Near-future quarterly financial projections

[jbcarioca]

As for modeling capex for factories and new models are concerned we should keep in mind that future GF are likely to cost ~$2 billion or so, and new models from Model Y, pickup, Roadster, Semi and new Model S and X will all be strongly derivative from most respects. Further, Tesla has just passed the point that they can command substantial Tier One investment with lower prices to boot. I never quote my own projections, but I think almost all of you are overstating capex because of these two factors.

We all will not believe the GM's coming our way next year.

 

[luvb2b]

to add some more commentary:

over the last few years tesla has averaged around $2,200 per vehicle in warranty reserves.

that number climbed to around $2,400 in 18q1 and to $2,900 in 18q2.

my guess is they had more initial quality issues with the model 3 than they thought and so increased reserving. however as quality problems clear they will be reserving at a high level on a much higher number of cars, and at some point that trend should reverse again perhaps temporarily to as little as 1.5k per car, and then longer term i think settle closer to $2k per vehicle. on a typical model 3 that could mean about an other 1.5-2.0% of gross margin potential vs 18q2, with a short term boost of as much as 3% possible.

Almost but not quite correct. COGS includes allocations to warranty reserves NOT warranty repairs. Warranty repairs debit to warranty reserves. Over time as actual experience results in either higher or lower actual warranty costs than allocated to the reserve adjustments are made, ordinarily by either decreasing future reserve credits in the first case or increasing them in the second case. If the differences are deemed material (they very rarely are) previous period statements are restated. Normally the errors are not substantial so with experience future reserve allocations are adjusted. Under only very rare circumstances (I personally know of none) is any adjustment been treated as an exceptional item. Such radically unusual cases as the BMW E32 750i and, more recently the Cadillac ATS and Mercedes CLA have produced wildly excessive warranty costs but IIRC none were recognized explicitly as exceptional items. The sole exception of which I know is the NSU Ro-80 because that vehicle drove the manufacturer out of business and they were bought by VW where eventually their design philosophy became Audi's.

Tesla has been a unique case since they had minimal experience with running gear but have eight year unlimited mileage warranty on Models S and X. It appears that they've had much better experience than they initially forecast, but that is not specifically disclosed. The basic warranty has been different since the majority of FOR data was easily inferable from supplier data and much of the most sensitive repair risks were parts that were already in production with Mercedes Benz models. Large exception: FWD. Then the various Performance models all have much higher insurance reserves than do other models, and that has plentiful relevant data from, among others, AMG and M. As it turns out the relationship between P and normal is less distinct than it is for both AMG and M, probably because of various ICE-specific high performance parts being more failure-prone than are the electrical equivalent.

NOTE: a disclaimer. Although I did work as a consultant to PWC and SRI International on auto industry practices regarding warranty and R&D issues, that was a long time ago and I used nothing at all sourced from any work I did for them either for specific clients or for multi-client projects.

sorry for all the space. Tesla has not had a problem with warranty costs, even when replacing drive units, 12v batteries and door components.

 

[luvb2b]

*edit* just thinking warranty reserve should be included here as well. will come back with a refresh in a bit.
*edit-edit* including warranty reserve as a non-cash cogs item will push cash gross margin far over 30% in q3 and doesn't seem sensible. so going to leave the discussion as it is for now.

i have been thinking over the discussion with @Reality below.

i think what it means is that model 3 cash gross margin is going to be near 30% in q3. that will drive ~650-750m of additional cash generation. along with some zev credit sales and further increase in the payables balance, i think it means my cash modeling is missing the mark by several hundred million (meaning, tesla will generate more cash than i am currently forecasting). this is on a per quarter basis, meaning i may have missed end of year cash balances by a half billion or more.

i don't want to go into what drives every single line below, but you'll get the idea of how the evolution of depreciation and subtraction of sbc + dep from cogs will impact cash gross margin.

