Thank you for your comment Devin. I agree, the FCF turned to be quite inaccurate because I considered ~30% margin as in FY21, which turned to be wrong. In the second part of the year Palantir invested very aggressively (hiring+client acquisition) compressing the FCF margin to 17-22% and growth slowed down substantially (so the '22 estimates). To sum up while I was estimating ~$500-550mn FCF the '22 FCF turned $420mn.
I tried to be conservative, but not enough.
In December I shared the article on Seeking Alpha with the updated analyst estimates:
Using a EV/FCF ratio assumes Free Cash Flow to Firm. But you are referencing buying the equity and attempting to calculate equity returns. To calculate the Free Cash Flow to Equity, you must properly attribute the impact of SBC to equity, which there are a few viable methods, but it doesn't look like it is being done here. From an equity perspective, PLTR is not generating positive FCF.
The dilution is already incorporated in the multiples as the EV of each year was calculated with increasing number of shares. I shared in detail how I incorporate the SBC and why here:
Feb 28, 2023·edited Feb 28, 2023Liked by Arny Trezzi
One key issue with this is that you are often referencing enterprise-level growth, or equity as a whole, and not per share growth - which is what you are actually buying. The example of buying a company as a whole seems to exclude the idea that you'd still need to pay employees in the future, whether with cash or stock. If the former, good bye adjusted FCF, with the latter, you dilute your ownership off a base of 100%. Ultimately, SBC, while viewed as a non-cash expense, is (as Aswath calls it) more akin to a like-kind expense, unlike depreciation. SBC is a very real operating expense, unless you don't think employees are part of operations.
Another concern is that not viewing SBC as a very real, recurring operating expense only obfuscates calculating other factors that influence real operations. And extrapolating X top-line growth and X FCF margin off an adjusted FCF base after adding back SBC is likely going to present a skewed result for an additional reason: you're removing lots of other needed assumptions regarding operations and effectively concentrating them into your singular assumption regarding dilution rate. I am all for simplified models, unique adjustments, and avoiding false precision, but this seems several steps in the wrong direction.
Your other piece citing issues with DCFs can be also applied to nearly all other forms of valuation as well: inputs matter a great deal. It's also key to differentiate between unlevered and levered DCFs, which can present useful perspectives.
In that piece, your Company A vs Company B example fails to highlight a simple truth. In it, you say, "Company B, on the other hand, needs to erode its cash balance, or issue new Debt, in order to pursue growth. New Equity would be detrimental to existing shareholders because it would cause dilution. New Debt would be costly because of the weak ability to generate cash flows."
Company A is financing an ongoing operational cost by doing what is described in that second sentence. It's also key to differentiate here, once again, between FCF to the firm vs FCF to the equity. Debt holders are much less concerned with FCFE as they have priority in the capital structure. Dilution of equity doesn't dilute them. However, dilution needs to be a very real concern for equity holders.
None of this is to claim PLTR isn't valuable, can become valuable, is trash, or anything else. I simply think that while valuation provides a wide range in freedoms in how we express our views, such freedoms also allow all kinds of skewing. When evaluating an asset with underlying operations, it's up to us to determine how to best illustrate those operations, and I don't think that's being done here.
I truly appreciate different perspectives and schools of thought, and I am enjoying reading your work. Thanks for taking the time to read this.
You are right Devin, on the article I could have specified that the numbers were all on the per share growth. I just wrote "These multiples include the effect of SBC in terms of the increased number of shares." to avoid further complicate the reading for the avg. reader, but would have been better underscoring it.
If you consider SBC as part of the firm operations and subtract them from the FCF you get the same numbers of my approach under the same dilution assumptions. SBC in the end is nothing else than $price * shares given.
I just consider that focusing on the dilutive effect of the number of shares shows a better picture:
- the company generates cash (which is true as the Net Cash increases)
- the amount of SBC accounting for the existing RSUs is fixed at the grant date. If stock moves the SBC stays the same. New RSUs will be affected by the current stock price. What you really care as a shareholder is the net dilution you get when the RSU vests, not the SBC that has been recorded in accounting.
- as a shareholder realising that I get ~25% diluted growth from a 30% growth and 5% dilution is imo much more intuitive than a FCF - SBC number.
Overall, doing FCF - SBC or focusing on the per share numbers doesn't change under the same dilutive assumptions. The truly important thing is to incorporate the dilution in the valuation, which most didn't (even WS analysts).
While I agree that unlevered vs levered FCF generally gives useful perspective, here I believe it would only complicate things, which already are. The difference would be
- net interest on cash (which I didn't consider for simplicity)
- equity holders are affected by SBC (in terms of number of shares) while the firm no.
