DeFi Investing Strategies for Canadian Investors

The financial argument for DeFi is easy to state: lending, market making and collateralised borrowing without an intermediary taking the spread. The Canadian tax argument is harder, and it is the part that determines whether a strategy is worth running.

Two things make DeFi different from holding crypto on an exchange. Every protocol interaction is potentially a disposition of property, and the CRA has published guidance on some of these activities and not on others. Both facts change what a return looks like, and both get worse on a cheap chain, where the transaction cost stops acting as a brake on how often you interact: what a Layer 2 does to your record keeping is the practical version of the same problem.

The three activities, and what each one triggers

The column that surprises people is the third row of the middle column. Providing liquidity is not a passive hold. The pool rebalances your position continuously, and the token you receive in exchange for your deposit is a different property from what you put in.

Where the CRA has spoken, and where it has not

This distinction matters more than any yield figure.

Published. Mining is treated as a business or a personal activity, decided case by case, and in most cases the scale and resources involved mean it is a business. Rewards from staking on a centralized exchange platform will generally be income at the time they are credited to your wallet on the platform, per the CRA’s mining and staking guidance. Disposing of crypto, including crypto-for-crypto trades, is a disposition requiring a gain or loss calculation.

Not published. Staking through non-custodial protocols. Liquidity provider tokens and whether receiving one is itself a disposition. Rebasing tokens. Unsolicited airdrops. Whether impermanent loss, realised on withdrawal, is a capital loss or a business loss.

These are not questions with hidden answers. They are open. The professional approach is to take a reasonable, defensible position, apply it consistently across every year and every wallet, and write down the reasoning at the time. Presenting an open question as settled is what makes a return indefensible if it is later examined. More on the specific liquidity-pool issues in DeFi liquidity pool tax in Canada.

Income or capital, and why it decides everything

Yield received in kind is income. That part is not really contested, and the shape holds when what you supply is hardware rather than capital, which is why being paid in tokens for compute also lands in income at receipt.

The harder question is what happens to the principal. If your DeFi activity looks like a business, the underlying crypto is inventory and every gain is fully taxable rather than half. Frequency of transactions, use of leverage, sophistication, time devoted and whether you hold yourself out as trading all feed into it. Someone running a leveraged yield strategy across five protocols with daily rebalancing is on much weaker ground claiming capital treatment than someone lending a stablecoin and leaving it there.

The consequence is arithmetic. Capital gains are included at one-half; the proposed increase to two-thirds was cancelled in March 2025, so that rate stands. Business income is included in full. On the same $50,000 of gain, the difference between the two characterisations is the difference between $25,000 and $50,000 of taxable income. The factors are laid out in business versus capital account.

Impermanent loss, done correctly

Impermanent loss is the most commonly miscalculated figure in DeFi writing, so here is the mechanism rather than a number pulled from somewhere.

A constant-product pool keeps the product of the two balances fixed. Deposit 1 ETH and 5,000 USDC when ETH is $5,000, and the product is 5,000. If ETH doubles to $10,000, arbitrage rebalances the pool until the ratio matches the new price:

The gap is $858, about 5.7% of the held position. Trading fees offset it, and whether they offset it fully depends on volume. The general shape: a 2x price move costs roughly 5.7%, a 4x move roughly 20%. Correlated pairs move less relative to each other, so a stablecoin pair carries far less of this than a volatile pair.

For Canadian tax the interesting question is what that $858 is. It is not a realised loss while you remain in the pool. It crystallises on withdrawal, and its character follows the character of the position: a capital loss if you are on capital account, a business loss if you are not. The CRA has not published a position on this, which is exactly why your treatment needs to be written down before you file rather than argued afterwards.

Yields, honestly

Protocol yields move constantly with utilisation and market conditions, and any number written down here is stale before it is read. What is durable is the ranking:

Lowest risk, lowest yield
  Stablecoin lending on an established protocol
    Stablecoin-to-stablecoin liquidity pool
      Blue-chip asset lending
        Volatile-pair liquidity provision
          Leveraged looping and yield farming
Highest risk, highest yield

Every step down that list adds a category of risk rather than just more of the same risk. Lending adds protocol risk to price risk. Liquidity provision adds impermanent loss. Looping adds liquidation risk, and it converts a market drawdown into a forced disposition at the worst possible price.

The tax point about that last one: a liquidation is a disposition. You realise a gain or a loss at a moment you did not choose, on terms you did not set, and you owe tax on any gain in a year when your portfolio just fell.

The record-keeping problem is the actual problem

A single yield-farming position can generate hundreds of taxable events a year, denominated in tokens, priced in another token, on a chain that does not produce a statement.

What you need per event, per the CRA’s guidance on determining value:

Records must generally be kept for six years from the end of the last tax year they relate to, and the burden of proof sits with you.

Reconstructing a year of on-chain activity retroactively costs multiples of what tracking it as it happens costs. Portfolio software helps and does not finish the job, because the categorisation of an unlabelled protocol interaction is a judgment call the software makes by guessing. The practical setup is described in crypto tax software for Canada, with the pooled cost base mechanics in ACB record keeping.

Risks worth pricing before yield

Smart contract risk. Audits reduce the probability of a bug; they do not eliminate it, and an audited protocol has still shipped exploitable code more than once. Treat capital in any single contract as capital that can go to zero in a block.

Bridge risk. Cross-chain bridges have historically been the highest-value target in the ecosystem. Every bridge crossing is an additional trust assumption and, in Canada, likely an additional disposition.

Oracle and liquidation risk. Collateral is liquidated on a price feed, not on the price you see. A brief feed dislocation can liquidate a position that was never actually underwater.

Counterparty opacity. Over-collateralisation protects lenders against borrower default, and it does not protect against contract failure, governance capture, or the protocol simply being abandoned. It is a different risk from a bank, not a smaller one.

Loss of keys. Self-custody moves the failure mode from institutional to personal. What happens to the tax position when access is genuinely lost is covered in lost keys and exchange collapse.

A sane way to run this

1. Size it as risk capital. A percentage of the portfolio you can watch go to zero without changing your plans.
2. Start with the simplest position. Stablecoin lending on an established protocol generates one income stream and few taxable events, which makes it a good place to learn what your record keeping actually has to handle.
3. Track from day one. Not from the day you decide the position is big enough to matter.
4. Write your positions down. One page per open question, dated, with the treatment chosen and the reasoning.
5. Model the after-tax return before entering. Yield taxed as income at a high marginal rate, minus gas, minus impermanent loss, is often a smaller number than the alternative you dismissed.

That last step kills more DeFi strategies than any risk analysis. A headline yield is a pre-tax, pre-fee, pre-loss number, and Canadian tax on income received in kind arrives in cash regardless of whether you sold anything.

If you have a year of DeFi activity to reconcile, or you want a defensible position taken on the parts the CRA has not ruled on before you file, that is worth sorting out properly.