Tax-Loss Harvesting: When Does It Actually Add Alpha?

The Deferral Illusion

Tax-loss harvesting (TLH) is one of the most marketed features in modern wealth management. Robo-advisors claim it adds 1% to 2% per year in after-tax alpha. Direct indexing platforms advertise it as their primary value proposition. Yet the academic literature tells a more nuanced story: TLH does not eliminate taxes; it defers them. The benefit is real but diminishing, and its magnitude depends critically on investor-specific variables that marketing materials rarely quantify. Constantinides (1984) first formalized the theoretical case, showing that optimal tax trading in the presence of capital gains taxes creates value through the time value of money. Arnott, Berkin, and Ye (2001) translated this theory into practical estimates, while Chaudhuri, Burnham, and Lo (2020) provided the most rigorous empirical evaluation to date. This article examines when TLH actually adds meaningful after-tax alpha and when it amounts to little more than complexity for its own sake.

How Tax-Loss Harvesting Works

The mechanics of TLH are straightforward. When a security in a taxable portfolio declines below its purchase price, the investor sells that position to realize a capital loss. This loss can offset realized capital gains elsewhere in the portfolio (or up to $3,000 of ordinary income per year in the US). The investor simultaneously purchases a similar, but not "substantially identical," security to maintain market exposure.

The critical insight is that TLH does not make losses disappear. It accelerates the recognition of losses into the present while deferring gains into the future. The replacement security has a lower cost basis than the original, meaning when it is eventually sold, the capital gain will be correspondingly larger. The economic benefit comes from the time value of money: paying taxes later is worth less in present value terms than paying them today.

Constantinides (1984) showed that this tax-timing option has positive value under essentially all realistic scenarios. The optimal strategy involves realizing losses as soon as they arise (short-term losses are particularly valuable because they offset gains taxed at higher ordinary income rates) while deferring gains as long as possible. In his framework, the after-tax benefit can be modeled as an interest-free loan from the government, where the loan principal is the tax that would have been owed on the realized gain.

Quantifying the After-Tax Alpha

The after-tax benefit of TLH depends on several interacting variables: the investor's tax rate, the investment horizon, the volatility of the underlying securities, the turnover rate, and the eventual disposition method. Berkin and Ye (2003) provided one of the most cited quantification frameworks, estimating that TLH adds approximately 0.75% to 1.50% annually in the early years of a portfolio's life, declining to 0.20% to 0.50% over 15 to 20 years as the cost basis erodes and fewer harvesting opportunities remain.

Chaudhuri, Burnham, and Lo (2020) conducted the most comprehensive empirical study, using actual market data from 2000 to 2018. They found that the median annualized TLH alpha for a high-bracket investor was approximately 1.08% over the first five years, 0.71% over ten years, and 0.42% over the full sample period. Crucially, they documented substantial variation across investors: the interquartile range of outcomes was nearly as wide as the median benefit itself, meaning that for some investors TLH added negligible value while for others it was highly significant.

The Deferral vs. Elimination Distinction

The most commonly misunderstood aspect of TLH is the distinction between tax deferral and tax elimination. TLH defers taxes; it does not eliminate them. The deferred tax liability remains embedded in the portfolio as a reduced cost basis.

Consider a simple example. An investor buys $100,000 of a stock that drops to $80,000. They harvest the $20,000 loss (saving $4,760 at the 23.8% long-term capital gains rate) and purchase a replacement security at $80,000. If the replacement subsequently rises to $120,000, the taxable gain is $40,000 (from the $80,000 basis), not $20,000 (from the original $100,000 basis). The investor eventually owes taxes on the additional $20,000 of embedded gains.

The benefit is the present value of the deferral. If the investor defers $4,760 of tax for ten years and the after-tax discount rate is 5%, the present value of the benefit is approximately $1,838; $4,760 minus ($4,760 / 1.05^10). This is real value, but it is substantially less than the $4,760 in immediate tax savings that marketing materials often cite.

There are two scenarios in which TLH can approach true tax elimination rather than mere deferral. First, if the investor donates the appreciated replacement securities to charity, the embedded gain is never realized. Second, upon the investor's death, the cost basis of securities in the US receives a step-up to fair market value, effectively eliminating the deferred tax liability. Arnott, Berkin, and Ye (2001) estimated that when combined with the step-up in basis at death, TLH can add 0.75% to 1.00% in true (non-deferred) annual alpha for long-horizon investors.

The Wash Sale Rule and Implementation Constraints

The IRS wash sale rule (Section 1091) prohibits claiming a loss deduction if the investor purchases a "substantially identical" security within 30 days before or after the sale. This 61-day window (30 days before, the sale date, and 30 days after) creates meaningful implementation challenges.

The "substantially identical" standard is not precisely defined in the tax code, creating ambiguity. The IRS has not issued formal guidance on whether two index funds tracking the same benchmark are substantially identical. In practice, most tax advisors consider funds tracking different indices (such as the S&P 500 and the Russell 1000) to be sufficiently different, though this interpretation has not been tested in court.

