Warren Buffett’s IBM Investment Through A British Lens

In November 2011, Warren Buffett revealed that he purchased $10.7 billion worth of IBM common stock to add to the ballast of publicly traded investments in the portfolio of Berkshire Hathaway. He has added to the IBM positions in the five years since the bulk of the initial investment, but the average price was $169 (this amount is constantly subject to revision as Buffett adds to the position).

People who are stirred by changes in stock price are well aware that shares of IBM have fallen to $122 from $169, and those who study the fundamentals of the business are well aware that this stock price fall can be tied to falling revenues. IBM generated $106 billion in revenues back in 2011, and only generated a little over $80 billion in revenues by the end of 2015. IBM’s ability to grow revenues has been an issue since the early 1980s, and decided to begin executing a strategy of permanent stock buybacks in 1993 to offset some of these fluctuations.

Since IBM began executing this strategy, shareholders have been rewarded with 12.11% annual returns since 1993. The S&P 500 Index, meanwhile, has returned 8.61% over the same time. A $10,000 investment into IBM grew into $132,000 while the same amount invested into the S&P 500 grew into $65,000 (both figures rounded to the nearest thousand). And here is the really interesting part: IBM only grew its revenues by 4.2% annualized since 1993. Due to dividends, buybacks, and a move toward products with higher operating margins, IBM has been able to enrich shareholders even while revenue results have failed to impress in their own right.

And here is something else that is interesting. For the first time in its corporate existence, the dividend has come to be a meaningful part of the total return portrait during Buffett’s holding period. From the 1990s through 2005, the dividend yield on IBM stock was under 1%. It crossed 2% in 2013. And now the dividend yield is 4.23%, something unprecedented in IBM’s history (at least dating back to 1956).

That means Warren Buffett has collected $19.15 per share in dividends on the earliest batch of shares that he has purchased for Berkshire. Although the price of the stock has declined 27% since he began buying it, he is only down 16.5% when you count the dividends that he has received (and only Buffett knows how he has deployed the IBM dividends at Berkshire, but it is also possible that the IBM dividends could have been used to purchase something that is now generating income of its own). Despite the unusually high amount of negativity, and hand-wringing over the share price, all it has been is a 16.5% loss.

Now, it is almost certainly true that business performance at IBM–especially the 25% decline in revenues–has been worse than Buffett anticipated or desired. But, at a price of $122, it really does not take much to go right for IBM to work out as an investment. It makes $13.25 per share, for a P/E ratio of only 9.2. It still makes $13 billion in annual profits, putting it at one of the thirty most profitable firms in the entire world.

What is attractive about IBM as an investment is that: (1) moderate, single-digit revenue growth can translate into double-digit total returns; (2) stagnant revenues can translate into single digit returns; (3) modest revenue declines can translate into breaking even; and (4) moderate or high revenue declines can translate into losses. The past five years have seen IBM perform according to Door #4.

For patient capital, a company paying out 4.23% in dividends, trading at barely over 9x earnings, and repurchasing 4% of its stock per year is interest. You got high dividends, large share repurchases, and attractive valuations. Those are powerful forces over the long haul. The starting point is so low that the odds of success are heavily tilted–it’s just like we talked about on Friday with American Express. You don’t buy American Express because you prefer its core business to Mastercard or Visa–you buy it because it’s so cheap you can’t help but do well. Someone who bought American Express at 11x earnings in 2011 has earned 7.77% annual returns since then, even though the operational performance at American Express hasn’t been great since then. It’s about the relationship between performance and expectations that are baked into the purchase price, and the current price of $122 for IBM is a real-time example of this phenomenon.

Back in 1988, Professors Power and Lonie from the University of Dundee published a paper titled “The Over-Reaction Effect–Some U.K. Evidence” to study the sequence of returns following underperformance. The thirty worst British stocks in the large-cap universe (defined as the bottom thirty performers out of 200 stocks included in the publication Management Today) delivered -9.96% annual returns from 1973 to 1977 (compared to 5.64% annual returns for the 200 firms in aggregate). Then, from 1978 through 1982, that same collection of thirty stocks returned 7.92% (compared to 17.40% from the index). And then, from 1983 through 1987, that same basket of stocks returned 30.84% annually (compared to 20.76% from the index). After that five year period of mediocre returns, there was a period of satisfactory returns followed by a period of exceptional returns. It speaks to the ability of large firms to have the resources to eventually buy its way out of problems, and it also speaks to how cheap these types of stocks get during periods when no one seems to want them.

When I was studying the work of Professors Power and Lonie in Great Britain, I saw similarities between those poor performing stocks from 1973 to 1977 and IBM today. It seems plausible that IBM is in a similar position where a breakout period could lay ahead simply because the stock is so cheap and there are still $13 billion in annual profits for the IBM Board to deploy each year.

IBM is paying out $5.20 per share in dividends. Assuming 5% dividend growth for the next five years, each share stands to collect $28.81 in dividends. That’s 23.61% of your purchase price returned as cash dividends, and this is an assumption towards the conservative end of the spectrum. When you include the large buybacks, large dividends, and the expectation of eventual improving fundamentals, the current price of $122 is very attractive. It should be instructive that the last time IBM got nearly this cheap came in 1993, and it went on to deliver 12.11% annual returns for the next twenty-two years despite encountering difficult operating environments within this measurement period.