A Distribution Waterfall Should Explain Itself Tier by Tier

A waterfall answer without an explanation is another model someone has to distrust.

Distribution logic can combine return of capital, preferred return, catch-up, carried interest, hurdles, investor-specific terms, and historical activity. The LPA and applicable side letters govern. The workbook is an implementation of those terms, not the authority itself.

What an agent can prepare

A useful waterfall agent can:

• retrieve the controlling provisions;
• map each provision to a calculation tier;
• ingest approved ownership and distribution history;
• run scenarios through the existing model logic;
• show the amount entering and leaving each tier;
• attach the relevant term to the calculation;
• flag missing or conflicting inputs.

The tier is the unit of explanation

For every tier, the review record should answer:

1. What capital entered this tier?
2. Which investors or classes participated?
3. Which rate, hurdle, or split applied?
4. What amount was allocated?
5. What balance moved to the next tier?
6. Which governing provision supports the rule?

That structure lets fund accounting and legal reviewers inspect the result without reverse-engineering a single output cell.

Do not let the agent interpret ambiguity alone

Fund terms are not always clean. Defined terms can conflict. A side letter can change one investor's treatment. Historical distributions can reflect a prior interpretation.

Those cases belong in an exception queue with the competing evidence attached. Legal and fund-accounting judgment remains human work.

Prove against known scenarios

Before the agent touches a live distribution, run it against previously approved scenarios. Compare tier entry, tier exit, investor allocation, and rounding. Record every difference and its resolution.

The best result is not merely faster math. It is a waterfall the team can walk tier by tier in front of an investor, auditor, or committee.