Tax-Loss Harvesting: The Tax Alpha Most Investors Misunderstand

Tax-loss harvesting works because the tax code lets you turn market pain into a current tax asset. You sell an investment at a loss, use that realized loss to offset realized gains elsewhere, and then buy a correlated replacement so your portfolio stays invested. The trick is simple in concept and easy to get wrong in practice. The IRS wash sale rule can disallow the loss if you buy back the same or a substantially identical security within 30 days before or after the sale, and that rule can reach across accounts in ways many investors miss ^[1].

The reason this matters is not theoretical. Wealthfront has published estimates suggesting tax-loss harvesting can add roughly 1.0% to 1.8% of after-tax alpha for typical portfolios in the early years, with the benefit declining over time as deferred gains catch up ^[6]. That is a meaningful number. But it is not a permanent return stream, and it is not available to every investor in every market regime. If you want the real edge, you need to understand the mechanics, the tax math, and the tradeoffs.

For investors who already think in terms of rebalancing discipline and asset allocation, TLH is best viewed as a tax overlay, not a separate strategy. It can also be easier to implement in portfolios that use broad ETFs rather than concentrated single stocks, which is one reason the mechanics matter more than the marketing.

What tax-loss harvesting actually does

At its core, TLH converts an unrealized loss into a realized loss. That realized loss can offset realized capital gains dollar for dollar. If losses exceed gains, up to $3,000 of net capital losses can generally be deducted against ordinary income each year, with the remainder carried forward indefinitely under current U.S. tax rules ^[1].

That sounds like a free lunch. It is not. The loss you harvest lowers your current tax bill, but the replacement security usually has a lower cost basis. When you eventually sell the replacement, some of the deferred gain comes back. In other words, TLH often creates a tax deferral, not a tax elimination. The value comes from the time value of money and, in some cases, from permanently converting what would have been short-term gains into longer-term gains or from using losses that would otherwise expire unused ^[2][3].

That distinction is where many investors get sloppy. They hear “tax alpha” and assume it means extra return with no tradeoff. The more accurate framing is that TLH can improve after-tax outcomes when the investor has taxable gains to offset, a sufficiently long horizon, and enough market volatility to generate harvestable losses. It is much less impressive when the portfolio is small, the tax rate is low, or the investor is likely to donate or bequeath the asset and receive a step-up in basis later.

The wash sale rule is where most mistakes happen

The wash sale rule is the gatekeeper. Under IRS rules, if you sell a security at a loss and buy a substantially identical security within 30 days before or after the sale, the loss is disallowed for now and added to the basis of the replacement shares ^[1]. The 30-day window is not just after the sale; it runs both directions. That catches investors who think they are safe because they waited a few days after selling, or because they bought in a different account.

What counts as “substantially identical” is not defined with a bright-line list, which is why investors should be conservative. Selling SPY and buying IVV is often treated as too close for comfort because both track the S&P 500 very closely. Selling SPY and buying VTI is more defensible because VTI holds the broader U.S. market, not just large-cap U.S. stocks. But even here, the right answer is not “anything different is fine.” It is “different enough that the IRS would have a hard time calling it substantially identical.”

If you want a broader framework for thinking about replacement risk, the logic is similar to what we discuss in ETFs vs. mutual funds and bid-ask spread: the wrapper matters, but the underlying exposure matters more.

A worked example: the full lifecycle of a harvested loss

Here is the cleanest way to see the economics. This example is illustrative, not actual performance data. It assumes a taxable investor in a 15% long-term capital gains bracket, no state tax, and no transaction costs. The point is to show the tax mechanics, not to forecast a real portfolio outcome.

Now compare two paths.

Path A: no harvesting. You hold the original shares from $100 to $110 and sell. Your taxable gain is $1,000. At a 15% capital gains rate, tax is $150.

Path B: harvest the loss. You sell at $80, realize a $2,000 loss, buy the replacement, and later sell that replacement at $110 for a $3,000 gain. If you have no other gains, the $2,000 loss offsets $2,000 of that gain, leaving $1,000 taxable gain. Tax is still $150. On a pure end-state basis, the tax bill can be the same.

So where is the benefit? In the timing. Under Path B, you deferred tax when the loss was harvested and paid later when the replacement was sold. If that deferral lasts years, the present value of the tax bill is lower. If you also had other realized gains in the same year, the harvested loss may have reduced current taxes immediately. That is the real source of TLH alpha: timing plus optionality, not magic.

