I perused Tesla’s Q3 10-Q and scrutinized the footnotes to figure out, to the extent possible, where Tesla manipulated GAAP accounting standards and outright “cooked” its numbers. Before I had a chance to analyze the 10-Q, others had already posted their findings on Twitter or in Seeking Alpha articles. In the analysis below, I’ve double-checked and confirmed the findings presented by others. In addition, where appropriate, I’ve added my findings to the previous work of others and explained how and why Tesla’s numbers are highly misleading, if not outright fraudulent.
Net income – Tesla reported GAAP income of $311.5 million. But what it did not disclose when it released its earnings report was that $189 million of that income was generated from selling regulatory credits – Greenhouse Gas (GHG) credits and ZEV (Zero Emission Vehicle) credits. Automakers in 10 States are required to sell a specified number of electric or hybrid vehicles within the State. Credits are earned for the number of emission-friendly vehicles sold. Automakers are required to maintain a level of credits based on each automaker’s overall vehicle sales volume within the State. GHG credits function in a similar way at the Federal level.
Some companies, like Tesla, generate more GHG and ZEV credits than required to be in compliance with the law. Companies with excess credits are allowed to sell their excess credits to car manufactures and other companies that manufacture carbon-emission equipment and do not generate enough credits to be in compliance with the regulation. Selling excess credits over the past few years has been a significant source of cash flow generation for Tesla. The money raised by selling these credits is accounted for as income under GAAP.
The problem is that, in its presentation of its Q3 earnings, Elon Musk and the CFO did not disclose that nearly 61% of its GAAP net income was derived from selling these credits. While Tesla referenced that $52 million was generated from ZEV credit sales in Q3, they did not disclose the $137 million GHG credit sales in the earnings press release or the analyst conference call. Rather, they postured as if the net income was generated thru cost-efficiencies and sales volume. The $137 million in GHG credit sales was buried in the 10-Q.
In the chart above, you can see that TSLA’s use of GHG credit sales has been inconsistent over time. In all probability, Musk chooses the timing and quantity of the credit sales based on when he needs to generate cash. It’s pretty obvious that he decided to unload a massive quantity in Q3 in order to help generate the GAAP net income and positive cash flow he had been promising for months.
Technically, the manner in which Musk utilized,and disclosed the use of, ZEV/GHG credits to manufacture income, is highly deceptive. Selling regulatory-derived environmental credits is a low-quality, unreliable source of income. As Tesla’s competition ramps up production and sales of EV’s, the supply of credits will escalate rapidly. This will drive down the resale value of these credits toward zero. And there’s always the possibility that regulatory requirements will be rolled back. Over time, this source of income and cash will disappear.
Warranty Provision – Every quarter companies that issue warranties have to take a warranty expense provision, which is an estimate of the quarterly expense that will be incurred under warranties on products sold by the company. The warranty provision hits the income statement as an expense. The idea is to match estimated quarterly warranty costs that will be incurred from selling products covered by the warranty each quarter. Warranty expense is part of the cost of goods sold. The information on warranty expenses is found in the footnotes (this is standard).
In Q3 this year, Tesla expensed $187.8 million, or $2,249 per car delivered, vs $118.6 million, or $2,913 per car delivered in Q3 2017. If Tesla had kept the cost per vehicle delivered constant, the provision for warranty expense in Q3 would have been $243.2 million, or $55.4 million higher than was expensed in Q3. In this case, Tesla’s cost of goods sold would have been $55.4 million higher and the gross profit would have been $55.4 million lower. This is part of the reason Tesla’s gross profit margin was much higher than anyone expected. It also translates into a $55.4 million net income benefit.
In Q3 2108, Tesla sold a little more than double the number of vehicles sold in Q3 2017. At the very least, and to be prudent, in Q3 this year Tesla should have at used at least double the warranty provision it used in Q3 2017. This is especially true since the Model 3 is in its debut model year and will likely require higher than expected warranty-based repairs. The probability of greater than expected warranty repairs for cars sold during Q3 is even higher when taking into account the high number of production difficulties the Company encountered – and about which Musk whined publicly.
Using a warranty expense estimation method simply based on doubling the warranty provision taken in Q3 2017 – given that Tesla sold more than double number vehicles, Tesla’s warranty provision expense would have been $237.2 million in Q3 rather than the $187.8 million recorded, which would have reduced net income by $49.4 million.
To be sure, the warranty expense provision can be adjusted based on using the actual amount of warranty costs incurred over time. But given the limited history of Tesla, and given that the Model 3 is a 1st-year production automobile with noted production and quality control issues, Tesla probably should have used a warranty provision that was higher on a per car delivered basis than the number used in Q3 2017. But, then again, Musk and his CFO were goal-seeking positive net income and thus likely decided to reduce the provision per vehicle delivered by nearly 23% and pray that they figure out a way to bury an increase in the actual amount spent on warranty repairs in future quarters.
Inventory Write-Down – An inventory write-down is recorded as an expense in the quarter in which it is taken. For a company like TSLA, an inventory write-down occurs for excess or obsolete inventories (unsalable cars, worthless parts and supplies) or when the carrying value of certain cars held in inventory is greater than the realizable value. The latter would primarily apply to cars taken back by Tesla under lease guarantees (keep this tidbit in mind for reference below) or cars held in inventory deemed unsalable because the cost of fixing manufacturing defects is greater than the gross margin generated from selling the car.
Over the last six quarters, Tesla’s inventory write-down as a percentage of total inventory has averaged 1.4%. In Q3 2017, the write-down was 1.1% of inventory; in Q2 2018 it was 0.9%). However, in Q3 Tesla’s inventory write-down was 0.4% of inventory. In terms of numbers, Tesla’s inventory expense in Q3 was $12.4 million vs $26.2 million in Q3 2017 and $24.6 million in Q2 2018. This chart shows the degree to which it appears as if Tesla purposely minimized the inventory write-down expense in Q3 2018:
(Kudos to @TeslaCharts for the charts he created illustrating the extreme inconsistencies in Tesla’s Q3 financial statements)
The effect of taking an inventory write-down that is far lower than the historical average reduces the cost of sales and thereby increases the gross, operating and net profits. If TSLA had used the historical average of 1.4%, the expense taken for the Q3 inventory write-down would have been $46.2 million, or $33.8 million more than the $12.4 million used. The reduced write-down had the effect of reducing cost of sales by $33.8 million and increasing gross profit and net income by $33.8. This also contributed to the large increase in the gross profit margin in Q3 vs historical quarters.
The inventory write-down charge was clearly an extreme outlier in relation to the historical application of this write-down over the previous six quarters. Make no mistake, the minimization of the inventory write-down expense in Q3 was a blatant effort to exploit accounting standards for the purpose of reducing GAAP expenses and thereby increasing GAAP income. The discrepancy between the Q3 charge vs historicals predictably was not addressed by the CFO or by analysts in the Q3 earnings conference call.
Tesla’s Actual Net Income? Telsa reported $311 million of GAAP net income. Of this, $83.2 million represents the highly questionable reduction in costs attributable to lower than usual warranty and inventory write-down expenses. Tesla also sold an unusually high amount of GHG/ZEV credits, which boosted net income by $189 million. While this is a source of actual cash income, it’s not a long-term sustainable source of income. Combined, these items accounted for $272 million – or 87.5% – Tesla’s GAAP net income in Q3.
In addition to the items presented above, Tesla “achieved” significant and highly questionable reductions in the expenses taken for R&D and SG&A. In Q3 Tesla recorded $350 million for R&D and $729 million for SG&A – $1.079 billion combined. In Q2 Tesla recorded $386 million for R&D and $750 million for SG&A – $1.36 billion combined. Tesla wants the market to believe that R&D and SG&A expense declined by $290 million from Q2 to Q3, despite the fact that Tesla’s overall operations were expanded to accommodate a large increase in vehicles sold in Q3 vs Q2. On average, over the last six quarters, R&D plus SG&A has been running at 39.5% of revenues. In Q2 2018, these charges were 33.84% of revenues. But in Q3 2018, R&D and SG&A dropped to 17.7% of revenues.
To be sure, there are “economies of scale” with respect to R&D and SG&A expenditures as revenues grow. But for R&D and SG&A to decline nearly 50% as a percentage of revenues from Q2 is simply not credible, unless Tesla intentionally drastically cut back on R&D and administrative/sales functions in Q3. Without question, Musk and his CFO played games with the R&D and SG&A expense accounts in order to reduce the charges expensed for these categories in Q3 vs the previous six quarters and especially vs Q2 2018.
It’s quite possible that Tesla loaded R&D and SG&A expenses into Q2 that technically belonged in Q3 knowing that it was going to report a big loss in Q2 ($717 million loss in Q2) anyway and had promised profitability in Q3. But it’s impossible to know if this occurred without having access to the inside books and bank statements. The stunning plunge as a percentage of revenues for these items in Q3 vs Q2 is the equivalent of asking us to believe in the existence of Santa Clause.
If we give the Company the highly doubtful benefit of synergies which reduced R&D and SG&A to just 20% of revenues – despite the fact that it has been running nearly double 20% over the last six quarters – the combined charge for these accounts would have been $1.219 billion rather than the $1.079 billion used by Tesla (note, at the very least it would have been reasonable to assume that the expense level at a minimum stayed flat vs Q2, meaning I’m being overly generous in my assumption). Under this scenario, Tesla’s operating expenses would have been higher by $140 million.
