A portfolio is normally measured by the return it earns. This measures the return it keeps. Enter an allocation and the tax characteristics of each sleeve, and the instrument returns the drag in basis points, the cost in cash, and what that costs over ten, twenty and thirty years of compounding.
Open the instrumentThe usual calculator asks for a return and a tax rate and multiplies the two. The answer is arithmetic, and it is almost always wrong, because it treats every dollar of return as though it were taxed the same way at the same moment.
It is not. A dollar of bond interest is taxed this year at the investor’s highest rate. A dollar of unrealised appreciation in a private equity position is taxed in eight years, at a lower rate, and until then it compounds. Between those two extremes sit every other kind of return a large portfolio owns, and the difference between them is worth more, over a long horizon, than most of the manager selection that occupies the rest of the process.
So this instrument asks for four things about each asset class rather than one: how much of the return is paid out as income, what share of that income is taxed at ordinary rates, how much of the accumulated gain is realised each year, and how much of that realised gain is short-term. Those four numbers decide the answer. Two portfolios with identical weights and identical gross returns can be hundreds of basis points apart on them alone, and the worked example further down this page shows two sleeves of identical size differing tenfold.
The concept itself — character, turnover and location — is set out in full in Tax Intelligence. This page is the measurement.
Seven asset classes are loaded as a starting point. They are illustrative, not a recommendation and not drawn from any dataset — every figure is meant to be replaced with yours. Add or remove lines as your allocation requires.
Nothing here is proprietary and nothing is hidden. The model is stated in full so that a reader who disagrees with it can say precisely where.
Each asset class is treated as producing two streams. Income — interest, dividends, distributions — is taxed in the year it arises. Appreciation is taxed only when realised, and only the part that is realised. The annual tax borne by a sleeve, expressed as a fraction of that sleeve’s own value, is therefore:
tax = r × [ i × Ri + (1 − i) × t × Rg ]
Ri = o × ORD + (1 − o) × PREF
Rg = s × ORD + (1 − s) × PREF
Portfolio drag is the weighted sum of the sleeve figures, and each sleeve’s contribution to it is what the attribution table reports. That is the whole of the headline calculation, and it is deliberately legible: a reader can reproduce any line of the attribution table on paper.
The ten, twenty and thirty-year figures are produced by running the portfolio forward one year at a time and tracking cost basis, because basis is what makes the second decade differ from the first. Income is taxed and the remainder reinvested. Appreciation accumulates. Each year, turnover realises a share of whatever unrealised gain has built up, tax is paid out of the portfolio, and the proceeds are reinvested with a stepped-up basis. The portfolio is rebalanced to its target weights at each year end, with basis carried proportionally, so rebalancing neither creates nor destroys basis.
This has a consequence worth stating plainly, because it is the opposite of what a simple model would suggest. The first-year drag is a floor, not an average. In year one, basis equals market value, so the only gain available to be realised is that year’s appreciation. In year fifteen, a sleeve realising a fifth of its gains each year is realising a fifth of fifteen years of accumulation. Drag rises as unrealised appreciation builds, which is why the realised after-tax compound rate over thirty years in the worked example below is lower than the first-year after-tax return.
The untaxed benchmark is a benchmark, not an alternative. No structure available to a taxable family reaches it, and any material presenting it as an achievable outcome is selling something. It is in the table to size the gap, because the gap is what any structural decision is bidding for a share of — and a structure that closes part of it at a cost is a different calculation entirely, with its own break-even.
The liquidation column settles the deferred bill. Deferral is not forgiveness. A portfolio that has compounded for thirty years on unrealised gains is carrying a liability, and the final column shows what remains if the whole position were sold on the last day at the preferential rate supplied. Most families never liquidate in full, and in several jurisdictions a transfer at death changes the answer again — but a model that quietly omitted the embedded liability would be flattering itself.
A model is an argument about assumptions. These are the ones being argued, and the places where a serious user should be careful.
You supply the two effective rates that apply to you. Nothing on this page asserts a rate, a threshold or a residence, and no default is presented as anyone’s actual position.