- any other difference?
Since we are talking about a software company with 2.6bn Cash that the real difference between FCF to firm vs FCF to equity is the dilution which affects the number of shares but not the FCF.
In other words, the operating reality I see is that Palantir as a company produces FCF which makes the Net Cash increase and can be used to push growth then they see the opportunity (software companies are all about optionality). Shareholders meanwhile don't benefit from that FCF, but are entitled only to (1-dilution) of that cash flow. (if Dilution was 0% they would get 100% of It).
Thank you so much for taking the time of expanding your answers Devin, really appreciated. Honoured for the opportunity to better explain my choices.
(1+0.2}^9 tells you the cash flow you would receive at 9th year with flows increasing by 20% per year, not the amount that you "cumulated" value you obtained by receiving the FCF.
Imagine you invest $2.000 for a company that generates $100/y = 20x it would take 20y to break even:
Year 0 = -2000
Year 1 = +100, still -1,900 to break even
Year 2 = +100, still -1,800 to break even
... year 20 = +100, with the flows I covered all my investment.
But if the flows grow 20%/y you have:
Year 0 = -2000
Year 1 = +100 flow, still -1,900 to break even
Year 2 = +120 flow, still -1,780 to break even
Year 3 = +144 flow, still 1,636 to break even
...
Year 8 = +358 flow, still 350 to break even
Year 9 = +429 flow, you cover your first 2,000 investment with the cumulated cash flows you received.
Since we talk about FCF remember that cash cumulates in the back account unless the company distributes it do shareholder, but in any case that is value that is generated.
https://www.tandfonline.com/doi/full/10.1080/01972243.2022.2100851 (Do you have this info. Important.)
Hi Carlos, I remember I read it all an offered some points of reflection. Thank you for sharing!
Would you be willing to break down how you are calculating FCF? Because this looks quite inaccurate.
Thank you for your comment Devin. I agree, the FCF turned to be quite inaccurate because I considered ~30% margin as in FY21, which turned to be wrong. In the second part of the year Palantir invested very aggressively (hiring+client acquisition) compressing the FCF margin to 17-22% and growth slowed down substantially (so the '22 estimates). To sum up while I was estimating ~$500-550mn FCF the '22 FCF turned $420mn.
I tried to be conservative, but not enough.
In December I shared the article on Seeking Alpha with the updated analyst estimates:
https://seekingalpha.com/article/4565128-palantir-buy-signals-a-dive-inside-my-playbook
Using a EV/FCF ratio assumes Free Cash Flow to Firm. But you are referencing buying the equity and attempting to calculate equity returns. To calculate the Free Cash Flow to Equity, you must properly attribute the impact of SBC to equity, which there are a few viable methods, but it doesn't look like it is being done here. From an equity perspective, PLTR is not generating positive FCF.
The dilution is already incorporated in the multiples as the EV of each year was calculated with increasing number of shares. I shared in detail how I incorporate the SBC and why here:
https://palantirbullets.substack.com/p/why-99-of-palantir-dcfs-fail
By focusing on the per share numbers you capture the "diluted growth"
One key issue with this is that you are often referencing enterprise-level growth, or equity as a whole, and not per share growth - which is what you are actually buying. The example of buying a company as a whole seems to exclude the idea that you'd still need to pay employees in the future, whether with cash or stock. If the former, good bye adjusted FCF, with the latter, you dilute your ownership off a base of 100%. Ultimately, SBC, while viewed as a non-cash expense, is (as Aswath calls it) more akin to a like-kind expense, unlike depreciation. SBC is a very real operating expense, unless you don't think employees are part of operations.
Another concern is that not viewing SBC as a very real, recurring operating expense only obfuscates calculating other factors that influence real operations. And extrapolating X top-line growth and X FCF margin off an adjusted FCF base after adding back SBC is likely going to present a skewed result for an additional reason: you're removing lots of other needed assumptions regarding operations and effectively concentrating them into your singular assumption regarding dilution rate. I am all for simplified models, unique adjustments, and avoiding false precision, but this seems several steps in the wrong direction.
Your other piece citing issues with DCFs can be also applied to nearly all other forms of valuation as well: inputs matter a great deal. It's also key to differentiate between unlevered and levered DCFs, which can present useful perspectives.
In that piece, your Company A vs Company B example fails to highlight a simple truth. In it, you say, "Company B, on the other hand, needs to erode its cash balance, or issue new Debt, in order to pursue growth. New Equity would be detrimental to existing shareholders because it would cause dilution. New Debt would be costly because of the weak ability to generate cash flows."
