Do you use Trailing or Forward price multiples?

[Milu]

Just curious if you use Trailing or Forward multiples when you are assessing the valuation of a stock? Personally I have always used trailing multiples as it feels a bit more conservative, but if you think about it logically then some sort of forward multiple should make more sense as that is the current expected 'yield' of the stock.

 

For example, Meta is currently 32 P/E which gives a earnings yield of 3.1%, forward P/E is roughy 23 which gives earnings yield of 4.3%. Obviously the forward earnings are based on estimates which could be higher or lower than what actually transpires.

 

For fixed income or real estate investors I assume they mostly care about the expected future yields and not necessarily the previous although they would factor these in as part of their general valuation approach.

 

What do you do?

[Spekulatius]

Forward multiples. Nobody drives just looking in the rear view mirror.

Edited April 23 by Spekulatius

[coffeecaninvestor]

Forward but I would use a conservative estimate of future earnings. The more predictable the industry/company the easier this is. 

[K2SO]

Use 'em both. There is no one magic metric, don't use PE in isolation either. 

 

[Gregmal]

Both are a waste of time. 

 

If Im concerned with the earnings power of something, I typically try to figure out what the next three years of earnings will look like. One year is such a small sample size; anything can happen in a year. 

[KCLarkin]

9 hours ago, Milu said:

Personally I have always used trailing multiples as it feels a bit more conservative, but if you think about it logically then some sort of forward multiple should make more sense as that is the current expected 'yield' of the stock.

Trying to be "conservative" is a mistake. You should try to be accurate. How many great and very reasonably priced companies have value investors missed because they were too "conservative". 

 

I agree with Gregmal, you should be looking out at least 3-5 years. Trailing earnings are already priced into the stock. Forward earnings are usually too (though Meta is an example of "conservative" investors missing forward earnings by a mile).

 

If you look at only forward earnings, every great company will look expensive. If you use trailing earnings, every value trap will look cheap.

 

Nvidia is a great cautionary tale. In January 2022, TTM (2022) was $4.44. actual fwd (2023) was $3.34. You could have been conservative and estimated 2024 earnings at $3 or $4 or $5. But actual 2024 earnings were $12.96. Sure, you might have missed the drawdown from $300 to $100. But you would also miss the run from $100 to $900. Conservatism comes at a price.

 

---

This also depends on your strategy. If you are buying a cyclical, you should be looking at the past 10-20 years. And then maybe overlay some thoughts on forward earnings. For a real growth stock, you should be looking out 5-10 years.

 

 

[Sweet]

2 hours ago, KCLarkin said:

Trying to be "conservative" is a mistake. You should try to be accurate. How many great and very reasonably priced companies have value investors missed because they were too "conservative". 

 

I agree with Gregmal, you should be looking out at least 3-5 years. Trailing earnings are already priced into the stock. Forward earnings are usually too (though Meta is an example of "conservative" investors missing forward earnings by a mile).

 

If you look at only forward earnings, every great company will look expensive. If you use trailing earnings, every value trap will look cheap.

 

Nvidia is a great cautionary tale. In January 2022, TTM (2022) was $4.44. actual fwd (2023) was $3.34. You could have been conservative and estimated 2024 earnings at $3 or $4 or $5. But actual 2024 earnings were $12.96. Sure, you might have missed the drawdown from $300 to $100. But you would also miss the run from $100 to $900. Conservatism comes at a price.

 

---

This also depends on your strategy. If you are buying a cyclical, you should be looking at the past 10-20 years. And then maybe overlay some thoughts on forward earnings. For a real growth stock, you should be looking out 5-10 years.

 

 

Great post.

[Eldad]

Trailing, it makes no sense to me to base my purchase on someone’s guess. Maybe if they just purchased a company I will add that companies PY into the earnings. 
 

Valueline does past 2 quarters plus next 2 estimates which is kind of interesting. 

[Gregmal]

44 minutes ago, Eldad said:

to me to base my purchase on someone’s guess.

Well why not base it on your own work?

[Eldad]

58 minutes ago, Gregmal said:

Well why not base it on your own work?

Yeah I mean obviously we all buy stuff that we believe will grow earnings in the future. I just find that my or anyone else’s guess on the next 12 is not reliable. 
 