This is not a tax computation and cannot substitute for one. It answers what taxation is costing a portfolio in compounding terms, given inputs you provide, which is a different question from what is owed and when.
Each sleeve compounds at its stated return every year. There is no volatility, no sequence risk and no distribution of outcomes. Drag is close to linear in return, so this matters less for the drag figure than for the terminal-wealth figures, which should be read as central cases rather than forecasts.
No loss harvesting, no carryforwards, no offsetting. A disciplined harvesting programme will reduce realised drag below what this model reports, particularly in the equity sleeves. The figures are therefore conservative in that specific direction.
Turnover is taken to reflect the trading a strategy requires, rebalancing included. If your rebalancing discipline is heavier than the turnover you enter, raise the turnover rather than expecting the model to infer it.
Liabilities are settled out of the assets themselves, and no external cash funds them. A family paying investment tax from operating income compounds differently, and better, than this shows.
The model cannot know what a fund actually distributed. For pooled and alternative vehicles, character and realisation are decided inside the fund and reported after the fact, which is precisely why those sleeves are the hardest to estimate and the most consequential to get right.
Nothing about the calculation depends on the existence of insurance, trusts or any other structure. It is a measurement instrument, and a reader who never reads another page on this site should still get an honest number from it.
A diversified allocation of the kind a family office might hold, run through the instrument at an ordinary rate of 40% and a preferential rate of 25%. Both rates are illustrative round numbers chosen to demonstrate the mechanism, not a claim about any jurisdiction. The figures below are produced by the same engine that runs above, on exactly these inputs.
| Asset class | Weight | Gross return | Income share | Ordinary | Turnover | Short-term |
|---|---|---|---|---|---|---|
| Public equity — index, low turnover | 30% | 7.0% | 25% | 0% | 5% | 0% |
| Public equity — actively managed | 10% | 7.0% | 25% | 0% | 60% | 30% |
| Investment-grade fixed income | 10% | 4.5% | 100% | 100% | 0% | 0% |
| Hedge fund strategies | 15% | 8.0% | 20% | 100% | 80% | 60% |
| Private credit | 15% | 9.0% | 95% | 100% | 0% | 0% |
| Private equity and venture | 15% | 12.0% | 0% | 0% | 12% | 0% |
| Cash and equivalents | 5% | 4.0% | 100% | 100% | 0% | 0% |
The portfolio returns 7.80% gross. Tax takes 147 basis points of it, leaving 6.33% in the first year — 18.9% of the gross return, and $735,866 in cash.
| Asset class | Weight | Drag on the sleeve | Contribution | Annual cost | Share of total |
|---|---|---|---|---|---|
| Private credit | 15% | 342 bps | 51 bps | $256,500 | 34.9% |
| Hedge fund strategies | 15% | 238 bps | 36 bps | $178,560 | 24.3% |
| Investment-grade fixed income | 10% | 180 bps | 18 bps | $90,000 | 12.2% |
| Public equity — index | 30% | 50 bps | 15 bps | $75,469 | 10.3% |
| Public equity — active | 10% | 137 bps | 14 bps | $68,338 | 9.3% |
| Cash and equivalents | 5% | 160 bps | 8 bps | $40,000 | 5.4% |
| Private equity and venture | 15% | 36 bps | 5 bps | $27,000 | 3.7% |
| Portfolio | 100% | — | 147 bps | $735,866 | 100% |
The instructive line is the comparison between the two 15% sleeves. Private credit and private equity are the same size in this portfolio. Private credit contributes 51 basis points of drag; private equity contributes five. The difference is not that one is a worse investment — private equity is carrying the higher gross return of the two. It is that private credit pays its return out as interest every year, and private equity does not pay it out at all until it exits.
