In Brief
Excessive leverage, deficient collateral, volatile pricing, and a liquidity crunch contributed to the downfall of several crypto lenders, hedge funds, and exchanges in 2022—yet these risks were opaque in these entities’ balance sheets, which failed to provide insight into the dangers they posed to their stability. The lack of a single accounting rule for crypto lending resulted in the emergence of blended accounting that obscured leverage risk, liquidity risk, and price risk via certain practices, including: offsetting receivables and liabilities, measurement mismatch of related assets, measurement mismatch of assets and related liabilities, inadequate breakdown of crypto assets, defective methods to determine cost basis, ambiguous accounting for customer funds, questionable accounting for in-house tokens, and inappropriate related-party accounting. Instead of using blended accounting that dilutes transparency, an industry-specific accounting is needed for the crypto sector.
* * *
In a report issued on February 16, 2022, the Financial Stability Board (FSB) (Assessment of Risks to Financial Stability from Crypto-assets, https://www.fsb.org/wp-content/uploads/P160222.pdf, (2022) p. 18) warned that crypto assets could reach a point where they represent a threat to global financial stability. Shortly after the FSB raised its red flag, a crisis hit the crypto sector.
In May 2022, the crypto market started crashing when stablecoins TerraUSD and Luna fell to zero and wiped out many other crypto entities. Crypto lenders, such as Celsius Network LLC, entered bankruptcy. Hedge funds, such as Three Arrows Capital, crashed in July 2022 after suffering huge losses when the price of numerous digital assets plunged. This culminated in the exchange FTX Trading Ltd. filing for bankruptcy in November 2022. Following the troubles at FTX, certain crypto-related banks failed, and the demise created chaos for other crypto lenders (hereinafter, “crypto turmoil”).
The burgeoning crypto industry has shifted from simple buy-and-sell trading to leveraging products similar to those in a traditional market. But a lack of standardized accounting undermined the usefulness of crypto companies’ financial statements. The discussion below will explore how the lack of crypto industry–specific accounting concealed risks and thus contributed to the downfall of crypto lenders and platforms in 2022.
Background
Professional investors, such as investment advisers and hedge funds, have taken part in the virtual currency market. The growth of institutional investment has shaped the way cryptocurrency markets function. First, the crypto-structured products which would magnify volatility and transmission of risks have grown to a remarkable level (Gregory Zuckerman, “Mainstream Hedge Funds Pour Billions of Dollars Into Crypto,” Wall Street Journal, March 9, 2022 https://tinyurl.com/hy4a3wsu). Additionally, the virtual currency lending extended to a professional investor who might borrow from or lend to its customers could increase spillover risk.
The AICPA and the IFRS Interpretations Committee (IFRIC) issued similar non-authoritative guidance in 2019 for crypto assets. The AICPA Practice Aid, “Accounting for and Auditing of Digital Assets,” advised that crypto assets be accounted for as intangible assets at cost and subject to impairment tests unless they are subject to specialized accounting guidance in the United States, such as for brokers and dealers.
On March 23, 2023, FASB issued an exposure draft addressing the accounting for and disclosure of crypto assets. The standard, ASU 2023-08, Intangibles—Goodwill and Other—Crypto Assets (Subtopic 350-60), was finalized on December 13, 2023. The guidance does not cover crypto lending activities.
During the crypto turmoil, some crypto lenders failed as a result of hidden risks that were not reflected in the financial statements. The blowups of Celsius and FTX exposed hidden risks that might have raised red flags for investors (Ben Foldy and Jean Eaglesham, “Public Crypto Firms Flash Accounting Red Flags,” Wall Street Journal, Dec. 8, 2022, p.B11). High-quality financial reporting promotes effective risk management, but the vulnerabilities and risks posed by troubled crypto lenders were hidden due to a lack of industry-specific accounting.
There was no single existing accounting rule that could be used for crypto lending transactions, as it is stated in the AICPA Practice Aid (updated in 2023, p. 24): “U.S. GAAP does not provide explicit guidance specific to the lending of crypto assets and accordingly, we understand the SEC staff considered all relevant guidance in U.S. GAAP, but did not base its accounting conclusions solely on a single FASB ASC Topic.” Consequently, blended accounting—which refers to a mixing of multiple accounting rules—was required. Yet, there was no rule about how to mix existing accounting principles properly when it came to crypto risks.
