Taxation and Planning for DeFi Liquidity Providers: Redefining Taxable Events in Liquidity Pools

I. Introduction

Decentralized finance (“DeFi”) is a market worth tens of billions of dollars that leverages blockchain technologies to offer financial services without relying on a centralized intermediary. The world of finance runs on liquidity, and DeFi is no exception. One key innovation of DeFi is a mechanism through which people can directly exchange digital assets, in the form of crypto tokens, without the need for a traditional market maker. This mechanism is known as a liquidity pool. Anyone can provide liquidity by depositing tokens into the pool, and these people are called liquidity providers. This paper discusses the federal tax consequences for liquidity providers and the viability of possible tax planning strategies.

There are no easy ways for the IRS to identify transactions in liquidity pools at scale and to tax liquidity providers who choose not to self-report. Neither the legislative nor the judicial branch has issued specific guidance on the taxation of liquidity providers.1 Academic discussions on this topic are also scarce.2 Consequently, despite significant transaction volumes, many liquidity providers remain under-taxed. A recently proposed regulation could alter this landscape. In August 2023, the IRS and Treasury proposed to expand the definition of “brokers” to include anyone who provides facilitative services to effectuate sales of digital assets.3 If this proposed regulation becomes law, certain platforms hosting liquidity pools will be treated as brokers and will have to report transactions to the IRS. As a result, it has become increasingly important to develop sound policies to tax liquidity providers. This paper primarily focuses on liquidity pools with protocols similar to those of Uniswap v2 and v3.4

In order to determine how to tax liquidity providers, we must know when they are deemed to have disposed of their tokens. Is it when they deposit tokens into a liquidity pool, or when the pool trades their tokens to fulfill exchange orders? If the deposit of tokens is considered a disposition event, liquidity providers would be taxed on every token at the time of deposit. If not, taxation would only occur when the pool sells liquidity providers' tokens to third parties, with taxes being imposed only on the tokens sold.

This paper argues that merely depositing tokens into a liquidity pool should not constitute a disposition of property, and thus should not trigger a taxable event. Liquidity providers should be taxed on tokens actually traded by the pool, rather than tokens merely deposited into the pool. This is because a liquidity pool lacks the right to freely dispose of the deposited tokens. Moreover, treating the deposit of tokens as a taxable event would encourage transactions lacking economic substance, given that the wash sale rule in Internal Revenue Code §1091 does not explicitly apply to crypto assets.5 Nevertheless, when some tokens are traded by the pool and some are not, this paper suggests that liquidity providers cannot choose which specific units of tokens are to be taxed first. Most likely, liquidity providers have to use the FIFO method to compute gains and losses.

This paper also discusses a tax planning strategy for liquidity providers to defer gain recognition on appreciated tokens. Instead of depositing appreciated tokens directly into liquidity pools, liquidity providers could deposit borrowed tokens, pledging their appreciated tokens as collateral to secure the borrowing. This strategy involves multiple legal risks. Liquidity providers could be deemed to have transferred ownership of the pledged tokens. If the borrowed tokens are fungible with the appreciated tokens, the IRS might also challenge this transaction for lacking economic substance, and for contravening the underlying purpose of Code §1259.

Accordingly, the rest of this paper is structured as follows: Section II provides a technical overview of liquidity pools. Section III analyzes how liquidity providers should be taxed, proposing two reasons why they should be taxed on the tokens actually traded rather than on the tokens merely deposited. Section IV explains why liquidity providers cannot identify specific units of tokens for taxation. Section V examines the tax planning strategy of depositing borrowed tokens while pledging appreciated tokens as collateral. Section VI concludes the paper.

II. How Liquidity Pools Work — A Technical Overview

To enable people to exchange crypto tokens, a liquidity pool has to have sufficient quantities of different tokens to fill exchange orders. In a centralized exchange, liquidity is provided by large institutions and market makers. DeFi platforms, however, do not have such intermediaries, by definition. Hence, liquidity pools have to rely on ordinary people, i.e. liquidity providers, to supply the tokens that fill exchange orders.

Why would anyone want to be a liquidity provider? When traders use a liquidity pool to exchange crypto tokens, they have to pay transaction fees to the pool. A portion, if not all, of these fees goes to liquidity providers as a reward for making their tokens available for trading. These transaction fees attract people to deposit tokens into liquidity pools and hence become liquidity providers. While transaction fees accumulate as liquidity pools sell liquidity providers’ tokens to external traders, liquidity providers can only claim their share of the fees when they cash out, i.e., when they withdraw tokens from the pool.6

Some liquidity pools allow liquidity providers to customize a price range, within which trading is allowed. For example, suppose there exists a liquidity pool for exchanging token A and token B. A liquidity provider deposits some tokens into the pool and specifies a price range of (0, 3]. If the exchange rate between token A and B exceeds 3, the pool cannot sell this liquidity provider’s tokens to fulfill exchange orders, until the exchange rate falls back within her specified range. For pools that do not support a customized price range, the default price range is usually (0, ∞). Importantly, even if the exchange rate falls within a liquidity provider’s price range, it does not necessarily mean that the liquidity pool can freely sell her tokens. The following example illustrates why a liquidity pool’s ability to trade tokens is limited by nature. For the sake of simplicity, I use a liquidity pool that follows the Uniswap v2 Protocol, where the price range is set at (0, ∞) for all liquidity providers.