Trailing 12 makes me wait for cheaper prices usually instead of cooking my own books to make the case right now because I am impatient. I can see it both ways, but I want to gear my buying to waiting for corrections, etc. 

[LC]

Both, or neither?

 

I look at the last call it 5 years earnings just to get my bearings on the business. Then I look forward - are prospects better, worse, or similar to the past few years? 

 

If I think the prospects are way better (something like Fairfax in 2021), but the market is pricing as if prospects are middling, well maybe there's an interesting investment. 

 

Ultimately it comes down to what the business will earn in the future, vs. what you pay today. We know today's price, the rub is in estimating future earnings. 

I think prior earnings history does provide some baseline at least to start estimating. 

[Gregmal]

Everyones gotta find their own process. And then the work begins refining it and evolving as an investors. In a world of "this is what youre taught"...everyone runs to the income statements. Personally, the first place I go when looking at an investment is the balance sheet. Why? Because Im not interested in crappy balance sheets. It only means...bad things. Poorly managed, declining profits, interest rate/refi risk...and many more derivative negatives. Whereas Ive made an absolute fortune, finding companies with pristine balance sheets, and "nothing really going on"....playing inflections or event driven stuff. CKX a good recent example. Nintendo falls into this bucket as well. When the balance sheet is a fortress, you have time to sit around and wait to see the cards turned. You can also employ your own leverage to put on the investment. Whereas some crapco with a huge debt burden, you sweat more and more with every tick of the clock. 

[Cod Liver Oil]

@Gregmal  I actually like the waiting for the balance sheet geyser to erupt.  Let’s see if Nintendo can explode from all the value building pressure!!!! 

Edited April 24 by Cod Liver Oil

[Milu]

14 hours ago, KCLarkin said:

Trying to be "conservative" is a mistake. You should try to be accurate. How many great and very reasonably priced companies have value investors missed because they were too "conservative". 

 

I agree with Gregmal, you should be looking out at least 3-5 years. Trailing earnings are already priced into the stock. Forward earnings are usually too (though Meta is an example of "conservative" investors missing forward earnings by a mile).

 

If you look at only forward earnings, every great company will look expensive. If you use trailing earnings, every value trap will look cheap.

 

Nvidia is a great cautionary tale. In January 2022, TTM (2022) was $4.44. actual fwd (2023) was $3.34. You could have been conservative and estimated 2024 earnings at $3 or $4 or $5. But actual 2024 earnings were $12.96. Sure, you might have missed the drawdown from $300 to $100. But you would also miss the run from $100 to $900. Conservatism comes at a price.

 

---

This also depends on your strategy. If you are buying a cyclical, you should be looking at the past 10-20 years. And then maybe overlay some thoughts on forward earnings. For a real growth stock, you should be looking out 5-10 years.

 

 

That's a good answer and I suppose it depends on the investors approach.

 

I've never quite aligned with the looking out 3-5 years approach. I think it is hard enough to estimate some sort of steady-state earnings number for one year out, that to believe you can do it with any degree of accuracy when looking 3-5 years into the future is not possible (in my opinion). Wars, Oil Shocks, Recessions, Pandemics, Tech Booms etc, how can any logical person look at Apple, Nvidia, Nike for example, and assess with any degree of accuracy what their earnings will be in 2029? Perhaps you will say that these aren't good examples and other firms are more predictable, but I'm not sure.

 

I take a more simple estimation approach, through researching the company, reviewing financial statements history and making adjustments, I come up with a single normalised earnings/free cash flow number that I feel has a reasonable chance of being accurate of being achieved over the next 12 months. I will use this estimate to determine what the stock is currently yielding (earnings/Price), and then lastly I will apply a margin of safety to account for situations where the actual earnings end up being less than my estimate. 

 

I don't believe anybody can estimate growth rates with any degree of accuracy (at least I certainly can't) so I try to buy at price where even if their was no growth at all over the coming years I'd still have an acceptable return. Similar to how I believe Alice Schroeder summarised Buffets approach, he like to buy at a 10% pre-tax yield and then just let the growth take care of itself. 

[KCLarkin]

3 hours ago, Milu said:

I take a more simple estimation approach, through researching the company, reviewing financial statements history and making adjustments, I come up with a single normalised earnings/free cash flow number that I feel has a reasonable chance of being accurate of being achieved over the next 12 months. I will use this estimate to determine what the stock is currently yielding (earnings/Price), and then lastly I will apply a margin of safety to account for situations where the actual earnings end up being less than my estimate. 

