Near-future quarterly financial projections

[luvb2b]

it was the first week of june. i think what he's doing is making some assumptions based on observations of the lot and or cars he counts during a specific time. if delivery flow improves and the lot is less full, he may observe times that are less busy and draw whatever conclusions he wants. get me to the guy who has the camera set up in the hotel across the street and the analyst who's counting the cars, that's who i would believe.

i see a lot of discussion around the 27k i put out for s/x. my estimate at this moment is simply based on guidance. i assume they are roughly at a natural rate around 25k s/x per quarter. the first quarter they were inventory constrained. the 2nd quarter they intentionally held back deliveries to the usa. now those deliveries will come through and we come back up 2k or so over 25k vs. 2k below in q2. also guidance is to reach for 100k deliveries, which would mean approx 56k delivered in the back 2h18. i modeled 55k. literally zero rocket science.

i can get a little better once i have 2 months of delivery figures from the usa and europe. but for now it's literally an educated guess based on guidance.

Do you remember the approximate date of your visit, so that we can cross-check it against skabooshka's daily production figure for that day?

I came to the conclusion that skabooshka is lying about the interpretation of the numbers pretty much all the time, including manipulation of the daily leak which he is rounding down and his withholding of the data strategically at the end of Q2, but I found no proof so far that he was fudging the absolute values. I think he used it as a 'carrot' to attract longs and to win credibility.

IIRC someone also summed up the skabooshka leak for Q2 and it came very close to the official figures? It's statistically be very hard to do a fake real-time leak of production data that would match up to the official figures in the end to such a degree.

Also note that skabooshka's numbers up to Aug 7 support the 27,000 Q3 S+X production estimate you are using, and they are also broadly consistent with the production data leaks to Fred Lambert which he claims came from a 'very reliable' source.

 

[Fact Checking]

you simply can't expand sales at this pace without having some inventory buildup.

Well, I'm basing it on measured Model 3 delivery times, 14 days on average from VIN assignment to delivery. (Model S/X inventory should be mostly unchanged.)

This is very short compared to S+X, and would mean that about 3 weeks of Model 3 production are in inventory.

Assuming a ~4k weekly run rate at the end of Q3 that would mean about 12k Model 3's in inventory - close to the 11.4k vehicles in transit at the end of Q2.

Which is why I suggest that 3.3b in inventory at the end of Q2 is broadly representative of the end of Q3 inventory, despite the 60% higher revenue, which is generated by completed sales in July and August - which Q2 didn't have. (April and May had comparably low sales.)

 

[Fact Checking]

i think what he's doing is making some assumptions based on observations of the lot and or cars he counts during a specific time. if delivery flow improves and the lot is less full, he may observe times that are less busy and draw whatever conclusions he wants.

Ok, that looks plausible too, I'll stop using his data. (I didn't use it for any of my Q3 inventory arguments.)

 

[Fact Checking]

this is common throughout the industry and is called talking one's book. longs are as guilty of it as shorts, and many are just as bad as some of the bad shorts, spreading various false rumors to be able to sell at higher prices etc.

There's a fundamental assymetry though: large institutional investors are mostly public, while shorts can stay hidden indefinitely in the U.S., there's no public reporting obligation for many entities.

This means that a big long talking their book is much more obvious (and can be discounted by investors) than those with short exposure talking their book.

 

[luvb2b]

let me address the article you quoted and just say flat out the research in the chinese article is garbage.

the article you cited includes a link to a published study in a reviewed journal which covers a much broader population of stocks in a many more countries and reaches the opposite conclusion. i know you said you reviewed the other studies. i won't take the time to review this one at 58 pages, but it seems the work they have done is much more detailed than the chiense authors.
http://compus.uom.gr/FIN145/document/Short_Selling/Price_Efficiency_and_Short_Selling.pdf

the statistical methods used by the chinese authors are dubious at best. in particular, the concept of comparing relative rank statistics of volatility using t-distributions doesn't have sound footing to me. i couldn't find a single reference about this method, the authors provided no references or technical support, and my guess is they just wagged and decided to use a t distribution with 2 degrees of freedom.