Example: Suppose there exists a liquidity pool for exchanging token A and token B. The pool’s only liquidity provider, Alice, deposits 100 of token A and 100 of token B into the pool. Recall from Section I that liquidity pools use mathematical formulas to adjust the exchange rate after each trade. Our pool uses a popular formula called Constant Function Market Maker (CFMM), which states:

In the above formula, $x$ refers to the quantity of token A in the pool and $y$ refers to the quantity of token B in the pool.7 In simple terms, the formula requires that the product of each token’s quantity should always equal a constant value k. In our example:

Under the Uniswap protocol, the pool presumes to hold the two tokens in equivalent value.8 Hence, the exchange rate that the pool offers to external traders can be determined by the ratio of quantities of the two tokens in the pool. Specifically, the exchange rate from token A to token B (that is, the amount of token B per token A) can be calculated using the following formula:

In our example, the initial Exchange Rate _A/B is 100 / 100 = 1.

After this transaction, the pool will offer a higher Exchange Rate _A/B to other traders. Intuitively, this is because the pool sees more demand for token A and less demand for token B. Mathematically, the new Exchange Rate _A/B after the transaction becomes 1.5625, because:

Seeing that the pool offers a higher Exchange Rate _A/B , other traders may sell token A and buy token B from the pool, which would drive Exchange Rate _A/B downward. Assume that the highest Exchange Rate _A/B ever reaches is 1.5625 and the lowest Exchange Rate _A/B ever reaches is 0.64.9 Before Alice withdraws tokens from the pool, the Exchange Rate _A/B keeps fluctuating between the upper bound of 1.5625 and the lower bound of 0.64, but never escapes this range. The graph below summarizes our assumption.

When the Exchange Rate _A/B is at its highest, there are still 80 of token A in the pool that remain untraded. Similarly, when the Exchange Rate _A/B is at its lowest, there are still 80 of token B that remain untraded. Regardless of how many trades happen through the pool, as long as the Exchange Rate _A/B stays within the bound, these 160 tokens simply sit in the pool idly, waiting for the liquidity provider to withdraw them. Naturally, we would ask: why would the pool not trade these 160 tokens?

The pool cannot trade these 160 tokens, because Exchange Rate _A/B has not covered the entire price range, which is (0, ∞) in our example. To provide more context, liquidity pools use ticks to measure changes in the exchange rate, with each tick representing a discrete and narrow increment in exchange rate (e.g., from 1.0001 to 1.0002).10 When liquidity providers deposit tokens into the pool, their tokens are distributed across all ticks within the entire price range. A token will be available for trading if the exchange rate crosses the tick to which the token is assigned.11 In practice, the exchange rate typically fluctuates within a limited range and rarely reaches extreme levels, such as 0 or infinity. Thus, a portion of the tokens in the pool may never be traded, since they are assigned to ticks that are not crossed.

In our example, the 200 tokens Alice deposits are distributed across the entire price range of (0, ∞). However, the Exchange Rate _A/B fluctuates only within the bound of [0.64, 1.5625]. Since the Exchange Rate _A/B never exceeds 1.5625 before Alice withdraws her tokens, 80 of her token A assigned to ticks in the range (1.5625, ∞) would not be traded. Similarly, since the Exchange Rate _A/B never goes below 0.64 before Alice withdraws, 80 of her token B assigned to ticks in the range (0, 0.64) would not be traded either. 12 By analogy, Alice can be seen as placing a sell limit order on each deposited token, with the limit price being the boundary of the tick to which the token is assigned. The pool sells a token only when its limit price is reached. This point will be further illustrated in the next section, which discusses liquidity providers’ tax obligations.

III. How Liquidity Providers Should Be Taxed — On Tokens Traded, Not on Tokens Deposited

Pursuant to Code §1001, a realization event occurs when there is a sale or other disposition of property. 13 Crypto tokens, as a form of virtual currency, are treated as property for federal tax purposes.14 Hence, absent any non-recognition safe harbors, liquidity providers will be taxed when they dispose of the tokens that they deposit into a liquidity pool.

Thus, the key question is when liquidity providers are deemed to have disposed of their tokens. Should the act of depositing tokens be viewed as a disposition event, liquidity providers would be taxed on every token at the time of deposit. However, this paper contends that the mere act of depositing tokens does not constitute a disposition of property and thus should not lead to a taxable event. Liquidity providers should be taxed on the tokens actually traded by the pool, not the tokens merely deposited. To support this proposition, this paper presents two arguments: (1) liquidity pools cannot freely dispose of the deposited tokens;15 (2) taxing untraded tokens disregards economic substance and promotes wash sales.16

A. Liquidity pools cannot freely dispose of the deposited tokens

To determine whether a disposition of property has occurred, courts and the IRS generally look at whether the ownership of the property has transferred.17 The primary factor in deciding ownership transfer is whether the transferor has granted the transferee an unrestricted right to further dispose of the property.18 If the transferee can freely dispose of the property, then the transferor has effectively transferred ownership and should be taxed on the transfer.

