Saturday 24 July 2021

Estimate of Tesla’s Bitcoin Holding and Analysis of 1Q2021 Bitcoin Sale

AA Uses Only the Most Accurate Equipment
for His Estimates

I thought I’d throw my chapeau into the ring of those analyzing Tesla’s Bitcoin Holdings.

And then for good measure provide my own analysis of the 1Q sale.

Summary

  • Numbers of BTC and original costs are estimates.

  • Tesla originally purchased BTC 46,561.73 at an average original cost of US$ 32,215.30 per “coin”. (My Scenario #2).

  • Using averages of my Scenario #1 and #2, in March it sold 4,466.64 BTC at an average price of US$ 60,895.95 per coin.

  • As of 31 March Tesla held BTC 42,901.81 at an average cost of US$ 31,621.36 per coin.

  • Under US accounting “rules” once Tesla takes an “impairment” on its BTC holdings, it cannot reverse it if fair value increases later. The only way to “capture” the increase is to sell BTC in which case the higher fair value over carrying cost is a component of net profit.

  • Because of these accounting “wrinkles”

    • Tesla may have an economic incentive to support BTC’s price because impairments flow through the income statement.

    • Or to sell BTC to generate a profit to offset impairment charges.


Introductory Comments

Don’t be followed by the apparent precision in my numbers below.

There isn’t sufficient information to achieve precision.

So my numbers and those of others cited below can only be rough estimates.

You’ll find those other estimates here and you can compare methodology and results.

First up is Shawn Tully at Fortune: 38,300 BTC held at 31 March 2021.

Second is Chuck Jones at Forbes: 42,902 BTC held.

I’ll be using Tesla’s 1Q2021 10Q as the source document for data.

If you’re interested in the US accounting treatment for digital assets, here’s a link to an AICPA publication on ASC-350.


What are Tesla’s BTC Holdings as of 31 March 2021?

From Note 3 in Tesla’s 10Q the fair market value of their BTC as of 31 March 2021 is some US$ 2.48 billion.

Using Yahoo Finance data, the closing price of BTC on 31 March was US$ 58,918.83.

That equals BTC 42,901.81. US$2.48 billion divided by US$ 58,918.83.

That number is in agreement with Chuck Jones’ calculation.

Shawn’s number differs because he’s using a profit of US$ 101 million on the March BTC sale, due to his including the US$ 27 million impairment as a component of the sales proceeds. Therefore, his cost of sale is US$ 171 million not US$ 144 million.

By my calculation the carrying value of Tesla’s BTC portfolio is US$ 31,621.36 as of 31 March 2021. US$1.331 billion divided by 42,091.81 coins.

But that is an adjusted cost after the US$ 27 million impairment. (Also disclosed in Note 3).

First cut.

To determine the original purchase price of the remaining BTC we have to add back the US$ 27 million impairment charge. That means the original cost of the Bitcoin remaining after the March 2021 sale but before impairment is actually US$ 1.358 billion.

Note the implicit assumption that the impairment was taken after the March sale.

On this basis the historic cost per Bitcoin is US$ 32, 262.81.

But another wrinkle.

As Shawn Tully points out, reconciling the balance of the BTC holdings results in a US$ 2 million difference.

That is, Tesla purchased US$ 1.5 billion sometime between 1 January 2021 and early February. The last purchase would have had to occurred some time prior to 8 February.

Why?

Tesla first announced the purchase in its 2020 10-K which is dated 8 February.

As per note #3 in Tesla’s 10Q Tesla recognized gains of US$ 128 million on the BTC sale and took a US$ 27 million impairment.

As per my understanding of the required accounting, the impairment is unrelated to the sale.

From the Consolidated Statement of Cash Flows, we see that Tesla received proceeds of US$ 272 million from the BTC sale. If the recognized gain on the sale was US$ 128 million, then the cost of the BTC must be US$ 144 million.

