Direct Indexing vs. ETFs: The Math Behind the Tax-Loss Harvesting Debate

In 2025, the S&P 500 climbed roughly 25%, meaning almost every diversified exchange-traded fund (ETF) investor saw their account balance rise—and almost no one harvested a tax loss. Yet in that same up year, direct-indexing managers quietly captured billions in tax-deductible losses for their clients.

This divergence exposes the structural limitation of the ETF, the vehicle that has dominated passive investing for three decades. When an investor buys an S&P 500 ETF, they own a single security. If the overall market rises, the fund rises, and the investor has no capital losses to claim, even though dozens of underlying companies within that index inevitably declined.[2]

"For most of the last three decades, index investing meant buying a fund," notes Forbes. "Direct indexing breaks that bundle apart." Instead of owning shares of a fund, the investor directly holds the 500 underlying stocks in a separately managed account built to mimic the index's weightings. This allows the portfolio software to scan for individual losers—selling them to book a tax loss, and immediately replacing them with correlated stocks to maintain the target market exposure.[1]

The financial impact of this granular control is substantial for high earners. A 2025 study analyzing real client accounts found that direct-indexed portfolios harvested an average of $18,281 in losses per account, compared to just $4,808 for ETF-only accounts. For an investor facing top federal and state capital gains rates, those harvested losses translate directly into thousands of dollars in reduced tax liability, which can be reinvested to compound over time.

Academic simulations confirm this advantage. A 2026 working paper from Harvard Business School examined the after-tax tradeoff between the two strategies, concluding that "direct indexing often outperforms ETFs for tax-sensitive investors, particularly over shorter investment horizons and when contributions are recurring rather than lump sum."[2]

However, the strategy carries a built-in expiration date known as "portfolio lock-in." As the market rises over a five-to-ten-year horizon, nearly all the individual stock lots in a direct-indexed portfolio appreciate significantly above their original purchase price. Once that happens, the opportunities to harvest new losses dry up, and the portfolio's "tax alpha" decays from an initial 1.5% to 2.5% down to less than 0.3% annually.

Furthermore, the complexity of managing hundreds of individual cost bases makes it difficult to exit the strategy. An investor who wants to move a mature, highly appreciated direct-indexed portfolio back into a simple ETF cannot do so without triggering massive capital gains taxes on the sale of those individual stocks.

As zero-commission trading and fractional shares have pushed direct indexing minimums down from millions of dollars to as low as $5,000, the strategy is no longer confined to family offices. The deciding factor is no longer access, but utility: an investor paying a 15% capital gains rate will mathematically underperform the ETF baseline once the direct indexing fee is applied, while an investor paying 54% in California will net thousands in tax alpha before the portfolio locks in.[1][3]