variances are notoriously hard to estimate and comparison of variances is subject to wide error. the usual accepted tests for comparison of variances involve chi-square distributions and the tolerance bands on those are very wide. using t distributions inappropriately will lead to garbage results as far as tests of statistical significance.

yet another problem (which is cleverly ignored) is that half the sample has its addition event on january 31, 2013. and if you're familiar with china, you know that china basically shuts down for the spring festival (chinese new year). dates vary, but in 2013 the spring festival happened the second week of february. so basically in half your sample you're trying to compare something during more normal trading vs. a period when the country is more or less shut down for 2 weeks including stock trading fully halted for 1 week. good luck expecting anything tremendously useful.

oh and how about the absolute value comparison? you can see the tremendous and highly significant drop in trading volume which is a function of the chinese new year holiday. and the actual volatility is flat - meaning no measured increase in volatility. they had to do some funny comparisons across holiday trading sessions to come up with something they can claim is useful.

it's just total garbage, and probably the reason this article was not published in any real finance journals.

if 99% of the utility of owning tesla is a limited supply of shares, you're in the wrong stock. the value of a stock like as large as tesla is primarily determined by sentiment and fundamental prospects. if you want to own stocks as collectibles where limited supply drives value, then you should focus on low float stocks where people are attempting corners.

a profitable short did not take returns from a long, he facilitated price discovery by making the market more efficient. he provided supply when there otherwise wasn't any, and he provided demand when there otherwise may not be any. in his absence the price would have gone even higher, and the following drop even lower.

i suppose since you don't like shorts you probably don't like options either. and if you really want a guilty party for the volatility in tesla, that's one place to look. there is more than 1 option outstanding for every tesla share, and the resulting hedge flow as the stock moves up or down exaggerates volatility more than any other factor.

let's also blend in some personal experience, i remember the day short selling was banned in banks in the united states. bank prices shot higher the next day. and that was the single best day to sell banks i have ever seen... the market was totally broken without the shorts and anyone who was dumb enough to buy in that rigged market got their heads blown off. you may prefer rigged markets like china where shorting is either banned or shorts are carted off to jail for being right, but please give me my free market here where shorts can trade and invest directly alongside me.

there are so many problems around your views of short selling that i insist that i must put an end to the discussion. i hope you understand.

Voting rights and dividends have little value for a closely controlled growth company like Tesla, so I think we can leave them out of the equation for now. I suspect Elon is only going to approve dividend payments the day the 100th Gigafactory is built, on Antarctica. Stock buy-back might happen much earlier: in 2-3 years Tesla might be generating a lot of cash they might not be able to spend immediately.

Anyway, the point is that shorting creates a new long position, which is 99% of the current utility of a share of TSLA.

I.e. the dilution is real: there's ~32 million more long shares, who can be panicked into selling and adding additional pressure, who will dilute existing shareholders and who add volatility.

The 'added liquidity' is only true as long as the short is wrong and is losing money. A profitable short is basically a middleman who took returns from a (weaker) long.

Understandably there's very little research into the harms of shorting in the western hemisphere, other than self-serving studies financed by the industry that benefits most from shorting.

But there was a recent natural experiment though, when China introduced short selling on a per equity basis which allowed to statistically control for the before and after effects of shorting, outlined in this paper:

Short selling and intraday volatility: evidence from the Chinese market​

"By focusing on the 30 trading days around the addition events, the results document statistically significant post-event increase in volatility relative to the overall market and absolute value of trading volume. Specifically, small-cap stocks experience the sharpest increase."​

Also note that the oft touted benefit of short selling deflating bubbles doesn't seem to be visible in the data:

"Our key finding is that short selling regulations facilitate the increase of intraday volatility relative to the overall market performance even the absolute value is almost invariant."​

I.e. according to their analysis the natural balance of bulls buying and bears selling was enough for efficient price discovery - shorting was not required.

They also succinctly point out:

"Overall, these results show that short sellers in China destabilize the market, which is in accord with the Hong Kong evidence (Chang et al. 2007) and inconsistent with the international evidence (Saffi and Sigurdsson 2011)."​

Not a surprise: I have checked those other studies - they are using mostly flawed methodology and in any case they don't observe natural data of healthy companies with/without short-selling, they mostly observe broken, distressed companies where short selling bans are implemented.

