The impact on U.S. equities in four numbers and three client questions
Thesis
When holding stocks in a taxable account, the timing of a sale can matter as much as the choice of what to own. Portfolios built in ways that force unnecessary sales may give up more of their return to taxes than an investor may realize. For a portfolio held for many years, three challenges tend to arise:
- How much gain can still be deferred?
- Can the portfolio still produce harvestable losses?
- What is the potential tax cost for diversification?
For suitable investors, a tax-aware long/short extension may be one way to address these challenges.
Four Numbers
| 347 bps Avg. annual federal tax drag across eight overlapping 30-year windows (1926–2025)1 | $2.95 M Tax gap on $1M over 30 years: 36.5% of terminal zero-tax wealth (1996–2025)1,2 | 71 bps/yr Extra tax drag from forced realization (top 500 vs. broad market, 1996–2025)1 | 55/60 Age at which step-up at death is worth more than half the embedded tax (male/female)1 |
Same Market, Same Thirty Years — Three Different Conclusions
Growth of $1 million, 1996–2025:
Four Findings from the Research1
The tax drag can be large and timing driven.
Across eight overlapping 30-year windows starting every decade from 1926 to 1996, federal taxes cost 165 to 538 basis points a year: worst in 1936–65 (538 bps), with the least in 1996–2025 (165 bps). If an investor were to sell after ten years instead, the drag ranged from 49 to 553 bps depending on the start year.
After tax, two clients in the same model may have had very different experiences.
Structure alone can invert a ranking.
Ang’s top 500 portfolio (the 500 largest US stocks by market cap, re-ranked monthly, a stand-in for a large-cap index) beat the broad market before tax, 10.32% vs. 9.91%, and lost after tax, 7.96% vs. 8.26%. Each time a stock fell out of the top 500 it had to be sold and the gain realized; its tax drag ran 236 bps a year against 165 for the broad market.
Long-only harvesting has a ceiling.
In a mature, passive portfolio, Ang puts the harvestable-loss inventory at about 30 bps of portfolio value, refilling mainly in drawdowns. He notes that active direct-indexing tax alpha (50–150 bps) is a different benchmark; either way, the supply a long-only book can draw for loss-harvesting on is finite.
Most of the benefit comes from waiting.
Holding the same portfolio through an ETF wrapper, so that nothing changes except when gains are taxed, and Ang’s after-tax return improves by 34 bps a year if the investor eventually sells and 49 bps if they hold. Krasner and Sosner (2024) reach the same conclusion for tax-aware long/short.3
The benefit comes mainly from postponing gains as opposed to generating more losses. If an investor were to defer to death, the step-up forgives the embedded tax altogether; Ang’s estimate cuts the effective drag from 165 bps to 127 for a 30-year-old and to 103 by age 70 (federal income tax only).
Three Questions Worth Asking
Does the portfolio ever have to realize its gains?
Much of the embedded tax may never come due. The step-up only arrives at death, so its value today depends on how far off that probably is. For a 30-year-old, Ang values the eventual forgiveness at roughly a quarter of the embedded tax; by about 55 for men and 60 for women it passes half.1 The closer the step-up, the less reason to realize gains now.
Where does next year’s loss come from?
In a mature long-only account, next year’s losses often don’t come, due to the embedded low-cost basis of the underlying holdings. Harvesting is front-loaded (Liberman et al., 2023, show realized losses peak in the first year and taper after),2 and the supply refills mainly when markets fall.
What would repositioning cost in realized tax?
That depends on the portfolio, which is why the question is worth asking. Lot selection helps less than people expect. On Ang’s broad-market portfolio, switching from FIFO to HIFO moved the after-tax return only 14 bps.1 How the portfolio is built matters more than how individual lots are sold.
A close review on any direct indexing statement
If the account is three or four years old and hasn’t been through a major drawdown, consider the realized-loss history.
A harvest that shrinks year after year usually isn’t underperformance. More often the loss inventory is simply running down, consistent with the front-loaded pattern in the published research.
Where an Extension May Fit
- Portfolios that have already harvested most of the losses available to them and now sit on large embedded gains where the client has a genuine need to diversify.
