The thing that surprises most people the first time they actually run the math is that a single set of trades does not have one right answer for how much gain they realized. You bought some coin at a few different prices over the year, you sold a chunk of it, and the profit you report depends entirely on which of those earlier buys you decide the sale came out of. That decision is the cost basis method, and it moves the number more than almost anything else you can control after the fact.
Let me make it concrete, because it stays abstract otherwise. Say you bought one unit at roughly 20k, another at roughly 40k, and a third at roughly 60k, three separate buys over a year. Later you sell one unit for 50k. What was your gain? If the sale is matched against the 20k lot, you booked a big gain. Matched against the 60k lot, you booked a loss. Same coin, same sale price, same wallet. The only thing that changed is which lot you assigned the sale to, and every one of those answers can be defensible depending on the method you are running.
What the four methods actually do
There are really four ways people match sales to buys, and they are simpler than the acronyms make them sound.
- FIFO (first in, first out) sells your oldest coins first. In a market that has generally gone up over the time you held, your oldest coins usually have the lowest cost, so FIFO tends to surface the largest gains in the near term. It is the default most tax authorities and most software fall back to when you say nothing.
- LIFO (last in, first out) sells your newest coins first. If you have been buying into a rising market, your newest lots have the highest cost, so LIFO tends to book smaller near-term gains. The catch is that it leaves your cheap old coins sitting in the account, waiting to be taxed later.
- HIFO (highest in, first out) ignores order entirely and sells your most expensive lots first, whatever their date. HIFO is the method built to minimize the gain you report this year, because it always reaches for the highest cost it can find to offset the sale.
- Specific identification lets you pick the exact lot for each sale, by hand, trade by trade. FIFO, LIFO, and HIFO are really just automated rules for doing specific ID a particular way every time.
That last point is the one worth sitting with. HIFO is not a separate universe from specific identification. It is specific ID with a rule bolted on that says always grab the highest cost. Which is why, in most places, if you can legitimately run HIFO, you can also run something smarter than HIFO.
Why specific identification usually wins
HIFO minimizes this year's gain, and for a lot of people that sounds like the obvious goal. But minimizing the current gain is not always the same as minimizing your total tax, and specific ID is the only method that lets you optimize for the thing you actually care about instead of a proxy for it.
A couple of examples of where blind HIFO leaves money on the table. Short-term and long-term gains are usually taxed at very different rates, so the lot with the highest cost is not automatically the lot that saves you the most, because a slightly cheaper lot you have held long enough to qualify for the lower long-term rate can beat a pricier lot you have only held a few months. HIFO does not know or care about your holding period, it just chases the highest number. Specific ID lets you weigh cost against holding period and pick the lot that actually costs you the least in tax.
The other place specific ID pulls ahead is loss harvesting. If you want to deliberately realize a loss to offset gains elsewhere, you want to sell a lot that is currently underwater, and that is often a high-cost recent buy, but not always. Specific ID lets you go find that exact lot and sell it on purpose. A fixed rule cannot do that with any precision.
The honest tradeoff is that specific ID is more work and demands better records. You are making a decision per sale rather than letting a rule run. If you are doing a handful of trades a year, that is fine and probably worth it. If you are doing hundreds, you almost certainly want software applying a consistent rule, and then HIFO is a reasonable rule to run.
The records you need to defend any of this
Here is where people get burned, and it has nothing to do with picking the wrong method. It is that whatever method you claim, you have to be able to show your work. For specific identification in particular, most tax authorities expect you to have identified the specific lot at or near the time of the sale, not reconstructed it conveniently a year later when you are doing your return. Deciding after the fact which lot you meant to sell is the thing that does not hold up.
The practical version of that is a per-lot ledger. For every acquisition you want to be able to produce the date and time you received the coins, the amount, the price or fair value at that moment, and the fees. For every disposal you want the date and time, the amount, the proceeds, the fees, and crucially which acquisition lot or lots it drew from. If you are running specific ID, the record that you selected that lot should exist from around when the trade happened, not be something you assert later.
A workflow that holds up:
- Export the full transaction history from every exchange and pull every on-chain wallet you used during the year. Missing one small venue is how a cost basis chain silently breaks.
- Reconcile the totals. The coins going out should trace back to coins that came in. If your records say you sold more than you ever acquired, you have a gap, usually a transfer between your own accounts that got read as a disposal.
- Pick your method and apply it consistently across the whole account for the year, not cherry-picked per trade after you already know the outcomes.
- Keep the exports, the reconciliation, and the lot-by-lot matching as your backup. If anyone ever asks how you got your number, that folder is the answer.
Consistency, and how to actually lock it in
You generally do not get to run FIFO on one sale, HIFO on the next, and LIFO on a third just because each happened to produce the friendliest number. The expectation is that you apply a method consistently, and if you switch methods between years there is usually a right way to do that which involves declaring the change rather than silently flipping. Rules differ by country and they do change, so this is exactly the kind of thing worth a short conversation with someone who does tax in your jurisdiction rather than trusting a stranger on the internet. Me included.
What locking it in looks like in practice is boring and that is the point. Choose your method before you start reconciling, not after you have seen which one gives the lowest bill, because choosing based on the outcome is what turns a legitimate election into something that looks like you reverse-engineered the answer. Set the same method in whatever software you use so the automated matching agrees with what you are claiming. Write down which method you used and why, keep it with that year's records, and use the same method next year unless you make a deliberate, documented switch.
The short version, if you only remember one thing: FIFO is the default and usually the least favorable in a rising market, HIFO minimizes this year's reported gain, and specific identification beats both when you have clean records and are willing to think trade by trade. The method matters, but the records are what let you actually use the method you picked. Get the ledger right first and the rest is just choosing a rule and staying with it.