Return on Equity

Return on equity (ROE) is usually considered something of a ‘Holy Grail’ for many investors, particularly those, like us with a “Value Investment” disposition.
ROE is the return (by way of profits) on the original capital used in the business (plus retained profits) shown as a percentage.  To demonstrate for those less familiar with the concept, imagine I decide to open a hot-dog stand on a street corner near my home.  Further, imagine that I need to invest $10,000 in the stand/equipment (cooler/cash-register/umbrella/bun-steamer/boiler etc) required to establish the business.  Now after deducting cost of sales (wages/buns/sauce etc) from sales (and depreciating the assets), assume at the end of 12 months I have profits of $2,000, then my ROE for that year was 20%.  Assuming all profits are paid out as dividends, and the business returns $2,000 again the next year, the ROE was stable. Assuming all profits are retained, then in order to maintain a 20% ROE, the business would need to earn $2,400 the following year.
From the above example, assuming 50% of profits are distributed and 50% retained, profit would need to grow at 10% pa in order to maintain the 20% ROE.  Generally speaking, for me, if ROE is likely to decline as a consequence of retaining earnings, then it is preferable to pay the earnings out as dividends.  Provided, however, that a business can maintain or grow their ROE whilst retaining earnings, it is far preferable that it retain the earnings, in fact, this is the sort of business we fervently seek.
What ROE focussed investors often fail to realise, however, is that when you buy shares on market, you are not paying the original equity price.  Assume in the ‘Hot-Dog Stand’ business above could be conservatively assumed to grow EPS at 10% pa (in perpetuity – or something like it) by retaining 50% of profits.  Such a business (provided we can be confident of the growth assumptions) would likely eventually trade at about something like 18x earnings (depending on many factors). On this basis, my $10,000 equity & $2,000 earnings above, the business would likely be valued at $36,000.  So very rapidly, the market would choose to pay $3.60 for every $1 of equity (due to the sound prospects of the business).  Assuming it makes that 10% growth for 10 years, pays out 50% of earnings and still trades at 18x, the investor who paid $3.60 (say per share for one or more of the 10,000 shares) after 10 years would own a stock priced at $8.49 and would have been paid $1.59 in dividends.  In this case, ‘over-paying’ for the equity was quite profitable, a return of about 10.85% pa (despite not having re-invested the dividends).  Depending on inflation levels, this might be a relatively sound result.
Imagine instead a business very much like the above, but it has had disappointing results and recently earned $500 on its $10,000 in equity (an ROE of 5% instead of the 20% above).  Now if due to its disappointing ROE and prospects, the business traded on say 8.5x earnings (Benjamin Graham indicates this is an approximately reasonable price for a business that is likely to neither grow nor shrink earnings), then instead of paying $3.60 per $1 for equity, I would be paying $0.425 for each $1 of equity (granted the original $1 payer has lost – that is not our problem).  Assume in this case, that for 10 years earnings grow at 5% pa and again, 50% of profits are paid out as dividends.  Based on similar factors as used above, a business that consistently grows it profits by 5% pa for 10 years would likely trade at about 11.8x earnings.  On this basis, our 42.5c shares would after 10 years be worth 91.5c and we would have received about 31.5c in dividends.  In such a situation, despite a significantly weaker ROE (in this example, in the 10th year, the company was still producing an ROE of less than 6.1%) and half the annual earnings growth, we would have earned 11.21% pa (again without re-investing dividends), more than the better ROE and growth business above.

So to put it simply, ROE is a useful metric, but must be used with caution, as like virtually every other financial metric available, we must be mindful of the circumstances around which we use it.  Just because someone, somewhere in the history of the company paid too much for the equity doesn’t mean at its current cost, the equity is overpriced.  Tony Hansen 13/03/2011