In 1996, roughly 8,000 domestically incorporated companies were listed on a US stock exchange. Today the figure is approximately 4,300 — a reduction of close to half. Over the same thirty years, the US economy roughly doubled in real terms, the population grew by a third, and the number of private equity-backed companies expanded from approximately 1,900 to more than 11,200. The arithmetic is strange. American capitalism produced more companies; fewer of them ended up publicly listed.
The conventional explanation is regulatory. Sarbanes-Oxley (2002), Dodd-Frank (2010), and the accumulating compliance overhead of public-company life are said to have made listing prohibitively costly relative to remaining private. JPMorgan’s Jamie Dimon has made the regulatory argument repeatedly, including in his 2024 letter to shareholders, where he describes the diminishing role of public companies in the American financial system as a serious trend that the total should have grown rather than shrunk. The argument has the virtue of identifying a real cost. Compliance, board governance, quarterly disclosure, and litigation exposure are non-trivial overheads.
The argument has the disadvantage of being only partly correct. Research published by Columbia Business School in November 2025, using a bunching-estimator approach to size-based regulatory thresholds, finds that direct regulatory cost accounts for approximately 7.3 percent of the IPO decline. Even removing all post-2000 regulatory cost, the authors conclude, would not change the underlying trend. Sarbanes-Oxley contributed. It did not cause.
The cause sits a few years earlier and is structural rather than disciplinary. The National Securities Markets Improvement Act of 1996 — known as NSMIA — exempted eligible private issuers from state-by-state blue-sky securities registration. It also amended the Investment Company Act of 1940 so that private funds could raise capital from far more than the previous 100-investor cap without becoming registered investment companies. Before NSMIA, a venture or private-equity fund that wanted to grow large faced a regulatory wall. After NSMIA, that wall came down. NBER analysis finds that 26 percent of firms first financed in 1994 went public via IPO. Of those first financed in 2000, only 2 to 3 percent did. The proximate driver of the listed-company decline is not what made public listing harder. It is what made staying private easier.
The JOBS Act of 2012 deepened the change, raising the shareholder-of-record threshold above which private companies must report publicly. A firm could now grow to a thousand or more accredited holders without crossing into public-company obligation. Combined with the steady expansion of the venture-capital and private-equity industries — and the growth of secondary markets that allow employees and early investors to sell positions before any public listing — the net effect is that a US firm with strong fundamentals can today raise tens of billions in private capital, distribute liquidity to insiders periodically, and never face a public-market quotation. OpenAI is the most visible example; it is not unusual.
The composition of the firms that do go public has shifted accordingly. According to Jay Ritter’s IPO Statistics, updated March 2026, 90 operating companies went public on a US exchange in 2025 — a fraction of the 624 listings recorded at the 1996 peak. The companies going public today are larger, older, and more profitable than those that listed thirty years ago. Median pre-IPO sales have risen substantially in real terms. The small-cap IPO — the kind that historically dominated the listing market and produced both Russell 2000 returns and the diversified equity culture of the 1980s and 1990s — has thinned to a residual.
This sounds, at one level, like a market that is simply working — capital flowing where it earns returns, with private vehicles outperforming public ones over rolling ten- and twenty-five-year windows by most credible measurements. Why does the structure matter?
It matters for two reasons. The first is concentration. A market with 4,300 listed firms produces a different distribution of investment opportunity than one with 8,000. The S&P 500 today is more concentrated in its top ten constituents than at any point since the 1960s; the universe outside the index is proportionally narrower. A retail investor who indexes broadly is, in effect, indexing a smaller and more concentrated economy than the same investor would have indexed in 1996. The second reason is access. The growth phase of a contemporary firm — when most of the equity appreciation typically occurs — now happens almost entirely inside private-market vehicles. Roughly 87 percent of US households are not eligible to invest in those vehicles under current accreditation rules. The growth equity that thirty years ago accrued partly to public investors now accrues almost entirely to private ones.
What the evidence says, with reasonable confidence, is that the listed-company decline is a consequence of deliberate policy choices made between 1996 and 2012, not of compliance overhead alone, and that the resulting system has concentrated the equity returns of the contemporary American economy in a small fraction of the population that can access them. Whether that is a problem depends on what the public-company structure was for in the first place. The answer was never only about price discovery. It was also about distribution. The distribution function is what has thinned.
