Great Migration Public to Private Equity
What this covers
Patrick talks with Michael Mauboussin about why the public equity market has thinned out and where the activity has gone. The starting point is a count: US listed companies rose from 1976 to a peak in 1996, then fell even as the economy grew, leaving a "listing gap." Counted by assets rather than companies, the decline is only about 5%, because public-to-public acquisitions preserve assets. Mauboussin also points to 720 IPOs in 1969 and to the 27 CFA charterholders now covering each public company, which bears on how hard it has become to stand out.
The middle of the conversation is about economics. Tangible investment once ran about double intangible; now the ratio has flipped. Work by Luminita Enache and Anup Srivastava splits SG&A to make intangible investment measurable in aggregate. Paul Romer on non-rival goods and Brian Arthur on demand-side scale explain why network effects produce 80 to 90% market shares. The more contested claim is that stock-based compensation is financing, which would cut operating cash flow by 15 to 20% for many companies.
The last stretch turns to private equity. Steve Kaplan's data show higher entry multiples going with lower subsequent returns, with 2019 averaging a record 11.5 times. Over 40% of companies credit future savings in adjusted EBITDA. Software's share of buyouts has risen sharply, and the CLO market funds much of the leveraged loan growth. Mauboussin notes that general partners cite 9 to 11% net returns as the asset class norm while each expects about 20% from their own fund. A shorter, contested exchange defends active management as a source of price discovery.
Mauboussin argues that the shift from public to private equity reflects structural changes in how companies grow and finance themselves, including intangible investment, stock-based compensation as financing, and network-driven concentration, so investors must change how they measure value and where they look.
- Public company counts peaked in 1996 and fell as IPOs dropped and M&A and buyouts absorbed companies, while firms now stay private longer.
- Intangible investment now exceeds tangible investment, and stock-based compensation functions as financing that standard free cash flow metrics miss.
- Network effects and increasing returns raise concentration and ROIC gaps, making it harder for small companies to scale as public firms.
Rising costs of being public and companies listing later shrink public markets relative to private ones.
- Counting assets rather than companies, the drop in public assets from 1996 to 2020 is only about 5%, versus a halving of public company count, because acquisitions of public companies by public companies and of private companies by public companies are treated as asset-preserving.
“rather than us having half as many public companies in 2020 versus 1996, the assets are only down about 5%”
- In 1969, before Nasdaq was founded in 1970, there were 720 IPOs, which equals about 20% of the number of public companies existing today.
“in 1969, there were 720 IPOs. Which is 20% of the number of public companies today”
- The number of US public companies rose from 1976 to a peak in 1996 and then declined through 2016 and 2019, even though the economy and population grew, creating a 'listing gap' relative to what size would predict.
“the data went from '76 to 2016, so it was 20 years up because of the peak year for number of public companies was 1996”
- The age of companies at IPO rose from just under eight years in the mid-1970s through the late 1990s to just under 11 years today, so venture-backed firms are staying private longer before listing.
“It was just a little under eight years ... today it's closer to just a shade under 11 years”
- Jay Ritter's work suggests small companies find it hard to become medium and large companies, so they sell themselves to large established buyers where economies of scope are greater.
“is this idea of economies of scope. Small companies are having a harder time transitioning to becoming medium and large companies”
- As of year-end 2019, US public equity markets were about $38 trillion, roughly 27 times the $1.4 trillion US buyout AUM and about 80 times the roughly $450 billion US venture capital AUM.
“public markets as of year-end 2019 ... were 38 trillion dollars”
- Academics argue the costs of being public, including regulatory, legal, and IPO costs, have risen more than the benefits of capital access and compensation flexibility, so companies must be larger to bear public listing burdens, creating a small company trap.
“they basically argue that the costs have gone up more than the benefits, which means that you need to be a larger company, to be able to shoulder these burdens”
- Annual IPOs averaged about 280 from the mid-1970s through 2000, but dropped to an average of about 115 per year from 2001 through 2019.
“the average was about 280 per year ... the averages dropped to 115”
Unadjusted accounting for intangibles distorts value, earnings, and free cash flow, hiding real investment.
