Square’s recent $2.9 billion IPO is either a huge win for its investors or a sign of very bad days to come, depending on whether you believe the “high-five” tweets from prominent venture capitalists (VCs) or the slew of hand-wringing articles — from sources like NPR, The Verge and the Wall Street Journal — essentially saying it marks the end of free-flowing startup capital.

The wildly contradictory responses to Square’s November 19 debut on the New York Stock Exchange, when the payments-processing company’s stock was priced below its previous funding round in the private markets, has left a lot of people in the start-up community (and beyond) wondering what’s going on.

My take is that the IPO is further evidence that the new rules of the hunt for “unicorns” — startups valued at more than one billion dollars — are creating a small class of “haves” and a much larger class of “have-nots.” The lucky ones are the earliest-stage investors that find tomorrow’s unicorns. The less lucky are often the late-stage investors — generally the providers of the bulk of the capital — and the talent that signs on as these companies grow to fighting strength and scale. I think this is likely to cause a rebalancing of late-stage valuations and an increase in transparency in private markets.

The basic facts are these two data-points: The company, which started with zero value six years ago, ended up worth $4.2 billion at the close of the market on the day of its IPO. Thus the high-fives. Yet at the moment of its public market debut, its lower-than-planned share price put the company’s value at just $2.9 billion, significantly lower than its $6 billion valuation at the time of its Series E funding round a little over a year earlier, implying the dreaded “down-round.”

And yet, the real substance to this story — in terms of how if affects the future — may have more to do with the many markers in between those selected pricing bookends. In other words, behind the magnitude of the value created by Square from inception (which is significant and real), how is it distributed and allocated across the various stakeholders, VCs and employees? What does that suggest going forward?

Let’s use Square as Exhibit A for this allocation-of-value story, which is really just beginning. For those wanting a spoiler alert, the old adage of “the early bird gets the worm” is apropos here — both for employees and investors — but very little of the capital or employees were there for the best spoils, which may have implications for the future.

Some quick math from the various public filings suggest that the investors did well — and pretty much as they were supposed to — with later rounds taking less risk and generating lower returns. These rounds of equity invested are labeled “A” through “E,” with each letter designating a subsequent discrete investment, generally by different investor groups. Each investor group that follows an earlier one is taking less risk because the company is further along in its development and, as such, is presumably expecting lower returns. But just “how low?” is the question.

Note that sometimes a round will have two closes with modestly different terms. While the B and C rounds each had two tranches of closing dates for Square, I collapsed them in the tables that follow, because their valuation levels and return profiles are very similar to each other. But the E round is very different, so that it is shown in the tables that follow by tranche.

Interestingly, the valuation in each round in Square’s case was essentially one multiple lower than the previous one: The B round was 5x the A round, C was 4xB, D was 3xC, and E was 2xD. While simple to remember, and probably not fully a coincidence, it might not be the best approach for the C/D/E rounds in the future.

Sources: Crunchbase.com, Amended S-1 filed 11/16/15

Translating the original investment of each round into compound annual growth rates, and taking into account time invested, shows the following:

The A and the B rounds generated compound growth rates over a five-to-six-year period of 72% to 95%. Considering that only about 25% of startups make it to a C round, this return seems about right; if 75% of investments go to zero for these early-stage investors, then the other 25% need to generate close to 100% in annual returns to deliver a weighted average overall return to their limited partners of about 25% to 30%.

If you think about it, in a different era the C, D and E rounds of Square might have been conducted in public markets, in which price discovery is based on inputs from a much broader investment base, and in which the terms of the deal are transparent to all current and future investors.

By the time of the C round, almost two and a half years after founding in February 2009, the company achieved its unicorn status, valuing it at $1 billion. It had expected revenue in the calendar 2011 year of $40 million (per an article in the New York Times). Although the C round may not have generated quite the returns hoped for at the time of investment, the Series C value looks to be roughly in the right valuation ballpark. At an 18% Compound Annual Growth Rate (CAGR), this round came in when the risk was substantially less than the previous rounds; and an 18% CAGR is nothing to sneeze at, compounding for four years.

Where things get really interesting is the D and the E rounds. Here, one might argue that proper private value discovery is much more challenging in closed private rounds than it would be in the public markets.

The D round is sitting at a 3% CAGR now, for its three years of capital invested, clearly well below the expectations at the time the investment was made. And it is even worse for the E-2 round, which is down 96%.