- Concentrated low-basis positions the client wants to hold for the step-up while taking stock specific risk down around them; the longer the client is likely to live, the more that risk matters.
The mechanics of an extension are relatively straightforward. A long position can produce a loss when its price falls; a short position can produce one when its price rises. Holding both means the portfolio may no longer need a broad market decline to find losses, while gains on the long side can stay unrealized.
Where it May Not Fit
Clients past about 60 who are already diversified, have no large outside gains to offset, and simply want to keep deferring; for them, waiting may be a reasonable answer.
Other routes worth weighing case by case: adding new money to the account, gifting appreciated shares to charity, or an exchange fund.
What it May Cost
- Financing spreads
- Cost of borrowing stock and covering dividends on the short side
- Higher turnover and fees
- Tracking error
- Short-selling risk, including the potential for unlimited losses on short positions
Any potential tax benefit depends on the individual client circumstances and should be discussed with the client's tax adviser.
Why Now?
Ang measures the 1996–2025 window at 165 bps, the lowest of the 30-year windows in his study, and cites analysts who expect fiscal pressure to push tax rates higher.1 If rates do rise, a loss realized under today’s rates and applied against gains taxed at tomorrow’s may be worth more. That is a tax outcome, not an investment return.
It is important to complete proper due diligence on investment managers. It is also important to consider the investable time horizon as building a supply of losses can take time.
Why Gateway?
Gateway’s ability to combine long/short extension with option overlays in a single account is a differentiated, integrated, and straightforward solution to address challenging investment scenarios.
Since 1977, Gateway has provided quantitatively driven equity strategies and options-based solutions for risk-conscious, tax-aware investors. As an affiliate of Natixis Investment Managers, LLC, Gateway focuses on leveraging its extensive experience in managing customized portfolios while Natixis offers dedicated, institutional-level relationship management services.
When holding stocks in a taxable account, the timing of a sale can matter as much as the choice of what to own. Portfolios built in ways that force unnecessary sales may give up more of their return to taxes than an investor may realize. For a portfolio held for many years, three challenges tend to arise:
- How much gain can still be deferred?
- Can the portfolio still produce harvestable losses?
- What is the potential tax cost for diversification?
For suitable investors, a tax-aware long/short extension may be one way to address these challenges.
Four Numbers
Same Market, Same Thirty Years — Three Different Conclusions
Growth of $1 million, 1996–2025:
$2.95 million. The federal tax bill on a lifetime of market returns.
Four Findings from the Research1
The tax drag can be large and timing driven.
Across eight overlapping 30-year windows starting every decade from 1926 to 1996, federal taxes cost 165 to 538 basis points a year: worst in 1936–65 (538 bps), with the least in 1996–2025 (165 bps). If an investor were to sell after ten years instead, the drag ranged from 49 to 553 bps depending on the start year.
After tax, two clients in the same model may have had very different experiences.
Structure alone can invert a ranking.
Ang’s top 500 portfolio (the 500 largest US stocks by market cap, re-ranked monthly, a stand-in for a large-cap index) beat the broad market before tax, 10.32% vs. 9.91%, and lost after tax, 7.96% vs. 8.26%. Each time a stock fell out of the top 500 it had to be sold and the gain realized; its tax drag ran 236 bps a year against 165 for the broad market.
Long-only harvesting has a ceiling.
In a mature, passive portfolio, Ang puts the harvestable-loss inventory at about 30 bps of portfolio value, refilling mainly in drawdowns. He notes that active direct-indexing tax alpha (50–150 bps) is a different benchmark; either way, the supply a long-only book can draw for loss-harvesting on is finite.
Most of the benefit comes from waiting.
Holding the same portfolio through an ETF wrapper, so that nothing changes except when gains are taxed, and Ang’s after-tax return improves by 34 bps a year if the investor eventually sells and 49 bps if they hold. Krasner and Sosner (2024) reach the same conclusion for tax-aware long/short.3
The benefit comes mainly from postponing gains as opposed to generating more losses. If an investor were to defer to death, the step-up forgives the embedded tax altogether; Ang’s estimate cuts the effective drag from 165 bps to 127 for a 30-year-old and to 103 by age 70 (federal income tax only).
