Very quick background – because of the PFIC limitations in holding index funds in an ISA, I’m trying a mostly “pseudo-indexing” approach in my and my wife’s ISAs, buying 40 individual stocks in the hopes of approximately matching the performance of the FTSE 100. They’re all UK listed stocks, avoiding foreign currency fees from Hargreaves Lansdown and any dividend withholding headaches from US stocks.
A couple small updates since September:
- We’d already maxed out our ISA allowances for 2021-22, so no new contributions, just holding the same 40 stocks in our S&S ISAs and seeing how they do.
- I changed the source of the monthly account fees (0.45%, capped at £45 a year, so £3.75 a month) from within our ISAs to newly opened taxable brokerage/general investment accounts (GIA, what HL calls a “Fund and Share Account”). This means that we keep £45 more a year within the ISA wrapper. For the moment, I just put £50 cash in the GIA, but am going to purchase £1,000 of two new stocks in each of them in January, in the hopes that the dividends will pay that £45 fee in perpetuity. This is a miniscule optimization, but there’s no downside so why not?
Honestly, aside from adding a few more months to the graphs below, there’s been very little change here since September. We’re still very close to the FTSE 100, so no complaints. Overall, if we’d put our money in a FTSE 100 ETF (CUKX), it would have grown 5.3% since we started our ISAs, which matches our actual performance (within £14, assuming equal fees).
Definitely more noise month to month, although by the end of December they’ve come together closely:
The performance of individual stocks is still all over the place, despite averaging out to something practically identical to the FTSE 100:
Is an ISA worth it?
The big question – after 9 months (a short trial in investing terms!), is an ISA worth it for a US citizen in the UK? Quick rundown of the pros and cons:
Update 20 November 2022: The 2022 Autumn Statement changes some of the calculations, see here for an updated analysis.
- UK tax-free dividends: you’ll never pay UK tax on dividends, while in a GIA, you’d pay tax on anything above the £2,000 annual allowance. You “only” need about a £50k ISA balance before this starts to matter.
- UK tax-free capital gains: you’ll never pay UK tax on capital gains, while in a GIA, you’d pay tax on anything above the £12,300 annual allowance. In practice, the allowance is generous enough that this is pretty manageable, but it could be a significant benefit for very large balances.
- Deeper awareness and understanding of investing: this is a double-edged sword, but I certainly feel like I’ve learned more about investing, the companies I’m invested in, and the broader British business world. It’s also fascinating, to me, to see the vastly disparate performance of individual companies, which averages out to something more modest in an index fund. You could see this without having any skin in the game, but that certainly helps drive interest.
- US taxable dividends and capital gains: compared to a GIA this is a wash, but clearly a disadvantage compared to a pension or IRA.
- No indexing: this one is difficult to quantify, but there is clearly an increased risk of underperformance (or overperformance!) due to being invested in a smaller number of companies. Even just compared to the relatively small FTSE 100, that’s 2.5x more companies than the 40 in our ISAs. Compared to something like a total world ETF (Vanguard’s VT, for example), that’s 9,289 stocks compared to our 40 – not even close. So far, I feel reasonably comfortable that my performance isn’t going to dramatically diverge from the FTSE 100 in normal circumstances, but if the next Amazon/Apple/Tesla is a UK company that we haven’t bought in our ISA, there could be a much more significant divergence.
- Skewed asset allocation: this is a constraint of how I have personally implemented my ISA, picking only UK stocks. This is for what I think is a good reason, avoiding currency conversion fees and dividend withholding headaches on US stocks, but it does mean that I’m overweight UK in my overall portfolio. I’m ok with a modest UK home bias, but not everyone will be. You could buy other country stocks too, but will have an even harder time mimicking a bigger index with only 20 stocks. You could buy more stocks, but the fees add up and each holding winds up tiny.
- Higher fees: £45 a year for each account, plus £1.50 for regular investing, £11.95 for on-demand dealing, 1%/min £1 for dividend reinvestment – it all adds up. Managed carefully, I can keep it around £125 a year per account, but that’s a lot more than the 0.08% you might pay for an index fund for an account with a modest balance (£16 on a £20k balance at 0.08%). However, if your account value is north of about £150k, you might actually save money compared to an index fund. If you do any sort of active investing, the fees add up really fast – not recommended.
- If you’re curious, that £125ish a year breaks down as:
- £45 account fee
- 19x £1.50 for regular investing (£1,000 for 19 stocks, once a year) = £28.50
- 1x £11.95 for the 20th stock, using up the “leftover” cash from the other 19 (if you contribute £1,000 but the stock price is £19 a share, you get 52 shares and £12 change – add up all that change and you’ll want to buy something with it rather than sitting there at 0% interest)
- 40x £1 dividend reinvestment fees (2 dividends per year, per stock) – this is technically a 1% fee, min £1 and capped at £10 per reinvestment, but none of my individual holdings are nearly big enough yet to throw off more than £100 per dividend = £40.
- In practice it’s slightly less because a) not all the stocks pay dividends and b) if a single dividend doesn’t reach the £10 threshold for HL to reinvest, it just stays in cash until the next dividend, so only £1 to reinvest 2 dividends.
- Total: £125.45
- If you’re curious, that £125ish a year breaks down as:
- Bookkeeping: keeping track of the purchases of 40 stocks and their dividends is more work than just buying one index fund with a few dividends a year. It’s not unmanageable, and HL’s records are pretty easy to use plus an Excel sheet, but it is a bit more work. If you were being efficient, you could probably get this down to an hour or so a year. For me, it’s now a part of my normal monthly financial updates, so a few extra minutes a month.