Backward Mapping
This article employs the backward mapping method, by which the end result of a troubled entity is mapped to analogous accounting rules, to identify problematic accounting treatments.
Backward mapping is used because—
- ▪ amid a lack of industry-specific accounting for the crypto sector, the IASB reiterated the use of existing accounting rules,
- ▪ there was no antecedent data because the crypto turmoil was unprecedented,
- ▪ most crypto lenders did not disclose their financial statements publicly, and
- ▪ the AICPA Practice Aid did not specify which accounting principles the SEC staff considered when commenting on crypto lending, and
- ▪ when there is no industry-specific accounting principle applicable, each entity might reach different conclusions. For example, the AICPA Practice Aid (2023, p. 21) states, “Because of the variety of facts and circumstances that may exist, it is impossible to provide a general rule or accounting for stablecoins.”
Backward mapping is used to map then-current existing accounting principles based on the troubled crypto lenders’ practices described in:
- ▪ the Celsius report from the independent examiner, Shoba Pillay, who was appointed by the bankruptcy court (United States Bankruptcy Court Southern District of New York. Chapter 11, Case No. 22-10964 (MG), filed January 31, 2023. https://tinyurl.com/2vnc8xjv);
- ▪ two interim reports issued by FTX’s new management, led by CEO John Ray III, and attached to bankruptcy court filings (The United States Bankruptcy Court for the District of Delaware. Case 22-11068-JTD Doc 1242-1, filed April 9, 2023, https://tinyurl.com/2rbw3ssj; The United States Bankruptcy Court for the District of Delaware. Case 22-11068-JTD Doc 1704, filed 26 June 2023, https://tinyurl.com/yn6mhdsv);
- ▪ relevant SEC enforcement actions; and
- ▪ other publicly available sources, such as Senator Sherrod Brown’s letter to the Treasury Secretary Janet Yellen (https://tinyurl.com/35w2b6pu).
Example
The following example is used to demonstrate how blended accounting helped hide risks. Company X is a virtual currency service provider that lent Bitcoin carried at a fair value of $50,000 to an investor, Company Y, which pledged Ethereum with a fair value of $52,000 as collateral. Company Y had the right to sell or repledge the borrowed Bitcoin, and Company X had the right to lend or repledge the Ethereum collateral.
The example above with Company X is based on the end result, which means the practice described in the Celsius Examiner’s Report:
They are presented in a way where it is difficult to understand which assets offset which liabilities (and vice versa). For example, institutional and retail loans receivable (asset) are recorded in the crypto asset in which they were lent out, but the offsetting liability (Institutional/User Collateral) is recorded in the crypto asset of the collateral provided by the borrower (e.g., a borrower taking a USDC loan secured by BTC). The same is true when Celsius is the borrower (liability) of coins secured by posted collateral (asset) of another coin (e.g., Celsius taking a DOGE loan secured by ETC). This mismatching of assets and liabilities create challenges in calculating surpluses and deficits by coin since the offsetting entry is often in a different crypto asset. (p.40)
This practice above is mapped to the analogous accounting rule, which is traditional market security lending accounting. Company X used blended accounting, which mixed multiple accounting rules. It initially applied security lending accounting, but also considered accounting for receivables. Subsequently, it applied AICPA’s intangible asset accounting treatment.
Assume that Company X initially made the following entries:
Entry I: To reclassify from virtual currency to loan receivable that Company Y had the right to sell or repledge, assuming both the carrying amount and fair value of virtual currency inventory were the same at the time of the transaction:
Dr. Virtual currency loan receivable $50,000
Cr. Virtual currency $50,000
Entry II: To record the receipt of collateral that Company X had the right to lend or repledge:
Dr. Virtual currency collateral received $52,000
Cr. Obligation to return collateral $52,000
Analysis
Measurement mismatch between a crypto asset and related liability.