Surprisingly, the investors of the first tranche of the Series E round were guaranteed a 20% return at the time of an IPO, through the issuance of extra shares if the IPO price were to be lower than the agreed upon share price. Consequently, because of the number of extra shares the Series E-1 investors received as a result of the $9-per-share IPO price, their cost basis per share was effectively cut in half, dropping from the original $15.46 per share to $7.50 per share.

Sources: Crunchbase.com, Amended S-1 filed 11/16/1

This means that the Series E-1 investors with the ratchet clause had a higher CAGR than the Series C and D investors, even though those investors put money in two or three years earlier and assumed more risk. The Series E CAGR is somewhat close to the Series B-2 CAGR, and that round was completed almost four years earlier.

Granted, these Series E-1 returns have accumulated for only one year, versus four years for the earlier rounds. Nevertheless, this “ratchet,” acting almost like an embedded put option tied to a discount to the IPO price, will result in a highly non-uniform return experience for the late-stage investors.

In a public market situation, this would not occur; regulations in place make sure all investors see the same information. and the terms of any public offering are transparent and harmonized across investors.

Another critical angle here is that it is in these late stages of capital raising that the bulk of the employees join. To wit, hiring talent is a primary use of the capital, funding salaries of the employees that will drive the scaling of the business, after product/market fit is found.

At the time of the A and B rounds, in which investors (and early founding employees) were well-rewarded for their risk, the company only had about 10 employees. By October 2012, around the time of the D round, the employee base had grown to 300. In the most recent three years, it has expanded to about 1,200. Most of those employees joined with the dream that they would share in the value creation.

The three most fascinating insights to how the invested capital is allocated in Square (and likely many other successful unicorns) are:

The tricky price discovery and generally lower returns for these late-stage private rounds is applicable to the bulk of the capital that was invested. In this case, the D and E rounds together represent a total of $402 million dollars of the $550 million raised, or 73%.
The return profiles of the D, E-1, and E-2 rounds are dramatically different, even though they were successive to one another. Incredibly, the E-1 round has a CAGR approaching the B-round profile. As the light shines on these variations, it will beg questions.
Nearly 80% of the employees joined the company and were awarded equity grants at prices near or above the IPO value. And there are no ratchets or liquidation preferences that will adjust downward the exercise prices on those equity grants.

So where do we go from here? And what is the trickle-down effect to earlier rounds of capital if the late stages are coming to a table with not enough food in returns? Is this the end of the late-stage private investing?

I don’t think so, but here is a stab at my take on likely changes ahead:

First, late-stage private capital is here to stay, but it is like a partying teenager now, and it will need to mature to an adult, with more curfews and controls.

I think it is a positive development to have more capital options for private companies that the public market (Incredibly, Yahoo went public only two years after it was founded, with $1 million of trailing sales). And the structural factors at work creating this class of capital — hedge funds and mutual funds wanting to get in before the IPO — is not going away

But for this late-stage private capital to stay viable, the allocation of value today is too large now in favor of the earliest rounds. A recent article in The Economist talks about how late-stage values are already coming down, dropping an estimated 25% in the past six to eight months, according to Fred Giuffrida of Horsley Bridge.

This value adjustment will come both as demand slows — many funds are shifting their investment priorities closer to the seed and A rounds — and also with more muted step-ups in value from prior rounds. This may lower returns somewhat in A rounds and/or produce very high values in certain cases (a.k.a. Vessel) as the market goes through this adjustment.

Second, I think it will also force more discipline on burn rates of startups, as they come to realize that huge premiums for later-stage capital may not be there. Perhaps that means we will also end up with fewer steps in the process, as entrepreneurs realize they may need to stretch their capital for more time in between rounds, and it may lead to a more uniform understanding of what the startup has to achieve from A to B and B to C.

Third, the ratchets and liquidation preferences that act as embedded put options in various late-stage rounds today, are likely to become more transparent and uniform across rounds of capital. It is wacky that the return profiles here across D, E-1, and E-2 are so vastly different, and late-stage investors will increasing require pass-throughs of future favorable terms to apply to their own rounds as the public market disclosure process turns a light on in this dark room.

Obviously, for the specific case of Square, what matters most from here is the rest of the story; how Square performs in the next few years will determine a lot in terms of how the bulk of the employees and the last-stage capital invested (including those that just invested in the IPO) fares.

But I think there are some lessons learned here, that we will be taught a few more times again for those unicorns that are successful enough to access public markets. And in those lessons, we will see some changes to values across stages, more transparency in terms, and a move to reduce burn rates for those lucky enough to receive the capital.

A version of this post also appeared on Re/code

Circling Square

Follow Sue