“Our view on share repurchases is that share repurchases are value neutral. It is not something that grows value. It does for the short-term holder, so maybe you can get a buck in the share price, but for the long-term holder it is not something that's value enhancing.”
That’s a quote from Gary Fayard; Fayard is the former CFO of The Coca-Cola Company (KO), a company that has spent nearly $14 billion on share repurchases over the past 36 months. That spending amounted to more than 50% of the company’s net income over that same period – spent on an action that, according to the company’s top financial executive, doesn’t add any value for the owners of the business. While it’s still early, my read on the new CFO is that she hasn’t brought a change in thought process to the role (hopefully I’m wrong).
I find this particularly astonishing because the company’s largest shareholder over the last quarter-century, Warren Buffett of Berkshire Hathaway (BRK.B), has been world renowned for his investing and business acumen – and he has a very different opinion on share repurchases than the executives at the company.
Warren Buffett (Trades, Portfolio) has discussed share repurchases multiple times in his shareholder letters over the years. The 1994 shareholder letter succinctly addressed the importance of this topic:
“Understanding intrinsic value is as important for managers as it is for investors. When managers are making capital allocation decisions - including decisions to repurchase shares – it's vital that they act in ways that increase per-share intrinsic value and avoid moves that decrease it. This principle may seem obvious, but we constantly see it violated. And when misallocations occur, shareholders are hurt.”
In 2011, he offered an example of the financial impact of repurchases when discussing IBM:
“Today, IBM has 1.16 billion shares outstanding, of which we own about 63.9 million or 5.5%. Naturally, what happens to the company’s earnings over the next five years is of enormous importance to us. Beyond that, the company will likely spend $50 billion or so in those years to repurchase shares. Our quiz for the day: What should a long-term shareholder, such as Berkshire, cheer for during that period?
I won’t keep you in suspense. We should wish for IBM’s stock price to languish throughout the five years.
Let’s do the math. If IBM’s stock price averages, say, $200 during the period, the company will acquire 250 million shares for its $50 billion. There would consequently be 910 million shares outstanding, and we would own about 7% of the company. If the stock conversely sells for an average of $300 during the five-year period, IBM will acquire only 167 million shares. That would leave about 990 million shares outstanding after five years, of which we would own 6.5%.
If IBM were to earn, say, $20 billion in the fifth year, our share of those earnings would be a full $100 million greater under the “disappointing” scenario of a lower stock price than they would have been at the higher price. At some later point our shares would be worth perhaps $1.5 billion more than if the "high price" repurchase scenario had taken place.”
Unsurprisingly, repurchasing shares at a lower cost can materially increase the number of shares repurchased at a set dollar amount and subsequently boost earnings / FCF per share.
As it relates to Coca-Cola, I think Mr. Fayard could argue that repurchasing shares is value neutral IF they purchased a stable number of shares each period (quarter / year) over a long period of time. If you settled on buying 10 million shares a quarter as the stock price moved from $10 to $20, back to $15, and then up to $25, you would spend $700 million over the course of the year for 40 million shares – at an average cost of $17.50 per share. To the extent that the long-term movements settle around an equilibrium that roughly approximates fair value (a fair enough assumption in my opinion), this action would be value neutral. Naturally, buying a set dollar amount would be even better (rather than ramping at higher prices), with an average cost of ~$15.60 per share in this scenario (knowing how much to purchase requires hindsight of average prices in the period in order to be exact).
For Coca-Cola, how do these scenarios compare to what the company has actually done? Well, here’s how much the company has spent on share repurchases over the past decade, in order from the most spent to the least spent: $4.8 billion, $4.6 billion, $4.5 billion, $2.9 billion, $2.4 billion, $2 billion, $1.8 billion, $1.7 billion, $1.5 billion and $1.1 billion.
