Every Portfolio Is Long/Short, Even If You Never Short a Stock

Two investors each own a total U.S. stock index fund and nothing else. The first thinks of herself as a passive investor who has taken no view on anything. The second knows he is running a large position against every company listed outside the United States. They hold identical portfolios. The second description is the more useful one, because it names the position that will determine whether the decision looks smart in twenty years.

Where the phrase comes from

In November 2017, Corey Hoffstein of Newfound Research published a piece arguing that the difference between any two portfolios can be expressed as a long/short trade. He ended it with a line he has used ever since: “For us, it’s long/short portfolios all the way down.”¹ Revisiting the idea in 2023, he was cheerfully blunt about its reception: “Several years ago, I started using the phrase, ‘It’s long/short portfolios all the way down.’ I think it’s clever. Spoiler: it has not caught on.”²

Credit where it is due in both directions. The phrase is his. The underlying decomposition is older. Cremers and Petajisto stated it plainly in the working paper that introduced Active Share: any portfolio can be decomposed into a 100% position in its benchmark plus a zero-net-investment long-short portfolio on top.⁵ Their own footnote traces it back further, to Cliff Asness in 2004, who discussed the same decomposition from the standpoint of tracking error.⁶ Grinold and Kahn had built the textbook machinery around benchmark plus active portfolio well before that.⁷ What Hoffstein added was a memorable way to say it and a habit of applying it to ordinary retail decisions.

The four words the slogan is missing

The phrase is incomplete without a qualifier: long/short relative to a benchmark. This is the part most explanations skip, and skipping it turns a useful tool into a meaningless one, because without a reference portfolio you could describe every investor as short every asset on earth they do not happen to own.

A 100% U.S. stock portfolio shows this well. Against a global equity benchmark it carries a large underweight to international stocks. Against a U.S.-only benchmark it carries none at all. Against a 60/40 portfolio it is a big overweight to equities financed by a complete underweight to bonds. Same holdings, three different sets of implicit positions, depending entirely on the yardstick.

Choosing that yardstick is a real decision with real consequences, not a reporting formality. Berk Sensoy found that almost one third of diversified U.S. equity mutual funds specified a size and value benchmark in the prospectus that did not match the fund’s actual style, and that these mismatched benchmarks still drove investor flows, consistent with funds choosing them strategically.¹³ For a household, the most useful benchmark is usually the policy portfolio you would hold if you had no views at all.

The arithmetic behind the claim

Let $p_i$ be your weight in asset $i$ and $b_i$ the benchmark’s weight. Both portfolios are fully invested, so both sets of weights sum to one. Define the active weight as the difference:

Because both sides sum to the same total, the active weights sum to zero:

Which means the positive active weights and the negative active weights are equal in total size:

The positive side is your implicit long leg. The negative side is your implicit short leg. They balance by construction, which is the formal version of a plain idea: you cannot overweight anything without funding it from somewhere.

A concrete case. You hold 60% U.S. stocks, 20% international, 20% bonds. Your benchmark holds 50%, 30%, and 20%. Your active weights are +10% U.S., −10% international, and zero in bonds. Whatever you tell yourself about owning both U.S. and international stocks, the decision you made was to bet that ten points of U.S. equity will beat the ten points of international equity you sold to fund it.

Active Share is the size of that trade

Cremers and Petajisto gave the total size of the implicit trade a name. Active Share is half the sum of the absolute active weights, with the half preventing double counting so the measure runs from 0 to 100%:⁸

In the example above, Active Share is 10%. Ninety percent of that portfolio is the benchmark wearing a different label, and ten percent is the part that will make the results differ. Hoffstein writes the whole thing as a single expression, portfolio equals benchmark plus a scaled long/short portfolio, and draws the useful distinction: “Active share simply defines the quantity. The active bets, expressed in the long/short portfolio, will determine the quality.”³

Because ordinary funds do not short, their weights cannot go below zero. The largest underweight a long-only fund can take in any security is therefore that security’s benchmark weight, which is why Active Share for such funds is bounded between 0 and 100%. Underweights are capped, overweights are not, and the cap binds hardest on the many small positions where a manager might most want to express a negative view.

