How Accurate is DeFi Income Calculation by Crypto Tax Software?

U.S. crypto traders are at significant risk of inaccurately reporting their DeFi trading income on their tax returns.

How Accurate is DeFi Income Calculation by Crypto Tax calculating Software?

Traders depend on online crypto gain calculation services (services) to calculate their gain and taxable income. However, calculating gain on DeFi trading is significantly more complicated than on centralized exchanges. These services are immature at calculating DeFi gain causing taxpayers to either pay too much or too little in taxes. These services claim expertise in DeFi and NFT income calculation, but they do not disclose the significant risk of inaccuracy.

My firm recognized the unreliability of these services in early 2021. We took a forensic accounting approach to calculate DeFi gain for our clients. We pulled all the blockchain addresses and manually reconciled each transaction by hand. We determined the price of unlisted coins and the tax treatment of protocols and transaction types. It was painstaking work, but we can defend these results vigorously in an IRS audit.

When we compared our income results to the online services, we were shocked at the wide diversity of results. We dug down into the transaction details and found significant differences.

Why do services not agree on DeFi income calculation?

The promise of peer-to-peer DeFi trading is that everything is recorded on the blockchain. No records are lost. Every service had the exact same information. Why do these results differ so much?

Currently, all these applications try to do DeFi calculations using their old analytical engine for doing centralized exchange (CeFi) trades. DeFi transactions are significantly more complicated. These services must completely rewrite their applications to do DeFi correctly.

Were a DeFi trader to limit his transactions to simple swapping on Uniswap, these services would be closer to the correct answer. When a trader uses more sophisticated DeFi investment applications, the service results will be less accurate.

Complexity in doing DeFi gain calculation is caused by limitations of the application, vague tax regulations, services coping to explosive change, and investor behavior. The burden of all these limitations falls upon the services to accurately calculate income. I know the founders of most of these firms. They are doing their best amidst rapid change.

We benchmarked their accuracy

Below are three clients we analyzed manually by our in-house forensic accountants. Some services are wildly high or wildly low on Total Income. Some services discover twice as many transactions on the blockchain as other services. There is no agreement on the number of tokens processed on the blockchain.

These three clients were active traders doing more sophisticated DeFi trades. These results are not indicative of what other investors will experience.

We recommend that traders pull the results from all services to document what each of these services would calculate. Save these results with your tax records to prove your due diligence. Then go with the service with a result closest to your personal estimation of income.

2020 Benchmarks

(as of 10/31/2021)

Case 1: In 2020, this trader primarily used these protocols: Curve(42%), Maker-dss(17%), Uniswap(8%), Balancer(7%), Compound(6%) and these coins: USDC(19%), DAI(15%), Curve.fi(12%), ESD(9%), WTBC(7%). At the end of 2020, the trader closed all open positions.

Case 2: In 2020, trader primarily invested in these coins: YFI(84%), LEND(4%), USDC(2%), CREAM(2%).

Case 3: This trader used a lot of exotic coins. Most popular coins: YFI(56%), XOR(9%), UNIFI(5%), BUIDL(5%), NYAN(4%)

Services claiming DeFi ability

Services claiming limited DeFi ability

TaxBit.com
CryptoTrader.Tax

To their credit, both TaxBit and CryptoTrader acknowledged upfront that their DeFi functionality was limited.

The other services claiming DeFi ability do not acknowledge any limited abilities on the main page. They do not make a disclaimer that the results are unreliable if unsupported protocols or tokens are used.

Total Proceeds and Total Cost

The Total Proceeds and Total Cost are the totals from the Forms 8949.

Note: Cointracker generated “ERROR” statements on an average of 10% of the lines on Form 8949 for all three benchmarks both in August and when run again in October. So all of their Total Proceeds are understated.

Other Income

This is the total income from all types of rewards that aren’t capital gains.

Total Income

This is the total gain (Total Proceeds minus Total Cost) plus Other Income. This is the amount on which taxes are paid. That a particular service has a result that is close to the manual calculation may be coincidental.

Total Transactions

This is the total number of transactions a service reports having found on the blockchain.

Total Lines on F8949

The IRS Forms 8949 are where every taxable capital gain event is listed. So, the Total Lines states how many taxable capital gains events occurred according to the service.

Zero Cost Basis Lines

This is the total lines on Forms 8949 showing a zero value for the cost basis. This usually means that the purchase price of a token could not be determined. Some services allow for a greater number of zero-cost basis lines as a shortcut to compensate for “loose ends”.

Gain with Zero Cost Basis

This is the total gain (proceeds-cost basis) from lines showing a zero value for the cost basis on Forms 8949.

Total Tokens

This is the total number of unique token types identified on the Forms 8949 and income statement. Most services don't count spam tokens, but they cant be hard to detect.

About the author

Clinton Donnelly is the founder of Donnelly Tax Law and CryptoTaxAudit. He is a leading expert on income taxes for crypto traders in the U.S. Preparing crypto returns since 2018, most of his clients are large traders. He has defended traders over 24 times in IRS audits of crypto income. He has authored five books and is a popular speaker.