…

est model 3 manuf deprciation

per model 3

blended manuf deprciation per 3

stock-based comp per model 3

estimated cogs per model 3

minus non-cash cogs

est cash cogs per model 3

est model 3 cash gm

est model 3 gaap gm

[TD2]luv q4-18e[/TD2][TD2]luv q3-18e[/TD2][TD2]Jun-18[/TD2][TD2]Mar-18[/TD2] [TD2]135,000[/TD2][TD2]125,000[/TD2][TD2]115,000[/TD2][TD2]95,000[/TD2] [TD2]2.25[/TD2][TD2]2.50[/TD2][TD2]4.02[/TD2][TD2]9.73[/TD2] [TD2]2.31[/TD2][TD2]2.84[/TD2][TD2]4.65[/TD2][TD2]12.61[/TD2] [TD2]-0.65[/TD2][TD2]0.35[/TD2][TD2]0.32[/TD2][TD2]0.50[/TD2] [TD2]45.24[/TD2][TD2]50.40[/TD2][TD2]53.29[/TD2][TD2]70.15[/TD2] [TD2]1.66[/TD2][TD2]3.19[/TD2][TD2]4.98[/TD2][TD2]13.11[/TD2] [TD2]43.58[/TD2][TD2]47.21[/TD2][TD2]48.31[/TD2][TD2]57.04[/TD2] [TD2] 33.1% [/TD2][TD2] 27.1% [/TD2][TD2] 15.5% [/TD2][TD2] -0.4% [/TD2] [TD2]22.0%[/TD2][TD2]16.0%[/TD2][TD2]4.5%[/TD2][TD2]-23.5%[/TD2]

part of an offline discussion i was having with @Reality which i felt is worthwhile to throw out for discussion. posted with permission.

@Reality asked: Interesting thing to note, but the "tesla positive gross margin for model 3" was excluding non cash items per the update letter. I missed that part. How does that change your modeling?

my reply thus far which requires some further development:

that's a good find.

here's one comment in the gross margin para:

Model 3 gross margin turned slightly positive in Q2 even though we were still ramping production and did not yet deliver any AllWheel-Drive or performance models.​

here's the other comment in the cash flow para:

Model 3 gross profit excluding non-cash items shifted from negative in Q1 to positive in Q2, driving significant improvement in cash profitability​

non-cash items would be sbc & depreciation i guess. they give us sbc and it's only a couple hundred bucks a unit, so we can ignore that.

these 2 statements together indicate it's not only depreciation math that is driving positive gross margin, but also real cash cost reductions (maybe lower parts or labor?).

so the new piece of information here is that excluding depreciation, the model 3 was negative gross margins 18q1. i think i had a -24.5% gm impact of depreciation and model gm at 20.3%, meaning i had estimated gm excluding depreciation as positive by several percentage points.

what this tells me is i probably modeled too much depreciation in 18q1 and/or model s/x gm was higher than i thought and model 3 gm was lower than i thought.

i'll have to go back and review the commentary around all of that.

 

[jbcarioca]

*edit* just thinking warranty reserve should be included here as well. will come back with a refresh in a bit.
*edit-edit* including warranty reserve as a non-cash cogs item will push cash gross margin far over 30% in q3 and doesn't seem sensible. so going to leave the discussion as it is for now.

Warranty reserve is treated as direct cogs charge, it actual;ly acts as a charge against current sales ( and a few charge to production rather than sales). either way the result is to charge the reserve of COGS or COS. The difference depends largely on whether they manufacturer uses distributors (COS) or not (COGS).

 

[jbcarioca]

to add some more commentary:

over the last few years tesla has averaged around $2,200 per vehicle in warranty reserves.

that number climbed to around $2,400 in 18q1 and to $2,900 in 18q2.

my guess is they had more initial quality issues with the model 3 than they thought and so increased reserving. however as quality problems clear they will be reserving at a high level on a much higher number of cars, and at some point that trend should reverse again perhaps temporarily to as little as 1.5k per car, and then longer term i think settle closer to $2k per vehicle. on a typical model 3 that could mean about an other 1.5-2.0% of gross margin potential vs 18q2, with a short term boost of as much as 3% possible.