Company A is financing an ongoing operational cost by doing what is described in that second sentence. It's also key to differentiate here, once again, between FCF to the firm vs FCF to the equity. Debt holders are much less concerned with FCFE as they have priority in the capital structure. Dilution of equity doesn't dilute them. However, dilution needs to be a very real concern for equity holders.
None of this is to claim PLTR isn't valuable, can become valuable, is trash, or anything else. I simply think that while valuation provides a wide range in freedoms in how we express our views, such freedoms also allow all kinds of skewing. When evaluating an asset with underlying operations, it's up to us to determine how to best illustrate those operations, and I don't think that's being done here.
I truly appreciate different perspectives and schools of thought, and I am enjoying reading your work. Thanks for taking the time to read this.
You are right Devin, on the article I could have specified that the numbers were all on the per share growth. I just wrote "These multiples include the effect of SBC in terms of the increased number of shares." to avoid further complicate the reading for the avg. reader, but would have been better underscoring it.
If you consider SBC as part of the firm operations and subtract them from the FCF you get the same numbers of my approach under the same dilution assumptions. SBC in the end is nothing else than $price * shares given.
I just consider that focusing on the dilutive effect of the number of shares shows a better picture:
- the company generates cash (which is true as the Net Cash increases)
- the amount of SBC accounting for the existing RSUs is fixed at the grant date. If stock moves the SBC stays the same. New RSUs will be affected by the current stock price. What you really care as a shareholder is the net dilution you get when the RSU vests, not the SBC that has been recorded in accounting.
- as a shareholder realising that I get ~25% diluted growth from a 30% growth and 5% dilution is imo much more intuitive than a FCF - SBC number.
Overall, doing FCF - SBC or focusing on the per share numbers doesn't change under the same dilutive assumptions. The truly important thing is to incorporate the dilution in the valuation, which most didn't (even WS analysts).
While I agree that unlevered vs levered FCF generally gives useful perspective, here I believe it would only complicate things, which already are. The difference would be
- net interest on cash (which I didn't consider for simplicity)
- equity holders are affected by SBC (in terms of number of shares) while the firm no.
- any other difference?
Since we are talking about a software company with 2.6bn Cash that the real difference between FCF to firm vs FCF to equity is the dilution which affects the number of shares but not the FCF.
In other words, the operating reality I see is that Palantir as a company produces FCF which makes the Net Cash increase and can be used to push growth then they see the opportunity (software companies are all about optionality). Shareholders meanwhile don't benefit from that FCF, but are entitled only to (1-dilution) of that cash flow. (if Dilution was 0% they would get 100% of It).
Thank you so much for taking the time of expanding your answers Devin, really appreciated. Honoured for the opportunity to better explain my choices.
To understand the importance of growth consider the 2 cases of buying an entire business at 20x EV/FCF:
no growth: it would take us 20 years to return on our investment;
20% growth: it would take us 9 years to return on our investment.
The 20% growth makes us “save” 11 years of investment.
Hi Arny, I am trying to calculate how u arrive at 9yrs.... (1+0.2}^9 is only 5.`16... To get to 20, I will need 17 yrs.
YR multiplication yoy
2 1.44
3 1.728
4 2.0736
5 2.48832
6 2.985984
7 3.5831808
8 4.29981696
9 5.159780352
10 6.191736422
11 7.430083707
12 8.916100448
13 10.69932054
14 12.83918465
15 15.40702157
16 18.48842589
17 22.18611107
18 26.62333328
19 31.94799994
20 38.33759992
21 46.00511991
thanks for sharing
Hi Irene, thanks for your question.
(1+0.2}^9 tells you the cash flow you would receive at 9th year with flows increasing by 20% per year, not the amount that you "cumulated" value you obtained by receiving the FCF.
Imagine you invest $2.000 for a company that generates $100/y = 20x it would take 20y to break even:
Year 0 = -2000
Year 1 = +100, still -1,900 to break even
Year 2 = +100, still -1,800 to break even
... year 20 = +100, with the flows I covered all my investment.
But if the flows grow 20%/y you have:
Year 0 = -2000
Year 1 = +100 flow, still -1,900 to break even
Year 2 = +120 flow, still -1,780 to break even
Year 3 = +144 flow, still 1,636 to break even
...
Year 8 = +358 flow, still 350 to break even
Year 9 = +429 flow, you cover your first 2,000 investment with the cumulated cash flows you received.
Since we talk about FCF remember that cash cumulates in the back account unless the company distributes it do shareholder, but in any case that is value that is generated.
Hope I clarified!
Thank you very much Arny... I understand now.!!.. appreciate that you took the time to reply!!
My pleasure Irene! When in doubt, I am happy to answer :D