 

With this approach, you are systematically undervaluing the best businesses. Here is an example from my own "mistake" in using this normalized approach:

 

MSC (industrial distributor) YE 2015:

Trailing earnings: $3.79

Foward earnings (actual): $3.77

Normalized Earnings: ~$4.00

Cheapest Price 2015: $55

Normalized PE: 13.75

 

FASTENAL YE 2015:

Trailing earnings: $0.89

Foward earnings (actual): $0.87

Normalized Earnings: ~$0.75

Cheapest Price 2015: $19

Normalized PE: 25

 

On normalized PE, MSM was almost half the price of Fastenal! But which was cheaper? Over the next 9 years, performance was:

 

MSC

EPS CAGR: 5.3%

Annualized Return: 7.6%

 

FAST

EPS CAGR: 10.6%

Annualized Return: 17.2%

 

MSC was pretty close to fairly priced in 2015. Fastenal was the real bargain. But that is easy to say in hindsight. How could you have known at the time? There were a couple pretty strong clues:

 

MSC

ROE 17%

10yr EPS Growth: 9%

 

FAST

ROE 29%

10yr EPS Growth: 12%

 

Fastenal was clearly the better business. If you asked 100 investors in 2015 which was the better business, 100 (including me) would have said Fastenal. But then fools, like me, bought MSC because it was "cheaper". 

 

 

 

[Eldad]

1 hour ago, KCLarkin said:

 

With this approach, you are systematically undervaluing the best businesses. Here is an example from my own "mistake" in using this normalized approach:

 

MSC (industrial distributor) YE 2015:

Trailing earnings: $3.79

Foward earnings (actual): $3.77

Normalized Earnings: ~$4.00

Cheapest Price 2015: $55

Normalized PE: 13.75

 

FASTENAL YE 2015:

Trailing earnings: $0.89

Foward earnings (actual): $0.87

Normalized Earnings: ~$0.75

Cheapest Price 2015: $19

Normalized PE: 25

 

On normalized PE, MSM was almost half the price of Fastenal! But which was cheaper? Over the next 9 years, performance was:

 

MSC

EPS CAGR: 5.3%

Annualized Return: 7.6%

 

FAST

EPS CAGR: 10.6%

Annualized Return: 17.2%

 

MSC was pretty close to fairly priced in 2015. Fastenal was the real bargain. But that is easy to say in hindsight. How could you have known at the time? There were a couple pretty strong clues:

 

MSC

ROE 17%

10yr EPS Growth: 9%

 

FAST

ROE 29%

10yr EPS Growth: 12%

 

Fastenal was clearly the better business. If you asked 100 investors in 2015 which was the better business, 100 (including me) would have said Fastenal. But then fools, like me, bought MSC because it was "cheaper". 

 

 

 

I will pay up for high ROE businesses and really that is all I will buy. But I will use trailing numbers only to arrive at my buy level. Seems like you are conflating using trailing EPS and buying crappy low P/E stocks. 

[Milu]

35 minutes ago, Eldad said:

I will pay up for high ROE businesses and really that is all I will buy. But I will use trailing numbers only to arrive at my buy level. Seems like you are conflating using trailing EPS and buying crappy low P/E stocks. 

Yes I am the same. Quality of business and management is the first hurdle a company has to pass before it meets my requirements, earning estimates and price to pay is the very last step so I would have already ruled out the bad businesses (low returns on equity, capital, excessive dilution, excessive leverage, inconsistent margins etc) before they ever get to the stage where I am estimating fair value. Doesn't mean I get it right every time but so far I've mostly avoided value traps. I don't think I have ever bought a business because it was 'cheap'.

[KCLarkin]

23 minutes ago, Eldad said:

I will pay up for high ROE businesses and really that is all I will buy. But I will use trailing numbers only to arrive at my buy level. Seems like you are conflating using trailing EPS and buying crappy low P/E stocks. 

I get that this can work when Mr. Market goes crazy. But intellectually, it makes no sense.