I think the increase is more related to sales mix than to Model 3 specifically because there was probably a higher proportion of more heavily options vehicles, partly visible in the rising ASP. Other things being equal the higher option mix higher the warranty cost, with a very large jump for high performance versions. 3q will be higher again, I think, partly because of a big jump in Model 3 Performance sales. If we had really accurate model mix data it would be easy to find out if I am correct. I suggest this rather than pure Model 3 because model mix is usually a larger influence than specific model (there are spectacular exceptions). OTOH, first production is always worse too, but if it were only model 3 then Q2 should have been lower.
Were there not and unusually high proportion of S and X P100D produced in the first half, delivered or not? If this logic is sound Q3 will have another rise, followed by a drift lower during 2019.

 

[luvb2b]

making a couple corrections and posting a revised table for discussion purposes. fixed the 18q4 sbc in cogs having the wrong sign.

to review briefly, i am trying to reconcile these statements from the past 2 letters:
here's one comment in the gross margin para:

Model 3 gross margin turned slightly positive in Q2 even though we were still ramping production and did not yet deliver any AllWheel-Drive or performance models.​

here's the other comment in the cash flow para:

Model 3 gross profit excluding non-cash items shifted from negative in Q1 to positive in Q2, driving significant improvement in cash profitability​

i took these two statements to mean that cash gross margin was slightly negative in 18q1 while gaap gross margin was positive in 18q2. i hope someone else tries to sit down and think this through because it makes my head spin with all these non-standard accounting terms tesla throws around.

…

est model 3 manuf deprication

per model 3

blended manuf deprctn per 3

stock-based comp per model 3

estimated cogs per model 3

minus non-cash cogs

est cash cogs per model 3

est model 3 cash gm

est model 3 gaap gm

[TD2]luv q4-18e[/TD2][TD2]luv q3-18e[/TD2][TD2]Jun-18[/TD2][TD2]Mar-18[/TD2] [TD2]135,000[/TD2][TD2]125,000[/TD2][TD2]115,000[/TD2][TD2]95,000[/TD2] [TD2]2.25[/TD2][TD2]2.50[/TD2][TD2]4.02[/TD2][TD2]9.73[/TD2] [TD2]2.31[/TD2][TD2]2.84[/TD2][TD2]4.65[/TD2][TD2]12.61[/TD2] [TD2]0.34[/TD2][TD2]0.32[/TD2][TD2]0.32[/TD2][TD2]0.50[/TD2] [TD2]45.24[/TD2][TD2]50.40[/TD2][TD2]53.29[/TD2][TD2]70.15[/TD2] [TD2]2.65[/TD2][TD2]3.17[/TD2][TD2]4.98[/TD2][TD2]13.11[/TD2] [TD2]42.59[/TD2][TD2]47.23[/TD2][TD2]48.31[/TD2][TD2]57.04[/TD2] [TD2]36.2%[/TD2][TD2]27.0%[/TD2][TD2]15.5%[/TD2][TD2]-0.4%[/TD2] [TD2]22.0%[/TD2][TD2]16.0%[/TD2][TD2]4.5%[/TD2][TD2]-23.5%[/TD2]

*edit* just thinking warranty reserve should be included here as well. will come back with a refresh in a bit.
*edit-edit* including warranty reserve as a non-cash cogs item will push cash gross margin far over 30% in q3 and doesn't seem sensible. so going to leave the discussion as it is for now.

 

[Singer3000]

Am interested to understand what drives the pace of reduction on cogs in your forecast. Clearly it's pretty key in determining the timing of when the SR RWD model will be released. Q1/Q2 19 seems to be the roughly accepted timeline but your figures suggest otherwise, unless there's a cliff edge in production costs in 2019.

 

[erthquake]

Am interested to understand what drives the pace of reduction on cogs in your forecast. Clearly it's pretty key in determining the timing of when the SR RWD model will be released. Q1/Q2 19 seems to be the roughly accepted timeline but your figures suggest otherwise, unless there's a cliff edge in production costs in 2019.

