A company can only do 2 things with the profit they make: give it to you (dividends and buybacks) or retain it for their own operations (acquisitions, new projects, paying off debt). Many companies do a little bit of both, and IBM is no exception.
Now, it may turn out that the company you’re examining has a cash retention policy. You might contact investor relations or find such information in the annual report. IBM indeed has a section in their annual report that details how much they’ve already returned in dividends and buybacks, and how much they plan to in the near future (part of their “RoadMap 2015”). Let’s just estimate what this will be from their actions over the last 3 years. I’m not saying this is better, but if you haven’t cracked the annual report yet it seems a decent first swag until your research gives you something better.
Rounding to billions, over the last 3 fiscal years IBM has reported 17+16+15 = $48B in net income. Dividends have totaled 4+3+3 = $10B and buybacks 10+13+12 = $35B. So, they’ve paid out 10/48 = 21% of net income as dividends and 35/48 = 73% of net income as buybacks. Assuming this will continue, we can start to add some color to our estimated return equation.
So until we know better, we’re looking at 94% of the earnings yield (E/P) being given to us, with the remaining 6% retained and reinvested by IBM.
Now, if I told you I had a new investment to sell you that paid 20 cents per year forever, what would you pay for it? We’re back to our old NPV equation: P = E / (r – g). But since the income isn’t growing, it simplifies to P = E / r. If I’m just some guy you met over the internet, you might use a discount rate of 15%, and not pay more than $0.20/.15 = $1.33. If you know me and trust me a little more, you might pay $0.20/.10 = $2.00.
I think you see where this is going. Each dollar IBM retains is a new investment. And it will earn their marginal return on capital (ROC), which should be discounted at their marginal cost of capital (r). The NPV of that is the new capital x ROC / r (just like P = E / r). We’ve just put a value on the active growth, so we know what active growth “yield” we’re getting. Let’s fill in a bit more: