Can anyone explain to me how the P/E ratio is a meaningful metric for a company's stock price and what it "should be" at when that company does not distribute earnings to the shareholders?
There are two ways to gauge the value of a company. One way (the way to which you allude) is to buy it and hope that in the future, someone thinks it's worth more than what you paid for it. Some people call this the Greater Fool theory.
The other way is to treat owning shares like owning part of a business. (You might hear about this as the Graham-Dodd or Graham-Dodd-Buffett model.) In this case you consider the value of the company as the current liquidation value of the company's assets plus the total amount of free cash the business can produce over its useful lifespan.
The math of that valuation gets a little bit interesting, because if you project the free cash growth into the future ten or twenty years, you can use a present value calculation to get a fair price for the company right now. Divide that by the number of outstanding shares and you get a target price per share.
One of the flaws of this method is that you need to have a sane sense of the growth rate of free cash, so you'd better base that on a stable and measurable history, and you have to verify that against sanity and the company's published plans.
Graham's real insight was saying "If you do all that work, also add a significant margin of safety to account for any flaws in your calculations."
Of course you shouldn't use reported earnings for these calculations; they're far too easy to manipulate under GAAP and accrual accounting. Even so, finding realistic numbers and ruling out most stocks as overvalued is relatively easy.
(I'm working on a financial analysis site right now.)
So does this mean that if you use the Graham-Dodd-Buffet model of valuation, a company that pays a dividend to shareholders would get a higher valuation (all other factors being equal) than the hypothetical same company that does not pay dividends?
Because if the company keeps the cash it still is part of the company worth calculation (unless investors expect managers to steal it in the future). Not issuing a dividend is also a signal that the company believes it can earn a better return on the cash then if it were returned to the shareholder.
Yes and no. Free cash paid out in dividends is free cash not invested back in the company. Presumably free cash invested in the company could help the company grow to produce even more free cash in subsequent years.
If you measure only the average growth in free cash for the trailing ten years, you're ignoring whether the company pays out dividends. You only measure how much free cash it generates. Dividends are irrelevant to the free cash growth rate.
If you measure something like a return on invested capital, you measure how much actual cash the business reinvests in itself and how effective it is at free cash growth from that reinvestment. That metric does account for dividends, because money paid out in dividends or used to buy back stock is obviously no longer available for reinvestment.
edit With that said, some investors value regular and reliable dividends more highly than the fluctuations of the market's semi-random valuation of a stock at any point in time. If you're confident that Coca-Cola will always pay, for example, 4% of what you paid for a share in dividends every year, that stability might be worth something to you.
That's psychology and harder to measure and predict than numbers from financial reports filed with the SEC. You might get some interesting data if you calculate the time value of that money--is KO more valuable because you can get $0.13 per share quarterly in dividends starting now rather than holding onto it for up to ten years to make even more money? That's the kind of decision individual investors have to make for themselves.
I treat dividends as bonuses rather than expectations, but that's my own investment philosophy.
There are two ways to gauge the value of a company. One way (the way to which you allude) is to buy it and hope that in the future, someone thinks it's worth more than what you paid for it. Some people call this the Greater Fool theory.
The other way is to treat owning shares like owning part of a business. (You might hear about this as the Graham-Dodd or Graham-Dodd-Buffett model.) In this case you consider the value of the company as the current liquidation value of the company's assets plus the total amount of free cash the business can produce over its useful lifespan.
The math of that valuation gets a little bit interesting, because if you project the free cash growth into the future ten or twenty years, you can use a present value calculation to get a fair price for the company right now. Divide that by the number of outstanding shares and you get a target price per share.
One of the flaws of this method is that you need to have a sane sense of the growth rate of free cash, so you'd better base that on a stable and measurable history, and you have to verify that against sanity and the company's published plans.
Graham's real insight was saying "If you do all that work, also add a significant margin of safety to account for any flaws in your calculations."
Of course you shouldn't use reported earnings for these calculations; they're far too easy to manipulate under GAAP and accrual accounting. Even so, finding realistic numbers and ruling out most stocks as overvalued is relatively easy.
(I'm working on a financial analysis site right now.)