Effectively, there is a cliff edge in the production costs for the SR RWD because of fewer and less expensive parts. Compared to the currently available RWD version ($49k+), the $35k SR RWD has a smaller battery, no premium upgrades package, and uses less LTE bandwidth. In addition, by the time the SR is produced in the beginning of 2019, the factory should already be at or near 8k/week so depreciation will be spread out over even more cars.

 

[Singer3000]

Ah yes thanks, of course. I’d better not make any important decisions today, my brain has obviously vacated the building.

Effectively, there is a cliff edge in the production costs for the SR RWD because of fewer and less expensive parts. Compared to the currently available RWD version ($49k+), the $35k SR RWD has a smaller battery, no premium upgrades package, and uses less LTE bandwidth. In addition, by the time the SR is produced in the beginning of 2019, the factory should already be at or near 8k/week so depreciation will be spread out over even more cars.

.

 

[Matias]

What do you think is the major reason for analyst consensus estimate for Q3 to be negative, while I understand, that consensus here is positive Q3 eps?

TSLA Analyst Opinion | Analyst Estimates | Tesla, Inc. Stock - Yahoo Finance

 

[luvb2b]

"i think they are just wrong."

What do you think is the major reason for analyst consensus estimate for Q3 to be negative, while I understand, that consensus here is positive Q3 eps?

TSLA Analyst Opinion | Analyst Estimates | Tesla, Inc. Stock - Yahoo Finance

 

[erthquake]

What do you think is the major reason for analyst consensus estimate for Q3 to be negative, while I understand, that consensus here is positive Q3 eps?

TSLA Analyst Opinion | Analyst Estimates | Tesla, Inc. Stock - Yahoo Finance

To be frank, I don't think any analyst on Wall Street does the kind of in-depth analysis that @luvb2b and others have done. And that is our advantage as investors in $TSLA.

 

[MitchJi]

hile gaap gross margin was positive in 18q2. i hope someone else tries to sit down and think this through because it makes my head spin with all these non-standard accounting terms tesla throws around.

If you email Tesla IR with a list of questions they should get back to you?
I just sent them the following:

Hi Martin,
I think it would be useful if you would add one or two brief sentences to the production numbers email that gave a rough idea of M3 profitability. Something like the following quoted from the Q2 ER letter would be excellent:
 Model 3 gross margin turned slightly positive in Q2, expecting roughly 15% in Q3.

A statement saying something like roughly 10-20% (a rough magnitude) would be excellent. When you put out the numbers for the last quarter the SP spiked and then immediately fell. Imo the reason it fell was because of the missing profitability information.

I’d also like it if you would announce your intention to do that.

If anyone else wants to chime in I’d appreciate the support. I’ll post it if I get a reply.

 

[omgwtfbyobbq]

to add some more commentary:

over the last few years tesla has averaged around $2,200 per vehicle in warranty reserves.

that number climbed to around $2,400 in 18q1 and to $2,900 in 18q2.

my guess is they had more initial quality issues with the model 3 than they thought and so increased reserving. however as quality problems clear they will be reserving at a high level on a much higher number of cars, and at some point that trend should reverse again perhaps temporarily to as little as 1.5k per car, and then longer term i think settle closer to $2k per vehicle. on a typical model 3 that could mean about an other 1.5-2.0% of gross margin potential vs 18q2, with a short term boost of as much as 3% possible.

I think it's for their production during the tax credit phase out as a whole. Prior to the credit phasing out, they could spend a little more time on build quality and everyone would still get the $7,500. As of this quarter, the more cars they get into customer hands, the more customers get the full or partial credit and the more cars Tesla sells. The increase in warranty reserves reflects Tesla's push to get cars into customer hands even if there are more defects than in the past.

 

[luvb2b]

i may have joked too soon.

your inquiry did have me reviewing my work vs a few others, and i discovered that i had still kept some opex adjustments assuming cost reductions due to workforce cuts. that idea was blown out of the water with the last report.

i'll make some adjustments to opex and try to post some revised estimates. i think i'd still be over $1 non-gaap though and so highest on the street.

