Why Startup Formation Keeps Declining

John Haltiwanger's University of Maryland work documents that US firm formation rates per capita have declined consistently from 1980 to 2020. Jorge Guzman and Scott Stern's MIT work shows the high-quality startup rate has held up better. The honest reading is that 'entrepreneurship' content has grown while actual entrepreneurship has not.

The rate of new firm formation in the US has declined for forty years. Business dynamism — the rate at which new firms enter, grow, contract, and exit — has fallen across nearly every sector. This is contrary to the popular image of a vibrant startup economy, and the empirical pattern is one of the most robust and underappreciated facts in US economic data.

The Data

John Haltiwanger and his collaborators have produced the most careful documentation. New firm formation as a share of total firms has fallen from roughly 12% per year in 1980 to under 8% in 2020. Young firms (under five years old) employ a smaller share of the workforce than they did in 1980. Job reallocation rates — the share of jobs that are created or destroyed in a year — have fallen across industries. These trends predate the 2008 crisis and have continued through the recovery, though the pandemic produced a short-term surge in formation that has since moderated.

The decline is broad. It is not concentrated in declining industries or rust-belt geographies. Manufacturing, retail, services, even technology have seen declining dynamism. The exceptions — software and certain biotech segments — are the popular image of the startup economy but are a small share of overall business activity.

Possible Causes

Several explanations have been proposed.

Aging workforce: older workers start new firms at lower rates than younger workers. The aging US population mechanically reduces firm formation. Estimates suggest demographics explain 25-40% of the decline.

Rising fixed costs: regulatory compliance, healthcare costs, and the costs of building scale capacity have all risen. Small firms face larger fixed-cost barriers than they did in 1980, which discourages entry. This is harder to quantify but is consistent with the patterns by firm size.

Incumbent power: dominant incumbents make entry harder through network effects, branding, supply-chain control, and outright acquisition of potential competitors. The rise of platform companies and the increased pace of M&A in the 2010s are consistent with this explanation.

Non-compete agreements: the spread of non-competes into industries and worker categories where they were rare before constrains the labor mobility that drives spin-out formation. Silicon Valley's historically permissive non-compete enforcement is a major reason software dynamism has remained relatively high.

Concentration of venture capital: VC funding has concentrated in fewer firms, fewer geographies (Bay Area, NYC, Boston), and fewer sectors (software, SaaS). Sectors and geographies outside that focus get less startup capital, which constrains formation regardless of demand.

Decline of small banks: small business credit historically came disproportionately from local community banks. The consolidation of US banking since 1980 has reduced the number of community banks substantially, which has tightened credit for small business formation.

What This Means for Productivity Growth

New firms historically have been the main source of productivity growth in the US economy. They introduce new technologies, new business models, and competitive pressure that pushes incumbents to improve. When firm formation slows, productivity growth slows.

US productivity growth has indeed slowed since the early 2000s, from about 2-2.5% annual labor-productivity growth in 1995-2005 to about 1-1.5% in 2010-2020. The declining-dynamism story is one of the leading explanations for this productivity slowdown, alongside the slower diffusion of new technologies and the maturing of the information-technology revolution.

If the dynamism decline continues, future productivity growth will likely be lower than past growth, with implications for wage gains, public-finance sustainability, and political-economy capacity to deliver rising standards of living.

The Regional Concentration

The remaining business dynamism is concentrated geographically. Bay Area, NYC, Boston, Seattle, and a few other metros account for disproportionate shares of new firm formation in high-growth sectors. Most of the country has experienced more substantial dynamism decline than the aggregate numbers suggest, with the high-growth metros masking the broader trend.

This regional concentration has political-economy effects beyond the productivity numbers. It contributes to the urban-rural and coastal-interior political divides, and it concentrates the gains from the modern economy in places that the broader political system does not represent in proportion to their economic weight.

The Industry-Specific Patterns

Software dynamism has remained relatively high — though it has declined from 1990s peaks. The reasons are partly demographic (the sector's workforce is younger than average) and partly structural (non-compete agreements are unenforceable in California, the dominant software cluster). The sector is the exception that highlights the broader pattern.

Retail dynamism has fallen sharply with the rise of platform e-commerce and big-box consolidation. The mom-and-pop retail formation that was a major source of new businesses in the 1980s has largely disappeared in much of the country.

Manufacturing dynamism has fallen as the sector has both consolidated and offshored. Even the surviving US manufacturing has become more concentrated in larger incumbents.

What Policy Could Do

Several policy interventions have been proposed:

Federal restrictions on non-compete agreements (the FTC 2024 rule in litigation).
Antitrust enforcement against killer acquisitions that absorb potential competitors.
Small-business credit expansion through SBA and community financial institutions.
Universal healthcare reforms that decouple insurance from employer relationships, lowering the risk of leaving an existing job to start a firm.
Skills-based immigration that increases the entrepreneur pool.

Each addresses one channel of the decline. None individually is large enough to reverse it. The decline is the product of multiple interacting trends, and reversing it requires multiple coordinated interventions across decades — exactly the kind of sustained policy commitment the US political system has had trouble producing.

The Honest Reading

Declining business dynamism is one of the most robust empirical patterns in US economic data, and one of the most consequential. It explains a substantial share of slowing productivity growth, contributes to widening regional inequality, and reduces the competitive pressure that pushes incumbents to improve. The popular image of a vibrant startup economy is mostly an artifact of the few metros and sectors where dynamism has been sustained. The broader US economy is becoming less dynamic, and the trend has continued through booms and recessions, both political parties, and multiple regulatory regimes. Reversing it will require substantial policy intervention; without it, the slowing dynamism will continue and the productivity, wage, and political-economy consequences will compound.