Anyway, while I agree that some level of shorting is probably beneficial to weed out the chaff, and if a quant algorithm thinks a company is overvalued I see no reason they shouldn't be able to short it, but the human trader driven value shorting that is happening in U.S. markets and which is perversely high for Tesla is dominantly parasitic: short hedge funds ganging/herding on companies and magnifying volatility to profit from it.

It's compounded by the ridiculously lax U.S. regulations that do not require large short positions to be disclosed while large longs are disclosed. This obscurity of true intent is used to hide lines of influence and conflicts of interest: in particular conflicts of interests between Wall Street analysts and the short exposure of their firms and clients.

And yes, shorts are a large source of Tesla FUD in practice, you can tell it by the pattern of the arguments - even though Tesla and Elon makes it arguably easier for them by committing unforced errors.

(Anyway, I suspect this is off-topic for your thread and I'm not going to press this argument.)

 

[luvb2b]

but inventory is not just vehicles in transit.
it also includes parts and materials needed to build the cars.
and you need a lot more of those when you're running 4k weekly vs 2k.

Well, I'm basing it on measured Model 3 delivery times, 14 days on average from VIN assignment to delivery. (Model S/X inventory should be mostly unchanged.)

This is very short compared to S+X, and would mean that about 3 weeks of Model 3 production are in inventory.

Assuming a ~4k weekly run rate at the end of Q3 that would mean about 12k Model 3's in inventory - close to the 11.4k vehicles in transit at the end of Q2.

Which is why I suggest that 3.3b in inventory at the end of Q2 is broadly representative of the end of Q3 inventory, despite the 60% higher revenue, which is generated by completed sales in July and August - which Q2 didn't have. (April and May had comparably low sales.)

 

[luvb2b]

although i disagree with using finished goods inventory as a proxy for total inventory, i believe you are correct that there are upside possibilities to cash flow from operations.

the way you've reached this conclusion doesn't appear correct on a quick look. accts payable declining would be a net negative vs my estimate on cash from ops. lower inventory a net positive, and the two effects roughly cancel the way you've presented them.

the accrued liabilities seem to track opex more than anything else, and your entire change in cash vs my estimate is being driven by (i guess) the change in accrued liabilities. but there's no reason for that to happen that i see.

I fully agree there! There's very little shorts will be able to do once Tesla starts executing - and I think they are executing pretty well by now.

BTW., sorry about derailing the thread which was mostly about financials - I've attempted to modify your Q3 projections with Tesla's guided stable (non-burst) end of Q3 inventory:

…

…

balance sheet

current assets

…

…

inventory

…

inventory

…

…

current liabilities

…

…

accts payable

…

accts payable

…

…

total current liabs

…

total current liabs

…

cash flow statement

cash flows from ops

…

accts pybl/accr liabs

…

accts pybl/accr liabs

…

…

net change in cash

…

net change in cash

…

cash & eq start

…

cash & eq end

…

cash & eq end

…

[TD2] luv q3-18e [/TD2][TD2] Jun-18 [/TD2][TD2] Mar-18 [/TD2] [TD2] 4,626,666 [/TD2][TD2] 3,324,643 [/TD2][TD2] 2,565,826 [/TD2] [TD2] -1,302,023 [/TD2][TD2] … [/TD2][TD2] … [/TD2] [TD2] 3,324,643 [/TD2][TD2] 3,324,643 [/TD2][TD2] 2,565,826 [/TD2] [TD2] 4,684,499 [/TD2][TD2] 3,030,493 [/TD2][TD2] 2,603,498 [/TD2] [TD2] -1,302,023 [/TD2][TD2] … [/TD2][TD2] … [/TD2] [TD2] 3,382,476 [/TD2][TD2] 3,030,493 [/TD2][TD2] 2,603,498 [/TD2] [TD2] 10,641,488 [/TD2][TD2] 9,141,362 [/TD2][TD2] 8,650,359 [/TD2] [TD2] -1,302,023 [/TD2][TD2] … [/TD2][TD2] … [/TD2] [TD2] 9,339,465 [/TD2][TD2] 9,141,362 [/TD2][TD2] 8,650,359 [/TD2] [TD2] 2,041,277 [/TD2][TD2] 591,737 [/TD2][TD2] 317,983 [/TD2] [TD2] -651,011 [/TD2][TD2] … [/TD2][TD2] … [/TD2] [TD2] 1,390,266 [/TD2][TD2] 591,737 [/TD2][TD2] 317,983 [/TD2] [TD2] 696,762 [/TD2][TD2] -436,470 [/TD2][TD2] -745,251 [/TD2] [TD2] +651,011 [/TD2][TD2] … [/TD2][TD2] … [/TD2] [TD2] 1,347,773 [/TD2][TD2] -436,470 [/TD2][TD2] -745,251 [/TD2] [TD2] 2,236,424 [/TD2][TD2] 2,665,673 [/TD2][TD2] 3,367,914 [/TD2] [TD2] 2,933,186 [/TD2][TD2] 2,236,424 [/TD2][TD2] 2,665,673 [/TD2] [TD2] +651,011 [/TD2][TD2] … [/TD2][TD2] … [/TD2] [TD2] 3,584,197 [/TD2][TD2] 2,236,424 [/TD2][TD2] 2,665,673 [/TD2]

