When buybacks destroy value
A buyback is a trade between the owners who sell and the owners who stay. Pay too much and the ones who stay lose. The arithmetic, Buffett's two tests, and the EPS trick that makes a bad deal look like a good one.
Someone wins every buyback
When a company buys its own shares, the cash goes to whoever sold them. The owners who hold on get a bigger slice of a company with less cash. Whether that leaves them richer depends on the price paid compared with what the shares are really worth.
Marlowe Beverages (illustrative) has 100 million shares. Its business is worth about $4,500 million and it holds $500 million of cash it doesn't need, so each share is worth $50. It spends the spare $500 million on buybacks.
At $40, the sellers took $500 million for shares worth $625 million, and the $125 million difference stayed with the owners who held on. At $80, the sellers took $500 million for shares worth $312.5 million, and the owners who held on paid the gap. Paying the same $500 million as a dividend would leave each holder with $45 of business and $5 of cash. Still $50, before tax.
So a buyback at fair value does no more for you than a dividend. Below value it does more. Above value it moves wealth from the owners who stay to the ones who leave.
Two conditions
Warren Buffett has set out the same two conditions for repurchases in several Berkshire Hathaway letters, 1999 and 2011 among them. First, the company should already have enough money for everything the business needs, with room to spare. Second, the shares should be selling below their intrinsic value, and that value should be estimated conservatively.
Plenty of boards skip the second one and talk about "returning capital" as if the price didn't matter. Buffett treats a repurchase as an investment decision. Nobody would approve a new factory without asking what it costs.
Earnings per share rise at almost any price
Here's why bad buybacks survive. Earnings per share go up after nearly every buyback, whatever the price, so the headline looks like success.
EPS rises because the shares bought earn 2.5% and the cash spent earned 2%. Each remaining share is worth less, because Ingleby paid $40 for something worth $25. A bonus scheme tied to EPS would still pay out.
The same mechanics make a flat business look like a growing one. If net income stays at $400 million for five years while the share count falls 5% a year, from 200 million to about 155 million, EPS climbs from $2.00 to about $2.58. That's 5.3% a year of per-share growth from a business that didn't grow. Fine, perhaps, but don't value it as growth. DL 230 shows how buybacks inflate return on equity in much the same way.
Borrowed money at the top of the cycle
Buybacks follow profits. Boards buy most when cash piles up in good years and stop when a recession dries it up, often just as the shares get cheap. S&P 500 buybacks set a record in 2007 and had fallen by more than half by 2009.
Add borrowed money and the damage can be permanent. Saltash Homebuilders (illustrative) has 50 million shares at $120, fifteen times peak earnings of $8 a share. It borrows $600 million and buys back 5 million shares. Two years later housing has turned. Earnings are $3 a share, the price is $45, and the lenders want their money, so Saltash sells new shares at $45 to repay them.
Saltash bought its own shares high and sold them low, and that owner's stake ended about 14% smaller. Several big US banks did something similar, buying back heavily in 2006 and 2007 and selling new shares far cheaper in 2008 and 2009. Financial Fortress covers the balance-sheet side.
Judging a buyback fairly
- Value is an estimate, which is why Buffett asks for a conservative one. A small premium to a rough estimate is a judgement call. Double the estimate is a mistake.
- Hindsight. A price that looks silly five years later may have been reasonable on what was known then. Judge the board on the reasoning it gave when it bought, if it gave any.
- The alternative matters. A buyback at fair value is neutral against a dividend and poor against reinvesting at a high return. Incremental ROIC in DL 230 is the comparison to make.
Checking a buyback record on Dividend Line
On the Buybacks tab, Buyback Efficiency divides the cash spent on buybacks by the fall in the diluted share count, which gives an average price per share retired, and compares it with today's price. Because the divisor is the net fall, a company whose buybacks mostly replace employee stock shows a high price per share retired.
The card Were the buybacks well-timed? plots that price each year against the five-year moving average of the share price, green below and amber above, and gives a share of smart purchases once there are three years to judge. Self-funded or debt-financed? compares cumulative buybacks with cumulative free cash flow over up to 15 years: up to 85% is Self-funded, up to 110% Stretched, and beyond that Debt-funded. The Accretion Engine sets net income growth beside EPS growth and shows the gap as the Pure Buyback Lift.
A price below the moving average, or below today's price, says the shares were cheap against their own history. Whether they were below intrinsic value needs your own estimate.
Takeaways
Sources for this lesson
- Warren Buffett, Berkshire Hathaway letter to shareholders (1999)
- Warren Buffett, Berkshire Hathaway letter to shareholders (2011)
- S&P Dow Jones Indices, quarterly S&P 500 buyback reports