Three Questions Worth Asking
Does the portfolio ever have to realize its gains?
Much of the embedded tax may never come due. The step-up only arrives at death, so its value today depends on how far off that probably is. For a 30-year-old, Ang values the eventual forgiveness at roughly a quarter of the embedded tax; by about 55 for men and 60 for women it passes half.1 The closer the step-up, the less reason to realize gains now.
Where does next year’s loss come from?
In a mature long-only account, next year’s losses often don’t come, due to the embedded low-cost basis of the underlying holdings. Harvesting is front-loaded (Liberman et al., 2023, show realized losses peak in the first year and taper after),2 and the supply refills mainly when markets fall.
What would repositioning cost in realized tax?
That depends on the portfolio, which is why the question is worth asking. Lot selection helps less than people expect. On Ang’s broad-market portfolio, switching from FIFO to HIFO moved the after-tax return only 14 bps.1 How the portfolio is built matters more than how individual lots are sold.
A close review on any direct indexing statement
If the account is three or four years old and hasn’t been through a major drawdown, consider the realized-loss history.
A harvest that shrinks year after year usually isn’t underperformance. More often the loss inventory is simply running down, consistent with the front-loaded pattern in the published research.
Where an Extension May Fit
- Portfolios that have already harvested most of the losses available to them and now sit on large embedded gains where the client has a genuine need to diversify.
- Concentrated low-basis positions the client wants to hold for the step-up while taking stock specific risk down around them; the longer the client is likely to live, the more that risk matters.
The mechanics of an extension are relatively straightforward. A long position can produce a loss when its price falls; a short position can produce one when its price rises. Holding both means the portfolio may no longer need a broad market decline to find losses, while gains on the long side can stay unrealized.
Where it May Not Fit
Clients past about 60 who are already diversified, have no large outside gains to offset, and simply want to keep deferring; for them, waiting may be a reasonable answer.
Other routes worth weighing case by case: adding new money to the account, gifting appreciated shares to charity, or an exchange fund.
What it May Cost
- Financing spreads
- Cost of borrowing stock and covering dividends on the short side
- Higher turnover and fees
- Tracking error
- Short-selling risk, including the potential for unlimited losses on short positions
Any potential tax benefit depends on the individual client circumstances and should be discussed with the client's tax adviser.
Why Now?
Ang measures the 1996–2025 window at 165 bps, the lowest of the 30-year windows in his study, and cites analysts who expect fiscal pressure to push tax rates higher.1 If rates do rise, a loss realized under today’s rates and applied against gains taxed at tomorrow’s may be worth more. That is a tax outcome, not an investment return.
It is important to complete proper due diligence on investment managers. It is also important to consider the investable time horizon as building a supply of losses can take time.
Why Gateway?
Gateway’s ability to combine long/short extension with option overlays in a single account is a differentiated, integrated, and straightforward solution to address challenging investment scenarios.
Since 1977, Gateway has provided quantitatively driven equity strategies and options-based solutions for risk-conscious, tax-aware investors. As an affiliate of Natixis Investment Managers, LLC, Gateway focuses on leveraging its extensive experience in managing customized portfolios while Natixis offers dedicated, institutional-level relationship management services.
1: Andrew Ang, “Uncle Sam’s Cut: A Century of the Federal Tax Drag on US Equity Returns” (SSRN 6847631). The title phrase “the return nobody got” is journalist Jason Zweig’s, quoted in Ang (2026), CFA Institute Research Foundation.
2: The front-loaded pattern of realized loss harvesting: Liberman, Krasner, Sosner, and Maia de Freitas (2023), The Journal of Beta Investment Strategies 14(3).
3: The finding that tax-aware long/short benefits derive mainly from deferring gains: Krasner and Sosner (2024), The Journal of Wealth Management.
Gateway does not provide tax advice. Tax treatment and rates can and do vary over time. Investment decisions should be made based on an investor’s objectives and circumstances and in consultation with his or her investment and/or tax advisors.