Four separate routes lead to zero, none of them exotic. Then the volatility arithmetic, the costs nobody totals, and a tax treatment that can leave a bill in currency against a position that has already fallen.
13 min read · Level 3
The debate about this asset class is conducted almost entirely at the level of whether it should exist, and I think that debate is a distraction from the thing that is actually taking money off people. The four routes to zero are real and they get all the attention. The ordinary year takes more. Somebody who never loses a key, never uses a venue that fails, never touches a fraud and never holds a protocol that breaks can still finish a year holding less than they started with, having paid 7.68% in costs on a position they turned over eight times, owing tax in currency on gains that arrived as assets and then fell, and unable to compute their own position because nothing was written down at the time. Not one line of that required an opinion about the technology. What makes it worth saying plainly is that the volatility, which is the part everybody talks about, is at least symmetric and at least visible - it works in both directions and it is on the screen. The costs, the disposals and the record-keeping are one-directional, invisible and administrative, and administrative losses are the only kind that can be eliminated entirely by decisions that cost nothing. If a reader takes one thing from this lesson it should not be a view about the asset. It should be that the cheapest improvement available here is to transact less and write down what you did, and that this is true whether the asset turns out to be the future of anything or not.
In most of this curriculum, a total loss is the extreme end of a distribution. A diversified equity fund going to zero requires an event outside anything in its history. A government bond going to zero requires a default. Here, total loss arrives by four ordinary routes, and they are independent of each other.
The key is lost. No institution holds a copy, no identity process recovers it, and the entry stays on the ledger permanently beyond reach. This is the only route in the whole curriculum where the holder is the sole cause and the loss is complete.
The venue fails. Lesson one arithmetic: an illustrative 35% recovery after four years turns 20,000 into 7,000, worth 5,758.92 discounted at 5%, and 65.0% of the holding never returns.
The protocol fails. A defect in the code, an economic design that does not hold under stress, or a network nobody continues to maintain. There is no operating company to restructure and no assets to distribute.
It was a fraud. The asset was created to be sold, the documentation described something that was never built, or the arrangement paid earlier participants with later participants money. Spotting-investment-fraud sets out the pattern, and this class has been unusually productive of it.
Then volatility, which is a different subject from loss and needs its own arithmetic. What a decline costs is not the decline; it is what is required to undo it:
A fall of 30% needs +42.9% to get back to where it started A fall of 50% needs +100.0% A fall of 80% needs +400.0% A fall of 90% needs +900.0% A fall of 95% needs +1,900.0%
That asymmetry is arithmetic and applies to every asset. It matters more here because declines of this size are within the ordinary range of what this class does, rather than being crisis events.
Then the costs, which are individually small and never totalled:
Trading fee on the way in, 0.25% Trading fee on the way out, 0.25% Spread paid across the quote, roughly 0.30% ------------------------------------------- Round trip 0.80% which is 80.00 on a 10,000 position
That is 6.67 times the 0.12% round trip a broad exchange-traded fund costs, and it excludes network transfer fees and excludes slippage entirely - and lesson three showed slippage alone at 1.3765% on a 300-unit sale into an illustrative book.
Because the tax treatment is where this class produces its most avoidable losses, and because it does so through a mechanism that has nothing to do with the price falling.
In the United States, digital assets are generally treated as property rather than as currency. The consequence is that a disposal is a taxable event, and the definition of disposal is wider than people expect. Selling for currency is one. Swapping one digital asset for another is a second. Using the asset to pay for something is a third. In each case a gain is computed against the cost of what left, in currency terms, on the day it left, and it is reportable for that year whether or not any currency was received.
That third category catches people because it does not feel like a sale. A purchase made with an appreciated asset is a disposal at the moment of purchase, with a gain to report and, frequently, no proceeds with which to pay the tax.
The second category is the expensive one. Work it through with clearly illustrative figures.
Asset D acquired for 10,000 Swapped for Asset E when D was worth 26,000 Gain realised on the swap 16,000 Tax at an illustrative 20% 3,200 Tax at an illustrative 25% 4,000
No currency was received at any point in this transaction.
