petit_bateau
Active Member
Let me throw something out for comment.
I suspect that a lot of folk would have liked to have sold TSLA at $400 in Oct-2021 and have bought back in at $120 or so in Dec-2022, with the benefit of hindsight and perfect timing and all that.
But without the perfect retrospectascope the thought may be niggling at folk to sell a little bit of TSLA from time to time in a rising market and derisk things that way. But when to sell, and when to buy.
With that in mind I've taken my success case model where Tesla does get to 20m by 2030, and maintains an automotive GM% of about 30% the whole duration (however unlikely that may be). This yields a forecast of net income per share, i.e. Earnings Per Share (EPS) for the next several years.
In the red boxed row the future share price is computed assuming a trailing PE of 40 x EPS, or the end-year historical is noted. So this means the historical PE varies (grey line on graph) but the share price is correct, whereas the future PE is pretty flat but the PE-driven share price forecast is likely to be wrong. I picked a trailing PE of 40x to drive the future forecast (red line on graph) and that corresponds to a 'future' PE of 25-35 (flattish grey line). (I'm explaining the difference between 40x and 25-35x which is a nuance. You can ignore this if it confuses you).
In the purple boxed row the share price is computed using the 10-year NPV of the EPS, and with the discount rate set to being the 10-year bond rate plus the typical equity risk premium. (I used the OECD data for US 10-year.) This means that the effect of rising Fed rates are correctly reflected in the purple box. (By way of comparison the green box holds the NPV discount rate invariant at 10%). The equity risk premium is the market-average - if anyone can point out sector specific equity risk data that is better I'd be interested.
I appreciate these are crude calculations, but I am a bear of little brain.
Next I've plotted them. Of course these are year-end numbers for each year and a lot can happen within a year. So to get a feel for how much can happen within a year I've overlaid the actual historical share price scaled to fit (that's the light green line). So the red line and the green line coincide perfectly at 12-month intervals if you look carefully.
Only the next few years are interesting so I've truncated share price at $600 (!!). The NPV calculation looks at the successive 10-years of EPS, but since I hold EPS flat beyond 2030 the NPV calculation gets a little screwy after 2025. But since I can't forecast that well it doesn't matter.
Remember this economic model for EPS is of course at the high end of the range, i.e. 20m by 2030 and maintaining 30% automotive gross margin (in a sense this is where FSD is taken into account). My model assumes success in energy at a lower GM%, and takes no account whatsoever of RoboTaxi or of Optimus.
The first observation I make is that PE on its own is not a sell signal. Selling at PE of over 100x can cause a great deal of missing out in a growth share such as TSLA.
A second observation is that when the market share price reaches double the NPV-driven share price that ought to be a sell signal as irrational exuberance is probably driving a share price bubble. Conversely when market share price is significantly below NPV-driven share price that may be a buy signal. This degree of deviation from NPV seems to be an actionable observation.
A third observation is that one has to have a somewhat sensible forecast to drive the NPV calculation in a EPS model such as this. That is of course the conversation that rational longer term investors are trying to have in the noise of all the short term traders, and we are all doing it through the mechanism of the share price. But one must have such a model if one is to be able to form a independent view using the NPV-deviation signal.
I guess a final comment is that this in turn relies on Tesla giving its shareholders sufficient information in a transparent and reliable and trustworthy and methodical manner that shareholders can form a rational view.
(P.S. I've looked at the PEG metric and it is not giving a reliable / useful indicator, that is why I am not showing it on the graph. You can see it in the table.)
I suspect that a lot of folk would have liked to have sold TSLA at $400 in Oct-2021 and have bought back in at $120 or so in Dec-2022, with the benefit of hindsight and perfect timing and all that.
But without the perfect retrospectascope the thought may be niggling at folk to sell a little bit of TSLA from time to time in a rising market and derisk things that way. But when to sell, and when to buy.
With that in mind I've taken my success case model where Tesla does get to 20m by 2030, and maintains an automotive GM% of about 30% the whole duration (however unlikely that may be). This yields a forecast of net income per share, i.e. Earnings Per Share (EPS) for the next several years.
In the red boxed row the future share price is computed assuming a trailing PE of 40 x EPS, or the end-year historical is noted. So this means the historical PE varies (grey line on graph) but the share price is correct, whereas the future PE is pretty flat but the PE-driven share price forecast is likely to be wrong. I picked a trailing PE of 40x to drive the future forecast (red line on graph) and that corresponds to a 'future' PE of 25-35 (flattish grey line). (I'm explaining the difference between 40x and 25-35x which is a nuance. You can ignore this if it confuses you).
In the purple boxed row the share price is computed using the 10-year NPV of the EPS, and with the discount rate set to being the 10-year bond rate plus the typical equity risk premium. (I used the OECD data for US 10-year.) This means that the effect of rising Fed rates are correctly reflected in the purple box. (By way of comparison the green box holds the NPV discount rate invariant at 10%). The equity risk premium is the market-average - if anyone can point out sector specific equity risk data that is better I'd be interested.
I appreciate these are crude calculations, but I am a bear of little brain.
Next I've plotted them. Of course these are year-end numbers for each year and a lot can happen within a year. So to get a feel for how much can happen within a year I've overlaid the actual historical share price scaled to fit (that's the light green line). So the red line and the green line coincide perfectly at 12-month intervals if you look carefully.
Only the next few years are interesting so I've truncated share price at $600 (!!). The NPV calculation looks at the successive 10-years of EPS, but since I hold EPS flat beyond 2030 the NPV calculation gets a little screwy after 2025. But since I can't forecast that well it doesn't matter.
Remember this economic model for EPS is of course at the high end of the range, i.e. 20m by 2030 and maintaining 30% automotive gross margin (in a sense this is where FSD is taken into account). My model assumes success in energy at a lower GM%, and takes no account whatsoever of RoboTaxi or of Optimus.
The first observation I make is that PE on its own is not a sell signal. Selling at PE of over 100x can cause a great deal of missing out in a growth share such as TSLA.
A second observation is that when the market share price reaches double the NPV-driven share price that ought to be a sell signal as irrational exuberance is probably driving a share price bubble. Conversely when market share price is significantly below NPV-driven share price that may be a buy signal. This degree of deviation from NPV seems to be an actionable observation.
A third observation is that one has to have a somewhat sensible forecast to drive the NPV calculation in a EPS model such as this. That is of course the conversation that rational longer term investors are trying to have in the noise of all the short term traders, and we are all doing it through the mechanism of the share price. But one must have such a model if one is to be able to form a independent view using the NPV-deviation signal.
I guess a final comment is that this in turn relies on Tesla giving its shareholders sufficient information in a transparent and reliable and trustworthy and methodical manner that shareholders can form a rational view.
(P.S. I've looked at the PEG metric and it is not giving a reliable / useful indicator, that is why I am not showing it on the graph. You can see it in the table.)
Attachments
Last edited: