I’ve always been a sceptic about the many claims that entrepreneurs, startup firms and venture capital are the “most” important drivers of improved quality of life in the US in my lifetime. I remain a sceptic. I believe that large firms deliver even more added value (driven by self-interest). I believe that the government and not-for-profit sectors play an equally important role. I believe that the government’s definition of the “rules of the game” and our culture’s influence on how people live their everyday lives are also very important. I’ll come back to the role of small firms in a separate blog post. I’ll try to tie together all of the strands of our amazing US labor market in another blog post.

But today, I want to tip my hat to the truly amazing role that startup firms play in driving the US economy and labor market.

https://www.bls.gov/bdm/us_age_naics_00_table1.txt

The US Bureau of Labor Statistics (BLS) tracks new jobs added (gains) and jobs lost (losses) each year in detail by firm and establishment (branch), including links to the year the firm was first created. At the total country level, we see that job creation and destruction follows the business cycle. The Great Recession, subsequent expansion and pandemic periods are obvious above. In the Clinton years, job gains averaged a great 15 million per year. The Bush, Jr. years showed still solid 13-14 million annual new jobs added. The Obama recovery increased new jobs from 10 to 13 million per year and Trump maintained this positive level at the end of a very long period of economic growth.

Job losses generally followed the pattern of job growth. Job losses are even more volatile. They peaked in the recessions of 2002, 2009 and 2021.

In the last 3 decades, the US economy added 2 million new jobs each year during periods of expansion.

Many economists, journalists and politicians claim that startups account for ALL new job growth. This is an “amazing” claim that deserves deeper analysis, investigation and description. I’ll chase this separately. The claim is “largely true”.

Startup firms delivered 4-5 million new jobs each year in the 1990’s. This declined to 2.5 million jobs during the “oughts”. It increased back to 3 million new jobs per year in the teens. In non-recessionary periods during the nineties and “oughts”, existing firms destroyed 2 million jobs each year. In the teens, the job destruction rates were much lower, roughly 1 million jobs per year. The startup jobs minus existing firm jobs number was typically 2 million net new jobs per year in positive economic years.

The BLS separates job gains from job losses and categorizes them by the firm’s first year of existence. Job gains at existing firms were roughly flat at 10 million per year. Startup firm new jobs declined from 4.5M in the nineties to 2.5 M in the “oughts”, recovering to 3M per year afterwards. Startup job creation declined from 45% to 31% of existing firm job creation, a one-third reduction.

The BLS provides data on the subtotal of all firms founded before 1993. These “mature” firms display a similar pattern. Annual job gains fell a bit from 4.5M in the “oughts” to 3M in the teens. Annual job losses fell even faster, from 6M in the “oughts” to 3.5M in the teens. Net job losses averaged 1.5M annually in the “oughts”, but just 0.5M in the teens.

The number of new jobs created by startups declined by one-third during this period, from 4.5 to 3 million per year.

The BLS data allows us to track the gains, losses and net jobs added by first year of existence for firms. First year startup jobs declined from 4.5M to 2.5M. The cumulative jobs created measured 10 years after startup is more positive. Cumulative new jobs, measured 10 years after startup, increased from 3M to 3.5M then dropped back to 3M in 2001. The startup classes of 2009-12, despite the Great Recession, report 2.5M net jobs added each year, as measured after 10 years.

Job losses have fallen much faster than job gains in the last 20 years, measured by a full decade of performance. The ratio of job gains to job losses has improved markedly. This ratio averaged 83% in the nineties, indicating that 1/6 new jobs was destroyed within a decade. This ratio has greatly recovered to the mid-90’s. indicating that new startups, in total, essentially maintain their initial jobs count a decade later. Other data shows that one-third of startups don’t exist after 10 years. Hence, this means that the successful remaining one-third have roughly tripled their employment in their first decade.

Firms die much faster than employment. There are many studies that claim that one-half, three-quarters, fourth-fifths or nine-tenths of employees at startup firms are eliminated in 10-20-30 years. These are mostly exaggerations.

Ten years after their founding, startups still employ 80-90-95% of their initial year hires. Job losses fell during the teens. Job gains grew rapidly after the Great Recession.

For the 18-year period where we have ten years of data on startup firms, we have a clear pattern of net employment decline at the end of the decade, on average. The ten-year retained employment level at almost 80% of the initial level is far higher than the 50% claimed by some commentators.

The years since 2009 show a clear pattern of startups maintaining 90-95% of their initial employment levels after 10 years.

Summary

The US economy typically added 2M net new jobs each year during periods of economic expansion. Historically, startups added 4M jobs annually to offset the 2M jobs eliminated by existing firms. The job destruction rate of existing firms has slowed. The jobs retention rate of startups has improved. Net, net, the US economy still generates 3M+ new jobs each year which essentially still remain a decade later. One-third of the firms are gone, but the winners employ 3 times as many as when they started.