The article never explains - how is "tax-loss harvesting" different from simply "loss"? There is a paragraph saying that it's complicated [1], but nothing in terms of actual explanation.
[1] "Its experts there couldn’t get their heads around AQR’s innovation"
I'll assume you know what tax loss harvesting means in its usual parlance (i buy 2 stocks at 50 dollars each, the first is worth 40 eoy, the 2nd is worth 60 eoy, i sell them both = I have no capital gains tax).
Where AQR innovates is in the following:
1. Suppose I am invested in my 401k into some sort of active etf. The past 10 years have been good, but I've noticed the performance is struggling lately. If I want to switch managers, I have to sell my stock and thus pay taxes on both the alpha (i.e manager skill) and the beta (i.e what the market did). But I don't want to pay the tax on the beta, I want to stay invested in the market swap my alpha.
2. The 2nd part is where it starts getting sketchy. From point 1 onward, we now want to separate our alpha and our beta. The first step is what everyone does: Leverage. Instead of owning a longly only mutual fund or active ETF, I now do the following: I buy a passive ETF (that behaves like the market) and then a long-short/beta neutral etf/strategy that will give me alpha. So for a 100 dollars invested, I now get 100 dollars of market exposure and then a varying amount of alpha exposure (it can range from 60 dollars in a conservative 130/30 or 100 dollars in a more aggressive 150/50). Key part: We solved the problem in point 1, I don't ever have to pay capital gains tax on my market until I genuinely want less exposure to the market.
3. Now comes the fun part: In a typical rising market, what would happen to our supposedly "market neutral" portfolio? On average, we would expect our longs to go up and our shorts to lose value. I could rebalance by selling some of my winners, but then I would be getting hit with the 35% short term capital gains tax which sucks. So what I do is something different: I don't do anything on the long side, I close out my short, open a new short and then carry forward my losses to the next year. With that, some time later, once I want to sell my long positions because they no longer have alpha, I can do it both using the 20% long term capital gains tax (ideally) and I will have accumulated tax losses from my shorts to further reduce exposure. And ideally I generate so many losses that I can offset a lot of my market portfolio gains.
Tl:dr It's basically pretty serious leverage and using constantly renewed short positions to keep delaying the realizations of capital gain taxes by creating tax losses. If you invest 100 dollars in a fund like this, the dream is that at the end of 10 years you have an 500 dollar portfolio with 300+ dollars of tax loss carry forward. The other fun part is when the initial 100 dollars comes from some sort of taxable event, like a sale business, and this strategy can actually cut taxes on that as well.
The article never explains - how is "tax-loss harvesting" different from simply "loss"? There is a paragraph saying that it's complicated [1], but nothing in terms of actual explanation.
[1] "Its experts there couldn’t get their heads around AQR’s innovation"
The principles for the strategy appear to be published in this paper https://www.tandfonline.com/doi/full/10.1080/0015198X.2019.1...
I'll assume you know what tax loss harvesting means in its usual parlance (i buy 2 stocks at 50 dollars each, the first is worth 40 eoy, the 2nd is worth 60 eoy, i sell them both = I have no capital gains tax).
Where AQR innovates is in the following:
1. Suppose I am invested in my 401k into some sort of active etf. The past 10 years have been good, but I've noticed the performance is struggling lately. If I want to switch managers, I have to sell my stock and thus pay taxes on both the alpha (i.e manager skill) and the beta (i.e what the market did). But I don't want to pay the tax on the beta, I want to stay invested in the market swap my alpha.
2. The 2nd part is where it starts getting sketchy. From point 1 onward, we now want to separate our alpha and our beta. The first step is what everyone does: Leverage. Instead of owning a longly only mutual fund or active ETF, I now do the following: I buy a passive ETF (that behaves like the market) and then a long-short/beta neutral etf/strategy that will give me alpha. So for a 100 dollars invested, I now get 100 dollars of market exposure and then a varying amount of alpha exposure (it can range from 60 dollars in a conservative 130/30 or 100 dollars in a more aggressive 150/50). Key part: We solved the problem in point 1, I don't ever have to pay capital gains tax on my market until I genuinely want less exposure to the market.
3. Now comes the fun part: In a typical rising market, what would happen to our supposedly "market neutral" portfolio? On average, we would expect our longs to go up and our shorts to lose value. I could rebalance by selling some of my winners, but then I would be getting hit with the 35% short term capital gains tax which sucks. So what I do is something different: I don't do anything on the long side, I close out my short, open a new short and then carry forward my losses to the next year. With that, some time later, once I want to sell my long positions because they no longer have alpha, I can do it both using the 20% long term capital gains tax (ideally) and I will have accumulated tax losses from my shorts to further reduce exposure. And ideally I generate so many losses that I can offset a lot of my market portfolio gains.
Tl:dr It's basically pretty serious leverage and using constantly renewed short positions to keep delaying the realizations of capital gain taxes by creating tax losses. If you invest 100 dollars in a fund like this, the dream is that at the end of 10 years you have an 500 dollar portfolio with 300+ dollars of tax loss carry forward. The other fun part is when the initial 100 dollars comes from some sort of taxable event, like a sale business, and this strategy can actually cut taxes on that as well.