Active Share measures difference, and difference alone

The original research found that funds with the highest Active Share beat their benchmarks and that funds with the lowest, the closet indexers, lagged.⁸ Petajisto later put the figure at roughly 1.26% a year net of fees and costs for the most active stock pickers through 2009.⁹

That result did not survive scrutiny in the form it was popularized. Working from the same sample, Frazzini, Friedman, and Pomorski of AQR reported that Active Share “correlates with benchmark returns but does not predict actual fund returns,” and that within a given benchmark it is as likely to correlate positively with performance as negatively. High Active Share funds tend to have small-cap benchmarks and low Active Share funds large-cap ones, so sorting on Active Share largely amounts to sorting on benchmark type. Their conclusion was that the evidence does not support using it as a manager selection tool.¹⁰

Here is the part worth sitting with, because it reconciles the two camps. The same AQR paper that dismantled Active Share as a selection signal endorsed a different use for it: “active share may be useful, for example, in evaluating fees. In general, fees should be commensurate with the active risk the fund takes: If you deliver index-like returns, you should charge index-fund-like fees.”¹⁰ That is precisely what Hoffstein uses it for, and it survives the critique intact.

The fee arithmetic this makes possible

You pay a fund’s fee on all of your money, but only the active portion is doing anything the index fund would not have done for you. So divide the extra fee by the Active Share to find what you are really paying for the part that differs. Hoffstein’s worked example uses two real funds: an S&P 500 value ETF charging 0.18% against an S&P 500 fund charging 0.04% is 0.14% of additional fee, and with an Active Share of 42%, the implied fee on the active bets is 0.333%.¹

The same arithmetic is brutal on closet indexers. A fund charging a full percentage point more than an index fund while differing from it by only 20% is charging roughly 5% a year on the only part of your money that is actually different. This is a sleeve-level version of the whole-portfolio cost hurdle in our guide to financial product marketing red flags, and it is usually the more damning number.

Active Share and tracking error answer different questions

These get conflated constantly. Active Share compares holdings at a point in time. Tracking error measures how much the return difference bounces around over time, the standard deviation of your return minus the benchmark’s.

Two portfolios can share an Active Share and behave nothing alike. Swapping one megacap technology stock for a similar one moves holdings a lot and returns very little. Moving a smaller slice from Treasury bills into emerging market equity moves holdings less and returns much more. Cremers and Petajisto found empirically that a tracking error of 4 to 6% can correspond to an Active Share anywhere from 30% to 100%, which is why they argued activeness needs measuring in two dimensions rather than one.⁸ They also noted that tracking error on its own has no reliable relation to fund returns.

Active Share is the holdings-space view of how active you are. Tracking error and factor regressions are the return-space view. Our guide to Fama-French factor analysis covers that second axis in depth; this guide is its counterpart on the holdings side.

Why “all the way down”

The logic repeats at every level of a portfolio, which is the part of the phrase that does the teaching.

Asset allocation. A 60/40 portfolio is 20 points long bonds and short stocks against an 80/20 benchmark, and 20 points long stocks and short bonds against a 40/60 one. Your policy allocation is itself an active position against every other allocation you might have picked. That does not make any allocation wrong, since horizon, spending, and risk capacity differ, but it does mean the choice is a trade-off rather than a neutral starting point.

