Should Repurchases be counted in FCF/yield per share?

[Carvel46]

Palantir,

 

In your example, you are not properly distinguishing between FCF and FCF/share.

 

After a stock buyback each share outstanding (not held in treasury), now has a larger % claim on the FCF. If it's a 30% buyback, FCF does NOT grow (now or in the future, based on the buyback)! FCF/share does grow!!

 

Try rewriting your example.

[Palantir]

That gets a little weird if the firm were to stop buying back shares.  You are going to have free cash flow jump from 700k to 1m in a single year, even though FCF isn't growing.  That might sound like I'm bickering over semantics, but if I was collaborating with you on a software project I would mention that your choice of naming for variables makes the code difficult for others to understand/maintain. 

 

It might compile and run just fine though, no argument there.

 

Let me ask you a question though.  How would you go about it if you were instead looking at a company that returned cash to shareholders only through dividends.  Would you ignore the dividend yield, or would you instead count the dividend and redefine "free cash flow"?

 

Maybe there is another way without redefining what free cash flow is -- that's all I'm trying to say.  I'm not saying that your end result would be wrong.

 

True, I am being lazy with the definitions. I would say the relevant metric for DCFs should be "Adjusted FCF" = FCF - Buybacks.

 

In your scenario, if the firm suddenly stops buying back, then Adjusted FCF = FCF, and then growth becomes 0, which essentially is the pre-buyback state, although now with higher FCF/s due to the buybacks.

 

Now if the firm returned all its FCF to shareholders via dividends, and assuming that no other cash was returned, then in that case, I would not revise the FCF figure downwards, as all FCF immediately flows to shareholders, and more importantly that cash is not reinvested to create growth in FCF/share. Therefore Adj. FCF = FCF, and growth is unchanged.

[hyten1]

Palantir,

 

i am a little confused, are you guys just looking at FCF/s in isolation? meaning at one point a company as $1  FCF/s and this company use the FCF to buy back share now suddenly its $1.2 FCF/s (since the number of share count has gone down for argument sake by 20%).

 

if you just look at FCF/s in isolation, you will be mislead in thinking there is 20% growth in FCF/s, but in reality there isn't (well there is depends on how you look at it, you now own more of the company, but the total FCF hasn't grown)

 

what exactly is the problem? i don't think you can just look at FCF/s in isolation without taking into account what is going on within the firm, if you are then there will be problems, just like any metric you look at in isolation, even the trusty PE ratio.

 

using FCF as sole metric in isolation can also mislead (this thread is a good example).

 

 

also look at the other way, what if company issue more shares and FCF stay the same, obviously same problem.

 

just my 2 cents

 

hy

 

That gets a little weird if the firm were to stop buying back shares.  You are going to have free cash flow jump from 700k to 1m in a single year, even though FCF isn't growing.  That might sound like I'm bickering over semantics, but if I was collaborating with you on a software project I would mention that your choice of naming for variables makes the code difficult for others to understand/maintain. 

 

It might compile and run just fine though, no argument there.

 

Let me ask you a question though.  How would you go about it if you were instead looking at a company that returned cash to shareholders only through dividends.  Would you ignore the dividend yield, or would you instead count the dividend and redefine "free cash flow"?

 

Maybe there is another way without redefining what free cash flow is -- that's all I'm trying to say.  I'm not saying that your end result would be wrong.

 

True, I am being lazy with the definitions. I would say the relevant metric for DCFs should be "Adjusted FCF" = FCF - Buybacks.

 

In your scenario, if the firm suddenly stops buying back, then Adjusted FCF = FCF, and then growth becomes 0, which essentially is the pre-buyback state, although now with higher FCF/s due to the buybacks.

 

Now if the firm returned all its FCF to shareholders via dividends, and assuming that no other cash was returned, then in that case, I would not revise the FCF figure downwards, as all FCF immediately flows to shareholders, and more importantly that cash is not reinvested to create growth in FCF/share. Therefore Adj. FCF = FCF, and growth is unchanged.

[Rabbitisrich]

Discounted cash flows are generally a measure of business values. When you start looking at per share DCF, you are mixing up business results with your actions as an investor. A shareholder may choose to hold during a buyback, or choose to sell. Similarly, a dividend recipient may decide to reinvest, or not. That is not right or wrong, but such choices result from expectations that should already be included in your discount rates.