I used the following crude assumptions to save time and effort:

• I hope the notation of the changes is obvious enough: I created an extra line for amounts I subtracted or added
  
• U.S. Model 3 delivery times average around 14 days, according to the Model 3 tracker
• Model S+X delivery times average more around 30 days, due to international deliveries
• Q3 Model 3 exit rate would be two 5k weeks: this creates roughly as many Model 3's "in transit" as the end of Q2 burst production created.
• This allowed me to simply use the Q2 inventory for end of Q3 as well.
• I assumed that about 50% of the inventory cost had a cash flow component: labor and immediate cash expenses. The remaining 50% is on a 60 days delayed payment schedule.
• Effects on net income should be near zero, unless I'm missing something.

Note that this is both sloppy and imprecise for various reasons - but I wanted to establish a basic direction for changes in the financials - did I get it right that cash flow and cash on hand would be improved significantly if Tesla exited Q3 with a similar inventory to Q2?

Or am I missing something important here? A sign error, or getting it wrong by an order of magnitude?

 

[Fact Checking]

but inventory is not just vehicles in transit.
it also includes parts and materials needed to build the cars.
and you need a lot more of those when you're running 4k weekly vs 2k.

So what I am looking at are the proportions of inventory components, about 80% of Tesla's Q2 increase in inventory was in finished goods - i.e. most likely vehicles in transit:

inventory

Raw materials

Work in process

Finished goods

Service parts

Total

[TD2] March 31, 2018 [/TD2][TD2] June 30, 2018 [/TD2][TD2] +$change [/TD2] [TD2] $902,190 [/TD2][TD2] $972,739 [/TD2][TD2] +$70,549 [/TD2] [TD2] $315,227 [/TD2][TD2] $350,443 [/TD2][TD2] +$35,216 [/TD2] [TD2] $1,125,665 [/TD2][TD2] $1,721,860 [/TD2][TD2] +$596,195 [/TD2] [TD2] $222,744 [/TD2][TD2] $279,601 [/TD2][TD2] +$56,857 [/TD2] [TD2] $2,565,826 [/TD2][TD2] $3,324,643 [/TD2][TD2] +$758,817 [/TD2]

Note that both March 31 and June 30 had very similar Model S+X in transit numbers: 4,060 vs. 3,892. This suggests that the changes in inventory levels are from Model 3 related inventory increases mostly.

BTW., a side note, this allows us to estimate the Model 3 cash CoG: if the +$596m increase in finished goods was due to the 11,166 Model 3's in transit, then the per unit average inventory value is $53,393. Given that the Model 3 ASP was around $57k at this point, am I correct to interpret this that the Model 3 was already close to break-even at the end of Q2, on a cash basis?

In terms of parts/materials inventory levels it's also an open question whether Tesla 'loaded' the pipeline for the 5k burst only and mostly emptied out their Model 3 inventory at the end of Q2? If yes then the Q2 inventory levels aren't really representative of the run-rate.

Tesla could also be doing the same for end of Q3 and empty out the Model 3 pipeline of partial inventory - if they can time the supply chain precisely enough.