Adding this $140 million in incremental expense to the $49.4 million warranty expense manipulation and $33.8 million inventory write-down manipulation implies that Tesla’s GAAP net income was overstated by $223 million. Using the historical experience for these expense accounts, including an overly generous benefit in the assumption I use for “normalized” R&D/SG&A, Tesla’s GAAP income as reported would have been $88 million instead of $311 million. Tesla’s $88 of net income as adjusted less the $189 million in income attributable to GHG/ZEV sales turns the $311 net income reported as net income into a $101 million loss.
In addition to the questionable accounting used by Tesla to generate $311 million of GAAP “net income,” Tesla engaged in questionable, if not problematic, balance sheet maneuvers to boost the level of cash presented at the end of Q3. The purpose of this was to create the illusion of solvency. In the Q3 10-Q, Tesla shows a cash balance of $2.96 billion. At the end of Q2 Tesla had $3.11 billion.
Tesla’s accounts payable jumped jumped by $566 million from Q2 to Q3. Companies will stretch out their bills in order to conserve cash. Tesla has made a habit out of dragging its feet on paying vendors, suppliers and service providers as evidenced by the large number of court filings from smaller vendors who are forced to get a court order for payment. The same dynamic applies to “other accrued liabilities,” which contains other short term liabilities for which payment has not been made (payroll, taxes, interest and smallish items).
While accounts payable and other accrued liablities will naturally rise with the organic growth of a company, the rise in Tesla’s payables year over year is nothing short of extraordinary. Through the first nine months of 2018, per the statement of cash flows, Tesla generated $1.6 billion in cash “financing” from “stretching out” its payables vs $170 million in the first nine months of 2017. While Tesla’s revenues nearly doubled over the same period, this amount of unpaid bills has a reason behind it. The net effect of withholding payment of its bills longer than necessary is that it makes the cash on Tesla’s balance sheet appear larger than otherwise. Accrued payables and other short term liabilities are the equivalent of a short term loan to a company. These liabilities should be treated as a form of short term debt.
Subtracting current liabilities ($9.78 billion) from current assets ($7.92 billion) shows that Tesla has negative working capital of $1.86 billion. Technically Tesla is insolvent, which explains the games the Company plays with its supplier/vendors.
Another curiosity on Tesla’s balance sheet was accounts receivable, which more than doubled, from $569 billon to $1.155 billion. In the footnotes under “credit risk,” Tesla disclosed that “one entity represented 10% or more of our total accounts receivable balance” at the end of Q3, whereas previously no entity represented 10% of receivables. In other words, one entity owed Tesla at least $115 million.
When asked about the big jump in A/R during the earnings conference call, the CFO dismissed it by claiming that the quarter ended on a Sunday. It’s beyond absurd that the analysts on the call accepted this answer without further interrogation. Subsequent to the release of the 10Q, a company spokesman told a reporter from the L.A. Times who had inquired about the 10% disclosure that the receivable was attributable to a large partner bank for car loans issued to U.S. customers. The spokesman said that “all of this receivable was cleared in the first few days of Q4.”
The inference was that Tesla sold $115 million or more worth of cars after 5 p.m. on Friday and over the last weekend of its quarter financed by one bank that could not be processed by the banking system. If this were truly the case, why not just state this as fact openly rather than leaving the market guessing what might have happened? 10% of $1.155 billion is considered “meaningful” under strict GAAP, which means this issue requires more detailed disclosure. The CFO’s vague response to the question about the issue reflects intentional obfuscation of the matter.
Unfortunately, we may or may not be able to figure out exactly what happened when the 10-K is released. I’m not optimistic that the Company will come clean. However, an analyst posted an assessment on Twitter (@4xRevenue) which seems to be a very reasonable explanation to this mystery. This analyst believes that the 10% receivable is from a lease partner (a bank) who has underwritten leases that contain Residual Value Guarantees from Tesla.
Tesla had been offering Residual Value Guarantees (RVG) on leases as an incentive to generate sales. The RVG is a guarantee from Tesla on the value of the car at the end of a lease. In order to stimulate lease-based sales, auto companies will guarantee the lease-end value of car at a level that is typically above the market value for that car at the end of the lease. It’s a “back-door” mechanism used to lower the monthly cost of a lease to the lessee.
If the receivable in question is from a bank that financed Model S&X leases, it means that a large number of vehicles came off lease at the end of Q3 and the bank was returning these cars to Tesla. The “receivable” is the guaranteed residualy value of these vehicles. It also means that Tesla likely will have a large cash payment (at least $115 million) to make to the bank that would be connected to the RVG. Based on actual market data, that the resale value of used Tesla’s has been declining rapidly. This being the case, Tesla has a large make-whole payment to make to the bank who represents at least 10% of the receivable. Tesla will then look to unload these used Teslas and recoup as much as possible, though it will be substantially less than the guaranteed make-whole made by Tesla.