Geography. A U.S.-only equity portfolio is an overweight to U.S. companies funded by an underweight to everything else, and it quietly carries currency, sector, and valuation positions along with it. Hoffstein has written about exactly this kind of unintended baggage, finding in one case that almost half of an intended value premium was eroded by a currency bet the investor never meant to make.⁴

Index selection. A fund is passive with respect to the index it tracks. Choosing the index is still a portfolio decision, made by somebody, under rules about eligibility, float, weighting, and reconstitution. An S&P 500 fund carries an underweight to U.S. small caps and to companies that fail the committee’s criteria. Our guides on what passive really means and the S&P 500 under the hood develop this. Broad market-cap weighting remains an excellent default: it approximates what all investors hold in aggregate, needs little trading, requires no forecasts, and costs almost nothing.

Factor tilts. A value fund is an overweight to cheap stocks financed by an underweight to expensive ones, which is economically a long value, short growth position sitting on top of the market. The long-only constraint matters here, because the fund cannot push any holding below zero.

Individual stocks. Owning twenty picked stocks means carrying an underweight to the thousands you left out, and that side of the ledger can dominate. Bessembinder found that across CRSP common stocks from 1926 to 2016, only 42.6% beat one-month Treasury bills over their lifetimes, while 1,092 companies, slightly more than 4% of the roughly 25,300 in the sample, accounted for all of the market’s net wealth creation of about $35 trillion.¹⁴ A concentrated investor has to do two things, not one: avoid the losers and also happen to own enough of the rare enormous winners. Broad diversification is largely insurance against an omitted winner becoming a ruinous implicit short. We work through that in why five researched stocks are not safer than an index fund.

Funds stacked on funds. A portfolio of a total market fund, a technology ETF, a dividend ETF, and a few individual names cannot be understood from the four line items. At the security level it may be heavily overweight a handful of megacaps held through several wrappers at once and underweight companies that pay no dividend. Our pieces on the tech bro portfolio and VOO plus QQQ plus SCHD show how ticker count hides this.

What the framework does not claim

An underweight is not a short sale. Three different things get called “short” and it is worth keeping them apart: holding a zero weight in something, holding less of it than the benchmark does, and holding an actual negative position by borrowing a security and selling it. Only the third involves margin, borrow fees, dividends owed to the lender, forced liquidation, and losses that can exceed your capital. In an ordinary long-only fund your loss is limited to what you put in. The word is an economic analogy about relative exposure, and calling an underweight a short describes the direction of the bet rather than the payoff structure. If you want the literal version, our guide to tax-aware long-short strategies covers the products that actually do it.

Omitting something is not automatically an error. You do not need to own every commodity, currency, collectible, or private placement in existence. The comparison has to be a plausible policy benchmark for the framework to say anything at all.

Beating the benchmark is not the goal. A retiree cares about funding spending, surviving drawdowns, keeping liquidity, and paying less tax. You can beat a benchmark and still fail, or trail it and fund every goal you had. Relative analysis is a complement to planning rather than a substitute for it, and our guide on why an index fund is not a financial plan makes that case.

Investing is not zero-sum. Relative performance among investors is close to zero-sum before costs, because your overweight is somebody else’s underweight. The underlying enterprise is not: companies earn profits and pay dividends, and all investors can do well at once. The zero-sum arithmetic governs who beats whom.

See it on your own weights

Enter your allocation, then switch the benchmark and watch the same portfolio produce a different set of implicit longs and shorts. That switch is the lesson.

What the evidence says about acting on this

Sharpe’s arithmetic is the backdrop: before costs the average actively managed dollar earns what the average passive dollar earns, and after costs it earns less.¹⁵ We develop that in the concentration guide, so one qualifier is worth adding here: Lasse Pedersen has argued the identity is not quite exact, since indexes rebalance and securities enter and leave, so someone has to trade even to stay passive.¹⁶

The costs of playing are large in aggregate. Kenneth French estimated that investors spent about 0.67% of the aggregate value of the U.S. market each year in the search for superior returns over 1980 to 2006, and separately that the typical investor would have raised their annual return by about 67 basis points over that period by holding a passive market portfolio. Those are two distinct findings that happen to round to similar numbers.¹⁷ Fama and French found that the aggregate portfolio of active U.S. equity funds looks much like the market before costs, with few funds producing enough benchmark-adjusted return to cover their fees, though Harvey and Liu later re-examined the same question and found more evidence of skill, so the literature is not fully closed.¹⁸¹⁹