With respect to Entry I above, if the lent virtual currency was borrowed from a customer, then the related liability would be the “borrowed virtual currency.” A mismatch between “virtual currency loan receivable” and “borrowed virtual currency” would appear if the former was evaluated at cost but the latter was measured at fair value subsequently. In addition, regarding Entry II, if the debits were measured at cost, subject to impairment test, but the credits were subsequently assessed at fair value, a mismatch would occur and Company X’s leverage ratio would be impacted.
In the face of turmoil in the crypto markets, SEC staff addressed the issue discussed in Entry I by requiring virtual currency loan receivables to be measured at fair value subsequently. Unfortunately, this treatment would not cover Entry II.
Measurement mismatch of related assets.
“Virtual currency loan receivable,” that is, the debit side of Entry I, and the underlying collateral that is the debit side of Entry II, were denominated in different virtual currencies. If both were not subsequently accounted for at fair value, the collateral-to-loan balance ratio would be in error and margin call could be missed when there was a depreciation or appreciation of the crypto assets.
Some crypto entities have many affiliates or related parties under the same group, and thus huge related-party transactions.
The collateral ratio for Company X is 104% initially (“Virtual currency collateral received,” $52,000, divided by “Virtual currency loan receivable,” $50,000). If Company X, subsequently extended intangible asset treatment to both items, however, the collateral-to-loan balance ratio would subsequently be inaccurate. For example, at the end of the reporting date, the fair value of Bitcoin was $60,000, but the fair value of Ethereum was $53,000; thus, the new collateral ratio should be 88% instead of 104%, and the margin call might be overlooked.
Inadequate separation of crypto-denominated assets on balance sheet.
Not all crypto-denominated assets are the same. A borrowed crypto asset is different from crypto collateral received. Both are also unlike a proprietary crypto asset inventory owned by the reporting entity. Inadequate separation and disclosure of all crypto-denominated assets, such as the use of a single balance sheet line item for all crypto asset holdings, further obscures risk.
For example, the first FTX Report stated that many transactions were grouped together and booked under the generalized account names “investments in cryptocurrency” (2023, p.18). As a result of this inadequate breakdown, it would not only be difficult to calculate the leverage used in terms of cryptocurrencies borrowed to fund loans, but also to gauge the collateral ratio, even if both debit sides were subsequently accounted for at fair value.
Flawed method to determine the cost basis of the units transferred.
The AICPA Practice Aid states that the first-in, first-out (FIFO) method is one reasonable and rational approach that could be used to determine the cost basis of the units sold (2023, p. 8). In addition, ASC section 350-60-50-2 will require disclosure of the method used to determine the cost basis for computing gains and losses; the examples cited include FIFO, specific identification, or other method used.
Neither FASB nor the AICPA have specified when FIFO, specific identification, or other methods should be used. In fact, using FIFO or a specific identification method is quite challenging for large crypto service providers.
It might be impractical to segregate the crypto assets an entity has received from different groups of investors, and therefore it might choose to pool all crypto assets together. Under such circumstances, the use of FIFO or specific identification method would be quite onerous, and the methods might thus be applied imprudently. The first FTX Report noted, “although the FTX Group consisted of many, separate entities, transfers of funds among those entities were not properly documented, rendering tracing of funds extremely challenging.” (2023, p. 15)
Some companies, like FTX, traded and had custody of clients’ assets. These companies were engaged in circular lending practices such as rehypothecation, where the same asset would be lent multiple times (Joshua Franklin, “U.S. Lawmaker Pushed Bipartisan Bill for Greater Regulation of Digital Currencies,” Financial Times, Nov. 29, 2022, p.10). In crypto lending activities, borrowed crypto assets or collateral received could be loaned out or used as collateral again. This would complicate the tracing of the source, especially in cases where all crypto assets were put in a single “omnibus” account instead of being situated across different online wallets with adequate records.
If FIFO or the specific identification method were applied inappropriately, realized gain and loss could be easily manipulated. The lack of separation between realized gains and unrealized gains represents a greater issue. Gains might be inappropriately distributed in the form of bonuses and dividends or other terms. In the worst case, since the crypto industry does not have the same capital requirements as banks to have an adequate buffer against insolvency or to mitigate systemic risks, an entity could buy back stocks. This could eventually leave only nominal capital on an entity’s books, meaning that a certain amount of unrealized gain in retained earnings might be the major component of its net worth. Once crypto prices dropped sharply, any entity with significant crypto-denominated assets and without a capital cushion would take huge losses—and might even go bankrupt.