There's also another approach, we can look at the Model S+X inventory levels, which dominated inventory in Q3 2017:

inventory

Raw materials

Work in process

Finished goods

Service parts

Total

[TD2] September 30, 2017 [/TD2] [TD2] $612,225 [/TD2] [TD2] $277,155 [/TD2] [TD2] $1,418,385 [/TD2] [TD2] $163,617 [/TD2] [TD2] $2,471,382 [/TD2]

In that case finished goods were about 57% of inventory, and there were 4,820 Model S+X vehicles in transit. If we take an average ASP of 103k and account them at 75% of ASP, that's about $372m of the finished goods. If that estimate is accurate, what does the other $1,046m of finished goods represent?

With about 100 Tesla stores and perhaps 3 showroom cars per store we've got another 300 cars accounted for - but that's only ~$40m of inventory. That's another 10,000 Model S/X's worth of inventory value, roughly, which is a lot of cars and cannot be explained by CPO inventory, right?

Do leased vehicles count as inventory perhaps?

I must be missing something basic ...

 

[luvb2b]

possibly some of the loaners are inventory. you'd have some solar and storage stuff in there too.

cash forecasting is very hard. it's hard enough to get an income statement close. trying to get a balance sheet or cash flow statement close is much harder, esp in a company with so many moving parts. and trying to get a model that correctly combines all of these into one cohesive unit is also extremely difficult. someone asked about my balance sheet numbers implying a profit... those balance sheet numbers are not even 100% properly tied to the income statement. so take them as i do: they are only a guide from which you can start your own thoughts.

i wouldn't get too hung-up on any particular detail. i could be very easily wrong on my inventory estimate - i can only say that it is derived using a days of inventory outstanding methodology, and by that method my estimate is reasonable. the method could be inappropriate, tesla could have inventory they can't sell, or your assumptions could end up being correct. i suggest it's best to take a structure you're comfortable with and work with that.

what we know for sure is easier to address. tesla had a solid inventory build in q2 and despite that suprised me with cash generatio. that cash seems to come primarily from margin on cars sold. if the inventory build wasn't so large cash generation could be much much better. that's a good development and i hope it continues.

So what I am looking at are the proportions of inventory components, about 80% of Tesla's Q2 increase in inventory was in finished goods - i.e. most likely vehicles in transit:

inventory

Raw materials

Work in process

Finished goods

Service parts

Total

[TD2] March 31, 2018 [/TD2][TD2] June 30, 2018 [/TD2][TD2] +$change [/TD2] [TD2] $902,190 [/TD2][TD2] $972,739 [/TD2][TD2] +$70,549 [/TD2] [TD2] $315,227 [/TD2][TD2] $350,443 [/TD2][TD2] +$35,216 [/TD2] [TD2] $1,125,665 [/TD2][TD2] $1,721,860 [/TD2][TD2] +$596,195 [/TD2] [TD2] $222,744 [/TD2][TD2] $279,601 [/TD2][TD2] +$56,857 [/TD2] [TD2] $2,565,826 [/TD2][TD2] $3,324,643 [/TD2][TD2] +$758,817 [/TD2]

Note that both March 31 and June 30 had very similar Model S+X in transit numbers: 4,060 vs. 3,892. This suggests that the changes in inventory levels are from Model 3 related inventory increases mostly.

BTW., a side note, this allows us to estimate the Model 3 cash CoG: if the +$596m increase in finished goods was due to the 11,166 Model 3's in transit, then the per unit average inventory value is $53,393. Given that the Model 3 ASP was around $57k at this point, am I correct to interpret this that the Model 3 was already close to break-even at the end of Q2, on a cash basis?

In terms of parts/materials inventory levels it's also an open question whether Tesla 'loaded' the pipeline for the 5k burst only and mostly emptied out their Model 3 inventory at the end of Q2? If yes then the Q2 inventory levels aren't really representative of the run-rate.

Tesla could also be doing the same for end of Q3 and empty out the Model 3 pipeline of partial inventory - if they can time the supply chain precisely enough.