The importance of the right to dispose of property has been underscored in an extensive body of case law. In the Supreme Court case Provost v. United States, the lending of stock to a broker—who was contractually allowed to short sell the stock—was deemed a taxable event.19 The Court reasoned that since the broker could short sell the stock, the broker must have "[acquired] dominion over [the] stock… with unrestricted power of disposition."20 As a result, despite the broker’s contractual obligation to return the stock at a later date, the Court found that the lender had transferred stock ownership to the broker and effectively disposed of the stock.21 The same reasoning was applied in Hall v. Commissioner by the tax court and the Ninth Circuit.22 The key question in Hall concerned the timing of ownership transfer for shares issued in the petitioner's name—whether ownership transferred in 1942 when the shares were issued on paper but retained by the petitioner’s company, or in 1943 when the shares were physically delivered to the petitioner. The tax court concluded that "the unfettered right of sale is one of the most important attributes of ownership."23 Since the petitioner was not able to sell the shares while they were held by the company's treasurer, he was deemed not to have owned the shares until 1943 and would not be taxed until then.24

On the other hand, if a transferor simply deposits property into the transferee’s account without granting the transferee the right to freely dispose of the property, the ownership of the property remains with the transferor, and no taxable event has occurred.25 This principle is underscored in Rev. Rul. 57-451, in which the IRS examines whether a disposition event occurs in three distinct scenarios.26 In the first scenario, a stockholder deposits stock into an agent’s account, allowing the agent to sell the shares only upon receiving further instructions. In the second, the stockholder deposits stock with the broker, authorizing the broker to lend out the stock certificates to other customers as part of the broker’s regular business activities. The third scenario is the same as the second, except the stockholder does not authorize the broker to lend out the stock certificates.27 The IRS determines that in the first and third scenarios, the stockholder retains ownership of the stock, because the disposal of the stock is contingent upon the stockholder’s directions. In the second scenario, however, ownership transfers from the stockholder to the broker. By lending out the stock to its customers in the ordinary course of business, the broker exercises full discretion to dispose of the stock and no longer acts as the stockholder’s agent.28 In all scenarios, the right to dispose of property serves as the key criterion in determining whether the ownership of the property has transferred and whether a taxable event has occurred.

At first glance, providing liquidity may appear more analogous to Provost and Hall than to the first and third scenarios in Rev. Rul. 57-451. When a liquidity provider deposits tokens into the pool, the tokens move from the liquidity provider’s blockchain wallet to that of the pool. Their ownership seems to have been transferred to the pool. However, a closer examination of how liquidity pools operate points to the opposite conclusion—the mere act of depositing tokens does not transfer ownership of the tokens to the pool, since liquidity pools lack the right to freely dispose of the tokens in them. Section II has demonstrated this point through a technical lens. The next three paragraphs further illustrate the reasoning in three separate scenarios.

First, a liquidity pool has no right to trade a liquidity provider’s tokens if the tokens’ exchange rate falls outside her specified price range. Recall from Section II that some pools allow liquidity providers to customize a price range within which trading is allowed. For example, suppose there exists a liquidity pool for exchanging token A and token B. A liquidity provider named Alice deposits 100 tokens into the pool and sets the price range to be [1.5, 2], meaning that the pool can only trade her tokens when the exchange rate is between 1.5 and 2. Now suppose the exchange rate never reaches 1.5. The pool cannot trade any of Alice’s tokens. Upon withdrawing liquidity from the pool, Alice would receive 100 tokens, the same amount as she initially deposited. By analogy, Alice can be seen as placing a sell limit order on each token deposited, with limit prices ranging from 1.5 to 2. None of her limit orders are executable, since the exchange rate does not even reach the lowest limit price. The pool simply serves as a holding vehicle for Alice’s tokens, eventually returning the untraded tokens back to her.

Next, even if the exchange rate falls within a liquidity provider’s price range, the liquidity pool still cannot freely trade all her tokens, unless the exchange rate spans across the entire price range. Recall from Section II that a price range is subdivided into ticks, which are discrete and narrow price increments. The deposited tokens are distributed across all ticks within the price range. Each token becomes available for trading when the current exchange rate enters the specific tick assigned to that token.29 Consequently, unless the exchange rate traverses the entire price range, some ticks remain unactivated, and their associated tokens remain unavailable for the pool to trade. For example, suppose liquidity provider Alice allows the pool to trade her tokens within the price range of (1, 2). From the time of her deposit until her withdrawal, the exchange rate fluctuates between 1 and 1.5. Thus, only her tokens assigned to ticks in the range of (1, 1.5) would be traded, but the pool could not trade her tokens assigned to ticks in the range of (1.5, 2). Alice’s situation is akin to that of the stockholder in Rev. Rul. 57-451 (scenarios one and three). She does not grant the pool an outright permission to trade all her tokens. Instead, she instructs the pool when her tokens can be traded. Applying the “limit order” analogy, only part of Alice’s limit orders are executed, and only that part should be subject to taxation.