Ambiguous accounting for customer crypto assets.
On March 31, 2022, the SEC published Staff Accounting Bulletin (SAB) 121 to address accounting for entities that have obligations to safeguard crypto assets held for their platform users (https://www.sec.gov/oca/staff-accounting-bulletin-121). SAB 121 requires affected companies to record a liability and the corresponding asset on their balance sheets at fair value.
In addition, according to the AICPA Practice Aid, a crypto custodian must book customer crypto assets on the balance sheet regardless of whether the entity controls customer crypto assets (2023, p.10). There are exceptions, such as for a broker-dealer that falls within the scope of ASC 940. Under the AICPA Practice Aid, “when a broker-dealer is an agent for a customer, the transaction should not be reflected on its statement of financial condition” (2023, p.13). As for how an entity determines that it is a broker-dealer within the scope of FASB ASC 940, unfortunately there is no answer in the AICPA Practice Aid since there is diversity in practice (2023, p.13).
As a result, some entities might treat a customer’s crypto assets as an off-balance sheet item. Furthermore, the practice of pooling all crypto assets together could provide more leeway for the mixing of customer funds with corporate funds. The second FTX Report states that FTX was “commingling and misusing vast sums of customer and corporate funds,” and “it is extraordinarily challenging to trace the source of funding for particular FTX Group transactions, or to differentiate between FTX Group operating funds and FTX.com customer deposits” (2023, p. 8). Accordingly, a customer’s funds might be used inappropriately while leaving no record on the books, which in turn would further obscure an entity’s leverage.
Problematic accounting for in-house tokens.
Celsius’s native currency, the CEL token, played a central role in its collapse. FTX’s in-house token, FTT, also contributed to its failure. Accounting for in-house tokens is uncertain and unclear, and at least two issues have arisen that would obscure leverage risks:
Questionable on-balance sheet treatment.
Assuming there is a market, how does the issuer account for its in-house token and what is the initial cost? A newly minted in-house token has no cost basis, but some entities might record an in-house token on the balance sheet at fair value. An inflated balance sheet would conceal actual leverage ratio and liquidity risk. For example, the Celsius Examiner’s Report states that, “In February 2019, Celsius added its Treasury CEL to its balance sheet at then-market price.” (p. 8) Treasury CEL consisted of Celsius minted at the time of the initial coin offering but not offered for sale. The lack of industry-specific accounting might have contributed to this questionable accounting of recording at fair value, even when no sale had taken place and the asset had no cost basis.
Questionable accounting for in-house tokens used in lending activity.
An in-house token or native token created by the issuer has no value unless it is sold or traded. An issuer might record the loan as an off-balance sheet item, because an in-house token is an intangible asset with zero cost basis. For example, FTX transferred its in-house token to a related party at no cost.
Even if there was a sale for an in-house token, there remains an inflation issue. Crypto assets do not have an inherent or underlying value. Their value derives from supply and demand. Since an issuer controls the supply side of in-house tokens, it is easy to inflate the value of the token, as stated in the Celsius Examiner’s Report citing one employee’s note: “We are using users USDC to pay for employees worthless CEL … All because the company is the one inflating the price to get the valuations to be able to sell back to the company” (p.10). In addition, FTX’s deadly mistake was to borrow against its own token. In his letter to the Treasury Secretary, Senator Sherrod Brown indicated that “FTX relied on its own proprietary crypto token leading to inflated valuations that further fueled irresponsible risk taking” (2022, p 2). Inflated collateral could trigger excessive leverage and cause disaster if or when the value of in-house tokens becomes worthless.
There was another issue with how to determine the market price of an in-house token. For example, when an in-house token was used as collateral, how should an entity determine its fair value? Accurate pricing could have prevented entities from overestimating the value of their collateral and liquidity, especially if there was no secondary market for any loans. For example, FTX used FTT as collateral to obtain billions of dollars of loans, which left FTX with few sources of funding when the price of FTT collapsed (Vicky Ge Huang, “Crypto Firms’ Tokens Scrutinized,” Wall Street Journal, February 13, 2023, p. B9).