The recent scorecards point the same way. S&P’s SPIVA report for year-end 2025 found 79% of active U.S. large-cap funds underperformed the S&P 500 that year, up from 65% in 2024 and the fourth-worst showing in the scorecard’s 25-year history.²⁰ Morningstar’s Active/Passive Barometer found 38% of active funds both survived and outperformed the asset-weighted average of comparable passive peers in 2025, and 21% did so over the trailing decade.²¹ A single year proves little, and the decade figure is the one that should shape expectations.

On implementation, the short leg is harder to capture than theory implies. Israel and Moskowitz found that long positions comprise almost all of the size premium, about 60% of value, and roughly half of momentum, with the importance of shorting varying by firm size and moving in opposite directions for value and momentum.²² Blitz, Baltussen, and van Vliet reached a similar practical conclusion, that most of the added value sits in the long legs and the short legs add little for most investors.²³ Momentum is the genuine exception, so the tidy claim that factor premia live entirely on the long side is too strong. For a DIY investor the upshot is that a diversified long-only fund captures most of what the research describes without any of the machinery of literal shorting.

A checklist for any deviation you hold

For each meaningful difference between your portfolio and your benchmark, answer these in writing. The exercise converts a preference into a claim that can be checked later.

1. What am I overweight?
2. What am I underweight to fund it?
3. Why should the first outperform the second?
4. Is the expected reward payment for a risk, a behavioral pattern, or a forecast about the future?
5. What do fees, taxes, and turnover cost me to hold it?
6. How long could it underperform while still being a good decision?
7. Can I actually hold it through that stretch?
8. What evidence would tell me I was wrong?

“International stocks look unattractive” becomes “I am overweighting U.S. equities and underweighting international ones because I believe valuation and profitability differences justify it, after accounting for the currency and sector positions that come attached.” The second version can be wrong, which is what makes it useful.

Who this framework helps, and who it does not

It is most valuable if you hold several overlapping funds, use factor or sector tilts, own concentrated employer stock, invest in one country by default, apply values-based exclusions, make tactical shifts, or compare advisors and model portfolios. In all of those cases the interesting risk is usually hiding in what you left out.

It matters less if you own a single target-date or balanced fund and plan to keep doing so. Knowing the fund makes allocation choices on your behalf is worth understanding, but running formal active-weight analysis on it adds work without changing any decision.

It can actively hurt if you use it to shop for a benchmark that flatters recent results, treat every underweight as a forecast that demands action, stare at relative performance when your objectives are goal-based, or conclude that because you are “already short” things you may as well use leverage or literal short positions. More analysis is not the same as more trading.

How Summitward helps

Summitward does not compute Active Share, and it does not look through your ETFs to their underlying holdings, so it cannot tell you your exact implicit position in a single company. Four things it does do bear directly on this framework.

The rebalancing page computes drift for every asset class as current weight minus target weight, under 5/25 bands. That is the same subtraction this guide is built on, measured against your own policy portfolio at the asset-class level rather than against an index at the security level, and for most households it is the more relevant comparison anyway.

The portfolio dashboard runs a real factor regression against Ken French’s data, with your choice of CAPM, three-factor, five-factor, or Carhart, reporting betas, t-statistics, p-values, alpha, and R-squared. That is the return-space answer to which bets you are making, and it is the honest substitute for a holdings-based Active Share. Its benchmark tab already reports your tracking error alongside alpha, beta, information ratio, and up and down capture, so the second dimension Cremers and Petajisto asked for is a tab away. Its concentration tab computes a Herfindahl index and the effective number of positions you hold.

The portfolio explorer accepts negative weights, and ships a 130/30 template holding 130% equities against a −30% cash position, showing net and gross exposure separately. If you want to see what a literal long/short portfolio does to your risk statistics rather than an implicit one, you can type it in.

