An AMC wants an ESG equity fund. Its exclusion list removes 18% of benchmark weight and raises expected tracking error to 3.1%. Design the investment thesis for a long-only manager, and show how you would stop the fund becoming a label on an ordinary portfolio.

The thesis is that companies managing the issues material to their industry carry less tail risk, and the fund expresses it by exclusion, then scoring and tilting, then engagement, with the holdings different enough to prove it. Exclusions alone give an active share of about 18%, an index fund with holes. The 3.1% tracking error is the cost of a real tilt and buys, at an assumed information ratio of 0.3, about 0.9% a year of expected difference, with a bad year near minus 6%. Honesty comes from measurable targets and a published record.

A restaurant that calls itself organic because it dropped two dishes from the menu has not changed what it cooks. An ESG thesis for a long-only manager has to say what the fund believes about returns and risk, not about virtue, and the honest version is that companies which handle the environmental, social and governance issues material to their own industry carry less tail risk: fewer fines, strikes, stranded plants and accounting surprises. That is a risk thesis, not a return promise, and the manager should say so. It then has three consequences for the portfolio: avoid what the mandate forbids, tilt towards companies that manage their material issues well, and use the fund's votes and meetings to push the ones it still owns. Each stage has to show up in the holdings, or it is a brochure.

Measure the difference. Active shareHalf the sum of the absolute differences between a fund's weights and its benchmark's. An index fund scores zero; a fund with no overlap scores 100 per cent. after removing 18% of the index and scaling up the rest is only about 18%, with a tracking error near 1.2%. At that distance from the index the fund cannot justify a 1.0% fee against 0.2% for the index fund: at an assumed information ratio of 0.3 its expected excess return is about 0.4% a year, less than the 0.8% extra it charges. This is the arithmetic behind the head of products' complaint about rivals: an exclusion-only fund is an index fund that costs more and owns less.

Tracking error is the fund's budget for being different. Spending it on a scored tilt, overweighting the companies that manage their material issues well and underweighting the laggards, takes active share to around 55%, a level at which the holdings would tell a reader what the fund is without the name. At an assumed information ratio of 0.3, 3.1% of tracking error buys an expected excess return of about 0.9% a year, which covers the fee gap with an information ratio of only 0.26; the cost is that a two-standard-deviation bad year trails the index by about 6 points. The 0.3 is an assumption for the illustration, not a forecast, and ESG tilts have had multi-year stretches on both sides of the index. The thesis document should carry that sentence, because the first bad year is when the fund gets tested.

With numbers the trustees can check every quarter. Active share above a floor, say 50%, so that the fund cannot drift back to the index. A portfolio carbon intensity at least 40% below the benchmark's, or whichever material metric the thesis names, measured the same way every time. A published voting record and an engagement log that names the companies, the issue and what changed, because engagement without a record is a claim. Scores built from company disclosures the auditor has checked, not from a vendor's black box, and a stated rule for what happens when a holding's score falls: a deadline to engage, then sell. The regulator has set categories for ESG schemes, minimum shares of assets that must follow the strategy, and disclosure and assurance requirements; the design should name its category and confirm the current rules before launch.

It will not promise a higher return than the index; the thesis is about risk and about owning what the mandate says it owns. It will not match the index in a year when the excluded sectors rally, and the board should expect to field that question. And it will not call itself ESG on exclusions alone, because the arithmetic shows that an exclusion-only fund is the index with holes, and investors paying an active fee for it would be right to complain. Said plainly at launch, those three limits are what make the fund credible when the first hard year arrives.