Over the past year we have read, extracted and modelled the NAV provisions in roughly a hundred fund-facing agreements: ISDA Schedules and their credit support annexes, prime brokerage and GMRA terms, fund finance facilities, and the prospectuses and offering memoranda behind them. A good share are public, filed with a regulator or an exchange. The rest came to us from monitoring teams who wanted to know whether their templates matched their documents.
This post is about the shape of what we found. It is not a legal survey. It is what a monitoring engine has to be able to represent if it is going to run the tests the documents actually describe.
The backbone is vanilla
The large majority of agreements rest on the same structure: a decline in NAV over one, three and twelve months, with escalating thresholds, tested on month-end figures. The numbers vary by fund type, but the families are recognisable. Managed futures and multi-strategy funds tend to sit around 20%, 30% and 50% over a month, a quarter and a year. Long-only and lower-volatility strategies sit closer to 10%, 15% and 25% or 35%. Most add a floor, usually a fixed currency amount or a percentage of the audited year-end NAV.
We call this the vanilla case, and it is worth saying clearly that a template handles it. Three windows, three thresholds, one basis, a floor. If your book were all vanilla, a well-kept spreadsheet with a strict owner would get you most of the way.
The variance is where templates fail
The book is not all vanilla, and the non-vanilla provisions are not exotic. They are ordinary drafting choices that a template cannot see.
- Flow adjustment. Most schedules exclude subscriptions, redemptions and distributions from the decline. Some do not, and test raw reported NAV. The same 12% move fires one and not the other.
- Measurement convention. A minority test "any reported NAV" in a rolling 30, 90 or 365-day window rather than month-end to month-end. Those fire on intra-month estimates and need every reported figure stored, not just the official one.
- Cure periods. Some triggers are events on the day; others give the fund a number of business days to restore NAV or provide an explanation before the event crystallises. The register has to hold the cure date and hold the breach open until it passes.
- Floors tied to audited year-end. A floor of 50% of the NAV in the last audited accounts moves once a year, when the audit lands, and the register has to re-base it then and only then.
- Adviser and key-person changes. A change of investment manager or the departure of a named individual is an event that arrives by notice, not by number. It has to be logged against the fund and routed like a breach.
- Cross-default and sibling clauses. A trigger in one fund's agreement that fires on an event in another fund under the same manager, or under the same umbrella. The link between funds has to exist in the register for the test to run.
- Umbrella inheritance. A new sub-fund added to a master agreement inherits the terms unless a side letter overrides them. That is two facts, and both have to be recorded.
What a register has to hold
Each item above is a field or a relationship, not a paragraph. Basis, measurement convention, window, threshold, cure period, floor and floor re-basing rule, event triggers, cross-fund links and inheritance from the umbrella with per-fund overrides. Every one is extracted with a citation to the page that set it.
A template monitors the agreement you expected to sign. A register monitors the one you signed.
The practical consequence for a monitoring team is a two-part vocabulary we now use in every conversation. Vanilla funds can be fast-tracked: extract, confirm, run. Non-vanilla funds need a second look at onboarding, because that is the cheapest moment to catch the clause that the template would have silently ignored for the next ten years.