There's also another approach, we can look at the Model S+X inventory levels, which dominated inventory in Q3 2017:

inventory

Raw materials

Work in process

Finished goods

Service parts

Total

[TD2] September 30, 2017 [/TD2] [TD2] $612,225 [/TD2] [TD2] $277,155 [/TD2] [TD2] $1,418,385 [/TD2] [TD2] $163,617 [/TD2] [TD2] $2,471,382 [/TD2]

In that case finished goods were about 57% of inventory, and there were 4,820 Model S+X vehicles in transit. If we take an average ASP of 103k and account them at 75% of ASP, that's about $372m of the finished goods. If that estimate is accurate, what does the other $1,046m of finished goods represent?

With about 100 Tesla stores and perhaps 3 showroom cars per store we've got another 300 cars accounted for - but that's only ~$40m of inventory. That's another 10,000 Model S/X's worth of inventory value, roughly, which is a lot of cars and cannot be explained by CPO inventory, right?

Do leased vehicles count as inventory perhaps?

I must be missing something basic ...

 

[luvb2b]

i think we work so hard on all these details, and really the question is much simpler.

is tesla making cars they can't sell and building inventory that is sitting on lots unsaleable?

or

is tesla making and delivering cars smoothly and well-matched to their order flow?

if it's the latter case, with most reasonable production scenarios we're going to be fine.

if it's the former case, please spare me the pain and shoot me now.

 

[GeorgeSymonds]

So what I am looking at are the proportions of inventory components, about 80% of Tesla's Q2 increase in inventory was in finished goods - i.e. most likely vehicles in transit:

inventory

Raw materials

Work in process

Finished goods

Service parts

Total

[TD2] March 31, 2018 [/TD2][TD2] June 30, 2018 [/TD2][TD2] +$change [/TD2] [TD2] $902,190 [/TD2][TD2] $972,739 [/TD2][TD2] +$70,549 [/TD2] [TD2] $315,227 [/TD2][TD2] $350,443 [/TD2][TD2] +$35,216 [/TD2] [TD2] $1,125,665 [/TD2][TD2] $1,721,860 [/TD2][TD2] +$596,195 [/TD2] [TD2] $222,744 [/TD2][TD2] $279,601 [/TD2][TD2] +$56,857 [/TD2] [TD2] $2,565,826 [/TD2][TD2] $3,324,643 [/TD2][TD2] +$758,817 [/TD2]

Note that both March 31 and June 30 had very similar Model S+X in transit numbers: 4,060 vs. 3,892. This suggests that the changes in inventory levels are from Model 3 related inventory increases mostly.

BTW., a side note, this allows us to estimate the Model 3 cash CoG: if the +$596m increase in finished goods was due to the 11,166 Model 3's in transit, then the per unit average inventory value is $53,393. Given that the Model 3 ASP was around $57k at this point, am I correct to interpret this that the Model 3 was already close to break-even at the end of Q2, on a cash basis?

In terms of parts/materials inventory levels it's also an open question whether Tesla 'loaded' the pipeline for the 5k burst only and mostly emptied out their Model 3 inventory at the end of Q2? If yes then the Q2 inventory levels aren't really representative of the run-rate.

Tesla could also be doing the same for end of Q3 and empty out the Model 3 pipeline of partial inventory - if they can time the supply chain precisely enough.

There's also another approach, we can look at the Model S+X inventory levels, which dominated inventory in Q3 2017:

inventory

Raw materials

Work in process

Finished goods

Service parts

Total

[TD2] September 30, 2017 [/TD2] [TD2] $612,225 [/TD2] [TD2] $277,155 [/TD2] [TD2] $1,418,385 [/TD2] [TD2] $163,617 [/TD2] [TD2] $2,471,382 [/TD2]

In that case finished goods were about 57% of inventory, and there were 4,820 Model S+X vehicles in transit. If we take an average ASP of 103k and account them at 75% of ASP, that's about $372m of the finished goods. If that estimate is accurate, what does the other $1,046m of finished goods represent?

With about 100 Tesla stores and perhaps 3 showroom cars per store we've got another 300 cars accounted for - but that's only ~$40m of inventory. That's another 10,000 Model S/X's worth of inventory value, roughly, which is a lot of cars and cannot be explained by CPO inventory, right?

Do leased vehicles count as inventory perhaps?

I must be missing something basic ...