What do you think is the major reason for analyst consensus estimate for Q3 to be negative, while I understand, that consensus here is positive Q3 eps?

TSLA Analyst Opinion | Analyst Estimates | Tesla, Inc. Stock - Yahoo Finance

 

[oneday]

*edit* just thinking warranty reserve should be included here as well. will come back with a refresh in a bit.
*edit-edit* including warranty reserve as a non-cash cogs item will push cash gross margin far over 30% in q3 and doesn't seem sensible. so going to leave the discussion as it is for now.

i have been thinking over the discussion with @Reality below.

i think what it means is that model 3 cash gross margin is going to be near 30% in q3. that will drive ~650-750m of additional cash generation. along with some zev credit sales and further increase in the payables balance, i think it means my cash modeling is missing the mark by several hundred million (meaning, tesla will generate more cash than i am currently forecasting). this is on a per quarter basis, meaning i may have missed end of year cash balances by a half billion or more.

i don't want to go into what drives every single line below, but you'll get the idea of how the evolution of depreciation and subtraction of sbc + dep from cogs will impact cash gross margin.

…

est model 3 manuf deprciation

per model 3

blended manuf deprciation per 3

stock-based comp per model 3

estimated cogs per model 3

minus non-cash cogs

est cash cogs per model 3

est model 3 cash gm

est model 3 gaap gm

[TD2]luv q4-18e[/TD2][TD2]luv q3-18e[/TD2][TD2]Jun-18[/TD2][TD2]Mar-18[/TD2] [TD2]135,000[/TD2][TD2]125,000[/TD2][TD2]115,000[/TD2][TD2]95,000[/TD2] [TD2]2.25[/TD2][TD2]2.50[/TD2][TD2]4.02[/TD2][TD2]9.73[/TD2] [TD2]2.31[/TD2][TD2]2.84[/TD2][TD2]4.65[/TD2][TD2]12.61[/TD2] [TD2]-0.65[/TD2][TD2]0.35[/TD2][TD2]0.32[/TD2][TD2]0.50[/TD2] [TD2]45.24[/TD2][TD2]50.40[/TD2][TD2]53.29[/TD2][TD2]70.15[/TD2] [TD2]1.66[/TD2][TD2]3.19[/TD2][TD2]4.98[/TD2][TD2]13.11[/TD2] [TD2]43.58[/TD2][TD2]47.21[/TD2][TD2]48.31[/TD2][TD2]57.04[/TD2] [TD2] 33.1% [/TD2][TD2] 27.1% [/TD2][TD2] 15.5% [/TD2][TD2] -0.4% [/TD2] [TD2]22.0%[/TD2][TD2]16.0%[/TD2][TD2]4.5%[/TD2][TD2]-23.5%[/TD2]

Is this based on 50,000 3’s produced in q3? And 60,000 3’s produced in q4? Is this also what your previous posted estimates are assuming?

 

[Fact Checking]

your inquiry did have me reviewing my work vs a few others, and i discovered that i had still kept some opex adjustments assuming cost reductions due to workforce cuts. that idea was blown out of the water with the last report.

Do you mean their guidance that SG&A is going to stay flat in Q3? This is what Deepak Ahuja said in the conference call:

"Deepak here. I mean, there are many factors. Clearly, the working capital benefit of the difference in the payable terms versus collecting cash is one of them. But also, it's our gross margin improvement on the business. With the – it's the higher volumes and the higher gross margins, I'm thinking higher gross profit, I'm stating the obvious here on Model 3. Our SNX volumes are increasing too in the second half. That's going to help us significantly. And all of our other businesses are improving their profitability."

"While our OpEx is staying essentially flat, so massive leverage in the business. So when you combine all of that, that's what is giving us the cash flow from operations to fund the rest of our business and grow cash. I'm stating the obvious, but just sort of summarizing the whole point. Yeah."​

I.e. they are guiding their Q2 SG&A of $750m to stay flat. (And that ignores the 'essentially' qualifier, which could still easily mean a couple of million dollars in SG&A increase ...)