This analysis would explain why Tesla’s payables and receivables were unusually high at the end of Q3. If this transaction had been processed before the end of Q3, Tesla’s accounts receivable would have been lower by the value of the cars being returned to Tesla under the RVG. The accounts payable would have lower by the amount Tesla owes to the bank. Tesla’s cash balance would have been lower by the amount that Tesla paid to the bank under RVG.
Recall that the Tesla spokesman said that this specific A/R was “cleared” in the first few days of Q4. Holding off on processing this transaction until after the quarter ended enabled Tesla to show a higher cash balance than it would have otherwise. It also kept the used Teslas out of Tesla’s inventory, which further enabled Tesla to manipulate the inventory write-down by taking a much lower write-down than historical write-downs. This is because the market value of the used Teslas received is lower than the amount Tesla paid under the RVG. This would have required Tesla to write-down the value of the used Teslas, thereby increasing the inventory write-down charge, increasing cost of goods sold, lowering the gross margin and lowering the amount GAAP “net income” reported.
This also explains why Tesla moved $73 million worth of cars out of finished inventory and into the PP&E account on the balance sheet. Tesla accounts for vehicles used as service loaners as part of PP&E. I don’t have a problem with that. But moving $73 million of these vehicles allowed Tesla to avoid including those vehicles as part of its inventory write-down expense. It also allowed Tesla to move the cars taken back under the RVG transaction described above without causing an unusual change in inventory that required explanation. In other words, it’s entirely possible, if not probable, that Tesla wanted to “make room” for the used Teslas.
The bottom line – Tesla pulled out every accounting manipulation available to it in order to produce the promised positive GAAP net income, positive cash flow, extraordinarily high gross profit margins and a higher quarter-end cash balance. It was accounting deception, and in some areas probable fraud, at its finest. The Wall Street ass-kissing analysts did nothing other than cheer the results and lob easy questions at management on the conference call. Many of them are likely clueless about the degree to which Tesla manipulated reality.
It will be very interesting to see how Q4 turns out for Tesla. Based on reports from China and Europe, car sales have fallen off a cliff in October. Norway reported the first week of EV sales, which showed that Jaguar i-Pace deliveries, new to the market, were 44 vehicles vs. just 11 for the Tesla models S&X combined (the Model 3 has not been approved for sale yet in the EU). In October the i-Pace sold 441 units vs 201 for the Tesla S&X. This gives us a valuable glimpse at the effect competition will have on Tesla’s sales. Soon the Audi e-Tron will be available. It will likely smother any demand for Teslas.
Tesla had to make a $230 million convertible bond maturity payment in a couple weeks. It then has to start figuring out how to generate enough cash to make another $930 billion convertible bond maturity payment in March. On the assumption that Tesla’s sales are highly negatively affected by competition and the economy, Tesla will have a hard time raising the money needed to refinance the March convertible bond payment. Accounts payable will also become a problem, especially if Tesla is unable to raise more cash selling ZEV and GHG credits. On top of this, the tax-credit that Tesla car buyers receive from buying a Tesla EV will soon run out. This will make buying a Tesla more expensive.
The above analysis is from my Short Seller’s Journal from November 11th. I also provided some ideas for shorting Tesla using short term and long term puts. You can learn more about this newsletter here: Short Seller’s Journal information. Note: one of my subscribers emailed me this morning that he just took $3500 in profits on January KB Home (KBH) puts that I recommended a few months ago.
Dave,
You have truly revealed one of the kings of fraud who has no clothes.
What will it take to do an article on “Clown Cramer”?
He is nauseating with him protesting too much for the blind bulls to stay invested in the market.
Excellent analysis. It will be interesting to see the adjustments for these expense provisions after year end.
Musk must be a “made man” of the Deep State criminal racket to keep out of jail. Unfortunately, there doesn’t seem to be any end to the rackets being run out of Washington and Wall Street. Must be a “national security” issue.
I read another article on Tesla today highlighting the change in Tesla’s “rules” (hah!) for recognizing revenue. They previously recognized revenue upon closing a vehicle transaction, starting at some point in Q3 they began recognizing revenue upon shipment. That apparently counts for north of $400 million in revenue, which of course is a one-shot deal since they effectively stole Q4 revenue and moved it into Q3 with this ploy. Granted, their production volumes are still increasing, so it is better than a wash, but the net effect of all these games will be an ugly Q4. Are there any rabbits left in Elon’s hat?
Yes – I didn’t even touch on revenue recognition but that’s a good point.