When compliance raises expenses 1%, no big deal. But if that number approaches 50% for small, nimble, well-leveraged tech start-ups, then public markets become closed to them.
SOX is an awfully big variable to not be considered.
Recall the high unemployment rate; the government has dumped money into several tried-and-true methods to stimulate growth and bring back jobs, but whatever growth we've seen in the last few quarters hasn't translated into expansion of operations. We can infer from this picture that companies are sitting on their money and spending frugally. They did so in the early 2000s as well. SOX explains why it's hard for a small-to-mid-cap to break out, but it doesn't explain why existing big businesses would restrict their spending.
My conjecture, which I don't have a watertight case for, is that business(as a whole) doesn't want to grow, or can increasingly meet growth targets via productivity alone, and so the whole rationale driving IPOs - and the capital/labor side of the economy - is broken.
Google, Microsoft, Nvidia, Apple...No Debt.
The existence of large cash cushions on these balance sheets is insurance against a rainy day, and changes the risk profile. If these companies had been levered during the 03-07 expansion, they would have exploded in value.
They didn't, because it makes no sense to lever a tech company to the hilt, even if you believe it makes sense to lever a manufacturing company to high heaven. Cashflows are too variable and too prone to disruption.
It takes a special kind of tech co to go public, and many (i'm lookin' at you pets.com) weren't that kind.
Never mind the crazy IPO levels of bubble years 1999-2000. Why can't we get back to the modest 100-200 IPOs a year of the early 1990s?
Your single-minded obsession with SOX is missing about a million other general and business-specific factors that would go into this decision, many of which represent a much higher dollar figure than sarbox compliance.
Heck, the costs of actually executing the IPO probably dwarf typical sarbox compliance costs (this statistic totally made up but plausible).
Those who do have a business model, outlook, and revenues that makes them attractive, like Google, can and do still IPO successfully, but there aren't many of them.
At the margin, though, SOX's costs cause private companies to delay going public, and provide a relative advantage to older/larger/already-public companies -- acquirers -- for whom compliance costs are proportionately smaller.
And since the cycle of IPOs depends on emulating prior successes, even a small initial effect could depress the follow-on rate for years.
Google is sui generis, the most successful venture-backed company of the past 30 years. If being 'like Google' is the yardstick to IPO, the system is seriously broken.
In the bubble, IPOs abounded and investors were buying whatever shit they could lap up if it had "e" in the name. Post-bubble, tech companies that survived have been successful (by definition) and have sought to absorb smaller companies both to gain technologies and to kill off competitors; whereas investors have been wary of new tech stocks. This alone could explain the shift from IPO to acquisition.
Also, a small, nimble, well-leveraged tech start-up surely doesn't want or need an IPO anyway. An IPO doesn't make sense if you don't need the cash (unless you VC would rather have an "exit" than profits†). It is hard for investors to evaluate the potential of a small tech company; acquisition makes much more sense.
† VCs demanding an exit in a fixed time frame might make an IPO impossible -- is this something that happens? Do any VCs accept dividends instead of a buyout?
VCs aren't the issue - the investors, the folks/companies who put money into their funds are.
VC funds are partnerships which divvy up the stuff bought with the money. The investors value liquidity.