(Note that is 9.6% of the original purchase amount and would seem confirm Tesla’s 1Q statements that it sold 10% of its original holding)

US$ 144 million plus US$ 27 million equals an expected US$ 171 million decrease in the balance of BTC from first purchase through 31 March 2021

But that amount is US$ 2 million more than the net change in BTC holdings—US$ 169 million.

Is this due to rounding? Or to vehicle purchases using BTC? Or a combination of both?

We don’t know.  Sadly, a question that might have shed light on this issue was not selected for the Q&A on Tesla's 1Q Call.

We also don’t know what the US$ 27 million impairment charge relates to.

Is it the original BTC purchase? Or BTC received for car purchases? Or both? Or something else?

Scenario #1

If we assume there were no material car purchases with BTC and use the US$ 1.358 billion figure above, the original historic purchase cost per “coin” is US$ 32,262.81.

Tesla would have had to sell 4,463.34 BTC to equal the US$ 144 million cost of BTC sold in March.

The original number of BTC bought would then be 46,555.15

You’ll notice this equals US$ 1.502 billion at the estimated historic cost above. Thus it includes the unexplained US$ 2 million “difference”

Scenario #2

Same assumptions as Scenario #1, but US$ 2 million assumed rounding differences is excluded. The original cost of the BTC purchase is US$ 1.356 billion. That gives an original purchase cost of US$ 32,215.30 per “coin”.

In this case Tesla sold some 4,469.93 BTC.

Under this second scenario, it would have originally bought 46,561.73 BTC


Key Accounting Considerations

Under ASC-350 and ASC-820, once fair value is lower than carrying value, Tesla must make a one way adjustment in carrying value via an impairment charge.

If fair value later increases, the impairment can not be reversed. (Question #6 pages 6-7 in the AICPA publication linked above)

However, on sale of BTC in the future, the difference between carrying value (reflecting any impairments) and sale proceeds will be recognized as “profit”. 

Thus, if fair market value has increased but not carrying cost, Tesla would recapture the difference between FMV and carrying cost in additional profit on the sale.


Implications of Accounting Rules

It would be interesting to see if Tesla or any of its senior officers announce BTC initiatives or tout BTC when the price of BTC appears in “danger” of declining below the carrying value in Tesla’s financials: US$ 31,621.36 as of 31 March 2021.

It will also be interesting to see if Tesla conducts any additional sales to offset any future impairments.


March 2021 BTC Sale

The Scenario #1 and Scenario #2 estimates for the number of BTC sold are very close. So let’s use the arithmetic average of both. That’s 4,466.64.

Using this number, the average price received on the March sale was roughly US$ 60,895.95. US$ 272 million divided by 4,466.64.

If you look at the Yahoo Finance historic prices for BTC, you’ll see several days that might be candidates for a sale, e.g. March 14th.

 

The Sad State of BCP and Cyber Risk Planning at Financial Market Infrastructure Institutions

Nine Years Thundering Toward the Station
Alas, Yet to Arrive

Here I am again making what no doubt could be labeled as an “overdue wake up call” by the chronically somnolent or perhaps as a “sobering fact” by the habitually intoxicated.

On 21 July the BIS Committee on Payment and Market Infrastructures published a joint report with the Board of International Organization of Securities Commissions on a level 3 evaluation of implementation of Principles for Financial Market Infrastructures (PFMI).

Before beginning my rant, a couple of notes.

The PFMI were issued in 2012, which would appear to be some nine years ago, if my arithmetic is correct.

The PFMI were issued to set standards for Business Continuity Planning (BCP) and Recovery of Operations for systemically important and therefore critical financial market infrastructure institutions:

  • payment systems (PS)

  • central securities depositories (CSD)

  • central counterparties (CCP),

  • securities settlement systems (SSS) and

  • trade repositories (TR).


To be clear we’re talking about payment systems, e.g., CHIPS, CHAPS, Fedwire not individual banks. For the other categories, some US examples: DTCC, NSCC, FICC, etc.

The rationale is to require these critical market infrastructure institution to have effective BCPs to restore service in the event of disruptions. In that regard think of power blackouts, 9-11, and yes cyber attacks.

