How to Value Stocks: A Defense of P/E Ratios
Why Free Cash Flow to Equity Is Inaccurate and P/E Ratios Make the Most Sense
Why buy a stock? For most people, it’s because you hope the price will go up. That’s, after all, how performance is measured -- marked to market.
But what does buying a stock give you? It’s the right to all future dividends. The return on investment, therefore, is the return from those future dividends per share:
Buy a share of company: negative P0 cash flow
Get future dividends: positive D1, D2, … , Dn cash flow
Given this fact of investing, why does almost no one talk about the present value of future dividends when trying to figure out what to pay for a stock? Because many companies do not pay dividends at all, and those that do may be far from a steady state.
Free Cash Flow to Equity
To get around the dividend estimation problem, some investors instead calculate the “free cash flow to equity” or “FCFE,” which is the cash available to pay dividends, repurchase shares, or let build up on the balance sheet. The problem with this method is it assumes both repurchases and cash on the balance sheet are financially equivalent to paying out a dividend. So it’s not quite accurate.
Capitalization of Earnings
While the FCFE method is a decent dividend proxy, it is more accurate to value stocks another way: estimate what a company will earn in either the current or future year and then capitalize those earnings based on the company’s growth prospects (while also factoring in any earlier dividends received).
For this method to produce an accurate result, at whatever point chosen, the company must have reached a steady state that allows the investor to reasonably capitalize earnings based on expected future dividends at that time. The return at this steady state is equal to the dividend yield plus an expected steady-state growth rate.
r = D/P + g
The growth rate equals the company’s sustainable incremental return on equity multiplied by the percent of earnings that it can sustainably reinvest.
g = ROE * retention ratio = ROE * (1 - payout ratio)
Rearranging the first formula, the future price, P, is equal to the expected dividend at that time, divided by the discount rate, r, net of expected future dividend growth, g.
r - g = D / P
(r - g)*P = D
P = D / (r - g)
Since dividends are equal to earnings multiplied by the payout ratio, this formula can be expressed as a function of earnings, E:
P = E * payout ratio / (r - g)
Therefore the multiple used to capitalize steady-state earnings -- the steady-state P/E ratio -- equals the payout ratio divided by the discount rate net of the growth rate. For example, a steady-state company with a payout ratio of 60%, a discount rate of 9%, and a growth rate of 4% would warrant a P/E ratio of 12. Since growth is a function of the ROE and the payout ratio…
ROE = g / (1 - payout ratio)
...the implied ROE on incremental equity capital is 4% divided by (1 - 60%), equal to 10%.
Dealing with High Reinvestment Companies
For companies that are still in full-on reinvestment mode, you have to make an assumption about when they’ll stop reinvesting every penny and start paying out dividends, and then when they’ll get to a steady state. This is really hard. When will Berkshire Hathaway pay dividends? Or Amazon? Or Google (Alphabet)? Harder yet, Carvana, Peloton, or an early stage biotech? It’s extremely hard to predict.
At least in public, investors get around this thorny issue by guessing at what an appropriate future earnings multiple might be in 5, 10, 15, or 20, based on earnings multiples they see in the market for lower growth companies. Or they just think about the company’s characteristics in the future and then estimate what might be an appropriate multiple to pay.
Example: At Home
At Home was recently bought out by private equity at $37 a share. In opposition to the buyout, their former largest shareholder CAS put out a valuation framework for the company.
… we estimate the Company’s earnings can grow to approximately $6.748 per share by fiscal year 2027.
We contend that by fiscal year 2026, the market could value At Home at 20x forward profits or even more given the following tailwinds:
The Company’s unlevered balance sheet;
The further remaining opportunity to continue expanding to 600+ stores;
The potential to expand revenue per store to $10 million plus, and;
The potential to further expand margins going forward.
In light of this, we estimate that At Home’s stock would be worth more than $135 per share by the end of fiscal 2026, which is less than five years from now. Discounting this $135 back five years at the conservative rate of 13% yields a fair take-out value of $70 per share or more today.
Here, CAS isn’t explicitly estimating future dividends or payout ratios. They’re just saying, hey, given these four characteristics (“tailwinds”), and presumably what they’re currently seeing for how those characteristics are valued, 20x will probably be a reasonable multiple of forward earnings in 2026.
The problem with this analysis is that I don’t think At Home will be at a steady state by 2026 or 2027. Frankly I don’t think CAS does either. If I had to guess, they were using a 5-year timeline since that’s what private equity firms focus on. Maybe they thought it would be more palatable to other investors, too (since CAS was trying to rally opposition to PE firm’s tender offer).