Third, even if the exchange rate does span the entirety of a liquidity provider’s price range and thus allows the pool to trade all her tokens, this outcome is unpredictable at the time of deposit and usually happens gradually over time. Thus, the tokens should still be taxed upon trading rather than upon deposit. To reason with an analogy, suppose Bob instructs his broker to sell 100 shares via limit orders, each at a different limit price. Even if all of Bob’s limit orders are eventually executed, these orders are still taxed at the time of execution rather than at the time of placement. This is because Bob did not grant his broker the right to dispose of his shares at the outset. Bob’s act of placing 100 limit orders is similar to Alice’s depositing 100 tokens. Whether each share or token will be sold depends on whether a specific price condition is met. Just as limit orders are taxed at the time of execution, liquidity providers should be taxed when tokens are actually traded by the pool.

Given liquidity pools’ constrained ability to trade tokens, one way to characterize the relationship between liquidity pools and liquidity providers is through the common law theory of bailment—to the extent that a token is not available for trading, a liquidity pool simply holds the token on behalf of the liquidity provider in a bailee capacity. In common law, the doctrine of bailment refers to transactions in which one person (the bailor) delivers property to another person (the bailee) with the mutual understanding that the property will later be returned to the bailor.30 The key factor in determining whether a bailment arrangement exists is whether the bailee is obligated to return the property to the bailor. Although a bailee often has no right to dispose of the property, courts may still recognize a bailment if the bailee’s right to disposal is limited. In Miami National Bank v. Commissioner, a taxpayer transferred stock to a broker to be held in a subordinated account.31 Under the subordination agreement, while the taxpayer could withdraw the stock at the maturity date, the broker was also authorized to sell the stock to satisfy the claims of its creditors.32 Despite the broker’s right to sell the stock in certain situations, the tax court still found a bailment arrangement between the broker and the taxpayer.33 In doing so, the court acknowledged that the bailee might be given the power to dispose of the bailed property under specified conditions.

A liquidity pool holds the untraded tokens in a bailee capacity. Similar to the broker in Miami National Bank, the pool has a limited right to dispose of the bailed property. It can trade a token when the exchange rate crosses the tick assigned to that token. This limited right, however, does not negate the pool’s bailee status, since the pool still has to return the untraded tokens to liquidity providers (even if the tokens returned might not be the exact tokens deposited, due to tokens’ fungibility). In the landmark case Coggs v. Bernard, Justice Holt laid out six categories of bailment, among which include a pledge of property. Liquidity providers can be seen as pledging to the pool the maximum number of tokens they are willing to trade. The pool may not trade all pledged tokens and will return the untraded tokens in proportion to liquidity providers’ contributions. Concededly, in the event that liquidity pools become insolvent, there is no contractual guarantee that liquidity providers would get their tokens back. The securities bar has also been debating whether crypto platforms hold clients’ assets as custodians or general creditors.34 Nevertheless, from a tax law perspective, as long as a liquidity pool cannot freely dispose of the tokens in it, it is not the real owner of the tokens.

There is one counterargument that needs to be addressed. When liquidity providers withdraw tokens from the pool, a portion of the withdrawn tokens might not have been traded by the pool. But even for these untraded tokens, liquidity providers do not necessarily receive the exact tokens that they deposited. This is because liquidity pools treat crypto tokens as fungible units and do not distinguish between which units are withdrawn by which liquidity provider. Given that, a legal formalist may argue that liquidity providers must recognize gain or loss on every deposited token, even if some tokens are not traded by the pool. This argument, however, can be countered by Treas. Reg. §1001-1(a). As per Treas. Reg. §1001-1(a), an exchange would not result in a taxable event if the two properties exchanged are not materially different.35 According to the Supreme Court's decision in Cottage Savings Association v. Commissioner, two properties are materially different if they embody legally distinct entitlements.36 Generally, each unit of token in a liquidity pool embodies the same legal entitlements. Fungible crypto units possess identical economic value and redemption right, similar to fungible stock units held in street names. Since token units are not materially different from each other, under Treas. Reg. §1001-1(a), an exchange between them should not constitute a taxable event. Consequently, liquidity providers should not be taxed on tokens that are not traded by the pool, even if the untraded tokens they receive upon withdrawal are not the exact ones they deposited.

B. Taxing untraded tokens disregards economic substance and promotes wash sales

Tax laws are to be applied with an eye to economic realities.37 Liquidity providers bear the economic risks and rewards of owning the tokens that the pool has not traded. If depositing tokens into a pool were considered a transfer of ownership and thus a taxable event, it would encourage liquidity providers to engage in transactions without economic substance, using liquidity pools as a venue for wash sales of crypto tokens.