There are a few CPO trackers listing quite considerable MS and MX new inventory in some countries, this one is showing 1400 in Holland alone

Tesla Holland CPO en nieuwe voorraad

That’s possibly a bad case, but Holland isn’t a big country, but maybe Europe is carrying a lot of stock at the moment,

 

[SPadival]

it was the first week of june. i think what he's doing is making some assumptions based on observations of the lot and or cars he counts during a specific time. if delivery flow improves and the lot is less full, he may observe times that are less busy and draw whatever conclusions he wants.

Also, skabooshka said he was rounding the numbers to protect his source. He was also pretending to get the numbers by counting the cars on the lot. Why does he have to round the numbers when his source is himself counting the cars!!

 

[Remus]

i think we work so hard on all these details, and really the question is much simpler.

is tesla making cars they can't sell and building inventory that is sitting on lots unsaleable?

or

is tesla making and delivering cars smoothly and well-matched to their order flow?

if it's the latter case, with most reasonable production scenarios we're going to be fine.

if it's the former case, please spare me the pain and shoot me now.

I guess we will still have to endure some pain. The reality seems to be somewhere in the middle, with many people anxiously waiting for their model 3s, it is the question of how many cars are delayed or even rejected due to logistic or quality issues.

From what I heard from friends picking up their cars, it is not as bad as the shorts want us to believe, yet we can't ignore the reported bad cases.

I have a bad feeling when the number comes out, it will annoy both the bulls and bears, it will be not as good as we hoped but not as bad as the bear hoped either.

 

[Matias]

i think we work so hard on all these details, and really the question is much simpler.

is tesla making cars they can't sell and building inventory that is sitting on lots unsaleable?

I really can't see, how this could be true, since rest of the world is still waiting for Model 3.

 

[GeorgeSymonds]

I really can't see, how this could be true, since rest of the world is still waiting for Model 3.

I think it's true for MS and MX, not for the M3. The days of everyone custom ordering MS and MX are long gone, there's too much evidence of cars being listed, and even if Hank is correct in saying the New Inventory sites are listing cars that are still being built or are in transit (and I'm not sure if thats him protecting his own interest or he's genuinely correct, I'll give him the benefit of the doubt), the fact some VINs are published suggests they're not attributed to a customer, otherwise all cars would appear at some point. Secondly, we read stories that owners are being approached to switch early, sometimes as early as 6 months before they're due. Active promotion in such a way, especially so early, suggests they have stock that needs to be sold.

 

[luvb2b]

I really can't see, how this could be true, since rest of the world is still waiting for Model 3.

it can be true in many ways. one way is that tesla anticipates "clearing the decks" as the tax credit rolls off in q4. so they build spec cars in advance on configs that might be desirable. the spec cars are sitting in inventory and waiting in various lots for a home. even if they don't find one soon, tesla anticipates that the last minute rush to get the $7500 credit will get those cars sold. however, that could mean some cars sit around all the way until december.

another way it can be true is that some of the cars have various quality issues. it may be too expensive to slow the line / stop the line to rework at that moment, so the car gets fully built and then pulled off and carted away to a model 3 storage area awaiting rework. with the number of cars that need rework they may not be able to keep up and so a buildup develops over at the rework facility or at service centers where cars are awaiting rework by local service.

there are probably other ways too.

 

[Fact Checking]

i think we work so hard on all these details, and really the question is much simpler.

is tesla making cars they can't sell and building inventory that is sitting on lots unsaleable?

or

is tesla making and delivering cars smoothly and well-matched to their order flow?

if it's the latter case, with most reasonable production scenarios we're going to be fine.

if it's the former case, please spare me the pain and shoot me now.

BTW., there was at least one fundamental mistake I made in my grandparent post estimate:

• Tesla doesn't have 100 stores and galleries, they meanwhile have 400 of them. Just 10-20 new inventory cars and CPOs per location, especially in far away countries where a delayed sale might be a lost sale, would explain much of the inventory.
• The Netherlands also has an assembly plant that must be buffering a couple hundred vehicles and related parts as well. It's unclear to me whether they are counting that as 'materials' or 'finished goods'.
  