In my opinion, 2030 is a more reasonable steady-state point for At Home given their growth plans. My best guess is that At Home’s 2030 earnings per share will be $10. At that point, I think they can start paying out most of their earnings in dividends since they will have tapped out their growth opportunities. Earnings growth therefore should converge to GDP, so about 4%. If they pay out 80% of earnings as dividends ($8 a share), that’s an implied ROE of 20% (ROE = g / (1 - payout), so 4% / 20%). At a 10% discount rate, the implied 2030 earnings multiple is 13.3.
P = E * 80% / (10% - 4%) = E*80%/6% = 13.3*E
At $10 EPS in 2030, that’s a fair stock price of $133 in 2030. If we discount that back nine years to 2021 at CAS’s 13% discount rate, the terminal value is worth $44 today. Adding on $6 to account for the present value of pre-2030 dividends gets us to a fair value of $50. So, still a significant premium to the takeout offer -- 35% -- but definitely lower than CAS’s estimated fair value of $70.
Potential Ranges of Terminal-Value Earnings Multiples
As shown above, the terminal-value earnings multiple is the payout ratio divided by the discount rate net of the growth rate:
P/E = payout ratio / (r - g)
Given these three variables: the payout ratio, the discount rate, and the growth rate, what are possible ranges for terminal-value earnings multiples?
The growth rate is going to be no more than the growth of the economy and could be lower, so between 2-4%. Since g is equal to the ROE multiplied by 1 minus the payout ratio, we can also narrow down the payout ratio given possible ROEs.
g = ROE * (1 - payout ratio)
ROE = g / (1 - payout ratio)
ROE * (1 - payout ratio) = g
payout ratio = 1 - g / ROE
Potential ROEs for a steady state company (note - this is the return on incremental equity capital) probably can’t get extremely high. Maybe 25% tops. Let’s use a range of 7-25%.
Within a 7-25% range for ROEs on steady-state incremental capital and a range of steady-state growth rates of 2-4%, we can then calculate possible payout ratios. Doing the math, the lowest possible payout ratio is 43%, representing a 4% growth rate and 7% ROE (57% * 7% = 4%), and the highest possible payout ratio is 92%, representing a 2% growth rate and a 25% ROE (8% * 25% = 2%).
The last variable is the discount rate. While this one is certainly up for debate, for a steady-state company, in my opinion a reasonable possible range (with treasuries where they are now) is 6.5% to 11.5%.
So to recap, here are the possible ranges for the three variables, or four including the ROE
payout ratio: 43%-92%
r: 6.5%-11.5%
g: 2%-4%
ROE: 7-25%
Of course, the variables also have to make sense when combined together. For example, in order for a payout ratio of 92% to produce a growth rate of 4%, the ROE on incremental capital would have to be 4% = 8% * ROE = 50% which is highly unlikely.
Combining the possible variables, we get the following earnings multiples and implied dividend yields using the high-end of sustainable growth:
The 8x - 15x range for the 9.5% discount rate and 4% growth rate is probably most representative of what investors are using today. Hence you see a lot of slower growing companies in competitive industries trading at these sorts of earnings multiples. The 8x multiple represents a company with a 7% ROE on incremental capital and a 43% payout ratio. The 15x multiple represents a company with a 25% ROE on incremental capital and a 84% payout ratio. (Both result in earnings/dividend growth of 4%).
While the possible earnings multiples for the groupings vary quite a bit, they all have the same dividend yield. This is because the dividend yield is always equal to to the discount rate minus the growth rate:
r = D/P + g
D/P = r - g
From a practical standpoint, the interesting thing here is that the discount rate typically has to be on the low side to justify a terminal earnings multiple in the teens or above. So if a stock has a P/E ratio higher than 12, either it’s not a steady-state company, or the uncertainty priced into the stock is pretty low.
Summary
The value of a stock is the present value of future dividends. Some investors use free cash flow to equity as a proxy for dividends, but this method ignores the effect of share repurchases and free cash flow retained on the balance sheet. A more correct method is to capitalize future earnings. The earnings capitalization method requires guessing when a company will reach a steady state, and then estimating the earnings, dividend payout ratio, and future dividend growth at that time. Given natural growth constraints at a steady-state, terminal P/E multiples should rarely be much greater than 13 unless the discount rate is relatively low.