- Deeper awareness and understanding of investing: seeing the sausage being made isn’t for everybody! If you’re the kind of person who will be upset by a 50% fall in an individual stock, and especially if you’d then be tempted to sell it, this is a real danger. Could also tempt some into stock picking and active investing – an ISA really isn’t the place for it due to the transaction fees and US tax. If you want to do active investing, a Roth IRA makes a lot more sense.
Quantifying the tax benefits
I did some quick modeling of the tax benefits. For a basic rate taxpayer, with UK dividend tax at 8.75% (from 06Apr22) and a 0% US dividend tax rate, you need to get annual dividends above the £2,000 UK dividend allowance before an ISA makes any difference – otherwise, a GIA is exactly the same from a tax perspective, and slightly lower fee (HL doesn’t charge the 0.45%/£45 account fee for a GIA). Assuming a 4% dividend yield, you have to get to a £50,000 balance before you save a penny. But get to £100,000 balance and you’re saving £175 a year on dividend tax (all UK – assuming you stay in the US 0% qualified dividends/long term capital gains rate).
The math works out similarly for a higher rate taxpayer, with UK dividend tax at 33.75% and US at 15%. Up to £50,000 at a 4% dividend yield, you’re under the UK dividend allowance, and paying US dividend tax at £300 a year – but if the money was in a GIA, you’d be paying the same US dividend tax on that, too. Above £50,000, you’d start paying UK dividend tax in a GIA, but since the UK rate is higher than the US rate, you’d have enough passive category foreign tax credits to offset all of the US tax on the dividends above £2,000 – note that you would still be paying the £300 US tax on the first £2,000 of dividends, since it’s not UK taxable. In an ISA, there’s no UK dividend tax so you’re just paying the US 15% on all the dividends. At £100k, you’re saving £375 a year on dividend tax in an ISA compared to a GIA, and it only goes up from there
Quick math: that’s £4k a year in dividends. For the GIA, first £2k is only US taxed at 15% = £300. Second £2k is UK taxed at 33.75% = £675. US tax on the second £2k is fully offset by FTCs. So total of £975 tax in a GIA. In an ISA, it’s US tax only on the full £4k of dividends at 15%, for £600, £375 less than the GIA).
In summary, if you expect your account balance to get above about £50k, an ISA will probably save you money every year on dividend taxes. Not massive amounts unless you have a big ISA balance (to save £1k a year in tax, you’d need an ISA balance of almost £200k), but if you saved £20k a year for 5 years and then stopped, over 30 years you’d pay about £49k in dividend taxes in an ISA, and £120k in a GIA with the same investments (assumes 4% dividend yield, 3% capital growth, reinvested dividends, higher rate taxpayer). The extra fees in the ISA would add up to a few thousand GBP, but you’re still well ahead of a GIA, less than half the taxes and fees.
Capital gains is tougher to model, and realistically I think most people would be able to stay under the £12,300 annual allowance by careful management with tax-loss and tax-gain harvesting, especially if also supplemented by Roth contributions or other pre-pension investments. But it’s there if you need it, and you might also be able to stay in the 0% US long term capital gains bracket, for a total of 0% on capital gains tax.
All that said, is an ISA worth it? Honestly, it’s probably marginal for most people. There are tax savings to be had, but it has risks due to reduced diversity and it does add complexity. It’s pretty clear an ISA is not worth it in these situations:
- A pension/SIPP is far more tax efficient, so all your investment intended for ages after pension access age is better in a pension (at least up to the Lifetime Allowance). ISA only makes sense for money you need before 55/58/whatever.
- A Roth IRA is also clearly more tax efficient, since it’s recognized by both the US and UK, and you can withdraw the contributions at any time, helping with the bridge from early retirement to pension age. There will be exceptions, but in general I would recommend maxing out a Roth IRA before touching an ISA.
- If you aren’t comfortable with the additional risk and complexity that comes with individual stocks, accept the tax hit and use an index fund in a US brokerage account (assuming you can get/keep one, possibly using a US address).
- If you plan on moving back to the US, an ISA probably isn’t worth the faff. You won’t care about the UK tax advantages once you move back, and there are no US advantages.
So who is an ISA good for? Assuming you’re a US citizen in the UK, of course:
- You expect to be able to get above £50k-ish in your ISA in fairly short order (e.g. max the ISA for 3 years), where the UK dividend tax advantages start to matter OR
- You aren’t able to get non-PFIC index funds in a GIA, so you’re forced into individual stocks anyway (for example, your brokerage enforces the MiFid rules preventing you from buying US funds and you’re unable to open a different one, unable to use a US address, etc.). If you’re going down the individual stocks pseudo-indexing route anyway, the only downside to an ISA vs something like Interactive Brokers is the higher trading fees. If you’re just doing buy-and-hold investing once a year, plus dividend reinvestments and the account fee, you can keep this close to £125 a year, which is offset by the dividend tax savings once you hit a £70k-ish balance OR
- You’re an investing nerd and want to optimize every tax efficiency you can, and find watching the sausage being made interesting more than scary.
- AND you don’t mind the extra bookkeeping, the pseudo-indexing reduced diversification, and potentially the impact on your asset allocation.
I’m in buckets 1, 3, and 4, so I plan to continue with my ISA investments. I’ll move cash from Premium Bonds to our ISAs in April (and rebalance with G fund bonds in my TSP, to keep overall equities vs cash+bonds allocation in balance) and invest in the same 40 companies again. Keeping to the same 40 companies also means the dividends will double, which increases the efficiency of dividend reinvesting by reducing the number of £1 reinvestment fees I pay compared to having 80 companies. For me, I’m happy with the ISA wrapper as a supplement to my core investments in my UK pension and Roth IRA, plus legacy investments in my TSP, but it’s not for everybody.