In your last forecast posted on Aug 2 you used $750m for Q2 (which matches the Q2 report) and $707m for Q3 SG&A, so your numbers would have to be increased by +$43m, i.e. a crude refresh of your Aug 2 figures would be:

- net inc to common $156m
+ net inc to common $113m​

I.e. an EPS reduction from $1.5 to about $1.1, right?

BTW., one additional piece of feedback regarding your latest projections: AFAICS you are using +$100m for Q3 and Q4 ZEV credits.

That looks unreasonably pessimistic to me. Here's Tesla's track record with ZEV credits, what I've been able to decode from their past financial statements:

year   | deliveries | ZEV sold | ZEV per quarter | ZEV per car
--------------------------------------------------------------
2016   | 76,230     | $302.3m  | $75.5m          | $3,965
2017   | 81,824     | $360.3m  | $90.0m          | $4,403
2018H1 | 70,720     | $134.3m  | ......          | ......

Notes and assumptions:

• I didn't find such decomposition of their ZEV cost structure anywhere, I deducted it from a series of disclosures and back-calculated it all - it might not be correct. Can post the original disclosures and the rules if there's interest or doubt about the accuracy of these figures.
• There's no per unit estimate for 2018 yet, because we don't know how many credits they sold, we only know the value. If I am reading their disclosures correctly then Tesla is only 'flushing' their ZEV credits down to zero at the end of year, so the intra-year buffering of ZEV credits can only be known if Tesla discloses it, and they rarely do.
• I assumed similar count of credits for the Model S/X and the Model 3, they have comparable average range.
• Obviously not all deliveries go to ZEV states, but I think we can safely assume that Model 3 delivery ZEV mix is at least as good as the historic Model S/X mix. Model 3 is heavily present in California and Tesla had the option to first serve Model 3 customers from ZEV states. With the Model S/X they didn't have that option, and those also have significant international deliveries.

From the table we can see that:

• Per unit effective ZEV credit market price remained stable around $4.0k, in fact there was a small increasing trend - which surprised me, I expected a drop due to the higher unit count.
  
• In 2018 Tesla has sold only $134.3m worth of ZEV credits so far, and probably accumulated around $148m in ZEV credits already from Q1 and Q2, using the historic ~$4k/unit rate.
• With 50k-55k deliveries in Q3 Tesla would earn an incremental +$220m in Q3 under those assumptions just from the Model 3 - with 75k total units delivered that contribution could be up to $300m.
• Even if we conservatively assume a drop in the market price for ZEV credits from $4k to $3k, that's 3*112k units, i.e. +$336m total

So that's an approximate $336m-$448m in possible ZEV credits in Q3, assuming $3k-$4k/unit market price and 55k deliveries plus the credits left over from Q1/Q2.

Note that even if we ignore buffering of Q1+Q2 of ZEV credits, the increased unit count of the Model 3 in Q3 should increase the ZEV income by significantly more than $100m: 75k units would map to $225m-$300m depending on the market price. Even if the per unit ZEV market price drops in half to $2k/unit, that would still be $150m, not $100m.

I.e. 3x-4x of the value you seem to be using if we go by the historic trend and 1.5x if we get very pessimistic.

Am I missing anything important there? (I haven't double checked the numbers in this post so any trivial math mistakes you see are likely mine.)

(BTW., a really stupid question, you are posting such nice tables, but I'm unable to find the markup syntax for that, and the editor doesn't seem to support it. What kind of syntax are you using to post tables? I used the CODE markup as a poor substitute.)

 

[mrdoubleb]

Do you mean their guidance that SG&A is going to stay flat in Q3? This is what Deepak Ahuja said in the conference call:

"Deepak here. I mean, there are many factors. Clearly, the working capital benefit of the difference in the payable terms versus collecting cash is one of them. But also, it's our gross margin improvement on the business. With the – it's the higher volumes and the higher gross margins, I'm thinking higher gross profit, I'm stating the obvious here on Model 3. Our SNX volumes are increasing too in the second half. That's going to help us significantly. And all of our other businesses are improving their profitability."