Now suppose Asset E then falls 70% before the tax is due:
Position after the fall 26,000 x 0.30 = 7,800 Tax owed, in currency, at the illustrative 25% 4,000 Tax as a share of what remains 51.28%
Nothing improper has happened and no rule was misapplied. A gain was realised in the year of the swap and is owed in currency; the position that gain was rolled into then fell, and the fall is a separate year event that does not undo the first. Somebody who made thirty such swaps across a year, each one a disposal, can arrive at the end of it holding less than they started with and owing tax on gains they never received in money.
Three things generalise beyond any one country. Treatment differs completely between jurisdictions - some tax these assets as property, some as a distinct category, some differently again depending on frequency of activity and whether the holder is treated as trading - and every one of those rules is under active revision. Record-keeping is the practical burden: every disposal needs a cost basis and a date, across every venue and every wallet, and nothing assembles that for you by default. And the regulatory position itself is unsettled: whether a given asset is a security, which regulator supervises which activity, and what a venue must do to hold client assets are all questions with different answers in different places and changing answers over time.
Every rate in this lesson is illustrative. Take the mechanism; never take the number. The one thing that is safe to carry across borders is the record-keeping, because every system that taxes this at all needs the same inputs.
This is educational information, not tax or accounting advice. Rules differ by country and change — get advice from a qualified professional in your jurisdiction before acting.
Educational, not personalised financial advice. Nothing in the Investor Academy is a recommendation to buy, sell or hold anything — what is right for you depends on your own circumstances and the rules where you live.
Sources: U.S. IRS - official tax information; UK Government - Capital Gains Tax guidance; OECD - tax policy and statistics; U.S. SEC - Investor.gov
One holder, one year, with nothing dramatic in it.
Start: 10,000 committed to Asset D, self-custodied on one device, with the recovery phrase written down once.
Activity. Over the year they make eight swaps between assets, each of which is a disposal in the United States, and pay a 0.25% fee each way with roughly 0.30% of spread. Sixteen trades of a position averaging 12,000:
Round trip cost 0.80% per swap, on an average 12,000 = 96.00 per swap Eight swaps 768.00 As a share of the starting 10,000 7.68%
Before any price move, before any tax, and before slippage, activity cost 7.68% of the starting position. The individual figures - a quarter of a per cent here, three tenths there - never looked like anything.
Tax. Three of the eight swaps were at a gain and five at a loss, and the net position for the year is genuinely a matter for their own jurisdiction and their own records, which they do not have, because eight swaps across two venues and one wallet were not recorded anywhere at the time. Reconstructing eight cost bases a year later, from exchange exports that use different formats, is the real cost of this and it is the one nobody warns about.
Price. The asset ends the year 62% below where it started. To return to the starting level it now needs a rise of 163.2%.
And the part that does not appear in any of the above. The device with the key on it still works. The phrase is still in the drawer. Nothing failed. This is a perfectly ordinary year in this class, and it contains no exchange collapse, no fraud, no lost key and no protocol failure - the four routes to zero all remained closed and the holder is still down heavily.
The reason to lay it out this way is that the four total-loss routes get all the attention and the ordinary year does most of the damage. A holder who guards against catastrophe and pays no attention to turnover, costs, record-keeping and the recovery arithmetic has protected against the rarest outcomes and accepted the common one.
Run the same year with one change: no swaps. The 768.00 of costs does not happen, no disposals occur, no records need reconstructing, and the price does what it does. That single decision was worth more than every risk control in the lesson, and it required no view about anything.
Treating a swap between two digital assets as a non-event. In the United States it is a disposal, the gain is computed in currency on the day, and it is reportable whether or not any currency was received. The illustrative case here produced 4,000 of tax owed against a position that had fallen to 7,800 - 51.28% of what remained.
Paying for something with an appreciated asset without noticing that it is a sale. The purchase and the disposal are one event, the gain is reportable, and there are no proceeds set aside to pay it.
Leaving record-keeping until the return is due. Every disposal needs a cost basis and a date, across every venue and wallet, and nothing assembles that by default. Reconstructing a year of activity from exports in different formats is the largest unpriced cost in this lesson and it is entirely avoidable by recording at the time.
