Shareholder buybacks have the odd distinction of being a topic that used to be legally contentious but irrelevant to the discourse, and have since completely reversed course. While we have records of stocks, bonds, puts, and calls dating back to the 17th century, it's hard to find documented evidence of a company buying its own shares back in public markets in the prewar period, though some corporate reorganizations implicitly took this form—forming US Steel meant cashing out Andrew Carnegie but keeping some other executives around.

In the postwar period, they started picking up. This paper has some data from the 50s and 60s, showing the total volume of repurchases rising from $46m in 1952 to $5.4bn—almost equal to total new issuance!—in 1968. There was a problem with this trend, though: it wasn't 100% clear that it was 100% legal. The SEC's guidance was ambiguous, so companies often structured repurchases as a formal tender offer process: if the stock's trading at $10, the company might offer to buy back a set number shares at $12, with some kind of pro-rata approach if the offering were oversubscribed (for example, if they offered to buy back 1m shares, and sellers offered 2m, each seller would get half of their stock repurchased).¹ The worry was that a company that buys back its own stock is potentially manipulating the price as well. And there are examples of this, albeit later: during Robert Maxwell's downward spiral into insolvency, he ordered his company's pension funds to dump other assets and support the price of his holdings. It's easy to imagine a scenario where a CEO cynically avoids getting a margin call by using buybacks to keep the stock high, basically transferring liquidity risk from the CEO's own balance sheet to that of the company.

There was also the risk that a company would trade on information that shareholders didn't have. This is a trickier situation, because it's hard to imagine any case where the company and its shareholders have identical information. On the other hand, it's very easy to imagine a scenario where the company's variant view is driven more by their assessment of the value of the business, not by how any given quarter is going. But this can be partly mitigated by doing more buybacks at times when information is close to symmetric (i.e. soon after earnings) and avoid them when the information gap is widest (after the quarter ends but before results get reported).

Instead of creating endless debates and billable hours over the finer points of securities law, in 1982 the SEC mercifully declared that open-market buybacks were generally allowed (they updated the rule in 2003 to require more disclosure around buybacks, too).

And companies tend to prefer buybacks to dividends:

1. They're more tax-efficient, since the dividend is fully taxed and the buyback is taxed only if the investor in question has a gain, and only to the extent of that gain. (Edit: this was incorrect—buybacks are also taxed by 1% at the corporate level, so while they’re still a relative winner in tax terms, they do face taxes. Thanks to @MetacriticCap for flagging this.)

2. Buybacks are more flexible, and can be sped up or slowed down depending on market conditions. Dividends have more signaling value about the company's long-term expectations, but also lock in an ongoing expenditure.

3. Most CEOs have a healthy enough ego to believe that even if the average stock is roughly fairly valued, their company is misunderstood and cheap. Given that CEO ego can be expressed by either spending money on expansion or spending money to buy back stock, and that the information advantage for buybacks is bigger, this is probably net healthy for shareholder returns.

The downsides of buybacks are also readily apparent. Companies have finite cash, and if they return it to investors they aren't using it to expand their business, create more jobs, develop better and cheaper products, etc. But that's just the first-order effect; the longer-term impact is more ambiguous. The shareholders whose stock gets bought back are going to do something with the money, either spending it (contributing to aggregate demand and thus boosting employment and investing indirectly) or by investing in something else. If a company can't find good places to reinvest its cash, then either a) the economy is out of investable (ROIC > WACC) opportunities, in which case they should return cash to shareholders who will at least enjoy spending the money, or b) there are good investing opportunities, just somewhere else, and statistically the money that flows out of a company with few investment opportunities is, on average, going to end up somewhere with better opportunities.²

One easy way to illustrate why buybacks are more fair than dividends is to propose two ways that a company could return cash to shareholders:

1. It could force everyone to sell 5% of their shares to the company at the current market price, whether or not they would prefer to own more, or

2. It could survey shareholders, find the 5% who are least excited, and buy the stock from them.

The second option obviously leaves everyone better-off. But the first option is economically equivalent to a 5% dividend yield, minus extra tax friction.

Two adjacent critiques of buybacks are that companies overpay for their stock and that they buy back shares opportunistically and then end up being unable to run their own businesses. There was a lot of discussion around this when airlines got bailed out during Covid: if they hadn't bought back so much stock, they wouldn't have needed so much money! But that's tantamount to saying that the right way to run an airline is to have a cash buffer sufficient to pay almost all of their operating costs while eliminating almost all of their revenue, for several quarters or years. You'd end up with airlines that kept more than half of their assets in cash! It seems much more optimal for the airlines to buy an insurance policy that covers disaster risk rather than to self-insure, and for those same airlines to expect some combination of a) lots of political condemnation when they inevitably get bailed out, and b) a bailout on terms that offer lots of equity to whoever provided capital. (If a company is worth bailing out, it's worth bailing out in exchange for warrants that make the government a substantial winner if it survives. And if it's not worth bailing out, it doesn't particularly matter whether the thing that killed it was a pandemic, terrorist attack, volcanic eruption, or tail risk yet undreamt of.)

It's harder to defend companies buying back their stock at high valuations and then seeing the share price crater. Charter Communications spent more money buying back shares over $600 each quarter than they spent buying them back in the last year with the stock under $400. But that's just one case of companies misallocating capital; Intel should have cut its dividend earlier, for example, and would Charter shareholders be that much happier if Charter had doubled down on its business instead of returning their funds a few years ago? It's easy to tell a story where buybacks wounded or killed a company, but often harder to tell a counterfactual story where some other decision, besides merely hoarding cash, didn't.

Buybacks are one tool among many, but they're popular among companies and investors for a reason. Most of the complaints that apply to them are applicable to dividends as well.. And if we limit both buybacks and dividends, the best option for companies is to just keep on growing—a critique of buybacks is, implicitly, a complaint that Big Business isn't nearly big enough.

Read More in The Diff

In The Diff, we’ve looked at buybacks from many angles:

• There are some companies that have persistently value-adding buybacks because their business is hard for anyone else to understand ($).

• You can divide buybacks into the ones meant to produce alpha and the ones that are a component of beta.

• Do valuations mean-revert over time? 

• Some of the biggest one-year gains in single stocks are caused by buybacks at low valuations followed by a change in sentiment about how long the business will last ($).Subscribe to The Diff

Join the discussion!

1  Today, game theory is more developed and while these tenders still exist, they generally give a range of prices and let shareholders specify one within that range, and then prorate if there's still excess supply.

2  This is easiest to illustrate by imagining an economy with two companies, SlowCo and GrowthCo. If SlowCo buys back stock, and its investors reinvest the money elsewhere, they're shifting capital to GrowthCo. If GrowthCo either issues stock, or declines to pay a dividend or do a buyback on the grounds that there are other assets out there for income-seeking investors, the total amount of capital in GrowthCo goes up, and the economy as a whole—remember, it's an economy with just two companies—necessarily grows faster as a result. More concretely, buybacks are bigger in brick-and-mortar retail than in e-commerce, while e-commerce businesses are more likely to IPO and raise secondary funding. So that’s a direct case where capital that was invested in, say, Macy’s gets transferred to shareholders who will reinvest at least some of it in Shopify.