"While our OpEx is staying essentially flat, so massive leverage in the business. So when you combine all of that, that's what is giving us the cash flow from operations to fund the rest of our business and grow cash. I'm stating the obvious, but just sort of summarizing the whole point. Yeah."​

I.e. they are guiding their Q2 SG&A of $750m to stay flat. (And that ignores the 'essentially' qualifier, which could still easily mean a couple of million dollars in SG&A increase ...)

In your last forecast posted on Aug 2 you used $750m for Q2 (which matches the Q2 report) and $707m for Q3 SG&A, so your numbers would have to be increased by +$43m, i.e. a crude refresh of your Aug 2 figures would be:

- net inc to common $156m
+ net inc to common $113m​

I.e. an EPS reduction from $1.5 to about $1.1, right?

BTW., one additional piece of feedback regarding your latest projections: AFAICS you are using +$100m for Q3 and Q4 ZEV credits.

That looks unreasonably pessimistic to me. Here's Tesla's track record with ZEV credits, what I've been able to decode from their past financial statements:

year   | deliveries | ZEV sold | ZEV per quarter | ZEV per car
--------------------------------------------------------------
2016   | 76,230     | $302.3m  | $75.5m          | $3,965
2017   | 81,824     | $360.3m  | $90.0m          | $4,403
2018H1 | 70,720     | $134.3m  | ......          | ......

Notes and assumptions:

• I didn't find such decomposition of their ZEV cost structure anywhere, I deducted it from a series of disclosures and back-calculated it all - it might not be correct. Can post the original disclosures and the rules if there's interest or doubt about the accuracy of these figures.
• There's no per unit estimate for 2018 yet, because we don't know how many credits they sold, we only know the value. If I am reading their disclosures correctly then Tesla is only 'flushing' their ZEV credits down to zero at the end of year, so the intra-year buffering of ZEV credits can only be known if Tesla discloses it, and they rarely do.
• I assumed similar count of credits for the Model S/X and the Model 3, they have comparable average range.
• Obviously not all deliveries go to ZEV states, but I think we can safely assume that Model 3 delivery ZEV mix is at least as good as the historic Model S/X mix. Model 3 is heavily present in California and Tesla had the option to first serve Model 3 customers from ZEV states. With the Model S/X they didn't have that option, and those also have significant international deliveries.

From the table we can see that:

• Per unit effective ZEV credit market price remained stable around $4.0k, in fact there was a small increasing trend - which surprised me, I expected a drop due to the higher unit count.
  
• In 2018 Tesla has sold only $134.3m worth of ZEV credits so far, and probably accumulated around $148m in ZEV credits already from Q1 and Q2, using the historic ~$4k/unit rate.
• With 50k-55k deliveries in Q3 Tesla would earn an incremental +$220m in Q3 under those assumptions just from the Model 3 - with 75k total units delivered that contribution could be up to $300m.
• Even if we conservatively assume a drop in the market price for ZEV credits from $4k to $3k, that's 3*112k units, i.e. +$336m total

So that's an approximate $336m-$448m in possible ZEV credits in Q3, assuming $3k-$4k/unit market price and 55k deliveries plus the credits left over from Q1/Q2.

Note that even if we ignore buffering of Q1+Q2 of ZEV credits, the increased unit count of the Model 3 in Q3 should increase the ZEV income by significantly more than $100m: 75k units would map to $225m-$300m depending on the market price. Even if the per unit ZEV market price drops in half to $2k/unit, that would still be $150m, not $100m.

I.e. 3x-4x of the value you seem to be using if we go by the historic trend and 1.5x if we get very pessimistic.

Am I missing anything important there? (I haven't double checked the numbers in this post so any trivial math mistakes you see are likely mine.)

(BTW., a really stupid question, you are posting such nice tables, but I'm unable to find the markup syntax for that, and the editor doesn't seem to support it. What kind of syntax are you using to post tables? I used the CODE markup as a poor substitute.)