To ensure that a disposition of property has economic substance, courts and the IRS frequently assess whether the economic risks and rewards of owning the property have transferred. In Lorch v. Commissioner, the tax court held that the pledgor remained the owner of the pledged property, since he “retained the benefit of any increases and bore the risk of any decreases in value.”38 Similarly, in memorandum AM 2007-004, the IRS evaluated whether a contract for a future sale of shares, coupled with a share lending agreement, constitutes a current sale of the shares.39 The memorandum concluded that the seller had effectively sold the shares to the purchaser, noting, among other factors, that the purchaser had the right to most of the gain and bore all the risk of loss.40 In Anschutz Co. v. Commissioner, the court held that the taxpayer transferred ownership of his stock and should be taxed on the transfer. Among other things, the court noted that the taxpayer transferred all risk of loss and a major portion of the opportunity for gain under a stock purchase agreement, thus significantly reducing his economic exposure to the stock.41 The court in Samueli v. Commissioner reached the same conclusion, holding that the transfer had reduced the taxpayer’s opportunity for gain and risk of loss.42 By contrast, liquidity providers assume the economic risks and rewards associated with the tokens that have not been traded by the pool. First, liquidity providers can withdraw these tokens at any time. There are no special contractual obligations they have to meet before exercising their right to withdraw. Second, any increase or decrease in the value of these untraded tokens will ultimately be borne by liquidity providers. This stands in contrast to the taxpayers in Anschutz and Samueli, who significantly reduced their economic exposure in the transferred properties.

Requiring liquidity providers to recognize gain or loss on every deposited token would promote transactions void of economic substance. For example, suppose a liquidity provider named Alice holds tokens with large built-in losses. She could deposit these tokens into a liquidity pool and withdraw them immediately afterward. If the act of depositing tokens is considered a disposition event, Alice could claim a loss deduction without giving up her economic interest in these tokens. This is a typical example of a wash sale. Code §1091 prohibits wash sales of stock or securities, but crypto tokens do not fall explicitly within the scope of Code §1091.43 In fact, liquidity pools can be a more desirable avenue for wash-selling crypto tokens than centralized exchanges, since the latter charge higher fees. More importantly, wash sales are not the only method for liquidity providers to “harvest loss” through liquidity pools. As mentioned in Section II, some pools allow liquidity providers to set specific price ranges for trading. If the current exchange rate falls outside a liquidity provider’s price range, none of her tokens would be traded. To take advantage of this functionality, Alice could deposit tokens with built-in losses, specify a price range that is unlikely to be reached, wait for more than a month to eliminate wash sale implications, and then withdraw all her tokens from the pool. If the mere act of depositing tokens were treated as a disposition event, she could claim a free loss deduction without relinquishing any economic interests in her tokens.

Treating the deposit of tokens as a disposition event also creates tension with legislative efforts to curb crypto wash sales. Legislation has been proposed to extend the scope of Code §1091 to include crypto tokens.44 Under the proposed legislation, taxpayers are prohibited from reacquiring a substantially identical position within 30 days before or after they dispose of their crypto tokens.45 However, liquidity providers may have legitimate business reasons to withdraw tokens shortly after depositing them into a liquidity pool (e.g., to take advantage of crypto price volatility).46 If depositing tokens were considered a disposition event, so too would withdrawing tokens from the pool. This puts liquidity providers at risk of being seen as reacquiring a substantially identical position if they withdraw tokens within 30 days of deposit, even if they have legitimate business reasons for making a quick withdrawal. By contrast, if the deposit of tokens were not treated as a disposition event, as this paper proposes, liquidity providers could withdraw tokens at any time without implicating the wash sale rule. This is because they would not be able to deduct losses on tokens that are merely deposited but not traded. The unavailability of loss deductions removes wash sale implications. As a result, liquidity providers can withdraw tokens at the time that best suits their business needs, while not accidentally conducting wash sales of crypto tokens.

In summary, Section III examines how liquidity providers should be taxed. It argues that simply depositing tokens into a liquidity pool should not constitute a taxable event, and liquidity providers should be taxed on the tokens actually traded by the pool. This approach aligns tax consequences with property ownership, ensuring that taxes would only be imposed when ownership transfers. It can also effectively prevent liquidity providers from wash selling crypto tokens, while imposing no artificial restriction on when they can withdraw liquidity from the pool.

IV. Can Liquidity Providers Select Tokens for Taxation?

Liquidity providers may deposit tokens that they purchased at different times and at different prices. Their adjusted basis and the holding period for these tokens can vary significantly. Supposing a liquidity pool does not trade all deposited tokens, can liquidity providers designate which tokens are traded and thus subject to taxation? If such designation were possible, liquidity providers could choose to be taxed on tokens with a higher basis, thus avoiding recognizing gain on tokens with a lower basis.

Neither courts nor the government have provided a formal opinion on the identification of crypto tokens. The only guidance available comes from the IRS’s Frequently Asked Questions on Virtual Currency Transactions (the FAQs).47 According to the FAQs, a taxpayer may designate which units of crypto tokens are to be taxed, but only if the taxpayer can specifically identify the involved units and substantiate her basis in those units.48 As per the FAQs, there are two ways a taxpayer can make a specific identification. The first option is to document the specific units’ unique digital identifiers, such as private keys, public keys, and addresses. Alternatively, the taxpayer can provide transaction information for all units of the token held in a single account, wallet, or address. For each unit, this information must include the date, basis and fair market value at acquisition, the date and fair market value at disposal, as well as the value of property received upon disposal.49

Liquidity providers are unlikely to meet the criteria for specific identification outlined in the FAQs. First, while crypto tokens can be non-fungible, the tokens deposited in liquidity pools are fungible with each other, similar to stock units held in street names. There are no unique digital identifiers for each token. Hence, the first method of identification suggested by the FAQs is not feasible. The second method of identification is also impractical. After a liquidity pool collects liquidity providers’ tokens, it treats all tokens as fungible units and usually does not distinguish which token comes from which liquidity provider. Simply by looking at a liquidity pool’s transaction records, there is no way to tell whether and when a specific token was sold and at what price. Given that liquidity providers cannot specifically identify individual tokens, the FAQs require them to use the FIFO accounting method.50 The FIFO method can be undesirable for taxpayers who acquired tokens at a low basis in early transactions. Under FIFO, they must start recognizing gain or loss from earliest-acquired tokens, which are likely to be tokens with the largest built-in gains.