 

You stumble across a company that yields FCFE, pre-buyback, equal to your discount rate. But they announce that 100% of FCFE will go towards repurchasing shares. Management has not suddenly caused DCF to plummet to 0. Every moment your percentage ownership increases is another moment that you have reinvested in the firm. The growth rate of FCFE is unchanged. Your FCFE per share increases because of your reinvestment choices, just as if the firm changed course and announced a 100% dividend option. The same logic that compelled you to maintain shares in a buyback program, would cause you to purchase shares in a dividend program.

[Kiltacular]

In the thread on Accenture PLC, one member stated a valuation method where the expected return on investment in a share = FCF Yield + Growth

 

I reposted the original basis for this discussion (which was amended later to add a "per share" idea, IIRC and spawned the rest of the thread). 

 

Just looking at the original, I think it is useful.  If the business can grow without using the free cash flow to do it (assuming cap-ex = depreciation), you've got a home run (if purchasing a good business / good price).  These businesses are typically light on tangible investment.  If you can get "growth" from the existing business without the needing to reinvest the true, free cash flow, you've got a winning business.  If you pay the free cash flow out, it is a dividend.  You still get the "growth" over the original share base -- together, the FCF as a dividend plus the growth will give you the return from the business.  If you decide to take the free cash flow and use it to reduce the share count that is also a return like a dividend (as Eric and Rabbit explained).  I think where the confusion is arriving is the idea that some portion of the "next year's" growth in per share profits is coming from reducing the sharecount rather than just the "growth" of the business w/out any FCF going anywhere but to the shareholder's pocket.  It is better to separate out of the repurchase so that you can see whether the business is growing without it -- such that the FCF is really FCF and growth is really growth (not just growth in per share profits because all the FCF was used to reduce shares). .

 

In any case, using FCF is the same (taxes aside) whether it is used for dividends or share buybacks.  The comparative value of each changes depending on the price paid for the buyback, as others discussed.  But, as long as they aren't the sole "source" of growth, then the original "equation" -- FCF + Growth -- still makes sense for total return. 

 

[Palantir]

 

if you just look at FCF/s in isolation, you will be mislead in thinking there is 20% growth in FCF/s, but in reality there isn't (well there is depends on how you look at it, you now own more of the company, but the total FCF hasn't grown)

 

what exactly is the problem? i don't think you can just look at FCF/s in isolation without taking into account what is going on within the firm, if you are then there will be problems, just like any metric you look at in isolation, even the trusty PE ratio.

 

Hyten, I agree, the issue is, when you do a DCF, what cash flows do you end up discounting?

 

Company 1 (No buybacks)

Makes 1M a year, no buybacks, also does not grow, makes 1M into perpetuity.

 

Company 2

Makes 1M a year, buys back 20%, does not grow FCF either, but grows FCF/s due to buybacks.

 

The way I see it, you have to value these very differently, for the first, you'd discount $1.00 every year with no growth in FCF/s. Whereas in the second you'd start with a base "Adjusted FCF" of $0.80 and project growth into perpetuity.

 

The issue really arises IMO, with firms like IBM, who are growing very slowly, but are buying back hella stock. So there is a lot of FCF/s growth. However, the firm is also basically using all of its FCF in buybacks, so you can't really use all of that FCF to value the firm.

[hyten1]

palantir, i am a little confused, my DCF skill is not the greatest, but generally

 

- you take FCF (not the per share value) from the firm, apply some discount rate and growth rate etc, you get a valuation for that firm

- you divide that value by the outstanding share count (current outstanding share count) that gets you a value per share

 

what the future holds in terms how much stock company buys back or issue etc etc is another matter.

 

or am i missing something

 

you are trying calculate FCF/s and discount that value? hmmm that sounds funky to me.

 

hy

 

 

 

if you just look at FCF/s in isolation, you will be mislead in thinking there is 20% growth in FCF/s, but in reality there isn't (well there is depends on how you look at it, you now own more of the company, but the total FCF hasn't grown)

 

what exactly is the problem? i don't think you can just look at FCF/s in isolation without taking into account what is going on within the firm, if you are then there will be problems, just like any metric you look at in isolation, even the trusty PE ratio.

 

Hyten, I agree, the issue is, when you do a DCF, what cash flows do you end up discounting?

 

Company 1 (No buybacks)

Makes 1M a year, no buybacks, also does not grow, makes 1M into perpetuity.

 

Company 2

Makes 1M a year, buys back 20%, does not grow FCF either, but grows FCF/s due to buybacks.

 

The way I see it, you have to value these very differently, for the first, you'd discount $1.00 every year with no growth in FCF/s. Whereas in the second you'd start with a base "Adjusted FCF" of $0.80 and project growth into perpetuity.