• Then here's also the other types of inventory you mentioned such as solar and maybe loaners.
• Also, I reviewed their inventory related disclosures, and they seem to be culling inventory actively and aggressively: $120m of write-offs in 2017 alone, which is about 10% of inventory.
• They also disclosed record new orders for Model S+X so if they had any excess inventory they would probably have an easy way flushing it out with a bit of a price reduction - but they aren't doing it because there's record S/X demand.

So the "shooting you" scenario has very low probability I think.

 

[Fact Checking]

is tesla making and delivering cars smoothly and well-matched to their order flow?

if it's the latter case, with most reasonable production scenarios we're going to be fine.

Just one more data point: 2017 Q3 and Q4 inventory appears to offer the best insight on the S+X cost structure, I believe.

Here's the two inventory reports and production+delivery figures compared:

inventory

Raw materials

Work in process

Finished goods

Service parts

Total

S+X produced

S+X delivered

S+X "in transit"

[TD2] September 30, 2017 [/TD2][TD2] December 31, 2017 [/TD2] [TD2] $612,225 [/TD2][TD2] $821,396 [/TD2] [TD2] $277,155 [/TD2][TD2] $243,181 [/TD2] [TD2] $1,418,385 [/TD2][TD2] $1,013,909 [/TD2] [TD2] $163,617 [/TD2][TD2] $185,051 [/TD2] [TD2] $2,471,382 [/TD2][TD2] $2,263,537 [/TD2] [TD2] 25,076 [/TD2][TD2] 22,140 [/TD2] [TD2] 25,930 [/TD2][TD2] 28,320 [/TD2] [TD2] 4,820 [/TD2][TD2] 2,520 [/TD2]

(With 40 units Model 3 was not at significant levels yet.)

There was a reduction in finished goods inventory levels of $404,476k, and a reduction of 2,300 units in transit, plus deliveries were up by 2,390 units.

If we assume that the production slowdown was mostly due to an end of year holidays shutdown of the factory, and assume that S+X production was otherwise maxed out and in rough steady state, then the per unit inventory value is $404,476,000/(2,300+2,390), which is $86.2k of a per unit cost.

This lines up with Tesla's other disclosures very well: if S+X gross margin was at around 22% at that time, then the average sales price of the Q3=>Q4 inventory changes was $110.5k.

This is broadly consistent with the estimated ASP of S+X average sales price, which was somewhere between $105k-110k in that time frame.

BTW., this also serves as a critique of net working capital estimates: the current Q2 inventory of about $1.72b is in reality generating a cash flow of around $2.23b (if we assume a net gross margin slightly above 20% of S+3+X inventory combined) - while payable liabilities are accounted at 100% value.

I.e. just based on the finished goods inventory the true working capital position of Tesla was +$510m higher on the asset side.

This factor of leverage should be even higher for raw materials and parts, where the gross margin is supposedly even higher - while for service parts it's lower. The true cash flow value of the complete $3.3b of inventory I'd estimate to be around $4.3b - i.e. one billion dollars better working capital position.

If we add to that that on the current liabilities side the resale guarantee of $674m is not a 100% liability but more like a 30% liability, Tesla's working capital improves by another ~$471m.

I.e. the -2.44b working capital deficit got reduced to below -$0.97b in Q2 already.

With 2-3 times as many Model 3's sold in Q3 the true working capital position should flip to positive already, especially if end of Q3 inventory levels are going to be below the $4.6b you are projecting.

 

[Boomer19]

was changed to aaxn and you want to guess who's on the ir team over there today?

sorry, i must ask..this piqued my interest.
who is on their IR team and what’s the relevance?

edit: taser / axon IR i meant

 

[Boomer19]

no fud has been as bad as lack of execution and elon's twitter fiascos. in fact resolving both of those would put a big smackdown on the fudsters, and yet these are things we can't seem to get.

you can see it happening in front of your eyes. here we are discussing shorting and shorts and how have they damaged the company etc etc etc... and we ignore how much of this is tesla's own doing. this is why it concerns me when tesla focuses on shorts vs being laser focused on executing well. i could even make the case elon's twitter insanities have been driven by tesla's lack of execution: lack of execution => elon works more & sleeps less => making elon cranky while on twitter.

couldn’t agree more