Setting aside whether TLH is saving money or simply borrowing it from future tax liabilities, how much is 3k in capital losses worth to you? At what point does the 0.25 expense ratio cost more than the benefit to you?

By my calculations, the breakeven AUM is around $240k, assuming you always have 3k cap gains to offset.

It’s more of an argument for how lots of individual holdings could beat an ETF, in USA anyway. If your commissions are free anyway. Lots of paperwork tho.

For all I know, these robo-advisers just buy you ETFs.

I believe the discussion is about what Wealthfront is doing to earn their 0.25% fee. "Tax loss harvesting" seems unpersuasive.

At $100k AUM they start “direct indexing,” buying shares as well as ETFs. I only started in 2020 but so far they’ve harvested 3.6x their own fees. I would not have spent time doing that, so it’s gravy.

Except you are still going to pay those taxes when you do sell, you are just delaying the taxable event so I wouldn't consider that money actually saved.

What is the situation where that is actually useful? Take me for example, I invested for 5 years in VTI and then sold it for a downpayment. What would their TLH do for me in that situation because I didn't need to offset cap gains in those middle years and the end year both get hit with cap gains.

Doesn't seem to do anything except charge a higher fee and create more irs paperwork in that situation.

True, if you cash out that soon you only get use of the tax money for a little while. I like TLH for the same reason I like a traditional 401(k). Right now I’m getting a Bay Area tech salary and RSUs, and I want to defer some taxes until I’m in a lower bracket.

If we’re talking about the first $3k of losses applied to income account instead of capital account, it is saving you taxes by turning capital losses into income losses where the tax benefits are greater.

This doesn’t do what you think it does. It’s tantamount to timing the market which generally is a losing strategy. Not sure if you’re being sarcastic

Tax loss harvesting really only works in the long run if you know which stocks won’t recover.

Their computer sells stuff and then buys stocks that are heavily correlated to the sold stock.

Later, it unwinds the imbalance to avoid realized gains and wash sales. Doing that without impacting long-term returns is one of their biggest value adds.

My point is that even if it was literally the same stock, unless you know that you bought at the relative bottom all you’re doing is needlessly incurring loses.

For example you’re better off just buying the dip, then selling, and paying taxes (even with harvesting) and repurchasing assuming the stock recovers and you don’t mind fronting the capital

You can do it using basket ETFs. The IRS doesn't consider them "exact" replicas. So if VOO drops, you can harvest using SPY and maintain the same exposure, then switch back to VOO a month + few days later.

It gets harder to pull that off once you’re a long-term holder

Hence the emphasis on cycling things around. You’re not trying to cut and run, but retain exposure while crystallizing the loss.

I’m not in US, but what are the rules about buying back a stock that you just sold?

In Canada, it’s a 30d wait for the loss to count, but you can buy back another similar company/index the next minute and your loss still counts.

If the price of oil craters and you sell your -10% Exxon and buy -10% Chevron, you’re not timing the market but you are crystallizing a loss.

Or change between Solactive and MSCI-based index funds because they’re “only” 95% identical.