There is one more wrinkle. If the harvested loss exceeds gains, up to $3,000 can offset ordinary income each year, and unused losses carry forward indefinitely ^[1]. That carryforward can be valuable for investors with lumpy gains, business income, or concentrated stock sales. It is much less valuable if you never realize gains and never use the carryforward.

When TLH adds real value — and when it does not

Wealthfront’s published estimates suggest TLH can add roughly 1.0% to 1.8% in after-tax alpha for typical portfolios in the early years, with the benefit fading as deferred gains accumulate ^[6]. That is directionally consistent with the academic literature: the value of harvesting depends on volatility, tax rates, holding period, and the investor’s ability to use losses against gains ^[2][3][4].

But the headline number is easy to misuse. A 1% to 2% after-tax improvement is not a guaranteed annual return boost. It is an estimate of tax benefit under certain assumptions. If markets trend smoothly upward with few drawdowns, there may be little to harvest. If the investor has no gains to offset and no ordinary income to absorb losses, the benefit is mostly deferred. And if the replacement fund is meaningfully different, the tracking error can swamp the tax edge.

There are also situations where TLH is simply not the best use of attention. If you are in a low tax bracket, the spread between current tax savings and future tax cost may be too small to matter. If the asset is earmarked for charity, harvesting may be unnecessary because donating appreciated shares can already eliminate embedded gains. And if the asset is likely to receive a step-up in basis at death, the deferred gain may never be taxed at all, which changes the calculus materially ^[1].

What investors get wrong about robo-advisor marketing

Robo-advisors often market TLH as a systematic source of tax alpha. Sometimes that is fair. Sometimes it is oversold. The honest assessment is that automated harvesting can be useful because it removes the behavioral friction that stops humans from acting when losses appear. It can also monitor many lots and many thresholds more efficiently than a person can ^[6].

But the marketing can blur three separate ideas: realized tax savings, deferred tax savings, and permanent tax savings. Those are not the same thing. A harvested loss that offsets a gain today is valuable. A harvested loss that merely postpones a gain until next year is still valuable, but less so. A harvested loss that gets washed out by a repurchase in the wrong account is not valuable at all.

That is why the academic literature is useful. Berkin and Ye’s work on tax management and HIFO accounting emphasizes that tax-aware portfolio management is about lot selection, realization timing, and replacement discipline, not just “selling losers” ^[2]. Arnott et al. also highlight that tax management can improve after-tax outcomes, but the benefit depends on the investor’s tax situation and the path of returns ^[3]. More recent empirical work has been more cautious about the size and persistence of TLH alpha than the marketing copy suggests ^[4].

A decision flowchart: should I harvest this loss?

Use this as a quick screen before you trade. It is intentionally conservative.

If you want a broader framework for avoiding false precision in portfolio decisions, the same discipline shows up in overfitting and backtest checklist. TLH is a strategy where small implementation errors can erase the edge. That is exactly the kind of thing systematic investors should respect.

Implementation checklist: how to do it without stepping on a rake

Here is a practical checklist you can use before placing a TLH trade. This is a process asset, not a recommendation.

For investors who want to go deeper into how orders and execution affect outcomes, the mechanics are similar to what we cover in life of a trade and order types explained. A tax decision still has to survive market microstructure.

So what should a serious investor actually do?

The right answer is not “always harvest” and it is not “never harvest.” It is to harvest when the tax benefit is real, the replacement is clean, and the account structure supports the trade. That usually means taxable investors with meaningful gains, enough volatility to create losses, and enough discipline to avoid wash sales across all accounts. It usually does not mean chasing every tiny dip or treating TLH as a substitute for good asset allocation.

My practical view: TLH is most valuable when it is boring. The best implementation is often the least dramatic one — broad, liquid funds, clear replacement rules, careful lot tracking, and a willingness to skip marginal trades. If you have to stretch to justify the trade, the tax alpha is probably smaller than the spreadsheet says.

And if you are still deciding whether the strategy belongs in your process, start with the account structure, then the tax bracket, then the replacement universe. In that order. The tax code rewards precision, not enthusiasm.

Closing thought: Tax-loss harvesting is one of the few places where a small amount of discipline can create a real after-tax edge. But the edge comes from respecting the rules, not from believing the marketing.