That being said, FAQs are not dispositive legal authorities, nor are they binding on the IRS unless published in the Internal Revenue Bulletin.51 Thus, a thorough discussion of the identification rules also requires examining relevant case law. The Supreme Court decision in Helvering v. Rankin established the ground rule for the identification of fungible properties in sales and exchanges.52 In Rankin, the taxpayer opened a margin account with his stock broker in 1926 and purchased 1200 shares of stock. In 1928, he engaged in multiple buy and sell transactions of the same stock, ending the year still owning 1200 shares.53 Under the Revenue Act of 1928, when shares of stock are sold from lots purchased at different dates and at different prices, and the identity of the lots cannot be determined, the shares sold shall be charged against the earliest purchases of such stock (i.e., the FIFO method shall apply).54 The key question is: could the taxpayer identify which specific shares were sold in 1928? Since the taxpayer traded his stock through a margin account, all certificates for his shares were issued under street names and commingled with other clients’ shares. Thus, identification by stock certificates was not possible. Nevertheless, the Court held that stock certificates are not the only means of identification.55 Specific identification can also be satisfied if the taxpayer has, through his broker, designated the securities to be sold as those purchased on a particular date and at a particular price.56 While the Rankin holding mainly applies to stocks, it fundamentally challenges the notion that fungible properties can be identified only by a unique identifier.

The question that naturally arises is: if a liquidity provider instructs a pool that she wishes to trade specific units of tokens purchased on a particular date and at a particular price, can she argue under Rankin that she has adequately identified the units to be traded and taxed? Unfortunately, liquidity providers cannot avail themselves of the Rankin decision. First, there is no practical way for them to communicate with liquidity pools which tokens they intend to trade first. Unless a liquidity pool actively offers this function, liquidity providers have no means to specify their preferred order of trading. Moreover, the Supreme Court’s decision in Davidson v. Commissioner established that actual delivery, not the taxpayer's intent, determines which properties have been disposed of.57 In Davidson, a taxpayer plans to sell specific shares and instructs their broker accordingly. The broker, however, ends up selling different shares by mistake.58 The Court held that the intended shares cannot be considered as sold. Taxpayers must calculate their gains based on the cost of the shares actually sold, not the shares they intended to sell.59 Similarly, even if a liquidity provider instructs a liquidity pool to trade certain units of tokens, these tokens would not be considered traded unless the pool actually trades them according to the instructions. Due to crypto tokens’ fungibility, liquidity pools cannot guarantee the delivery of specific units for any particular trade. Likewise, liquidity providers cannot prove that the units they intend to trade are actually traded by the pool. As a result, they have to use the FIFO method to calculate their gains or losses.60

Given liquidity providers’ inability to designate specific tokens for taxation, this paper suggests that they select a narrow price range to provide liquidity. This approach improves capital efficiency by reducing the percentage of untraded tokens. However, if the price range is set too narrowly, liquidity providers could miss out on opportunities to earn fees when the exchange rate exceeds their narrow range. Therefore, liquidity providers who set a narrow price range should be prepared to regularly adjust their range to stay current with the latest exchange rate.

V. Tax Planning — Depositing Borrowed Tokens while Pledging Appreciated Tokens as Collateral

Due to the volatility of crypto prices, taxpayers may hold appreciated crypto tokens with a large built-in gain. For these taxpayers, providing liquidity could incur substantial tax costs. Are there any strategies they can use to defer gain recognition? This section discusses one such strategy, in which taxpayers deposit borrowed tokens into liquidity pools while pledging appreciated tokens as collateral to secure the borrowing.

The best way to explain this strategy is through a simplified example. Suppose that a liquidity provider, Alice, holds some units of token A that have appreciated over time. If Alice were to deposit these tokens directly into a liquidity pool, she would have to recognize gain on at least some of the tokens by the time she withdraws liquidity from the pool. Instead, Alice could borrow some new units of token A from a crypto lender, pledging her appreciated tokens as collateral. She could then deposit the borrowed tokens into the pool, effectively creating a short position. When she withdraws liquidity from the pool, she could use the tokens withdrawn to repay the lender, close her short position, and retrieve the appreciated tokens she pledged as collateral. Assuming her pledging of tokens does not constitute a disposition event, Alice would not have to recognize gain on the appreciated tokens. She would only recognize gain on the borrowed tokens she deposits into the pool. That gain would be capital to the extent the borrowed tokens are considered capital assets in her hands.61 Meanwhile, pursuant to the Richardson v. Commissioner decision, Alice would only recognize gain when she closes out her short position, i.e., when she returns the borrowed tokens to the lender.62

Though seemingly attractive, taxpayers who adopt this strategy face two major legal risks. First, the act of pledging tokens could be treated as a taxable event, in which taxpayers transfer ownership of the appreciated tokens to the lender. Second, the transaction could be viewed as a constructive sale that lacks economic substance. The IRS might even apply Code §1259 constructively to require gain recognition on the appreciated tokens. If either risk materializes, taxpayers would be deemed to have disposed of the appreciated tokens and must recognize gains accordingly.