- Srivastava and Lev found that when intangibles are adjusted, 30 to 40% of companies in the Fama-French value and glamour categories move out of their category, and the adjusted value signal is much stronger.
“up to 30% to 40% of the companies in each of those categories leaves the category. And they claim that they find a much stronger signal”
- Intangible assets differ from tangible ones because they are non-rival, usable by many at once, and partially excludable, and Romer showed that non-rival goods that are partially excludable allow companies to earn huge excess rents, even though no laws of economics have changed.
“What Romer talked about was these non-rival goods that are partially excludable, which allow companies to generate huge excess rents”
- The common definition of free cash flow as operating cash flow minus capex is extraordinarily misleading because it omits many items, so analysts must carefully trace where accounts such as stock compensation are recorded.
“free cash flow is just cash flow from operations minus capex. And that is extraordinarily misleading”
- Walmart went public in the early 1970s and had negative free cash flow every year for its first 15 years as it built stores, requiring external financing, which contrasts with today's capital-light intangible companies.
“Walmart went public in the early 1970s, and in their first 15 years, they had negative free cash flow every year”
- Stock-based compensation should be treated as a financing item rather than added back to cash flow from operations, and doing so would reduce operating cash flow by 15 to 20% for many companies, with the ratio much higher for smaller and younger firms.
“By the way that should not be an add back to cash flow from operations. That should be in the financing section”
- Capitalizing intangible investment would materially raise both earnings and invested capital, so multiples such as price-to-earnings and price-to-book would send a very different signal about value.
“if you capitalize these investments, it will have a very material impact on earnings, which will go up a lot”
- Dan McCarthy and Pete Fader have extended customer-based corporate valuation, making older ideas more rigorous, and such tools can give certain investors valuable insights into intangible-heavy businesses.
“Dan McCarthy and Pete Fader have done this great work on customer-based corporate valuation”
- A rival good can only be used by one person at a time, like a laptop or a book, while a non-rival good such as software, a recipe, or any digital product can be used by everyone simultaneously.
“rival goods are goods that one person can use at a time. ... a non-rival good is one that we can all use at the same time.”
- Whatever the accounting presentation, free cash flow itself is unchanged, so intangible adjustments matter for seeing investment and return on investment, which in turn informs future profitability and free cash flow.
“the thing that doesn't change throughout this is free cash flow. Free cash flow is unperturbed.”
- Companies use adjusted EBITDA figures that include future cost savings when they talk to creditors, so a company reporting 125 instead of 100 can persuade lenders to lend more.
“these are the kinds of numbers they're using when they're talking to their creditors as well. So will you lend to us? Our EBITDA is not 100, it's 125”
- Separating SG&A into a maintenance portion needed to run the business and an investment portion makes intangible investment measurable in aggregate going back to 1970, but it is still very hard to do for individual companies.
“segregate SG&A into a component that's maintenance, the SG&A we need to run the business, and investment SG&A.”
- Many intangible-heavy companies appear capital-light because they raise capital internally from their own employees through equity pay, a major change from 40 or 50 years ago when firms like Walmart had to raise external capital.
“a lot of the sense that these things are super capital-light is because in a sense, they're raising capital internally with their own employees.”
Network effects concentrate market share and returns among superstar firms.
- In traditional businesses the top companies hold roughly 40 to 50% market share, while in network effect businesses the leading firms reach 80 to 90% market share, levels never seen before.
“top companies have 90% market share, 80% market share, things that we had never seen”
- The Herfindahl concentration index has moved in a U-shape that mirrors the inverted U in public company counts, falling in the 1970s to 1996 and rising in the last 20 years, largely due to M&A.
“that actually is a U. So it's almost a perfect mirror of that. So Herfindahls we're high in the 1970s, they dropped in 1996, and then they're back up”
- When operating margins are sorted into quintiles, the bottom three quintiles fluctuate with the economic cycle without structural change, and the rise in averages is driven by the top quintile, so averages mask a highly skewed distribution.
“the top quintile, in particular, pulls everything up. So the average is, again, these averages belie these weird sets of distributions”
- Five stocks (Facebook, Amazon, Apple, Microsoft, Google) added roughly $1.8 trillion of market cap through July 30, 2020, more than the entire aggregate US buyout industry and multiples of the venture capital industry.