In fact, to determine an appropriate input is quite a challenge, since in-house tokens do not necessarily have an active market. The Celsius Examiner’s Report (2023, p.9) stated: “CEL had limited utility … because there was no market to deploy CEL outside of Celsius’s platform. In 2022, Celsius employees routinely discussed that CEL was ‘worthless,’ stating that its price ‘should be 0,’ and that Celsius should ‘assume CEL is $0 since we cannot liquidate our current CEL position.’”
Inappropriate related-party accounting.
Some crypto entities have many affiliates or related parties under the same group, and thus huge related-party transactions. For example, on FTX’s 2021 year-end balance sheet, the largest asset was a “related-party receivable” valued at $1.2 billion, which accounted for 44% of its assets. FTX’s liabilities included a $362 million “related-party payable.” (Jonathan Weil, “FTX Disclosed Related-Party Transactions but Didn’t Name Names,” Wall Street Journal, Nov. 18, 2022, https://tinyurl.com/5x3a6du2).
Individually or collectively, the items discussed above would lead a crypto entity’s leverage risk to be hidden, and furthermore its liquidity risk and price risk would be obscured.
The same issues that affect transactions with third parties also apply to related-party transactions. For example, Alameda Research LLC, FTX’s related party, used FTX’s native token, FTT, as collateral to get hundreds of millions in loans. It is further stated in the SEC complaint against FTX officers that “Those loans were backed in significant part by Alameda’s holdings of FTT—an illiquid crypto asset security that was issued by FTX and provided to Alameda at no cost” (Securities and Exchange Commission v. Caroline Ellison and Zixiao “Gary” Wang, U.S. District Court S.D.N.Y., Case 1:22-cv-10794, Document 1, filed Dec. 21, 2022. https://tinyurl.com/3savd59x, p. 3). A lack of industry-specific accounting might have contributed to the transfer of FTT from FTX to Alameda at no cost since an in-house token is an intangible asset with no cost basis. Off-balance sheet treatment is a disputable accounting treatment complicated by the fact that an entity’s crypto lending was conducted through another related party. The leverage would be hidden unless there were a consolidated balance sheet on the ultimate parent’s level.
Intercompany transactions should be conducted at arm’s length basis based on relevant supporting documents. The risk opacity issue could be exacerbated by the inherent risk of related-party transactions, such as a lack of supporting documents and transacting at non-arm’s length basis. The first FTX Report states: “transfers of funds among those entities were not properly documented” (2023, p. 15). In addition, the amounts “due to” and “due from” FTX did not balance across the relevant entities (2023, p. 17). Related parties conducting transactions on a non-arm’s length basis might contribute to such imbalance. For example, as stated in the SEC complaint against FTX officers, “Assets and liabilities of all forms were generally treated as interchangeable, and there were insufficient distinctions between the assignment of debts and credits to Alameda, FTX, and executives” (2022, p. 19).
Netting of loan receivables and obligations.
Collateral might be rehypothecated when the borrower gives the lender rights to use the collateral. In the traditional market, a reporting entity may offset receivables and payables under an enforceable master netting arrangement, or when a right of setoff exists as defined in ASC 210. But from an accounting perspective, cryptocurrency is different from a security. Therefore, a crypto lender might use offsetting accounting regardless of whether the conditions to qualify had been met.
In the above example, if Company X used collateral posted by Company Y to lend again, the debit side of Entry II, “virtual currency collateral received,” would be removed from the balance sheet. Though there would be a new cryptocurrency loan produced, the new loan asset might be offset with the credit side of Entry II, the liability “Obligation to return collateral.” This would result in only the net amount of many transactions being reflected on the balance sheet. The complex offsetting process could further mask which receivable offsets which obligation. Consequently, excessive leverage and inadequate collateral information would be hidden on the balance sheet.