 

[coffeecaninvestor]

2 hours ago, KCLarkin said:

 

With this approach, you are systematically undervaluing the best businesses. Here is an example from my own "mistake" in using this normalized approach:

 

MSC (industrial distributor) YE 2015:

Trailing earnings: $3.79

Foward earnings (actual): $3.77

Normalized Earnings: ~$4.00

Cheapest Price 2015: $55

Normalized PE: 13.75

 

FASTENAL YE 2015:

Trailing earnings: $0.89

Foward earnings (actual): $0.87

Normalized Earnings: ~$0.75

Cheapest Price 2015: $19

Normalized PE: 25

 

On normalized PE, MSM was almost half the price of Fastenal! But which was cheaper? Over the next 9 years, performance was:

 

MSC

EPS CAGR: 5.3%

Annualized Return: 7.6%

 

FAST

EPS CAGR: 10.6%

Annualized Return: 17.2%

 

MSC was pretty close to fairly priced in 2015. Fastenal was the real bargain. But that is easy to say in hindsight. How could you have known at the time? There were a couple pretty strong clues:

 

MSC

ROE 17%

10yr EPS Growth: 9%

 

FAST

ROE 29%

10yr EPS Growth: 12%

 

Fastenal was clearly the better business. If you asked 100 investors in 2015 which was the better business, 100 (including me) would have said Fastenal. But then fools, like me, bought MSC because it was "cheaper". 

 

 

 

Assuming you have the business quality aspect correct the terminal value of a business will have a much greater impact than your entry multiple unless you are pay some kind of crazy multiple but even then you just need a really long holding period (ie buying MSFT in 2000). Which is why some of WEB's best investments have been modest growers, but for a really long time (GEICO, AXP, KO). 

[CorpRaider]

Yes

[Intelligent_Investor]

EV/FCF, P/E or P/whatever are shit metrics because it doesn't take into account capital structure, they basically only work if the company is cash neutral

[mattee2264]

If a company has a conservative capital structure such as a debt/equity ratio of less than 50% then P/E and P/FCF should work just fine.

 

As for the OP's original question I've always felt you should anchor on what is real i.e. reported earnings. These should then be normalized for one-offs, cyclical factors etc. And then you consider qualitative factors such as growth prospects, competitive position etc in deciding what an appropriate multiple should be. Or if you have confidence in predicting the next few years earnings you can go a few years out and then apply an appropriate multiple. But as others have suggested better to come up with your own forward estimates rather than rely on Wall Street estimates as Wall Street isn't interested in accurate valuation it is interested in selling stocks. and no one sells stocks by indicating they are overvalued and forward multiples are great at making the valuation of fast growing stocks look more reasonable. And they play the same game with the S&P 500. At the end of every bull market Wall Street are still proclaiming that valuations are reasonable and forward PE ratios around in line with the historical average. 

 

[ValueArb]

It doesn’t matter. If you think a business growing earnings 15% a year is worth a 25 PE on trailing earnings, then it’s worth a 22 PE on forward PE. Gregmal has it right. Look at next 3-5 years of earnings to make your valuation.

 

I would just counsel to make your own estimates and not rely on analysts. The only reason forward PE was created was as a Wall Street sales tool. They know clients might view an over 20 PE as “expensive”, so be using forward earnings they can turn that trailing 20 to 30 PE into a sub 20 forward PE and claim it’s “cheap”, while ignoring the risk that those earnings won’t actually happen.

[scorpioncapital]

it's somewhat of a tautology using multiples , as the multiple you use feeds back into the growth or lack of it. Actually the multiple is quite difficult. In another macro-world, and the world can flip anytime, a very different multiple can be in vogue, even for the best of companies. If your threshold is too low you'll never buy anything. If too high, you risk capital loss. The margin of safety probably leans to a lower multiple , but that should be viewed relatively in terms of type of business too. You could use longer term historical averages, but then again businesses change. I would say the type of business, quality of management, the growth rate, the industry should impact the level of the multiple. Also how that business has changed over time and risks about how it might change in the future. It's very subjective and a bit of an art. I would say at the end of the day, except for radical errors in some range of multiples (which may be quite wide), you will have to go through any unpleasant events before finding out what you should have done. Lots of long term capital gains have been accrued by going through volatile contractions and expansions in the multiple. Selecting the right company probably more important than anything, if you can do it.

Edited April 28 by scorpioncapital