Epistemic status: I came up with an idea for excluding stocks in a direct index, googled around and found out that it's already a product with a startup offering it for a fair price. I have not used this long, and may discover new points of concern as I get more experience with it.

There have been a few news stories recently about people lobbying index providers to exclude SpaceX. Could you accomplish the same thing without negotiating with index providers? This post explains the costs and benefits of direct indexing, and compares the services available.

Direct Indexing is a financial product that behaves like an ETF except you hold the basket of stocks directly instead of paying the ETF provider to do it for you. That gives you more flexibility in what and when you trade.

The mechanics of rebalancing a direct index varies from provider to provider. I think they're all doing some form of the following:

1. They run a solver periodically, minimizing a loss function.
2. That loss function penalizes: excessive trading, tax costs, tracking error.
3. Solving is subject to regulatory constraints like wash sale rules.

If the direct index provider has the ability to send the balancing transactions for all their customers at once to the market makers on the other side, it is likely they'd get a very good price. The flow is predictable, uninformed, and partially offsetting, so market makers are able to compete hard for it.

I believe direct indexing is much more viable these days than it used to be. If your account is limited to buying whole shares you'd need ~$23m of S&P 500 exposure to bother buying your first share of NVR, a ~$6000 stock with a 0.0265% weight. Now that fractional shares are commonplace, this constraint is gone and it makes it much easier for individuals to track a direct index at scale.

Some direct index providers let you exclude sectors and stocks from your index by tweaking your rebalancing algorithm. The hope is that customers can use this feature to:

- exclude sectors they're already exposed to via their employment
- exclude Elon Musk companies
- exclude meme stocks
- express ESG or anti-ESG opinions
- express negative opinions about single stocks without being exposed to the complexities and convexities of short selling

As you add more and more exclusions you'll stop tracking The index and instead track Your index. I find this to be a nice balance for most people between broad diversification and wanting to express opinions about certain companies being bad for their portfolio.

Direct Indexing also gives you the ability to "harvest" tax losses. That's the idea that if you sell stocks that have lost money since you've bought them, then you can reap a capital loss to offset other capital gains. You will still owe that tax eventually -- tax loss harvesting provides tax deferral, not tax avoidance.

The amount of harvested gains you get to keep is the growth of your deferred tax, after tax. Quantitatively, if your holding decreased X% and your marginal tax rate is T% and you keep the holding for N years compounded at R%, this saves you X%*T%*(1-T%)*(1+R%)^N. In words, you keep X%*T% money that you otherwise would have paid in tax, that money grows by a factor of (1+R%)^N, and then you get to keep the part of that growth which isn't taxed: (1 - T%) of it. For me, T%*(1-T%) comes out to ~25%. So I get to keep about a quarter of harvested losses, compounded over my holding period.

It's worth pointing out that unless you have an alternative source of capital gains every year, you'll be limited by a maximum of $3000 of losses deducted against ordinary income. If you happen to be compensated in stock of a publicly traded company you can combine tax loss harvesting with selling enough of your concentrated position to generate capital gains offsetting your capital losses. It's a neat party trick, but if you find yourself in that position and your concentrated position(s) is/are > $1m, ask your wealth manager about a Section 351 or Section 721 exchange instead.

A valuable pattern for Direct Indexing is that if you regularly add money from your job and give away money to tax-deduction eligible charities, you have a source of freshly purchased securities to harvest and a source of highly appreciated individual stocks to give to charity. Donating appreciated stock means you never pay capital gains on them, and this turns your working hours into more money for the charity. If your capital gains are highly taxed, getting this right over time can be worth quite a bit to your charities.

Most of the tax losses will be harvested early in the life of your position. As stocks generally go up, you'll have fewer and fewer opportunities to lose money relative to where you bought things. This means the product looks like it easily pays for itself early in its life, then it stops being worth the fees, and then you realize it's hard to get out.

It's hard to get out because if you decide to go back to ETFs you'll need to either pony up all those deferred capital gains, give up on tracking your index, do a section 351 exchange, or figure out some way to keep tracking the index manually. This moat should make you very sensitive to fees on AUM that you're paying for the service.

I don't know of any Direct Indexing provider which has a flat fee structure. As far as I found, they all charge a fraction of AUM every year. There are also indirect costs that come from frequent trading but I've found US retail equity trading to be remarkably efficient.

You can technically do it yourself with Claude and a python script. It's a bit more complicated than it first appears (corporate actions?), but possible! However, the tax consequences of getting it wrong can be Very Expensive and there's no good way to check you've done it right. I recommend using a professional provider, but being sensitive to how much they charge for the service. All you're paying them for is making sure that their Claude did the right thing and you should not have to pay a steep price for that.

There's two classes of providers for this type of service: Retail Wealth Management companies like Fidelity and Schwab, and startups like Wealthfront and Frec.

If you already have a big account at one of the big Retail Wealth Management providers, it is possible you can negotiate them into providing you competitive Direct Indexing pricing. However, the fees publicly listed on their website are not competitive -- Schwab charges 0.35%, more than 10x what VOO charges. Fidelity charges a similar amount.

Wealthfront offers direct indexing at a fair price of 0.09% of AUM. But as far as I can tell, they don't offer exclusions. You can read more about their tax loss harvesting methodology here if you're interested.

Frec has the most compelling offering. I've been using their product for about a week. They charge 0.09% of AUM for Direct Indexing, which is 3x VOO. For my particular tax and life situation, the extra 0.06%/year is more than offset by tax loss harvesting and I have not seen a lower rate from anyone else. They seem to get excellent prices on the transactions they've done, and I've seen them provide < $1 of slippage for a $100k index buy. I recommend sticking to the S&P 500 Direct Index product because both the fees and transaction costs are low.

They have published a whitepaper explaining their tax loss harvesting methodology. It has enough detail to convince me they aren't doing anything crazy.

Frec also offer other products which I find less compelling: Diversify, Portfolio Margin, and Levered Long Short.

Diversify strategically sells concentrated positions, and seems plausibly good for some people. I talked about this in the Tax Loss Harvesting Strategies section.

Frec's Portfolio Margin rate is an excellent EFFR + 1%, and for my tax situation it's cheaper than the risk free rate. But unless you knew what volatility drag was before reading this sentence, know how to compute the optimal Kelly leverage for a portfolio, and know why you shouldn't lever up to the "optimal" point, you should almost certainly stay the hell away from this thing. Few things destroy wealth more quickly and completely than too much leverage.

Levered Long Short seems like it might help harvest tax losses faster than long-only direct indexing. But I have the sneaking suspicion that that's because it's better at losing money. I don't buy that factor tilts are a source of edge for retail -- if they were then wouldn't all the big asset managers have bid them up until all the edge was gone by now? "Value" tilts don't pass the sniff test, the top few stock holdings look like hot garbage and it sure seems like they're cheap for a reason. "Growth" tilts are just tech stocks and seem like a good way to juice returns in exchange for more concentration risk and less diversification. It also requires portfolio margin, which, again, you almost certainly shouldn't use.

If you make an account at Frec and know me personally, tell me before you sign up and I'll send you a referral code that gives us both $250. I'm not going to include a code in this article because I'm not a shill. I'm just some guy who happened to see a product he likes and is willing to explain why.