Digging into Tesla’s reporting on PP&E

In the FT last week, Dan McCrum and Stephen Morris crunched some numbers from Tesla’s financial statements. The punchline is straightforward: “Compare Tesla’s capital expenditure in the last six months of 2024 to its valuation of the assets that money was spent on, and $1.4bn appears to have gone astray.”

The FT piece nicely illustrates a couple of core ideas used by sophisticated analysts. The first is using multiple financial statements to extract information about what is going on with a business. By comparing the picture presented in the balance sheet with that shown on the statement of cash flows, we can collect some pieces of the jigsaw puzzle that is Tesla’s business.

The second idea is the old notion of digging into the footnotes. The inscrutable “Property, plant and equipment, net” (PP&E, net) seen on the face of the balance sheet is broken out in Note 7 into components such as “AI infrastructure” and “Construction in progress” (at original cost) along with accumulated depreciation.

Dan and Stephen compare the increase in gross PP&E (Note 7) with the amounts shown on the cash flow statement next to “Purchases of property and equipment excluding finance leases, net of sales” and conclude that $1.4 billion seems to be missing, with the biggest gap appearing to occur in the fourth quarter (Q4).

Over the weekend, accounting experts Olga Usvyatsky and Francine McKenna took up the baton with a post on their respective Substack feeds—Deep Quarry and The Dig—that refines the FT analysis.

PP&E acquired using liabilities

One refinement that Olga and Francine make is to account for amounts found in the disclosure of non-cash investing and financing activities that is a required supplement to the cash flow statement. “Acquisitions of property and equipment included in liabilities” totalled $1,410 million for Tesla in 2024. While adjusting for this amount—which is effectively negative in Q4—narrows the Q4 gap, it makes the gap for the remaining three quarters even bigger.

Many financial analysts argue that it is better to focus on cash flows than on accounting numbers. However, such arguments often stem from common myths about cash flows. One myth is that cash flow numbers are more difficult to manipulate than those relying on the kind of judgment that is needed with accounting measures such as bad debt expense. Another myth is that cash flow statement numbers are “accounting free” in the sense that there are no arbitrary accounting standards that need to be applied to figure out cash flows.

The reality is that accounting standards need to be applied precisely because cash flows are so susceptible to manipulation. Suppose an acquirer of equipment says to its supplier: “How about you ship the equipment to me in Q3 and I pay you in Q4?” In Q3, there’s no cash flow. In Q4, there would be a cash flow, but could it perhaps be buried in the operating activities portion of the cash flow statement under “decrease in accounts payable”?

Accounting standards handle such situations by requiring the disclosure of such transactions in Q3 in the supplement discussed above and also requiring the Q4 cash flow to be classified as an investing cash flow.

PP&E paid for in advance

A second refinement made by Olga and Francine picks up on the possibility that payment can be made to a supplier for equipment before it is delivered. Such payments give rise to assets—future economic benefits—but there is no PP&E until it is delivered. As such, these amounts will show up on Tesla’s balance sheet under “Prepaid expenses and other current assets”. Again, these cash flows should be classified as investing cash flows rather than operating cash flows. Olga and Francine note that “Tesla does not disclose in its 10-K whether there are any PPE prepayments, and if so, where they are recorded.” But one can get a sense of non-operating changes in “Prepaid expenses and other current assets” by looking at the cash flow statement. While the balance sheet amount increased by $1,974 million, the amount reported on the cash flow statement suggests an increase of $3,273 million. This would suggest that any investing portion of prepaid assets might have decreased by an amount reflected in the difference. Decreases like this will occur when the physical assets are delivered and the asset is reclassified to PP&E. (I provide an example to illustrate this in the ‘Example’ tab here.)

Putting these two pieces together arguably makes the gap identified by the FT larger: in one analysis, I calculate a gap of over $3.8 billion! Additionally, while the FT flags Q4 as anomalous, I see gaps throughout the year and perhaps the biggest gap is in Q1. (While Olga and Francine flag Q4 as anomalous, this relates to the shift of amounts from one component of PP&E—construction in progress—to another—AI infrastructure, which has zero net impact on PP&E totals.) My spreadsheet can be found here.

What next?

Dan and Stephen in the FT point out that “aggressive classification of operating expenses as investment can be used to artificially boost reported profits.” The classic case of this was WorldCom, which incorrectly classified expenses as additions to assets, overstating PP&E by about $20 billion. Tesla does not seem to fit the WorldCom pattern because the issue seems to be amounts we might expect to see in PP&E not showing up there. A scan of Tesla’s balance sheet does not suggest a lot of places to “hide” amounts of the magnitude discussed.

The FT notes that “Tesla did not respond to requests for comment” and speculates that “perhaps the assets will show up next quarter.” Olga and Francine also suggest that perhaps “time will tell - as we said, one-time cut-off differences usually reverse in the next reporting period.”

But why would Tesla force investors to wait to find out whether the more acceptable explanations hold? Why have analysts make guesses about undisclosed numbers? It should be a straightforward matter to produce the kinds of schedules that firms reporting under IFRS produce as a matter of course and file them on Form 8-K.