Offsetting alone could cause disaster if no adequate controls are in place. This is because rehypothecation creates interdependent exposure that could create market panic and systemic risk. A sudden price fluctuation could lead to a speedy decline in collateral value and trigger a margin call. For example, failing to meet margin calls from its lenders contributed to Three Arrows Capital’s collapse, and its bankruptcy filing triggered a downward spiral that troubled many crypto investors (MacKenzie Sigalos, “From $10 Billion to Zero: How a Crypto Hedge Fund Collapsed and Dragged Many Investors Down With It,” CNBC, July 11, 2022, https://tinyurl.com/nh7aj95b). Netting would not only prevent an entity from keeping track of its leverage risk but also conceal rehypothecation abuse, given that there are no specific limits on how much collateral can be rehypothecated, unlike in the traditional market.
Revealing Hidden Risks
Individually or collectively, the items discussed above would lead a crypto entity’s leverage risk to be hidden, and furthermore its liquidity risk and price risk would be obscured. For example, when an entity records crypto assets as intangible assets, it can record gains only when selling those assets. Cost information could not show how volatile the price is. In addition, stakeholders would not know the entity’s aggregate exposures when some crypto assets are recorded at cost as intangible assets, and others are booked at fair value, analogous to the accounting for securities held for trading. In addition, when an entity manages huge digital asset positions, the asymmetry of liquidity is dangerous. For certain loans, borrowers have a prepayment option and lenders have a call option at any time. To manage mismatches of liquidity risk and duration, fair value would provide more accurate information.
Finally, because assets denominated in a virtual currency were highly sensitive to the fluctuation of its price, an entity’s viability and profitability were contingent upon that volatile price. The profits generated from earning assets include the bid-offer margin that originated from trading activity; net interest income, or the spread between lending and borrowing, generated from a lending business; and sales of digital assets. However, the price risks could not be easily gleaned through a crypto entity’s financial statements, which only reflected cost information.
In the aftermath of crypto turmoil, FASB, SEC, and AICPA opined on certain accounting issues. The above intangible assets treatment issue has been addressed by FASB. FASB’s guidance will require an entity to measure certain crypto assets at fair value and recognize fair value changes in net income. But this fair value treatment applies only to certain crypto assets; it does not appear to cover crypto lending activities and in-house tokens.
In December 2022, at the AICPA & CIMA conference on “Current SEC and PCAOB Developments,” the SEC staff described an example of a basic crypto lending transaction and the staff’s views on accounting and disclosure. Based on the example from the AICPA Practice Aid, a lender should derecognize the lent crypto assets because it no longer controls them and recognize a crypto asset loan receivable at the fair value of the lent crypto assets initially and subsequently (2023, p.25). This example did not, however, require the borrower to post collateral. The SEC and AICPA have not opined upon the accounting for collateral received, which was a leading issue during the crypto turmoil. Many crypto lenders and investors were saddled with huge losses as loan collateral plunged in value.
Former FTX CEO Sam Bank-man-Fried blamed mistaken accounting of the exchange’s liquidity and leverage for the company’s collapse (Stephen Foley, “FTX Collapse Puts Focus on Digital Audit Groups,” Financial Times, Nov. 14, 2022, p. 6). Many factors contributed to the hiding of financial risks during the crypto turmoil, but blended accounting played a critical role. The author believes that a lack of industry-specific accounting principles for the crypto sector is partly responsible for the 2022 crypto turmoil, and piecemeal accounting guidance from FASB, SEC, AICPA cannot fully address the risks posed.
Applying industry-specific accounting consistently within crypto could help address global regulators’ concerns about identifying and quantifying the risks to financial stability from crypto assets (FSB, 2022, p. 18). Not only does the reporting entity itself need high-quality financial information to perform effective risk management, but its customers, investors, bank partners, and global regulators also need high-quality reporting to be prepared for a triggering event.
Accounting standards have struggled to keep pace with the ever-evolving crypto. As the sector continues to expand, for example, into the derivatives market (Philip Stafford, “Digital Currency Industry Makes Push into Regulated Derivatives,” Financial Times, Feb. 22, 2022, https://tinyurl.com/vuawvb4d), more research and analysis is needed to inform the decisions to be made by regulators and standards setters.



























