When PS is appropriate

Example 6 — Asset management — highly integrated portfolio teams

Annex II Ch.II · Example 6 · ¶26–33 · ASSET Co

OECD Guidelines 2022
At a glance
FUND Co (independent) hires ASSET Co to manage two-country mirror funds. ASSET Co is a nominee — actual portfolio management is done jointly by Company A (Country A equity) and Company B (Country B equity) through a shared investment committee. Comparables exist for the bundled service but cannot split the fee between A and B. Method?
FUND Co
Independent · pays AUM fee
AUM fee
ASSET Co (parent)
Nominee · no functions, assets, risks
split
Companies A & B
Joint portfolio management · share fund-performance risk
  1. Independent FUND Co hires ASSET Co to manage mirror funds and pays a fee based on combined AUM.
  2. ASSET Co contracts with Companies A and B to actually provide the portfolio management; ASSET Co performs no functions, contributes no assets, assumes no risks.
  3. Company A's specialists handle Country A equity; Company B's specialists handle Country B equity.
  4. A joint investment management committee (equal numbers from A and B) decides composition of the funds.
  5. Risk: retail investors withdrawing from poor-performing funds — Companies A and B share assumption of that risk.

Portfolio management itself isn't unique — there are bundled-fee comparables. But those comparables don't tell you how to split the fee between A and B. A and B's work is tightly integrated, and they share the risk of fund underperformance.

What's the right method for paying A and B?

OECD analysis · Profit split (A and B); ASSET Co = 0
Per Annex II Ch.II · Example 6 (¶33), operations of A and B are so highly integrated and interdependent that a one-sided method cannot determine an arm's length outcome for either. Profit split is the most appropriate method; the FUND Co fee forms the revenue pool to be split between A and B. ASSET Co — performing no functions, holding no assets, assuming no risks — earns zero. A mechanical 50/50 ignores relative value of A's and B's contributions; separate TNMMs lack data on how to allocate between them.
Key takeaway
High integration plus shared risk can make profit split appropriate even when the underlying service itself is not unique. The trigger is whether you can reliably price either party in isolation — not whether comparables exist for the bundle.

Related: ¶2.137–2.143 (high integration · shared risk) · ¶2.135 (when one-sided methods cannot reliably reward each contribution)

Test yourself on this example.Open it in the interactive walkthrough →