DeFi lending pools pose tax avoidance risks as users delay sales

Article by: Andjela Radmilac

Compiled by Saoirse, Foresight News

Imagine an investor who bought Ethereum for $1,000, and its price later rose to $4,000. Now, the investor wants to cash out $1,000. If they sell one-quarter of their Ethereum, they will receive cash, but under U.S. tax rules for investment digital assets, this will trigger a $750 capital gain.

However, DeFi offers an alternative solution: holders can deposit their entire Ethereum as collateral into a lending protocol and borrow $1,000 worth of a USD-pegged stablecoin.

This loan is not considered taxable income, and the Ethereum holder can still benefit from any future price appreciation; the holder can obtain spendable, fiat-convertible funds without selling the original asset.

While this approach can help holders save a significant amount in taxes, it also introduces vulnerable risks. This $1,000 loan initially has a loan-to-value ratio of 25% (with collateral valued at $4,000); if the price of Ethereum drops to $2,000, the loan-to-value ratio doubles to 50%, and with ongoing interest accruing on the debt, this ratio will rise even further.

When the ratio exceeds the protocol's upper limit, the code initiates collateral liquidation: external traders can repay part of the loan to acquire a portion of Ethereum at a discount.

Although the borrower has delayed the taxable sale, the lending pool bears the risks associated with collateral price fluctuations, debt size, and whether the borrower is willing to voluntarily close their position before liquidation.

An individual's tax decision effectively becomes part of a shared credit market funded collectively by other users, where most lenders see only a string of wallet addresses.

Lisa De Simone from the University of Texas at Austin, Peiyi Jin from the National University of Singapore, and Daniel Rabetti from the National University of Singapore examine this relationship in a working paper on tax planning and DeFi credit risk.

The subject of study is Venus, a DeFi lending protocol on the BNB Smart Chain, which uses smart contracts to enforce rules and allows users to stake crypto assets and borrow other tokens.

Study period: November 12, 2020, to July 31, 2022, covering 15 major tokens on the Venus platform. The study encompasses approximately 13 million transactions and generates 1.36 million daily borrower observations (one record per active wallet per day). Approximately 3% of traders experienced default as defined in the paper.

Default in DeFi differs from mortgage delinquency and requires clarification. The paper defines default as a loan whose loan-to-value ratio remains above Venus Protocol’s 60% threshold for at least seven consecutive days, during which the borrower neither makes additional deposits nor takes out new loans; the total accumulated default exposure amounts to $133.34 million. A risky loan is counted repeatedly across multiple reporting dates, so this total represents the cumulative daily exposure, not the actual principal loss from a single event.

The traditional version of this strategy is called "Buy, Borrow, Die." Investors purchase assets, wait for them to appreciate, and then use the assets as collateral for loans to fund their lifestyle—without ever selling the assets or realizing gains.

Continuous borrowing can defer capital gains taxes for many years; U.S. estate-related regulations stipulate that heirs receive a stepped-up cost basis upon inheriting assets, erasing most of the tax liability on gains accumulated during the original holder’s ownership period.

In the past, this was an exclusive practice for the wealthy: private banks served only high-net-worth clients with sufficient collateral and strong resilience to market downturns. Banks could review the client’s overall financial situation, determine loan limits, negotiate terms, and work together to mitigate risks before collateral was forced to be liquidated.

DeFi completely transforms this business, which previously relied heavily on manual due diligence, into code. The software does not assess the overall qualifications of customers; it simply reads the assets in wallets and applies the same collateral rules to all users.

The service barrier has been significantly lowered, but at the cost of building the entire system on over-collateralization: borrowers must initially pledge assets worth more than the loan amount.

According to the Venus parameters described in the paper: up to $6,000 in debt can be borrowed against $10,000 in eligible collateral. Borrowing the full amount leaves almost no price decline buffer; borrowing only $2,000 provides a substantial safety margin. Both accounts are continuously monitored by code via oracle-reported market prices.