Regarding the first risk, whether pledging property as collateral constitutes a transfer of ownership depends on specific circumstances. As discussed in Section III, courts evaluate a variety of factors to determine if and when property ownership has transferred. The most important factor is which party has the right to freely dispose of the property in question. The right to dispose of property can take many forms. For example, lenders would be deemed to have the right to dispose of the collateral if they can loan the collateral to third parties. If liquidity providers allow lenders to rehypothecate their pledged tokens in any way, the ownership of the pledged tokens would likely be deemed to have transferred to the lender. Consequently, liquidity providers would have to recognize gain on the pledged tokens, negating the intended benefit of this tax planning strategy.63

Lenders' policies on collateral substitution can also influence whether the ownership of the pledged tokens is deemed to have transferred. Courts and the IRS are less likely to find a transfer of ownership if the pledged property can be substituted with cash or other assets of equivalent value.64 For instance, Rev. Rul. 2003-7 describes a case where a shareholder and an investment bank enter into a variable prepaid forward contract. Under this contract, the shareholder receives a fixed cash amount from the bank and agrees to deliver a variable number of shares at a future date.65 The shareholder pledges the maximum potential shares to be delivered to a third-party trustee, preventing the bank from disposing of the pledged shares.66 The IRS concludes that the shareholder retains ownership of the pledged shares. This is not only because the bank is unable to dispose of the pledged shares, but also because the agreement gives the shareholder an unrestricted right to reacquire the pledged securities by delivering cash or other shares of equivalent value.67 Similarly, when liquidity providers pledge appreciated tokens as collateral, they need consider whether the lender allows collateral substitution, e.g., whether the lender allows them to replace the pledged tokens with tokens of equivalent value. If the lender prohibits any form of collateral substitution, liquidity providers are at higher risk of being deemed to have transferred ownership of the collateral. In that situation, they would have to recognize gain on the collateral, again negating the benefit of this tax planning strategy.

Now shifting gears to the second major legal risk: the IRS might challenge this tax planning strategy for lacking economic substance and apply Code §1259 constructively to require gain recognition on the appreciated tokens. In this scenario, the IRS would view taxpayers as having constructively sold the appreciated tokens and argue that the rationale behind Code §1259 should apply to crypto assets equally. To better understand this risk, it is helpful to start with an overview of constructive sales.

Taxpayers have traditionally resorted to constructive sales to defer recognizing built-in gains. One popular method for conducting a constructive sale is called “short sale against the box” (SSB). One high-profile example of SSB features Estée Lauder, the founder of the world’s premier cosmetic brand, Estée Lauder Company. Using a trust as her agent, Estée Lauder borrowed 5.5 million shares of the company’s stock from her son, and sold these shares short at the company’s IPO.68 Under the tax law at that time, realization of gain does not occur until the short seller returns the borrowed stock to the lender, i.e. until the short seller closes out the short position. As a result, Lauder was not required to pay tax at the time of the short sale. If Lauder returned the borrowed shares to her son after her death, her son would inherit the stock at a tax-free, stepped-up basis, eliminating the built-in gain on the shares.69 However, the enactment of Code §1259 renders the Estée Lauder transaction no longer viable. Under Code §1259, if a taxpayer holds an “appreciated financial position” and then enters into a short sale of the same or substantially identical property, the taxpayer shall recognize gain as if such appreciated financial position were sold.70

On the surface, Code §1259 does not prohibit the tax planning strategy outlined in this section. Code §1259 requires the taxpayer to hold an “appreciated financial position,” but the definition of an “appreciated financial position” does not expressly include crypto tokens.71 This omission makes it theoretically possible for taxpayers to carry out the abovementioned tax planning strategy (i.e., to pledge appreciated tokens as collateral, borrow the same type of tokens from the lender, and provide liquidity using the borrowed tokens). In reality, however, this tax planning strategy is by no means free of legal risks.

First, legislators may expand the scope of Code §1259 to include crypto tokens in the near future.72 Even without an official expansion, the government might challenge the legitimacy of the proposed tax planning strategy with a substance-over-form argument. The argument would be that the principles underlying Code §1259 should apply to crypto assets, given the minimal public benefit in allowing liquidity providers to delay gain recognition. Even without Code §1259, the IRS may constructively apply Code §1233(b) to treat gains from a short sale as short-term capital term.73 Furthermore, this tax planning strategy may also violate the straddle rules, which prevent taxpayers who hold offsetting positions in certain assets from accelerating losses while deferring gains.74 Ultimately, it is difficult for taxpayers to prove that borrowing the same type of tokens as those they already own has business purposes other than to minimize taxes. This lack of economic substance makes the outlined tax planning strategy even more risky.