“five stocks: Facebook, Amazon, Apple, Microsoft and Google ... their market cap went up roughly by $1.8 trillion”
- The gap in return on invested capital between the biggest and smallest companies was about 15 percentage points 20 years ago and is now closer to 30 to 35 percentage points, a widening that reflects superstar firms.
“that gap was, call it 15 percentage points 20 years ago, and now it's closer to 30, to 35 percentage points”
- Beyond classic supply-side economies of scale that lower unit costs, Brian Arthur's demand-side view holds that as more users adopt a good, willingness to pay rises, so platforms that do not raise prices give away consumer surplus.
“as more users use your good, the willingness to pay goes up, because it becomes more valuable for them”
- Brian Arthur's increasing returns idea holds that under certain conditions, especially with network effects, one company can capture the spoils, contrary to the microeconomic view that returns on capital migrate toward the cost of capital through competition.
“return on capital tends to migrate toward the cost of capital ... Brian heretically came out with this idea of increasing returns”
- Historically, monopolies like AT&T and utilities were regulated in exchange for capped profits, but some modern network companies become natural monopolies without meaningful regulation, prompting concerns about unjust profits, and breakups have often been good for shareholders.
“there've been cases where these monopolies have broken up, and it's actually been pretty good for shareholders”
- Enache and Srivastava split SG&A into maintenance and investment components and found a large upswing in intangible investment since 1970, so claims that companies have stopped investing miss a major component.
“segregate SG&A into a component that's maintenance, the SG&A we need to run the business, and investment SG&A”
- Back-of-envelope estimates put US intangible investment in 2019 at about $1.8 trillion, compared with roughly $700 billion of capex and $400 billion of R&D.
“the intangible investments in 2019 for U.S. companies, was about $1.8 trillion dollars. Capex was probably about $700 billion”
- The idea of network effects is old, as AT&T's 1908 Annual Report laid out network effects clearly, though their applicability to value creation is greater today as intangibles grow.
“I actually just was looking at the other day at the 1908 Annual Report for AT&T. They actually laid out network effects beautifully”
- Carol Corrado's work shows that tangible investment was about double intangible investment in the 1970s, but today intangible investment is about 1.5 times tangible investment, a complete flip within a couple of generations.
“Tangible investment was double that of intangible investment ... today intangible investments are 1.5x tangible investments”
High buyout entry multiples and flawed return metrics make excess buyout returns hard to earn.
- Dan Rasmussen's analysis found that historically, when buyout entry multiples exceed roughly 10 times EBITDA, there is a warning sign that generating excess returns becomes very difficult.
“typically over 10 times, at least historically, over 10 times there's a warning sign that it's very difficult to generate excess returns”
- IRR assumes capital returned during the investment period is reinvested at the IRR itself, an unduly optimistic assumption, and it lacks a benchmark for judging whether a given number is good; modified IRR gets closer to realistic figures.
“The biggest one is the assumption that capital that's returned during the period, earns the IRR rate of return”
- Kaplan's data show an almost perfectly downward-sloping relationship between buyout entry EV/EBITDA multiples and subsequent PME, so the more a buyout firm pays, the lower its future public-market-equivalent return; the average multiple paid in 2019 was a record 11.5 times.
“almost a perfect looking downward sloping chart. Which is to say, the more you pay, the less the PMEs are in the future”
- Buyout exits used to be mainly sales to strategic buyers with IPOs significant in the 1980s, but recently the largest exit channel has been secondary transactions, where one buyout firm sells to another, recycling capital.
“this big surge has been so called secondary transactions. Where the buyout firm sells the company to another buyout firm”
- Almost 90% of surveyed institutional investors have penciled in 8% or higher returns for alternatives, and 40% of them have penciled in 14% or higher, expectations that look difficult to meet given low public-market returns.
“almost 90% said they have penciled in 8% or higher returns for alternatives. 40% of them ... have penciled in 14% or more returns”
- In 2019, a record share of companies (north of 40%) used EBITDA adjustments that credit future cost savings, and a couple of years later many of these companies miss their adjusted EBITDA forecasts by a meaningful amount.