When collateral depreciates beyond the threshold, liquidators can repay part of the debt, acquire the collateral at a discount, restore the account, and earn a reward. This mechanism is designed to protect the pool before collateral value falls below the debt level. However, rapid asset sell-offs and insufficient market liquidity can undermine liquidation effectiveness, and blockchain network congestion may prevent liquidators from executing operations in time.

Tax motivations further complicate borrower-side risk: reducing risk requires executing trades, repaying debt, or selling part of an appreciated position.

Borrowers who take out loans to defer taxable transactions often delay closing their positions; this tendency becomes more pronounced the higher the unrealized gains or the closer the holding period is to qualifying for lower long-term capital gains tax rates.

Stablecoins greatly enhance the appeal of this trading strategy. Dollar-pegged tokens convert volatile collateral assets into usable dollar purchasing power. Traders can continue pledging assets like Ethereum and borrow USDT or USDC for other purposes.

Large holders are using their crypto assets as collateral to borrow stablecoins, gaining liquidity while maintaining price exposure to their underlying assets.

At the time the loan is originated, the collateral ratio displayed in the agreement appears normal. However, the agreement cannot account for factors such as the borrower’s originally very low cost basis in Ethereum, the substantial profits that could be realized upon selling, or the significant tax benefits that could be gained by holding for a few more months. These factors can significantly influence the borrower’s behavioral decisions when the loan becomes at risk.

Researchers need to distinguish between tax-driven behaviors and intrinsic market fluctuations in the cryptocurrency space; the Infrastructure Investment and Jobs Act, effective November 15, 2021, provides an opportunity for study.

Section 80603 of the bill expands the information reporting obligations for digital asset brokers. Traders anticipate that their significant on-chain activities will soon be reported to the Internal Revenue Service (IRS).

The bill alters traders' expectations regarding third-party tax reporting, constituting an external shock event: the behavior of potential U.S. taxpayers will be affected, while international users remain unaffected by this policy change.

This tax reporting system took a long time to implement. Starting January 1, 2025, custodial brokers must report gross proceeds from buy, sell, and exchange transactions using Form 1099-DA; subsequent IRS rules further require that, beginning January 1, 2026, the cost basis of certain transactions must also be reported.

The above regulatory rules apply only to institutions that actually hold users' assets; non-custodial DeFi services are currently not covered by regulation.

However, the research value of this paper stems from market participants' expectations at the moment the bill was enacted in November 2021—namely, that future transaction records would be reported, rather than waiting until official tax forms were finalized.

Researchers compared data before and after the bill was enacted (long before formal regulation took effect) to capture market reactions to the prospect that future trades would be subject to regulatory oversight, rather than passive responses to already filed tax forms.

The blockchain itself does not disclose users' nationality or tax residency status; researchers can only infer which wallets are likely to belong to U.S. users based on characteristics such as transaction activity concentrated during U.S. business hours, unusual behavior during U.S.-specific holidays, and holding U.S.-regulated dollar-stablecoins.

Each inference dimension carries a potential for misjudgment; therefore, the paper provides multiple calculation methods and employs strict criteria combining several conditions.

The entire research logic is equivalent to: the same financial system experiencing a policy event shock. Two groups of users face the same token price movements and identical Venus protocol rules; only the presumed U.S. user group is highly attentive to this newly reported regulation. This allows us to isolate behavioral changes caused by tax implications from market-wide price fluctuations.

Main model estimation results: After the law takes effect, borrowers presumed to be associated with the U.S. experience a 24.5% reduction in the probability of asset trading compared to international users. Borrowers using stablecoin liabilities see an additional 23% decline in trading activity. This aligns with the paper’s logic: users rely on stablecoins to obtain cash while keeping their collateral assets pledged.

In this article, "liquidity" is defined narrowly at the wallet level: the probability that a borrower executes any asset transaction on a daily basis. Common public understanding of liquidity typically refers to exchange depth, bid-ask spreads, and the cost of large sell pressures; however, this study measures user trading activity and whether appreciated assets are "locked up."