Direct indexing platforms address the wash sale constraint by holding individual stocks rather than funds. When harvesting a loss on a particular stock, the platform can purchase a different stock in the same sector or with similar factor characteristics. This approach dramatically increases the number of harvesting opportunities because individual stocks are far more volatile than diversified indices.

Direct Indexing vs. ETF-Based Harvesting

The rise of direct indexing has fundamentally changed the TLH landscape. By holding individual securities rather than funds, direct indexing platforms can harvest losses at the individual stock level, capturing far more tax alpha than fund-level approaches.

Berkin and Ye (2003) estimated that direct indexing harvests approximately three to five times more losses than ETF-based approaches in the first decade. The advantage stems from the dispersion of individual stock returns: even when the market is flat or rising, many individual stocks decline, creating harvesting opportunities that are invisible at the fund level. In a 500-stock portfolio tracking the S&P 500, roughly 40% to 50% of individual stocks may be trading below their cost basis at any given time, even during bull markets.

However, the incremental benefit of direct indexing must be weighed against its higher fees. At a typical fee of 0.25% to 0.40% per year, direct indexing needs to harvest at least $2,500 to $4,000 more in annual losses per $1 million portfolio than an ETF approach just to break even on fees. For portfolios below $500,000, the fee differential frequently consumes most of the incremental tax benefit.

When TLH Does Not Work

TLH is not universally beneficial. Several common scenarios reduce or eliminate its value entirely.

Tax-exempt accounts (IRAs, 401(k)s, Roth accounts) derive zero benefit from TLH because gains and losses within these accounts have no tax consequences. Given that tax-advantaged accounts represent roughly 35% of US household financial assets, this is a significant limitation.

Low-bracket investors receive minimal benefit. At a 0% long-term capital gains rate (applicable to single filers with taxable income below roughly $47,000 in 2025), TLH has essentially no value. Even at the 15% rate, the present value of deferral is modest, and the added complexity and tracking costs may exceed the benefit.

Highly appreciated portfolios with very low cost bases have diminishing harvesting opportunities. After years of bull market returns, the number of positions with harvestable losses shrinks, and the portfolio becomes increasingly tax-constrained. Arnott, Berkin, and Ye (2001) documented this "basis erosion" effect, showing that TLH alpha declines quasi-exponentially as the portfolio ages.

Frequent trading strategies that already generate substantial short-term gains may find that TLH merely shifts the timing of tax payments without generating meaningful present-value savings, particularly if the harvested losses are used to offset short-term gains that would have been realized regardless.

Investors who plan to hold positions until death receive less incremental benefit from TLH if they already expect to benefit from the step-up in basis. TLH in this case primarily accelerates the tax benefit rather than creating new value.

The Diminishing Returns Problem

One of the most important findings in the TLH literature is that its benefits decay over time. This occurs for two reinforcing reasons.

First, cost basis erosion: as losses are harvested and replacement securities are purchased at lower prices, the portfolio's aggregate cost basis declines. Over time, a larger fraction of the portfolio consists of embedded gains, and fewer positions carry harvestable losses. Chaudhuri, Burnham, and Lo (2020) showed that the percentage of positions with harvestable losses drops from roughly 45% in Year 1 to approximately 15% by Year 15 in a typical equity portfolio.

Second, the present value of additional deferral shrinks as the deferral period extends. Deferring a tax payment by one additional year when you have already deferred for 20 years adds very little incremental present value compared to the first year of deferral.

These two forces combine to produce a sharply declining benefit curve. For a high-bracket investor, TLH alpha may exceed 1.0% in the first three years, but it typically settles to 0.2% to 0.4% by Year 15 and becomes negligible beyond Year 25.

Practical Framework for Evaluating TLH

The decision of whether to implement TLH should be based on a clear-eyed assessment of its costs and benefits. The costs include management fees (for direct indexing), increased tax-reporting complexity, wash sale monitoring, tracking error relative to the desired benchmark, and the risk of making suboptimal investment decisions to accommodate tax considerations.

The academic consensus, synthesized across Constantinides (1984), Arnott, Berkin, and Ye (2001), Berkin and Ye (2003), and Chaudhuri, Burnham, and Lo (2020), suggests that TLH adds meaningful after-tax alpha (0.50% to 1.50% annually) for high-bracket investors in the first decade of a portfolio's life, particularly when implemented through direct indexing on portfolios exceeding $500,000. For lower-bracket investors, shorter horizons, smaller portfolios, or tax-exempt accounts, the benefit is marginal to nonexistent, and the added complexity may not be justified.

The most honest framing of TLH is as an interest-free loan from the government, not as free money. The loan has real value, but that value depends on how long you hold it, what interest rate you would otherwise pay, and whether you ever have to repay the principal. For the right investor in the right circumstances, it is one of the few genuine sources of after-tax alpha. For everyone else, it is an expensive solution to a modest problem.

References