Frequently asked questions

Am I really short the stocks I do not own?

Not literally. You have no borrow, no margin, and no obligation, and your loss is capped at what you invested. Relative to a benchmark that owns them, your zero weight behaves like a short position in the sense that you gain when those stocks lag and lose when they lead. It is a statement about relative exposure rather than about legal ownership.

Does a high Active Share mean a fund is worth its fee?

It means the fund is different, which is a precondition for being worth an active fee rather than evidence of being worth one. AQR’s work found Active Share does not predict fund returns once benchmark differences are accounted for. The defensible use is the fee test: divide the extra fee by the Active Share to see what you pay for the part that differs.

Which benchmark should I use?

For a household, use the policy portfolio you would hold with no views: your intended stock and bond mix, at broad market weights, with your real constraints. Choosing benchmarks after the fact to make results look better is the main way this framework gets abused.

Is market-cap weighting neutral?

It is the closest thing to neutral available, because it is what all investors hold in aggregate and it requires no forecast. It is still a rules-based portfolio, and against a different weighting scheme it carries active positions of its own. Being an excellent default and being free of all choices are different things.

Should I use this to justify shorting or leverage?

No. Literal shorts bring borrow costs, margin calls, forced liquidation, tax complexity, and losses that can exceed your capital. The research on factor implementation suggests most of the available premium in size and value sits on the long side anyway. A diversified long-only portfolio expresses the lesson without the machinery.

Key takeaways

• Because your weights and the benchmark’s both sum to 100%, the active weights sum to zero, so every overweight is financed by an underweight of equal size.
• The decomposition only exists relative to a stated benchmark. Change the benchmark and the same portfolio has a different set of implicit longs and shorts, which makes benchmark choice a real decision.
• Active Share sizes the implicit trade. It measures how different you are, and AQR found it does not predict returns once benchmark differences are controlled for, though the same paper endorses it for judging whether a fee matches the activity being delivered.
• Divide the extra fee by Active Share to see the cost of the part that actually differs. A fund charging a point more while being 20% different is charging roughly 5% on the only money doing anything.
• What you leave out can dominate results: only 42.6% of U.S. stocks beat T-bills over their lifetimes, and about 4% produced all the net wealth creation, so omitting winners is the expensive failure.
• Use it as a diagnostic. It is an argument for writing down both sides of every decision, not for shorting, leverage, or more trading.

Related guides

• Fama-French Factor Analysis the return-space counterpart to this guide’s holdings-space view.
• Five Stocks Are Not Safer Than an Index Fund what happens when the omitted side of the ledger contains the winners.
• Passive Investing Is a Label somebody chose the index rules before you bought the fund.
• Compensated vs. Uncompensated Risk which relative bets have historically paid for themselves.
• How Much in Emerging Markets? the single-sleeve version of this arithmetic, with a checkup tool.
• Portable Alpha and Return Stacking what happens when the funding decision is made explicit.
• Tax-Aware Long-Short the products that take literal short positions, and their real costs.