So talking about ZEV credits... I am not sure how certain we can be in any numbers around those. Yes, they are eligible for lots and lots of credits due to producing 250-300k pure EVs this year. The problem is, as Elon explained multiple times, it`s not that easy to sell them. Many competitors rather produce compliance cars then to have to pay Tesla and even when they find buyers they only get about half the actual value. (I seem to recall Elon saying they get .50c in the dollar).

Not saying they won`t sell any ZEV this year and we can`t count with some income, just wondering how much we can "plan" with that. I`ve always considered it the cherry on the cake.

 

[Fact Checking]

Many competitors rather produce compliance cars then to have to pay Tesla and even when they find buyers they only get about half the actual value. (I seem to recall Elon saying they get .50c in the dollar).

Yeah, so this is why I tried to establish historic patterns for the past couple of years, and Tesla seems to have earned more credits per unit, despite:

• growing sales
• growing international sales

I.e. the assumptions I made in my estimates are pretty conservative.

While it's true that there's probably a market saturation effect at some point, ZEV credits scale with unit count both on the producer (EV) and the consumer (ICE) side, and I think it's still safe to say that the combined Tesla unit count is way below the unit count of ICE cars in the relevant states, even if we factor compliance cars and asymmetric sales tactics of ICE carmakers into the equation.

I.e. Tesla's ZEV income should scale up with their ability to generate ZEV credits this year, i.e. should scale up with delivered units count.

Also note another factor: Tesla's ability to generate significant amount of ZEV credits actually increases their market value (as long as they don't saturate demand for ZEV credits): them selling to one ICE manufacturer over others gives that ICE manufacturer a competitive advantage.

As long as Tesla's ZEV credits were peanuts they could even artificially lower the price to hurt Tesla, but the hundreds of thousands of ZEV credits expected this year amount to some real money that ICE carmakers have to pay otherwise: over a billion dollars.

If Tesla makes ZEV business with just one carmaker then that can be around ~500m of differential income for that carmaker assuming a 50% rate - something they will not ignore in today's automotive environment, especially if some other ICE carmaker could gain that advantage.

This might explain why the per ZEV credit effective sales price appears to have increased from 2016 to 2017.

 

[mrdoubleb]

Yeah, so this is why I tried to establish historic patterns for the past couple of years, and Tesla seems to have earned more credits per unit, despite:

• growing sales
• growing international sales

I.e. the assumptions I made in my estimates are pretty conservative.

While it's true that there's probably a market saturation effect at some point, ZEV credits scale with unit count both on the producer (EV) and the consumer (ICE) side, and I think it's still safe to say that the combined Tesla unit count is way below the unit count of ICE cars in the relevant states, even if we factor compliance cars and asymmetric sales tactics of ICE carmakers into the equation.

I.e. their ZEV income should scale up with their ability to generate ZEV credits, i.e. with unit count.

Also note another factor: Tesla's ability to generate significant amount of ZEV credits actually increases their market value (as long as they don't saturate demand for ZEV credits): them selling to one ICE manufacturer over others gives that ICE manufacturer a competitive advantage.

As long as Tesla's ZEV credits were peanuts they could even artificially lower the price to hurt Tesla, but the hundreds of thousands of ZEV credits expected this year amount to some real money that ICE carmakers have to pay otherwise: over a billion dollars.

If Tesla makes ZEV business with just one carmaker then that can be around ~500m of differential income for that carmaker assuming a 50% rate - something they will not ignore in today's automotive environment, especially if some other ICE carmaker could gain that advantage.

This might explain why the per ZEV credit effective sales price appears to have increased from 2016 to 2017.

Very interesting... thanks for doing the research. Historically, as i recall, there were some quarters when they didn`t sell any ZEV as they didn`t like the price or demand was low. But I imagine they would want to pull out all the stops this quarter (and next) to achieve profitability, so they will not leave anything on the table. Shorts, of course, will claim they are only profitable due to ZEV, so it would be nice if they were in the black even without this, but whatever. In the words of our great poet Taylor Swift, "haters gonna hate".