Taxes: Long-term equity investing abroad, what needs to be tracked?

[tempset]

I saw this thread the other day related to foreign exchange gains among other posts on the topic relating to possible "ghost gains" when the yen weakens when compared to the original purchase time of an asset.

As an American who needs to invest outside of Japan, I'm looking to clarify specifically what I'll need to track for the future withdrawals and tax filing considerations.

/@kristhuy's comment on that thread linked above caught my attention and I thought perhaps I've been overestimating what I need to track, but I'd rather ask and clarify if possible.

In a hypothetical situation as follows (assuming no/minimal fees):

• Exchange rate of 145 JPY to 1 USD
• Convert 1,000,000 JPY to, let's say, $6,895 USD
• Send $6,895 USD to US investment account
• Buy 75 units of some ETF "X" at $91.50 per unit

... some number of years pass ...

• Sell 20 units of said ETF "X" at $150 per unit
• Send $3,000 USD to Japan account
• Convert $3,000 USD to JPY at 170 JPY / USD
• Now have an additional 510,000 JPY in Japan yen bank account

In this scenario, I originally thought I'd need to track the lifecycle of every yenny (ie. initial foreign exchange rate, exchange rate at time of ETF purchase, purchase price of ETF, sale price of ETF, exchange rate at sale, and final exchange rate after USD transfer) and calculate the various deltas across the contained events. Is this accurate or can we more simply track just the USD-based gain of the asset, exchange rate at asset sold (for gain calculation), and rate when converted to yen for the potential forex gain?

Or is it also necessary to track the exchange rate at time of asset purchase so the gain calculation can include the delta in exchange rate (ie. (20 * $150 * 170 - 20 * $91.50 * 145) rather than (20 * $150 - 20 * $91.50) * 170)?

Sorry for the wall of text, appreciate any insight!

 

[kristhuy]

@tempset The trick is to remember that only the disposal of foreign currency can trigger a foreign exchange gain/loss, and the disposal of foreign currency always triggers a foreign exchange gain/loss. The sale of an ETF, for example, cannot trigger a foreign exchange gain. Whereas the purchase of an ETF can trigger a foreign exchange gain, if you use foreign currency to make the purchase.

So, in your example:

Convert 1,000,000 JPY to, let's say, $6,895 USD

This is not a taxable event, but it does trigger a recalculation of your USD cost basis. So assuming you hold no other USD at that time, your USD cost basis will be 145JPY/USD.

Buy 75 units of some ETF "X" at $91.50 per unit

This constitutes the sale of USD, thus you must compare the JPY value of the USD at the time of the sale to your USD cost basis, to calculate the size of the foreign exchange gain/loss. In your example, you appear to assume that the exchange rate has not changed between the purchase of the USD and the sale of the USD. In that case, assuming you hold no other USD, your foreign exchange gain/loss would be zero.

The ETF purchase also triggers a recalculation of your cost basis in ETF "X". Assuming you hold no other shares in the ETF, your cost basis will be ~13,268JPY/share.

Sell 20 units of said ETF "X" at $150 per unit

To calculate your capital gain on the sale of the ETF, you must compare the sale price (in JPY) to your cost basis. Assuming that the exchange rate at the time of the sale is 170JPY/USD, your sale price will be 25,500JPY/share. Accordingly, your taxable capital gain will be 20 x (25,500 - 13,268) = ~244,640 JPY.

The ETF sale also triggers a recalculation of your USD cost basis. So assuming that the exchange rate at the time you sold the ETF was 170JPY/USD, and you hold no other USD at that time, your USD cost basis will be 170JPY/USD.

Convert $3,000 USD to JPY at 170 JPY / USD

This constitutes the sale of USD, thus you must compare the JPY value of the USD at the time of the sale to your USD cost basis. If your USD cost basis is 170JPY/USD, as it appears to be from your example, your foreign exchange gain/loss would be zero.

 

[tempset]

@kristhuy Thank you very much, that makes a lot of sense.

Apologies if this is searchable and that I didn't include this in my original post, but from what I think I've read, Japan doesn't use per-unit cost basis for equity but instead averages all cost basis for a specific asset, is that correct?

If so, how does that cost basis change when some portion of the assets have been sold?

For example:

• 2020 buy 10 of ETF "X" at 5723 JPY each
• 2021 buy 20 of ETF "X" at 6432 JPY each
• 2022 buy 15 of ETF "X" at 7140 JPY each

If the initial assumption above is accurate regarding average cost basis usage, the average cost basis from the purchase examples would be (10 * 5723 + 20 * 6432 + 15 * 7140)/(10 + 20 + 15) or ~6510 JPY

If I sell 5 units, is my cost basis still 6510? (I believe so)

If I buy 10 more units after my 5 unit sale at 7500 JPY each, is the cost basis just including the new sale in the average?

Something like: (10 * 5723 + 20 * 6432 + 15 * 7140 - 5 * 6510 + 10 * 7500)/(10 + 20 + 15 - 5 + 10) or ~6708 JPY

I included the sale removing shares from the cost basis, which changes the output to be slightly lower than if the sale wasn't included. Is that correct?

 

[kristhuy]

@tempset

Japan doesn't use per-unit cost basis for equity but instead averages all cost basis for a specific asset, is that correct?

I don't understand the distinction you are making. Your cost basis is determined by the average purchase price of all shares you hold. It is both "per-unit" and an average.

how does that cost basis change when some portion of the assets have been sold?

It doesn't. Only purchase transactions affect your cost basis.

If I sell 5 units, is my cost basis still 6510?

Yes.

is the cost basis just including the new sale in the average?

Yes, the new cost basis would be (40 x 6,510 + 10 x 7,500)/50 = ~6708

I included the sale removing shares from the cost basis

You don't really need to do that. Your equation still gives the right answer, but it's unnecessarily complicated.

All you need to do is multiply the number of shares you held before the purchase by your per-share cost basis, add to it the number of shares you purchased multiplied by the per-share purchase price, and divide the result by the total number of shares you held after the purchase. (See the equation above.)

 

[manishk012]

@kristhuy

You don't really need to do that. Your equation still gives the right answer, but it's unnecessarily complicated.

It is the correct way to do it if you're going to put this into an Excel file with a row for each purchase and sale. So for a sale you'd put a row with a negative quantity and set the FMV to the cost basis at the time of sale.

 

[kristhuy]

@manishk012 I guess. But it depends how you set up your sheet... I would prefer to track quantity and cost basis separately. No need to update the cost basis when a sale occurs, just the quantity. Then when you have a new purchase, you calculate a new cost basis by reference to the current quantity and the current cost basis.

 

[tempset]

@kristhuy

I don't understand the distinction you are making. Your cost basis is determined by the average purchase price of all shares you hold. It is both "per-unit" and an average.

My wording here was confusing, apologies. The distinction I intended to make was average of all shares as cost basis vs. each share maintaining its own individual cost basis from its purchase date (ie. in my example, the shares from 2020 have a cost basis of 5723, the shares from 2021 have a cost basis of 6432, etc.).

The latter is how I believe it works for US tax filing, which is why trading platforms enable different selling strategies (FIFO, tax optimized, etc.). Of course always happy to be corrected.

 

[kristhuy]

@tempset Ah ok I see what you meant now. Yeah, you're right that in Japan your cost basis at any given time is the same for all shares, regardless of when they were acquired.