Proponents of this tax planning strategy may argue that it does not contravene the purpose of Code §1259. Legislative history shows that the reason for enacting Code §1259 is mainly to prevent taxpayers from “completely eliminat[ing] risk of loss…without disposing of the property in a taxable transaction.”75 The transactions that Code §1259 targets are those similar to the Estée Lauder transaction, which completely eliminates taxpayers’ economic interest in their appreciated financial positions. By contrast, liquidity providers still retain full economic interest in their appreciated tokens unless they default on the loan. They continue to own these appreciated tokens and bear the corresponding economic risks and rewards. Moreover, unlike Estée Lauder who had secured a lump sum payment when she short sold the borrowed stock, liquidity providers do not receive any fixed payment when they deposit the borrowed tokens into liquidity pools. The payoff from liquidity pools can be highly contingent, which distinguishes it from the risk-free transactions that Code §1259 bans. Nevertheless, liquidity providers’ tax planning strategy is still subject to Code §1259 risks. Although liquidity providers are supposed to take back the collateral after they pay off the loan, nothing prevents them from simply defaulting on the loan and surrendering the collateral when they have sufficient capital losses to offset the built-in gains in the collateral. If their plan from the outset were to dispose of the collateral, such transaction would not be much different from that of Estée Lauder.

In light of these legal risks, what can liquidity providers do to protect themselves? This section suggests an alternative tax planning approach. Liquidity providers would still deposit borrowed tokens into liquidity pools, but the tokens borrowed would be of a different type than the tokens pledged as collateral. For example, suppose there exist two liquidity pools: the first pool exchanges token A and token B (A/B pool) and the second pool exchanges token C and token D (C/D pool). Both pools offer similar rates of return. Further suppose that Alice, a liquidity provider, holds a large number of token A with substantial built-in gains. To maximize tax efficiency while reducing legal risks, Alice should avoid providing liquidity to the A/B pool, which would require a deposit of token A. Instead, Alice could pledge her tokens A as collateral to borrow tokens C and D, and then deposit these borrowed tokens into the C/D pool. Since Alice borrows different types of tokens than those she pledges as collateral, her borrowing would not be deemed a constructive sale. It would simply be an ordinary transaction, in which she uses liquid assets to secure loans. Alice’s approach can be an attractive option for liquidity providers: It is more tax-efficient than depositing appreciated tokens directly into liquidity pools. Meanwhile, it involves fewer legal risks than the tax planning strategy initially introduced, which would require borrowing the same type of tokens as the collateral.

Notably, this alternative tax planning approach is not risk-free. Borrowing a different type of token does not guarantee that the transaction has economic substance. In the world of cryptocurrency, the prices of many tokens closely track each other, or are even pegged to each other at a 1:1 ratio. Continuing with the above example, suppose that Alice only wants to provide liquidity to the A/B pool. She finds that the price of token E closely tracks that of token A. As a result, she first pledges her appreciated tokens A to borrow tokens E. She then exchanges the borrowed tokens E for new tokens A, and uses these new tokens A to provide liquidity to the A/B pool. Upon exiting, she withdraws some tokens A from the pool, exchanges them back for tokens E, repays the lender, and eventually retrieves her appreciated tokens A. Since the price of token E closely tracks that of token A, the exchanges between token A and E do not create bona fide financial risks; they are merely bells and whistles to complicate the transaction. As a result, the concerns of constructive sales equally apply to this transaction, even if the tokens Alice borrows are technically different from the tokens she pledges. In addition to concerns of constructive sales, using borrowed tokens to provide liquidity also involves inherent financial risks. Since the token mix in liquidity pools is constantly changing, liquidity providers may not get back the same number of tokens they deposit into the pool. In that case, they may have to buy additional tokens to fully repay their loans and redeem collaterals.

Finally, there are other strategies worth exploring that are not covered in this section. For example, taxpayers may consider replicating liquidity pools’ payoffs by shorting options, as the payoff of providing liquidity is akin to that of shorting a straddle or a strangle.76 Any strategy may come with legal or financial risks. Taxpayers need to carefully assess potential risks to make well-informed decisions.

VI. Conclusion

This paper discusses the tax consequences for liquidity providers in decentralized finance. It contends that the mere act of depositing tokens into liquidity pools does not constitute a taxable event. Liquidity providers should be taxed on the tokens actually traded by the pool, not on the tokens merely deposited. Nevertheless, due to the fungibility of crypto tokens, liquidity providers likely have to use the FIFO accounting method to report which specific token units are traded by the pool. Finally, this paper presents a tax planning strategy for taxpayers holding tokens with significant built-in gains and analyzes associated legal risks.

Taxation on liquidity providers must be reasonable to be effective. Given the anonymity of blockchain transactions and the decentralized nature of liquidity pools, liquidity providers can hide their transactions more easily than ordinary taxpayers. An overly harsh tax policy would discourage liquidity providers from reporting their transactions to tax authorities, resulting in lost revenue and unfair treatment for those who comply with the law.