Sources

1. Hoffstein, C. (2017). It’s Long/Short Portfolios All The Way Down. Newfound Research, November 6, 2017. Source of the phrase and of the Active Share fee arithmetic.
2. Hoffstein, C. (2023). Portfolio Tilts versus Overlays: It’s Long/Short Portfolios All the Way Down. Newfound Research, April 12, 2023.
3. Hoffstein, C. (2018). Thinking in Long/Short Portfolios. Newfound Research, March 5, 2018. Portfolio = benchmark + a scaled long/short portfolio.
4. Hoffstein, C. (2018). Factor Investing & The Bets You Didn’t Mean to Make. Newfound Research, January 16, 2018. Source of the unintended currency bet finding.
5. Cremers, K. J. M., & Petajisto, A. (2006). How Active Is Your Fund Manager? A New Measure That Predicts Performance. Yale ICF working paper, SSRN 891719. States the benchmark-plus-long/short decomposition on page 1.
6. Asness, C. S. (2004). An Alternative Future. Journal of Portfolio Management 30(5), 94–103. Cited in Cremers and Petajisto’s footnote 1 as an earlier statement of the same decomposition, framed through tracking error.
7. Grinold, R. C., & Kahn, R. N. Active Portfolio Management. The textbook treatment of benchmark plus active portfolio, residual risk and return, and the information ratio. Cited as background.
8. Cremers, K. J. M., & Petajisto, A. (2009). How Active Is Your Fund Manager? A New Measure That Predicts Performance. Review of Financial Studies 22(9), 3329–3365. Active Share definition, the 0 to 100% bound for long-only funds, and the finding that a 4 to 6% tracking error spans a wide Active Share range.
9. Petajisto, A. (2013). Active Share and Mutual Fund Performance. Financial Analysts Journal 69(4). The 1.26% net figure for the most active stock pickers through 2009.
10. Frazzini, A., Friedman, J., & Pomorski, L. (2016). Deactivating Active Share. Financial Analysts Journal 72(2). The critique, and the endorsement of Active Share for evaluating fees.
11. Sensoy, B. A. (2009). Performance evaluation and self-designated benchmark indexes in the mutual fund industry. Journal of Financial Economics 92(1), 25–39. Almost one third of funds specify a mismatched benchmark.
12. Bessembinder, H. (2018). Do stocks outperform Treasury bills? Journal of Financial Economics 129(3), 440–457. CRSP 1926–2016: 42.6% of stocks beat one-month T-bills; 1,092 firms account for all net wealth creation.
13. Sharpe, W. F. (1991). The Arithmetic of Active Management. Financial Analysts Journal 47(1), 7–9.
14. Pedersen, L. H. (2018). Sharpening the Arithmetic of Active Management. Financial Analysts Journal 74(1). Why the identity is not exact once indexes rebalance.
15. French, K. R. (2008). Presidential Address: The Cost of Active Investing. Journal of Finance 63(4), 1537–1573. The 0.67% of market value and the 67 basis point figures are separate results.
16. Fama, E. F., & French, K. R. (2010). Luck versus Skill in the Cross-Section of Mutual Fund Returns. Journal of Finance 65(5), 1915–1947.
17. Harvey, C. R., & Liu, Y. (2022). Luck versus Skill in the Cross Section of Mutual Fund Returns: Reexamining the Evidence. Journal of Finance, doi 10.1111/jofi.13123. Finds more evidence of skill than the 2010 study.
18. S&P Dow Jones Indices. SPIVA U.S. Year-End 2025 Scorecard. spglobal.com. 79% of active large-cap U.S. funds trailed the S&P 500 in 2025.
19. Morningstar. Active/Passive Barometer, year-end 2025 (published February 19, 2026). morningstar.com. 38% of active funds survived and outperformed the asset-weighted average of comparable passive peers in 2025; 21% over ten years.
20. Israel, R., & Moskowitz, T. J. (2013). The role of shorting, firm size, and time on market anomalies. Journal of Financial Economics 108(2), 275–301. Long positions comprise almost all of size, 60% of value, and half of momentum profits.
21. Blitz, D., Baltussen, G., & van Vliet, P. (2020). When Equity Factors Drop Their Shorts. Financial Analysts Journal 76(4). Most added value comes from the long legs.

Editor’s note

Educational content, not investment advice. The calculator works at the asset-class level, so it cannot see bets held inside a fund such as sector or single-stock concentration, and its benchmark presets scale MSCI ACWI country weights as of May 29, 2026 to each preset’s stock and bond mix. Quotations were verified against the linked Newfound Research posts and journal sources as of July 2026. The phrase belongs to Corey Hoffstein; the decomposition it describes predates him, and the sources above credit it accordingly.