Here is the link to the BIS CPMI page on the PFMI which contains additional details.

In addition to the PFMI, you will notice that there are also an additional eight guidance papers on implementation of the PFMI.

Among those there is a 2016 guidance paper on cyber risks. Applying the same arithmetic as above, that would appear to be five years ago.

The July joint BIS CPMI/IOSCO-OICU IMSG (Implementation Monitoring Standing Group) reviewed the business continuity planning practices at a sample of 38 FMIs from 29 jurisdictions during 2019-2020.

The sample comprised 14 PSs, 15 CSDs/SSSs, five CCPs and four TRs.

The study was conducted by reviewing responses to a questionnaire.

That is, based on assertions made by the respondents rather than an on site investigation.

If you’re like me, you might find that a bit chilling given the results. 

If you're willing to self-certify to failure, isn't it likely the the failure is even more egregious?

So what were the findings?

1.2.1 Timely recovery in the event of a wide-scale or major disruption

The IMSG has identified one serious issue of concern, which is that the business continuity management of some, and potentially many, FMIs does not seem to “aim for timely recovery of operations and fulfilment of the FMI’s obligations, including in the event of a wide-scale or major disruption”, as expected by the Operational Risk Principle (Principle 17). Furthermore, based on the information provided by the participating FMIs, there are doubts about whether their business continuity plans are designed to “ensure that critical information technology (IT) systems can resume operations within two hours following disruptive events” and “enable the FMI to complete settlement by the end of the day of the disruption, even in case of extreme circumstances” as expected by KC6. [That’s Key Consideration 6 in Principle 17]. Given this is a serious area of concern, the CPMI and IOSCO expect the relevant FMIs and their supervisors to address this as a matter of the highest priority.

Given that the PFMI were issued some 9 years ago and implementation is still deficient, the use of the term “highest priority” is perhaps both an indication of importance as well as a bit of sarcasm. That being said, the IMSG only “expects” this to be done. No doubt as they have expected implementation over the past nine year.

The IMSG’s findings continue:

While almost all of the surveyed FMIs indicated that they have business continuity plans (BCPs) designed to meet this requirement, there is evidence that leads the IMSG to question this. In terms of specific evidence:

  • A few of the surveyed FMIs do not explicitly aim for the 2hRTO, even for wide-scale physical (noncyber) disruptions.

  • One of the surveyed FMIs acknowledges that its secondary site does not have a distinct risk profile from that of its primary site.

  • A small number of FMIs stated that they did not have alternative arrangements to allow for the processing of time-critical transactions. Of those that did have such arrangements, some relied solely on manual and paper-based alternative arrangements.

  • A few FMIs indicated that they do not have specific plans to mitigate potential widespread staff unavailability. This suggests that these FMIs may have difficulty completing settlement if this were to occur.

Mighty disappointing” to use a technical financial market term.

Inverse kudos to the respondent that apparently will rely on manual paper-based systems. Systemically important FMIs are likely to be ones that process “lots” (that’s another technical financial term) of transactions.

But you ask what about cyber attacks?

1.2.2 Cyber risk

Principle 17 states that “[a]n FMI should identify the plausible sources of operational risk, both internal and external, and mitigate their impact through the use of appropriate systems, policies, procedures, and controls…” The IMSG has identified one issue of concern, which is that a few FMIs in the sample did not provide specific BCP objectives with respect to cyber risk. Among the FMIs that have specific BCP objectives with respect to cyber risk, only a few explicitly acknowledged the breadth and depth of potential cyber attacks and the complexities of cyber risks that their BCPs may not be able to cover.

While not as serious as the previous risk in the assessment of the IMSG, I think this qualifies as very serious.

  • The probability of a cyber attack may be higher than some of the other risks of disruption and the impact much greater.

  • If computer networks are hacked, critical information even that at backup sites may be unavailable or destroyed.

Particularly, if the attack is the work of a state actor.

That would be a different kettle of fish than a natural disaster. Or even a 9-11 style attack.