Most senior roles are pretty clearly delimited: CFOs keep track of money, CMOs make products famous, heads of engineering make sure those products work, and heads of R&D figure out what the next products should be. And if there's something that none of them are specifically responsible for, it's the CEO's job to determine what that is and to get it done. That makes "CEO" a funny sort of job, because its responsibilities exist through a process of elimination.

As the epigraph above makes clear, one of the responsibilities that falls on the CEO is capital allocation. It’s a vague term, but that's because it's an open-ended project.

Think about it like this: a mature, functioning company generates more cash than it consumes. And since some of this cash could be returned to shareholders, for many public companies the principle capital allocation question is deciding between dividends and share buybacks (and, within the share buyback question, to decide on timing). In any case, some fraction of a company's profits will typically be reinvested in the business. As Buffett points out later in the same letter, it doesn't take many years of earnings retention for a CEO to be personally responsible for most of the capital that a business has invested. If the money is spent well, that's a happy result for shareholders; if poorly, it's not.

The narrowest question in capital allocation is to ask: is it better for shareholders if we keep their cash and do more of what we're doing now, or if we return it? Some businesses produce fairly predictable unit economics, so the capital allocator can say something like:

• If we spend an incremental $1m on marketing our SaaS product, we'll probably get about X new customers; the net present value of the profits these customers produce will be roughly $Y. If $Y is greater than $1m, we should spend the money.

• Or, equivalently: it takes about $1m to start up a new location for one of our restaurants. These typically produce some semi-predictable annual return, factoring in both direct costs like rent and wages and indirect costs like a slightly higher administrative burden at the corporate HQ.

• Even financial companies can get in on the fun: a venture fund might estimate the number of cold-outreach emails a new associate can send, the conversion rate between those cold emails and completed investments, and their expected fee impact.

But the interesting choices CEOs have do not usually boil down to turning a dial labeled "growth" or "shareholder returns." Instead, they're operating in conditions of extreme uncertainty, and making decisions whose payoff is far in the future and dependent on external factors.

Taking the examples above, it's straightforward to tweak them in a way that makes the outcomes much harder to predict:

• $1m on marketing could be swapped with $1m to develop a new feature. Customers are asking for it! On the other hand, they're already customers, so maybe they don't need it (many companies reach an equilibrium where their customer base is defined by people who insist they want something but obviously don't want it enough to switch to someone who has it).

• A restaurant could slow their expansion plans slightly in order to research and develop a new menu item. At scale, this is a bigger deal than any one location: for a 500-restaurant chain, for example, a new item might need to add just 0.2% to the average check size to be worth more than opening up a new restaurant. But there's also a cost to menu complexity, and those changes are hard to A/B test—which means that simply scaling units also means scaling the sample size for new tests.

• For a fund running a stock-picking strategy, it's relatively straightforward to estimate how much alpha that strategy could produce if it were expanded into a new industry, or into a new geography. But entirely new strategies can be complementary to existing ones: a company with a good quant team will have a much better idea of what risks the discretionary traders are really taking (stay tuned for a future piece about this!), but won't know what the real returns are. And "strategy drift" is generally a bad word in finance; if someone has spent years investing in small-cap energy companies and suddenly starts talking about hyperinflation and the inevitable collapse of the euro, the default assumption is that they've had a tough couple years in their main business.

Good capital allocators have to be relentlessly good at predicting the future state of the world. Whenever there's a big shift in technology, or a change in how markets are structured, there's a period where the change is obvious but the lead time for reacting to it is long.

You can see echoes of this—and get a rough proxy for a company's ability to execute—by looking at which big businesses launched NFT strategies when, and especially by looking at which ones got into NFTs well after the hype had died down. (Interestingly enough, one of the first public companies to get into NFTs was Liberty Media, whose Formula One business launched a crypto-related game in 2019. It was shut down last year.)

Good capital allocators also have to be ruthlessly accurate at assessing their company's own strengths. The more distant a capital allocation decision is from just scaling up the core business, the more other companies could conceivably do exactly the same thing. For example, if a company is getting into some new industry, it has to either explain why it's going to be more profitable than incumbents or why the incumbents are going to be more profitable in the future—because shareholders already have the option to allocate capital into that industry, by selling their stock and investing in the incumbents.

This dynamic gives us another way to think about dividends and buybacks: they are multi-billion dollar expressions of a quiet, monastic sense of humility—an admission that the company using them is about as big as it can sensibly be, and that its management and culture doesn't give it any particular edge in new areas. The alternative is growth through inertia, either expanding in the same industry past the point where there are economies of scale, or expanding into other industries for no reason other than available cash and whim.

Capital allocation is partly a matter of money, but it's also a matter of attention and time. When good CEOs get praised as good capital allocators, it's really a way of saying that they know how to focus—on what they can improve in their existing business, on what else that business could turn into, and on what shareholders could do with the money if it were returned to them.

Thanks to Rahim Nathwani for suggesting this a while back.

A Word From Our Sponsors

Meco is a distraction-free space for reading and discovering newsletters like Capital Gains and The Diff, separate from the inbox. Plus it’s completely free.

• Meco connects to your email, allowing you to instantly move your newsletters to the app (and clear your inbox).

• The experience is packed with features to supercharge your reading including the ability to group newsletters, set smart filters, bookmark your favorites and read in a scrollable feed.

• Over 10k readers are now using the app.

Head over here to try Meco today!

Read More in The Diff

And in The Diff we’ve written a few pieces about capital allocation:

• A comparison between how Amazon and Google do it ($).

• A look at Liberty Media, which has an amazing track record of generating cash and putting it into the right businesses.

• And a profile of Constellation Software, which buys up tiny vertical software companies and has built a compounding machine from this ($).

And later this week, we’ll be looking at a financial intermediary that shifted from just helping its clients make deals to betting on their outcomes. You can subscribe today to read it.

Share Capital Gains

Subscribed readers can participate in our referral program! If you're not already subscribed, click that Get Started button below and we'll email you your link; if you are subscribed, be sure to check out the email version of this edition.

Help us learn more about you!

We'd like your feedback so we know what to cover, so if you haven't already, please fill out this one-minute form.