Borrowers with higher book gains and higher loan-to-value ratios exhibit these behavioral patterns more prominently. In December each year—especially during the final week—investors commonly defer gains to the next tax year, leading to a decline in trading activity; trading activity then rebounds once holdings have been held for a full year, qualifying for the lower U.S. long-term capital gains tax rate. These behaviors confirm that the observed changes are driven by tax motivations, not merely coincidental effects of legislative deadlines.

Researchers estimate that U.S. borrowers in the sample defer capital gains taxes by an average of $3,357.42 per year, approximately 17% of their portfolio size during the same period. This estimate is based on the assumption that the wallets belong to U.S. taxpayers, with holdings reconstructed from on-chain data and applicable tax rates applied; it represents only a rough order of magnitude for the sample as a whole.

The final part of the study analyzes how reduced trading activity affects loan quality. Borrowers with low trading frequency tend to maintain high-risk positions for longer periods, miss repayment opportunities, and fail to add collateral before ratios reach critical thresholds. Tax preferences that originally occurred outside Venus ultimately translate into outstanding debt within the protocol.

There is a risk of reversed causality: default itself can lead users to abandon their wallets and cease trading.

Researchers used instrumental variable analysis to isolate the decline in trading activity caused by legislative shocks (an external factor independent of the protocol itself), thereby eliminating reverse causality bias.

Model estimates: A 1% increase in illiquidity due to tax factors leads to an 11.2% rise in the number of defaulting accounts and a 39.6% increase in the volume of defaulted loans. A one-standard-deviation increase in illiquidity corresponds to an additional $350 in default debt per borrower, or 2.7 times the model’s baseline value.

The percentage appears high due to statistical limitations: the results apply only to the subset of the sample affected by this policy event and sensitive to taxation, and cannot be generalized as a multiplier for all DeFi loans—they are intended solely to explain how motivations influence credit outcomes.

This reveals inherent blind spots in automated lending. Smart contracts read collateral prices, debt balances, accrued interest, and liquidation thresholds, but completely ignore the borrower’s cost basis and tax motivations.

Two wallets with identical Ethereum collateral and loan terms are viewed as having exactly the same risk by the Venus system; however, one borrower is willing to sell assets to reduce risk, while the other goes to great lengths to avoid selling. This creates a borrower selection bias issue.

Overcollateralization can withstand ordinary price fluctuations; however, borrowers who are extremely reluctant to liquidate appreciated assets will continue to maintain their debt and reduce trading as the safety buffer thins, concentrating risk among groups whose motivations the protocol cannot identify.

When liquidation proceeds smoothly, external participants repay debts and dispose of collateral, ensuring lenders suffer no losses. If liquidation fails, losses are absorbed collectively by the protocol reserve, token holders, and liquidity providers according to the protocol’s loss allocation mechanism. Individual tax preferences ultimately become financial consequences borne jointly by all liquidity providers.

The paper also points out that taxation is just one of many sources of risk.

This study focuses solely on a single protocol during the 2020–2022 bull-bear cycle; user U.S. identity is inferred from behavior; default is defined specially, counting only accounts with long-term high loan-to-value ratios. The study also confirms that sharp fluctuations in collateral prices and flaws in the liquidation mechanism similarly contribute to Venus loan issues.

Other protocols with greater liquidity and different collateral parameters will yield different results; the performance in a market environment with more sophisticated liquidation bots will differ from that of Venus during the sample period.

Even so, this paper provides a perspective that traditional lending data cannot achieve: complete observation of the entire process—collateral, debt, user behavior, liquidations, and abandonment—at the individual wallet level.

DeFi moves lending strategies traditionally held by private banks onto public blockchains, making them accessible to the general public. However, automating the replacement of credit evaluators does not eliminate human subjective motivations. Tax considerations, obsession with appreciating assets, and reluctance to sell are transmitted through layers of code, ultimately borne by everyone who provides liquidity to the pool.