 

The issue really arises IMO, with firms like IBM, who are growing very slowly, but are buying back hella stock. So there is a lot of FCF/s growth. However, the firm is also basically using all of its FCF in buybacks, so you can't really use all of that FCF to value the firm.

[Palantir]

palantir, i am a little confused, my DCF skill is not the greatest, but generally

 

- you take FCF (not the per share value) from the firm, apply some discount rate and growth rate etc, you get a valuation for that firm

- you divide that value by the outstanding share count (current outstanding share count) that gets you a value per share

 

what the future holds in terms how much stock company buys back or issue etc etc is another matter.

 

or am i missing something

 

you are trying calculate FCF/s and discount that value? hmmm that sounds funky to me.

 

 

That is one way to do it, but what if the firm is persistently diluting shares and/or persistently buying back shares? If you are a continuing shareholder, the valuation will be very different, because as you note, FCF does not change, but FCF/s will change in the future. Hence, I prefer to do it on a per share basis, which is more applicable to continuing shareholders.

[ERICOPOLY]

That gets a little weird if the firm were to stop buying back shares.  You are going to have free cash flow jump from 700k to 1m in a single year, even though FCF isn't growing.  That might sound like I'm bickering over semantics, but if I was collaborating with you on a software project I would mention that your choice of naming for variables makes the code difficult for others to understand/maintain. 

 

It might compile and run just fine though, no argument there.

 

Let me ask you a question though.  How would you go about it if you were instead looking at a company that returned cash to shareholders only through dividends.  Would you ignore the dividend yield, or would you instead count the dividend and redefine "free cash flow"?

 

Maybe there is another way without redefining what free cash flow is -- that's all I'm trying to say.  I'm not saying that your end result would be wrong.

 

True, I am being lazy with the definitions. I would say the relevant metric for DCFs should be "Adjusted FCF" = FCF - Buybacks.

 

In your scenario, if the firm suddenly stops buying back, then Adjusted FCF = FCF, and then growth becomes 0, which essentially is the pre-buyback state, although now with higher FCF/s due to the buybacks.

 

Now if the firm returned all its FCF to shareholders via dividends, and assuming that no other cash was returned, then in that case, I would not revise the FCF figure downwards, as all FCF immediately flows to shareholders, and more importantly that cash is not reinvested to create growth in FCF/share. Therefore Adj. FCF = FCF, and growth is unchanged.

 

That still fails my code review.

 

I would suggest you merely add back in the number of shares repurchased.  After all, you are willing to ignore a dividend.  So why not ignore the share repurchase?  This is more straightforward.

 

You can then think of the share repurchase and dividends as the same thing -- cash return yield to investors.  That's what they are.

[JSArbitrage]

Hyten, I agree, the issue is, when you do a DCF, what cash flows do you end up discounting?

 

Company 1 (No buybacks)

Makes 1M a year, no buybacks, also does not grow, makes 1M into perpetuity.

 

Company 2

Makes 1M a year, buys back 20%, does not grow FCF either, but grows FCF/s due to buybacks.

 

The way I see it, you have to value these very differently, for the first, you'd discount $1.00 every year with no growth in FCF/s. Whereas in the second you'd start with a base "Adjusted FCF" of $0.80 and project growth into perpetuity.

 

The issue really arises IMO, with firms like IBM, who are growing very slowly, but are buying back hella stock. So there is a lot of FCF/s growth. However, the firm is also basically using all of its FCF in buybacks, so you can't really use all of that FCF to value the firm.

 

FCF is FCF, it doesn't matter whether the cash is used on buybacks and not.  Buybacks are irrelevant to the definition of FCF.  The two companies above are worth the same amount in a DCF analysis.  Or to think of it another way, the difference between the two examples above is basically cash versus PIK.  Company 1 gives you $1M/cash, Company 2 gives you $800K/cash plus $200K stock.  Assuming the stock is valued rationally, you're not gain or losing value in example two.  You would be indifferent to the method of payment (excluding any externalities such as taxes, etc.)

 

I never compare companies using a per share metric because the denominator (shares) can be whatever the company wants.  A company with FCF/share of $100 is not worth twice as much as a company with a FCF/share of $50.  One company could just do a split and they'd suddenly be equal.  What matters is the FCF multiples.

 

 

[Palantir]

I am not a developer, I don't really follow your comment.

 

Because the dividend immediately accrues to shareholders and can be valued, buybacks are cash that continuing shareholders do not see, and cannot be valued.

