Yield on Cost: Useful Concept or Inconsequential Statistic?

This article is authored by Adam Mead, president of Mead Capital Management.

The concept of yield on cost is simple to explain but its usefulness as a tool is less clear – at least to me. My goal in writing this memo is to offer some of my thoughts on the subject, and to solicit feedback from others.

The Concept

Yield on cost is simply a security’s current dividend or earnings expressed as a percentage of the price paid for the security. A simple example will illustrate: Imagine you paid $100 for a share of a company that this year earns $20 and pays out 50% or $10 as a dividend. In this case your dividend yield on cost would be $10 / $100 = 10%; similarly, your earnings yield on cost would be $20 / $100 = 20%. This appears to be a great return. But what if you paid that $100 five years ago? Does that matter? What if there were an opportunity to invest in another security today, using the proceeds raised from selling your current investment, that would increase your absolute dollar returns (both earnings and dividends) but decreased your yields on cost – would that matter?

Deciphering the Master

I find myself with these questions as a result of pondering a couple of Warren Buffett’s comments on the subject. The first, from the 2010 Berkshire Hathaway Chairman’s Letter (emphasis Buffett’s):

Coca-Cola paid us $88 million in 1995, the year after we finished purchasing the stock. Every year since, Coke has increased its dividend. In 2011, we will almost certainly receive $376 million from Coke, up $24 million from last year. Within ten years, I would expect that $376 million to double. By the end of that period, I wouldn’t be surprised to see our share of Coke’s annual earnings exceed 100% of what we paid for the investment. Time is the friend of the wonderful business.

The second is from a December 15, 2016 Fortune Magazine article. Buffett was responding to a question about Berkshire Hathaway’s seemingly low returns from BH Energy:

We paid $35 a share for the utility [in 2000]. And this year it’ll earn something around $30 a share after tax.

So, is yield on cost just an interesting but otherwise meaningless statistic, or is there real insight to be had? To use Buffett’s example of Coke, its 1995 cost was $1,299 million. (I’m ignoring the fact that Berkshire’s stake was built up to this level starting in 1988, and instead assuming it was all acquired in 1995.) Using Buffett’s assumptions, in 2020 Coke would pay Berkshire $752 million and Berkshire’s share of Coke’s total earnings would amount to $1,299 million. Berkshire’s Coke investment would therefore have a 100% earnings yield on cost, and a 58% dividend yield on cost.

We can adjust the yields on cost above to account for the time value of money. Assuming a 10% discount rate over 25 years the present value factor is 0.923, or 9.23%. The resulting earnings yield on cost would be 9.2%, and the dividend yield would be 5.3%. Both seem more “reasonable” compared to the double and triple digit returns assuming original cost.

Another way to look at it would be to invert the situation and look at the 1995 cost as a 2020 future value. Making this adjustment (10%, 25 years = 10.8 times) Berkshire’s cost increases to $14,074 million which, again, appears more “reasonable” to have claim to $1,299 million in earnings.

Buffett has given us clues over the years into his thinking. I’m referring to his “business-as-a-bond” mindset. Buffett and some value investors consider equity as simply a bond without known coupons. Using this framework, the yield on cost concept becomes somewhat clearer. Assuming your going-in earnings yield is satisfactory, like a bond investor holding to maturity would, anything above and beyond your starting yield could be considered gravy.

In 1995, the 30-year Treasury Bond yield ranged from about 6.0% to 7.8% – I’ll assume an average of 6.9%. If Berkshire, instead of buying Coke, used its $1,299 million to purchase a bond yielding 6.9% it would still be paying Berkshire about $90 million a year in 2020. Adjusting for the same present value as above (10%, 25 years) that sum diminishes to a paltry $8.3 million. Instead of purchasing a “risk free” 6.9% that would decline by 3.1% per annum compared to an estimated 10% discount rate, Buffett instead went with the “risky” business (Coke) that had the ability to increase its output over time.

Still Confused

So how useful as a tool is the concept of yield on cost? Under the bond-as-a-business framework the idea seems to make sense. However, when viewed from the vantage point of an equity holder the notion appears to ignore one of the fundamental ideas of finance – namely, the time value of money. The ideas should reconcile. I’m still confused. What say you?

This commentary is the opinion of the author. It is neither investment advice, nor an offer to buy or sell securities, nor a substitute for your own thinking.