Totalling none of the costs. A 0.25% fee each way plus roughly 0.30% of spread is a 0.80% round trip - 6.67 times a broad exchange-traded fund 0.12% - and eight swaps on an average 12,000 position cost 768.00, or 7.68% of a 10,000 start. None of the components ever looked like anything.
Confusing volatility with the four routes to zero. A 62% fall is not a permanent loss and needs 163.2% to undo; a lost key is permanent and needs nothing at all to undo it because nothing can. Volatility-vs-permanent-loss is the distinction, and this is the class where both are genuinely present at once.
Assuming the regulatory position is settled. Whether an asset is a security, which regulator supervises a venue, and what protections apply to client assets have different answers in different places and changing answers over time. Any statement that begins "it is now regulated" needs a jurisdiction, a date, and a scope.
Applying any rate in this lesson to your own position. Every one is illustrative and no country system works exactly this way.
Lost key, venue failure, protocol failure, fraud. They are independent, so a control for one does nothing for the others. Self-custody removes the second and creates exposure to the first; diligence addresses the third and fourth and nothing else. Write the four lines and the answer beside each.
Date, what left, what arrived, the currency value of each on that day, the venue, and the fee. Sells, swaps and purchases made with the asset all count. This takes a minute per transaction now and can take a weekend per year later, and the later version is frequently incomplete.
Gain equals the currency value of what leaves, less its cost. On the illustrative case: 26,000 less 10,000 is a 16,000 gain, and at an illustrative 25% that is 4,000 owed in currency out of a transaction that produced none. Knowing that before the swap makes it a decision.
The tax is fixed at the moment of the disposal and the position can fall afterwards without reducing it. Holding the liability in the same volatile asset that generated it is how 4,000 of tax becomes 51.28% of what remains. Move it out at the time.
Fees both ways plus the spread plus the slippage from lesson three. At 0.80% per round trip, eight swaps on an average 12,000 position cost 768.00 - 7.68% of a 10,000 start, before any price move. Then ask what the same year would have cost with no swaps at all.
A 30% fall needs 42.9%; 50% needs 100.0%; 80% needs 400.0%; 90% needs 900.0%. Doing this arithmetic in advance is what makes a position size real, because it converts a percentage that sounds survivable into the return that would be required to undo it.
Before any swap, before spending an appreciated asset on anything, at the start of a tax year to set up records rather than at the end to reconstruct them, and when sizing a position that could plausibly fall by most of its value.
Do not apply any figure in this lesson to your own position - every rate is illustrative and no country system works exactly this way. And do not let a tax charge prevent a disposal that is right on investment grounds; a charge is the cost of a correct decision, not a reason to avoid making one.
Holders who move between assets frequently, without converting to currency at any point, commonly reach the end of a tax year having realised substantial reportable gains and holding no currency with which to settle them. The mechanism is that each swap is a disposal at the currency value on the day, and the value on the day is a real number regardless of whether any currency changed hands. When the assets held at year end have fallen from the levels at which those disposals occurred, the liability and the remaining position have moved in opposite directions - and the liability is fixed while the position is not. The outcome is frequently described afterwards as a tax problem. It is more accurately a sequencing problem, and the sequence was decided at each swap, months before any bill existed.
Move the tax out of the volatile asset at the moment the gain is realised. The liability stops moving at that instant and the position does not.
Cost figures in this class arrive one transaction at a time, in fractions of a per cent, at moments when attention is on the price rather than on the fee. A quarter of a per cent is genuinely small and nobody would describe it as a consideration. Eight round trips on an average position turn it into 7.68% of the starting amount, which is larger than most of the annual figures anybody quotes about anything, and it is paid whether the year went well or badly. What makes it hard to see is not the arithmetic but the reporting: the fees appear on individual confirmations, spread across a year, on a dozen separate screens, and there is no statement anywhere that totals them. The only way the number ever gets seen is if somebody adds it up deliberately.
Total a year of transaction costs once, as a single figure against your starting position. It is the one number in this class nobody will ever show you.
Create a free account to read the rest and apply it to your own investing.