 

Adding back shares doesn't really fix the problem that the cash is being spent from cash that would otherwise go to shareholders, it's an approximation, but it would be more accurate to just subtract the buyback from FCF/s.

[Palantir]

 

 

FCF is FCF, it doesn't matter whether the cash is used on buybacks and not.  Buybacks are irrelevant to the definition of FCF.  The two companies above are worth the same amount in a DCF analysis. 

 

Not true, it does matter whether the cash is used on buybacks or not. From the POV of a shareholder, the shares are worth different values, and that's the relevant metric, not the value of the firm. You are not the sole owner of the firm.

[lessthaniv]

That gets a little weird if the firm were to stop buying back shares.  You are going to have free cash flow jump from 700k to 1m in a single year, even though FCF isn't growing.  That might sound like I'm bickering over semantics, but if I was collaborating with you on a software project I would mention that your choice of naming for variables makes the code difficult for others to understand/maintain. 

 

It might compile and run just fine though, no argument there.

 

Let me ask you a question though.  How would you go about it if you were instead looking at a company that returned cash to shareholders only through dividends.  Would you ignore the dividend yield, or would you instead count the dividend and redefine "free cash flow"?

 

Maybe there is another way without redefining what free cash flow is -- that's all I'm trying to say.  I'm not saying that your end result would be wrong.

 

True, I am being lazy with the definitions. I would say the relevant metric for DCFs should be "Adjusted FCF" = FCF - Buybacks.

 

In your scenario, if the firm suddenly stops buying back, then Adjusted FCF = FCF, and then growth becomes 0, which essentially is the pre-buyback state, although now with higher FCF/s due to the buybacks.

 

Now if the firm returned all its FCF to shareholders via dividends, and assuming that no other cash was returned, then in that case, I would not revise the FCF figure downwards, as all FCF immediately flows to shareholders, and more importantly that cash is not reinvested to create growth in FCF/share. Therefore Adj. FCF = FCF, and growth is unchanged.

 

That still fails my code review.

 

I would suggest you merely add back in the number of shares repurchased.  After all, you are willing to ignore a dividend.  So why not ignore the share repurchase?  This is more straightforward.

 

You can then think of the share repurchase and dividends as the same thing -- cash return yield to investors.  That's what they are.

 

This.

 

First, understand the companies ability to generate FCF. Place a value on that FCF independent of dividends/repurchases.

 

Then, in a secondary step,  adjust your view for managements capital allocation choices. If management is making smart decisions allocating this FCF it's worth more to you as minority shareholder than the same FCF being generated by a company with a management team who is making poor decisions in allocating it.

 

 

[BTShine]

I'll throw my hat in the ring:

 

I ignore buybacks in "modelling".

 

I just calculate FCF or Owner Earnings or whatever you want to call it.

 

What the company does with that cash afterwards I treat as a different story.

 

So I separate the analysis into two sections: what the business operations generate, and what management does with those proceeds afterwards.

 

Smart.

 

"What the company does with that cash afterwards I treat as a different story."

 

What they do with the cash adds another opportunity for the company to either create, or destroy, value. 

[BTShine]

In general, I think everyone is making this too complicated.  Once you've grasped the concept of FCF and share buybacks vs. dividends you'll view all of this in a very simple manner.  It won't require paragraphs of explanation, etc. 

 

 

[JSArbitrage]

 

 

FCF is FCF, it doesn't matter whether the cash is used on buybacks and not.  Buybacks are irrelevant to the definition of FCF.  The two companies above are worth the same amount in a DCF analysis. 

 

Not true, it does matter whether the cash is used on buybacks or not. From the POV of a shareholder, the shares are worth different values, and that's the relevant metric, not the value of the firm. You are not the sole owner of the firm.

 

Let's think through this.  Company 1 gives you $1M.  Company 2 is giving you (a) $800K cash and (b) $200K worth of stock.  That's all a buy-back is doing.  This is ignoring any effects of taxes and frictional costs.  A company *does not* create value for a shareholder with a buy-back (ignoring under/overvaluation of stock in our example.)  It is simply buying, dollar for dollar, a part of the business.  $1 of FCF gets $1 of stock.  You're getting $1 either way. 

 

 

[Palantir]

Let's think through this.  Company 1 gives you $1M.  Company 2 is giving you (a) $800K cash and (b) $200K worth of stock.  That's all a buy-back is doing.  This is ignoring any effects of taxes and frictional costs.  A company does not create value for a shareholder with a buy-back (ignoring under/overvaluation of stock in our example.)  It is simply buying, dollar for dollar, a part of the business.  $1 of FCF gets $1 of stock.  You're getting $1 either way.

