Tax Loss Harvesting 2 – Elm Wealth
Today’s post is a return to the topic of Tax Loss Harvesting, as seen through the eyes of Elm Wealth.
Contents
Elm Wealth
Elm Wealth (Elm Partners Management, LLC) is an independent, registered investment advisor (RIA) that specialises in low-cost, systematic asset allocation.
They were founded in 2011 by Victor Haghani – famous as a co-founder of high-profile fao;e hedge fund Long-Term Capital Management (LTCM) – and manage around $2 billion for high-net-worth families and institutional clients.
Victor wrote a very good book called The Missing Billionaires. Elm has three papers on TLH, one each from 2024, 2025 and 2026.
Is the Juice Worth The Squeeze?
The 2024 paper is called “Is the Juice Worth the Squeeze?” and looks at Direct Indexing (DI).
Wealth managers like [DI], as they can charge extra fees by putting their clients into DI programs rather than steering them into index funds offered by Vanguard or Blackrock. Investors have been so convinced of the benefits of DI that Goldman Sachs, Morgan Stanley and JPMorgan manage upward of $300 billion in these programs.
Elm recommends Segmented Investing (SI):
Segmented ETF investing involves building your desired portfolio using low-cost sector and (optionally) international ETFs. Segmented ETF investing is expected to generate a similar amount of tax losses as Direct Indexing, but without DI’s costs, risks and limits on diversification.
The main problems of Direct Indexing stem from the difficulties of attempting to harvest single-name losses, while tightly tracking a benchmark index and complying with the Wash Sale rule.
This is also my view of DI – it seems more difficult than just using a few ETFs (for Elm, sector ETFs) to achieve TLH.
Expected harvest
How much you should expect to gain depends on the volatility of the stocks and your time horizon.
The vertical dotted line represents the average volatility of US stocks over the five years before the paper.
- Individual stocks are more volatile than indices, which is why DI was developed.
Note the diminishing harvest of each year the stock is held.
- Still, over ten years you might harvest 55% of tax losses.
The problem remains accurately tracking an index without re-introducing capital gains (by falling foul of the Wash Sale rule).
- Since DI trades are predictable, it’s possible that those on the other side of the trade are sophisticated (e.g. hedge funds), which means that DI tracking error would likely include some negative expected return.
Also, since the market tracks upwards, eventually you run out of losses to harvest – all your positions are in profit.
Segmentation
Segmentation involves breaking down your target portfolio into constituent parts:
Since multiple ETFs are usually available for each segment, when a loss is realised, the ETF can be immediately replaced with another (eg. from a different provider).
Volatility of ETFs is somewhat lower, so you are trading off some of the loss harvest for lower tracking error.
- The risk that the tracking error is systematically negative is also largely removed.
A TLH program using sector ETFs would deliver about 70% of the capital losses that a DI program could potentially generate – but remember that most DI programs do not deliver on their full potential, because most programs put a limit on tracking risk.
Elm believes they deliver less than 70% of their potential, making segmentation at least comparable.
- But the losses will be lumpier because of the lower volatility.
Fees are also lower for segmentation (though I’m not aware of retail DI services in the UK).
- And segmentation offers a more diversified base portfolio than DI.
Out of the frying pan
The second paper is from 2025, and is called Out of the Frying Pan and Into the Fire.
- It looks at leveraged direct index tax loss harvesting (LSDI), which:
Promises to transition an appreciated single stock holding into a diversified portfolio like the S&P 500 in about ten years, without realizing any capital gains along the way.
For a £1M position which is mostly gains, the approach would be to go long £1M of stocks and also short £1M of stocks, using quant factors to target alpha.
- US tax rules mean that the longs and shorts can’t match exactly.
- Elm is sceptical that alpha will actually be generated.
Longs that fall are sold, and shorts that go up are bought back.
- Each year as losses are harvested, a matching amount of the original £1M position is sold.
Elm calculates that after 10 years you should have £3.6M in longs and £1.8M of shorts.
- You would also have paid out £0.18M in fees and financing costs (10% of your portfolio value).
- In the US, this is more than half of the original CGT you were exposed to.
And you are stuck in the program, paying fees indefinitely, and tracking the index badly.
- After 20 years, the fees exceed the original CGT, and you are still stuck in the program with a CGT liability.
If you die within 20 years, the CGT rebasing on death could bail you out.
As with the first paper, Elm recommends ETF segmentation instead – creating the ETF portfolio by selling the large position up front (actually in sections, using an optimiser tool), and then applying tax management afterwards.
Robbing Peter to Pay Paul
The third paper, from April 2026, is called Robbing Peter to Pay Paul.
- This one allows for the possibility of 0.5% pa of alpha within the LSDI program.
Even [then] there’s only a 50/50 chance that the LSDI program leaves you better off than the simple sell-and-reinvest alternative.
For LSDI to be attractive, you need to be confident of a higher rate of alpha, or be close to death (and the CGT step-up).
- US investors can also move to a lower-tax state, an option not yet available in the UK.
Conclusions
Elm are critiquing DI and LSDI programmes not available to UK private investors,
- Running your own TLH programme would not attract the same fees.
But it would run into the same problems.
- Over 10 years, you can only expect to get 39% of your cash (55% * 70%) out without paying tax.
- You can add £30K of annual CGT allowances to that.
To get the remaining 61% – plus around 80% in gains (minus another £30K in s CGT allowances) – out over another 10 years, means that your starting pot can’t be very large.
- My calculations suggest you can only clear down a £33K pot (though total withdrawals would be around £74K).
So if the annual CGT allowance stays at £3K, TLH is not very attractive in the UK.
- Not because advisor fees replace your tax savings, but because the tax savings are not very large to begin with.
That’s it for today.
- Until next time.
Mike Rawson
Mike is the owner of 7 Circles, and a private investor living in London. He has been managing his own money for 40 years, with some success.