The second observation is that the largest holding is not the largest problem. Index equity is 30% of the portfolio and the fourth-largest contributor. Hedge funds are half its size and contribute more than twice as much, because 80% turnover and a majority-short-term character combine into the worst pairing in the model.
| Horizon | Untaxed benchmark | After tax | Forgone | Tax paid | After-tax CAGR | If fully liquidated |
|---|---|---|---|---|---|---|
| 10 years | $105,963,822 | $89,959,182 | $16,004,639 | $11,722,321 | 6.05% | $86,288,778 |
| 20 years | $224,566,630 | $159,212,570 | $65,354,060 | $34,459,764 | 5.96% | $151,091,699 |
| 30 years | $475,918,766 | $280,847,233 | $195,071,533 | $75,307,470 | 5.92% | $265,584,264 |
Over thirty years the portfolio pays $75.3 million in tax and finishes $195.1 million below the untaxed benchmark. The $119.8 million between those two figures is not tax. It is the return the tax would have earned had it stayed invested, and it is the reason drag is measured as a reduction in compounding rate rather than as an annual bill.
Note also the drift. The first-year after-tax return is 6.33%; the realised compound rate over thirty years is 5.92%. Nothing in the portfolio changed. Unrealised gains accumulated, turnover kept realising a fixed share of a growing stock of them, and the drag widened.
A drag figure is a diagnosis, not a prescription. Three broad readings, and the honest limits of each.
A portfolio of low-turnover equity and municipal-style income can land here. At this level a structure priced in tens of basis points a year would very likely cost more than the drag it removes, and the analysis worth doing is on fees and allocation instead.
Real but moderate. The interventions that cost nothing usually come before the ones that cost something: favouring return of a better character where the investment case is genuinely equal, reducing recognition where reducing it is free, and reconsidering which sleeve sits in which account. Asset location is the discipline that covers this.
Portfolios weighted toward private credit, active hedge fund strategies and taxable fixed income reach this level readily. Here the cost of a structure designed to reduce the drag becomes a genuine question rather than an academic one — but it is a separate calculation, with its own break-even, and it produces a negative answer more often than the market that sells such structures tends to admit.
That second calculation is the subject of PPLI Economics and Break-Even, and the cost side of it is itemised in PPLI costs and economics. A high drag figure is a reason to run that analysis. It is not, by itself, a reason to conclude anything.
The model divides return into ordinary income, short-term gains and long-term gains because tax codes do. The following establish those distinctions in United States federal law, which is the framework most readers of this page work within. Other jurisdictions draw the lines differently, and the instrument is built to accept whatever rates result.
Defines short-term and long-term capital gain by holding period: gain on a capital asset held for not more than one year is short-term; gain on an asset held for more than one year is long-term. This is the distinction the short-term input represents.
uscode.house.gov — 26 U.S.C. § 1222Imposes an additional tax on net investment income, including interest, dividends, annuities, royalties, rents and capital gains, above stated thresholds. It is one of the components a reader should fold into the effective rates entered at the top of the instrument.
uscode.house.gov — 26 U.S.C. § 1411Requires a holder to include original issue discount in income as it accrues, on a constant-yield basis, whether or not cash has been received. It is the reason a private credit sleeve can generate a tax liability ahead of the cash to pay it, and a reason to enter a high income share for such strategies.
uscode.house.gov — 26 U.S.C. § 1272Provides that a capital gain dividend from a regulated investment company is treated by the shareholder as gain from an asset held more than one year, regardless of how long the shareholder held the fund. It is why an investor in a pooled vehicle can be taxed on realisations made inside the fund, and why the turnover input for such sleeves belongs to the fund rather than to the investor.
uscode.house.gov — 26 U.S.C. § 852The administrative treatment of investment income, capital gains and losses, and fund distributions, in the form the Internal Revenue Service publishes for individual filers.
irs.gov — Publication 550Included for readers who follow the structural question from here: it is one of the requirements a compliant insurance-based structure must satisfy before it changes the tax treatment of anything held inside it.
ecfr.gov — 26 C.F.R. § 1.817-5Read personally by a senior specialist. Never routed into a sales funnel. A written reply, usually within one business day.
Confidential PPLI ReviewYour information is submitted over an encrypted connection and handled in accordance with our Privacy Policy. PPLI.com does not sell personal information. Any external introduction is made only with your permission.