“in 2019 we had a record year of companies that have EBITDA adjustments, and that was something, I think it was north of 40% or more”
- Return on capital measures how many dollars come back per dollar invested but ignores timing, so a 1.5x return in 12 months is good while the same multiple over five years is not, and it also lacks a benchmark.
“if you give me a dollar and I give you $1.50, and I do that in 12 months, that's pretty good. But if I do that over five years, it's obviously not very good”
Buyout gains rest on operational improvement that may not survive sponsor resale.
- In the 1980s most venture-backed exits were IPOs, often done quickly after founding, while today the vast majority of venture exits are sales of the business to other companies.
“the vast majority of exits were IPOs ... today the majority of exits, the vast majority are actually sales of the business”
- The biggest recent growth in buyout exits comes from secondary transactions, where one buyout firm sells a company to another, so capital is recycled, and this raises the question of whether operational improvements continue after the second owner takes over.
“this big surge has been so called secondary transactions. Where the buyout firm sells the company to another buyout firm”
- The core premise of buyouts has two parts: better governance through smarter, more engaged boards and better incentives, and operational improvement, and the second part may not persist when a company is passed to another sponsor.
“one is a governance thing if there's better governance. And I think that is probably really true. Smarter boards, more engaged, better incentives”
- Lower interest rates can justify some increase in the multiples buyout firms are willing to pay, since debt service is central to the buyout model, but high multiples are also seen across all industries, so rates explain only part of the rise.
“interest rates clearly are lower. And a lot of the buyout industry is about servicing debt ... There may be some reasonable increase in multiple you'll be willing to pay.”
- A CLO packages many leveraged loans and slices them into tranches with a cash-flow waterfall, so the AAA tranche receives cash first, and this packaging is what allowed the loan market to grow so fast.
“you take a package of these loans, and then you break them into tranches, and there's a waterfall for cash streams. So AAA get all the first stream of cash”
Pension shortfalls and allocator career risk push capital toward private and leveraged credit.
- The leveraged loan market has grown to be as large as the high-yield bond market, with CLOs absorbing about half of leveraged loans and packaging them into tranches that are sold to investors.
“the leverage loan market now, is as large as the high-yield market. So that's a really striking finding and the CLO market now is about half the size”
- Depending on the estimate, underfunded US pensions total somewhere between $1.5 trillion and $6 trillion, creating pressure for institutions that face growing liabilities and cannot count on higher cash flows.
“underfunded pensions, in United States are somewhere between one and a half and six trillion dollars”
- Allocators face an agency issue in which committing to venture funds is hard to evaluate for 5 to 10 years, while a public equity decline is visible within months, so career risk pushes CIOs toward illiquid alternatives they cannot easily judge.
“We don't really know if it did well or poorly for 5/6/7/8/9/10 years”
- CalPERS' assumed plan return compared with Treasury yields shows the implied equity risk premium moved from roughly zero in 1960 (4% return versus 4% 10-year yield), to about negative 6 points in 1981 (8% return versus 14% 10-year yield), to about 6 points now (7% planned return versus 70 basis point 10-year yield).
“in 1981 ... the 10-Year was yielding about 14%, and the planned returns were 8%. So there was a -6% risk premium”
Private-market access and institutional investment judgment often fall short of expectations.
- Even top venture funds like Benchmark, with fund sizes of $400 to $500 million, are not needle movers for large institutions, so access to the best funds generates dollars that do not matter much at institutional scale.
“their latest fund is the same size as their last few funds, which is between 400 and 500 million dollars”
- Public Market Equivalent methods, developed by academics about 15 years ago, allow individuals to access private asset classes, but individuals may end up in the weaker parts of the industry, which would be poor outcomes for them.
“In some ways that's an exciting development, because that means individuals can access these asset classes”
- In a Journal of Applied Corporate Finance roundtable, general partners said the asset class would earn 9 to 11% net, yet each believed their own fund would earn closer to 20%, showing they know base rates but do not apply them to themselves.
“they say, the asset class will earn somewhere between 9% and 11% net. He goes, But our fund, we think we're closer to 20%”
- Mauboussin's team studied roughly 50 investment committees of endowments and pensions, building a cognitive diversity scorecard with Scott Page and applying dimension reduction and k-means clustering to test whether committee cognitive diversity matters.