 

Not really. The company is giving you $200K in stock, but in a DCF model, we value cash flows, and cash flows are reduced in the near term for the potential of higher cash flows down the line.

 

Now even if you use your example, you're getting 200K in stock, but if you use higher level of FCF/s growth, it would also trigger double counting because you're adding the intrinsic value of the shares received to the discounted cash flow that is expected of them!

[AZ_Value]

Seems like we should just use Buffett's "owner earnings" which is the amount of cash left over after taking care of operational costs.  Everything after that is an allocation decision.

 

Racemize summarized it well.

 

The free cash flow you should be using to discount is just the cash flow the business generates that belongs to owners, that is all.

 

You need to make an assessment of how management allocates the cash that belongs to you after you've estimated what that cash is and assigned a value to it.

 

There is a cognitive dissonance if you're fundamentally treating differently allocation decisions like buybacks vs dividends because like others have pointed out, they're the same thing except that one party choses to take their distribution in a bigger ownership of the company vs cash. And if your management is using your cash to buy back shares of a significantly overvalued company then do what you do when you want to discount management actions you're not pleased with in your evaluation, and as someone pointed out, should you really be looking at buying a company whose stock is overvalued anyways.

 

But all this is separate from estimating the FCF that the business generated.

 

Your argument about it being a company that has a long history of continuously repurchasing their shares is not valid since repurchases are a distribution to all shareholders just like a dividend, to some in cash to others in % ownership of the business.  To be consistent you would have to look at companies like P&G that have continuously paid and raised their dividends for 40+ years non stop and also adjust their FCF by their dividends in your model.

[AZ_Value]

 

You stumble across a company that yields FCFE, pre-buyback, equal to your discount rate. But they announce that 100% of FCFE will go towards repurchasing shares. Management has not suddenly caused DCF to plummet to 0.

 

 

This is a great common sense check. If you were to find a company that has stayed under the radar for a long time and remained undervalued while generating good cash flows and management has consistently allocated 100% of that to buying back shares would your DCF model show us 0 then since your intention is to discount (FCF - Buybacks)

[Palantir]

I disagree, I think that's exactly what the implication is of doing a 100% buyback. However your future cash flows become higher in exchange for reducing CF₀.

[racemize]

The more I read this thread, the more I think I just shouldn't use DCFs at all, which would alleviate this question entirely.  I much prefer projecting owner earnings and then putting an appropriate multiple on it at the end.

[AZ_Value]

I disagree, I think that's exactly what the implication is of doing a 100% buyback. However your future cash flows become higher in exchange for reducing CF₀.

 

Maybe it's me but I'm not sure it makes much sense.

 

If in your view you should adjust for "the cash that you will not be able to tap into until later" before discounting your stream of cash flows then why stop at share buybacks why not the other allocation option called "nothing" (i.e. the Buffett way). If management retains everything and it all goes to retained earnings, you don't see a penny until: 1) They eventually give you some via dividends or 2) you sell your shares which will have presumably increased in value to reflect the higher book value. Why not then adjust for that every year since you're technically foregoing cash flows now with the idea that they will be higher in the future if the dollars retained grow into more dollars.

 

Maybe this all goes back to the fact that you are combining in your DCF model a per share analysis now and in the future which I'm not sure how it can be done without some heavy guesswork

[Palantir]

The reason being, I think the DCF model has some necessarily simplifying assumptions, eg, that cash flows are paid straight to investors, or that there is no distinction between the investor and the "firm".

 

If the firm just generated 1M in FCF and did nothing with it, it would form a part of the firm's cash account and be an element of value for the company, however if it was used to repurchase shares, it would certainly create value for shareholders, but the way I see it, that value would be counted in your future projections of per share FCF.

[maxthetrade]

Buybacks are basically like any other non maintenance cap-ex and should be treated equally in a DCF. Whether a company buys its own stock, aquires another company or spends the cash on any other growth cap-ex is irrelevant, in all cases the cash can't be taken out in the year the Investment is made but future FCF will (hopefully) be higher. If you discount FCF and project future growth from growth cap-ex you're double counting.

 

 

[BTShine]

The more I read this thread, the more I think I just shouldn't use DCFs at all, which would alleviate this question entirely.  I much prefer projecting owner earnings and then putting an appropriate multiple on it at the end.

 

Agree.  The more variables one includes in his analysis, the more likely it is he makes a mistake.