Unsurprisingly, the highest spend has come from more recent years ($4.8B in 2013, $4.6B in 2012, and $4.5B in 2011): the company generated $10.5 billion in cash flow from operations last year, compared to $5.9 billion in 2004. However, this doesn’t hold at the low end of the range: the lowest two years for repurchase spend, at an average ($1.3B), or less than 50% of the ten-year average ($2.72 billion a year, inclusive of the two years), were in 2008 ($1.1B) and 2009 ($1.5B). Looking at 2008, we can see that the stock took a dive in October; despite the sizable dropoff in the stock price, 98% of the company’s full year repurchase spend occurred prior to the fourth quarter.
Over the past ten fiscal years, Coca-Cola’s repurchase spending has totaled $27.2 billion; it has repurchased 901 million shares (gross) in that period, at an average cost of ~$30.20 per share. Over the same ten years, using weekly stock prices (Yahoo Finance), the average price per share has been ~$24.40 per share; obviously this figure doesn’t account for the rising spending in recent years, which weighs the company average cost to more recent (higher) prices.
How about if we tighten the time frame to begin at the start of 2008? Over that period, the average weekly price for the stock has been ~$28.50 per share. For the first three years (from 2008 through 2010), the stock traded at an average weekly price of ~$23 per share. Coca-Cola’s repurchase spend over that period was at an average cost of ~$30. In that period, they only spent $5.6 billion – enough to purchase 186 million shares. In the ensuing three years, the company would spend $13.6 billion on 369 million shares – at an average cost approaching ~$37 per share. As a result, the six-year spending was $19.2 billion – at an average cost of nearly $35 per share; the average cost was more than 20% higher than the weekly average price for the stock over the same measurement period.
By the way, this doesn’t only hold in dollar terms (which is affected by rising net income / FCF over time); it holds as a percentage of net income as well: Coca-Cola spent an average 41% of net income on share repurchase in the ten year period from 2004 – 2013. There are only two years where the total was below 30% of net income: 2009 (at 22%) and 2008 (at 19%).
Conclusion
Mr. Fayard’s comments on the conference call apparently didn’t cause much consternation among the analysts – no one asked him about his response on that call or has asked the new CFO for clarification (at least not publicly). My conclusion is that the repurchase over the past ten years has been well below its potential (on the basis of timing, not as a reflection of average prices within the periods – though I think you can make a case on that front as well). Considering the dollar amount spent on repurchases over the past decade, it’s astounding to think that this topic is almost never discussed – and that management’s ill-conceived notion about the value of share repurchases has not been questioned.
With the help of two ending variables (901 million shares purchased and $27.2 billion spent), we can test two alternatives that Coca-Cola might want to consider: buying a set number of shares a year, or buying a set dollar amount a year. Using the same average purchase prices per year as actually reported, KO could’ve repurchased 90.1 million shares a year over the past decade for a total cost of $26.4 billion – saving shareholders $810 million. They company also could’ve repurchased $2.72 billion worth of stock a year and repurchased a cumulative 967 million shares – 7.3% more than they actually purchased. This would have a material impact on intrinsic value per share.
Those routes require knowing – or at least making an estimate – at the level of activity you’re likely to engage in over the coming period (in this case, a decade). Another option doesn’t require such knowledge: just like many companies do for dividends, Coca-Cola could set a repurchase payout ratio (I’ll do so as a percentage of net income). We would still be spending $27.2 billion over the course of the previous decade; in this scenario, the total number of shares repurchased would increase by ~3% (gross), relative to the company’s actual result.
While these numbers may sound small, they’re material: at the end of 2013, Berkshire’s 400 million shares of KO accounted for 9.08% ownership in the company; if the number of shares out was 5% lower at year end, Berkshire’s stake would account for 9.56% of the company – and add an additional ~$40 million to their share of look-through net income.
Remember that the alternatives discussed don’t require making intelligent / active repurchases (buying when the stock is cheap); this is simply moving towards an approach that approximates “average” – a bar that the company’s repurchase spend hasn’t cleared in the prior decade.
Clearly, there’s a need for some reconsideration of Coca-Cola’s repurchase strategy; they might take comfort in knowing that plenty of their peers in corporate America are in the same boat…