“about 50 investment committees of real endowments and pension funds”
Venture and buyout persistence stems from firm skill and successful managers' follow-on advantages.
- To test whether VC success lies in the firm or the individual, one should examine how partners who left leading venture firms perform on their own.
“look at partners who have left these leading venture firms, and to see how they do on their own”
- As skill improves and more money and smart competitors enter, it becomes intrinsically harder to generate excess returns and distinguish oneself consistently, a pattern that is clearly true in public markets and probably true in buyouts.
“as skill improves, and more money comes in, more competition comes in, it's harder to grind out excess returns”
- Since 2000, past performance has continued to predict future performance in venture capital, while persistence in buyouts has moderated, with some measurement-dependent persistence remaining.
“Since 2000, persistence remains in the venture capital industry”
- In buyouts, top-quartile funds successfully raise follow-on funds while bottom-quartile funds drop out or participate far less, so successful managers become the main competitors in subsequent rounds, although new entrants keep the story incomplete.
“if you're successful in a sense it winnows out the weaker guys, and the stronger guys have to compete with one another”
- Venture firms with early luck can build reputations that attract the best startups, and those startups' prestige in turn reinforces the firm's success, creating a positive feedback loop that sustains persistence.
“this idea called preferential access ... a venture firm does well, and it could do well originally by luck”
Active management and analyst scrutiny keep public prices informative despite indexing.
- Public companies had fewer than one CFA charterholder per company in 1976, but in 2019 there were 27 CFA charterholders for every US public company, meaning far more analysts now scrutinize each firm.
“in 1976, there were less than 1 CFA charter holder, for every public company ... today there are 27 CFA charter holders”
- Indexing is a sensible solution for many people, but active management will not disappear because it provides price discovery, which indexers benefit from, and liquidity, and fees paid to active managers effectively subsidize indexing.
“Essentially the fees paid to active management subsidize the indexing industry”
- It is not clear that breaking up dominant companies would hurt shareholders, since prior monopoly breakups such as AT&T have sometimes been good for investors.
“it's not super clear that breaking up these companies would be bad for shareholders. In fact, there've been cases where these monopolies have broken up, and it's actually been pretty good for shareholders.”
- A revised edition of Al Rappaport's Expectations Investing, written in the spirit of the original book, would find that most of the original ideas have stood the test of time over the past 20 years.
“I'm happy to say that a lot of this stuff stands up pretty well. It stood the test of time.”
Software and cheap debt are reshaping buyout financing beyond hard collateral.
- Software's share of buyout deals rose from roughly 6 to 7% to about 17 to 20% within five years, challenging the traditional view that only hard assets and tangible collateral can support leveraged buyouts.
“In the last five years, we went from something like 6 or 7% of the deals being in software, to now just under 20%”
- A high-yield issuer recently sold debt at below 3%, roughly 200 to 225 basis points over the 10-year Treasury yield, a level Michael found hard to believe.
“a high-yield issuer, issued debt at below 3%. It's 200 basis points or 225 basis points over the 10-year Treasury note yield”
- Although buyout multiples have risen, leverage has not increased much because lenders monitor debt-to-EBITDA ratios, so the higher purchase prices are largely financed with more equity.
“as the multiples have gone up, there has been some more leverage, but a lot of it is just more equity going in”
Investors should work with reality and pursue less-crowded, exploratory niches.
- A durable career advice is to do something other people are not doing, such as less-trafficked markets or geographies where one can be the smartest person in the room, accepting that such niches are often smaller and capacity-limited.
“try to find something that's not what everybody else is doing. The challenge with that is, it's often the case that those are smaller asset classes”
- Investors should avoid lamenting policy moves such as Fed actions or indexing trends and instead play the cards they are dealt, working with the world as it is rather than as they wish it were.
“You have to deal with the world as it is, not as you want it to be”
- Organizations must decide how to allocate time, energy, and capital between exploiting what they already do well and exploring new opportunities, and this balance should change over the life of a business.
“how much allocation of my time, and energy and capital do I spend on just exploiting what I'm doing well, and how much do I allocate to exploring something new”