983), while other recent measures show surges in business applications and billions of dollars of output from small firms, meaning the narrative of steady, monotonic decline is not fully supported by the presented evidence [2] [3] [4].
1. What the most direct historical signals say: employer-firm counts dipped, but the record is patchy
The Small Business & Entrepreneurship Council and related summaries point out that the number of small businesses with payrolls hit 9.25 million in 2020—the lowest level since 1983—after peaking at 9.71 million in 2018, a fact that is often cited as evidence of a long-term deterioration in employer-firm counts relative to the 1980s baseline [2]. That signal is important because employer-firm counts are one of the clearest measures of the universe of firms with employees, but the reporting makes clear the series is volatile—affected by recessions, classification changes and the pandemic—so a single low point does not by itself prove a steady downward trend since the 1980s [2].
2. Broader measures complicate the simple-decline story: tiny firms still dominate counts and output
Other reputable trackers show that “small” firms remain the vast majority of U.S. businesses by count and contribute substantial GDP: small businesses have been reported to make up roughly 99.9% of U.S. firms in recent SBA-derived summaries and historically accounted for between 43.5% and 50.7% of U.S. GDP in the late 1990s through 2014 window—numbers that undercut a narrative that small firms have unequivocally faded as a share of all activity [5] [3]. Moreover, contemporary data on new business applications and patenting activity from sources such as Pew and the U.S. Chamber show rising entrepreneurial interest and innovation intensity among smaller employers in the 2019–2023 window, which further complicates claims of consistent decline [4] [3].
3. Structural change: more scale-ups, rising large-firm productivity, and measurement blind spots
Analyses from McKinsey highlight a parallel dynamic: productivity and growth concentrated in larger firms can create market opportunities that allow some MSMEs to scale into larger enterprises, and in sectors like ICT a meaningful fraction of today’s large firms were MSMEs two decades ago—an outcome that can reduce the share of firms classified as small even as the economy generates new small entrants [6]. McKinsey also underscores that productivity comparisons and sectoral shifts matter, and that available microdata are often incomplete—forcing researchers to rely on subsector averages rather than full firm-level longitudinal tracking—so apparent falls in share could reflect measurement limits as much as real structural loss [7] [6].
4. Verdict and evidentiary limits: mixed evidence, not a settled linear decline
Taken together, the sources show mixed evidence: there are documented low points in counts of employer firms (notably 2020) and legitimate structural pressures that can reduce the share of small firms, but there are also persistent dominance by small-firm counts, robust small-business GDP contributions in available windows and surging new-business applications in recent years—so the hypothesis of a gradual, monotonic decline in the proportion of MSMEs since the 1980s is not conclusively supported by the provided reporting [2] [3] [5] [4] [1]. Crucially, the dataset required to settle the question—consistent, long-run, firm-level shares by size across the entire 1980s–2020s period—is not present in these snippets, and the Census SUSB is the right reference for anyone seeking a rigorous year-by-year test [1].