A person's economic lifecycle looks something like this: you're born a net consumer of resources who produces no outputs that have any market value whatsoever, then, in young adulthood, you'll earn a bit, but you’ll also need to spend more (e.g. education, the frictional cost of on-the-job training, and later on the cost of a house and kids—all of which involve some cash-burning).¹ But at some point, if all goes well, you'll be making more than you spend, and your earnings will rise faster than your spending, so savings will start to pile up. Which is good, because not too long after that you'll either choose to stop working, or you’ll be unable to find any job that's worth the cost of not enjoying your retirement, so you'll downshift, and slowly spend your savings.

The corporate lifecycle is similar, but with a twist: companies are typically net consumers of cash early in their life, reach a pleasant middle age where they produce more cash than they can invest at an acceptable return, and then, again if all goes well, elegantly manage their terminal decline, cutting expenses faster than revenue declines until they can fully liquidate.

As with humans, earning more than you need to spend at a sufficiently young age is either a sign that something's going very well or that it’s going very poorly.² There are some companies that have pulled it off, but a rapidly-growing company that turns a profit very early on deals with one of two nice-to-have problems:

1. It can't raise outside funding to grow even faster, or

2. It can raise money but has internal constraints on finding ways to deploy it faster than it comes in.

Very early eBay is a decent case study of this: they were profitable from month one, despite charging fees based on the honor system. The huge volume of collectibles users were willing to buy and sell, and the fact that eBay didn't need to handle inventory or delivery meant that their costs were low. Meanwhile, the product was going viral within collectibles communities, and quickly became tied to the very-viral-indeed Beanie Baby boom.

Early Microsoft (disclosure: I'm long shares of no-longer-especially-early Microsoft) had a different set of constraints. An early experience with slow-to-pay customers made them cautious, software wasn't on the VC radar as a standalone business, and they were incredibly picky about talent but there weren't many professional programmers to go around (that last constraint was their real problem). So they had a headcount-light, work hours-heavy model where revenue persistently outran their ability to hire coders they were happy with. This, too, is a nice problem to have.

But in an abundant funding environment, it's a dangerous problem to have, because the company with the second-best product by a small margin can raise enough money to have the best marketing by a wide margin. This is a common enough risk that it's practically the fiduciary duty of an early-stage company CEO to outgrow their current market share in order to win their market, and then figure out what the right cost structure is after.

That process is unpleasant in many ways. Downshifting from maximum growth to prudent growth means paying more attention to expenses, putting more processes in place, and thinking more about just how much customers might be willing to pay. All of this is less fun than building new products and selling them to customers who are getting an amazing deal, but the funding required for doing that at scale is predicated on the cash flow generated from later, more corporate stages of a company's existence.

Where that growth rate settles varies significantly across different companies and industries. IBM is two thirds of the size it was twenty years ago, in nominal revenue, partly because of spinoffs and sales, but also because some of its core businesses are in secular decline. Investor returns have been mediocre, but they've managed to be positive over the last decade in part because for a while what IBM did with the substantial cash it still generated was mostly buying back stock. As a result, IBM's revenue per share is actually up over the same twenty-year period, from $53.26 in 2004 to $67.30 in the last twelve months.

IBM isn't a happy story for investors, but there are happier variants. Over the last ten years, shares of Dillards have returned 19%. Sales declined slightly over that period, but costs dropped faster, and they managed to buy back 63% of their shares outstanding. When the market correctly judges that a company is dying, but incorrectly judges how fast that will happen, two things hold true:

1. The company doesn't have many internal opportunities to reinvest its profits—the world is not clamoring for new Dillards locations at whatever malls still exist, but also

2. The stock is cheap, so the return on buying it can be quite high.

When that second scenario works, it works very well. 20% EPS growth can be manufactured from a 20% incremental return on equity, or from a 0% incremental return on equity and a policy of continuously buying the stock at 5x earnings. And then, if the business does perk up a bit, if only because of a cyclical swing at the right time, the stock in question becomes a cheap stock with a high per-share growth rate.³

This, too, doesn't last forever. The grimmest part of a company's life is when it's truly dying and the question is when to pull the plug. The graceful way to do this is to pay off debt and liquidate, but some companies handle the buyback phase by levering up, essentially running a capital structure arbitrage trade between optimistic bondholders and pessimistic equity investors. That can leave a company with zero financial flexibility once their operations start to weaken, and can force them to divest more and more assets at increasingly distressed prices, either to avoid bankruptcy or as a result of it.

And then, it's gone. The last employee cashes their last severance check, the last odds and ends are sold, and the business ceases to exist. Which is a perfectly fine thing for a business to do. Corporations are a way to put together people and assets to accomplish some common goal. Some of those goals will last a long time, especially if they're defined broadly—American Express and Wells Fargo still move money around, just not by stagecoach—but in some cases, the purpose will no longer matter and there won't be any reason to keep those valuable resources stuck together in a suboptimal way.

Just like in ecosystems, when one way to organize and direct a particular chunk of matter collapses into entropy, the resources it used don't disappear, but get repurposed. They usually get acquired by something younger, hungrier, and more eager to grow. It's not a pleasant cycle when you're on the losing end of it, but it's also an inevitable one. Specific employees, founders, and investors lose from the process, but we collectively come out ahead. The only thing worse than a company failing is a company that's failing to accomplish anything good for the world, but hoarding talent and assets that could be better used elsewhere.

Read More in The Diff

One good way to think about this model is to look at case studies from companies at different ages. For example:

• Here’s a piece on Snowflake, which prudently stayed in the cash-burning phase at scale ($)

• Here’s a look at a very early-stage cargo blimp business.

• This piece in 2022 is thinking out loud about how Meta ought to invest or return its cash ($).

• We examined when it’s optimal to run a business that is guaranteed to go under some day ($), with an emphasis on airlines.

• And here’s an analysis of WeWork as it slid into bankruptcy ($).

1. In accounting terms, buying a home is just changing what your assets are denominated in—you’re just going from having cash for a deposit to owning a house with an associated mortgage liability. But in terms of your use of scarce resources, buying a house is in one sense the peak of your burn rate: you're removing a fixed asset that's presumably worth a multiple of your annual pay from circulation. If renting and mortgages were forbidden, the cycle would be very different, with much later family formation and a period of net savings starting right after school and continuing until then.

2. Specifically, it means either that someone is a business prodigy who can earn an adult-level wage while working part-time in school—like Warren Buffett’s newspaper delivery empire, or oil entrepreneur J.D. Allen, who was out-earning his parents in high school—or it means that someone is getting so little support from their family that they’ve had to drop out of school prematurely to get a job that keeps them fed and sheltered.

3. Plenty of algorithms will be smart enough to differentiate between those two kinds of EPS growth, since one of them vanishes when the stock price rises while the other can easily accelerate if a higher stock price makes it easier for companies to recruit with equity. But some algorithms—and some individual investors!—will see the earnings growth rate, the stock price that matches it, and extrapolate both.

Join the discussion!