Promote crystallisation is a rule in a fund’s LPA that turns some portion of the GP’s carried interest from “floating with future outcomes” into a fixed allocation as of a stated date, often using a valuation or a transaction. A promote reset is a rule that changes the hurdle, catch-up, or promote split after a stated event, so the next dollars follow a new waterfall. Both sit inside the partnership profit allocation and distribution waterfall; they are not fees, and small drafting choices move cash timing, clawback exposure, and reported net IRR.
Most modelling mistakes come from treating crystallisation as a simple carry payment date and treating resets as a simple new hurdle. In practice, each one changes the carry base, the preferred return math, and the capital account ledger. If you miss that, you can get the “right” answer for the wrong contract-and the wrong cash.
Why crystallization and resets matter to returns and governance
Crystallization and reset clauses matter because they change how, when, and on what base carried interest is calculated. That shows up immediately in GP liquidity, LP distributions, and the fund’s reported performance metrics. In addition, these clauses often appear in moments of conflict risk (continuation funds, extensions, recaps), so precision in modelling is not just a finance exercise-it is a governance requirement.
Definitions and boundaries you must pin down early
Start with what “carried interest” really means in the LPA
Carried interest is the GP’s share of partnership profits after LPs receive return of capital and the preferred return, under the waterfall the LPA actually uses. European (whole-fund) waterfalls concentrate clawback risk at the end; American (deal-by-deal) waterfalls can pay carry early, so every early allocation deserves extra skepticism. If you need a refresher on mechanics, start with the promote and waterfall mechanics overview and map it back to your document.
Crystallization can mean three different things
Crystallisation can mean three different things, and the LPA rarely uses those labels. As a result, the first step in any model is to identify which version the agreement actually implements and how it interacts with the capital account ledger.
- Economic crystallisation: The promote percentage on a defined profit base becomes fixed as of a date. The key question is whether future losses can unwind it, or only trigger clawback.
- Accounting crystallisation: The fund books allocations to GP and LP capital accounts based on fair value marks, even if no cash changes hands. That affects reporting, audit questions, and partner statements.
- Cash crystallisation: The GP receives cash distributions tied to the crystallised amount, often with escrow or holdbacks.
Many funds use a hybrid: the allocation hits capital accounts now, while cash dribbles out later after reserves, lender constraints, or escrow. That hybrid approach is where models most often drift away from what the contract actually does.
A promote reset is a switch in the incentive engine
A promote reset is a contractual switch in the incentive math after an event-refinancing, recap, continuation sale, or amendment. It can change the pref rate, the catch-up, the promote split, and the measurement period start date. Managers sometimes use “reset” to describe a new asset-level incentive plan, but the modelling logic here is fund-level: what happens to the fund waterfall, and what happens to the balances that feed it.
What these mechanisms are not: they do not replace a clawback; they do not equal recycling; and they do not equal a NAV loan. Those tools affect cash timing and accounting, but they only change promote economics if the LPA says they do.
Where you see crystallization and resets in real life
Crystallisation shows up most often when someone wants to close a chapter without selling everything for cash. In those moments, the clause is usually doing two jobs at once: (1) creating certainty for the GP on value that has “effectively” been realized and (2) creating a governance framework for LPs around valuation and conflicts.
Continuation funds and GP-led secondaries are the obvious cases. The old vehicle transfers assets at an agreed price, and the GP wants carry certainty on what has been “sold,” while LPs want the valuation and conflicts handled cleanly.
Long-hold assets with partial liquidity come next. A recap or dividend may create a large distribution while the main value remains tied up. Some agreements crystallise promote on part of the value so the GP doesn’t wait a decade for the final sale.
Fund restructurings and amendments also invite crystallisation. A GP may offer fee reductions or a term extension and ask for a promote adjustment in exchange. The bargain lives or dies on the definition of the crystallised base and what happens if marks reverse.
Resets often travel with step-down or step-up promote schedules, new hurdles after partial liquidity, or a two-pocket structure that isolates “before” and “after” economics. If you hear “reset,” assume administrative complexity is coming, because someone has to keep score.
The modeling mindset: state variables, not one-off formulas
A reliable model treats crystallisation and resets as state changes. You build the waterfall once, then you show how balances evolve over time. Anything else is a spreadsheet that happens to print numbers.
At minimum, track these balances explicitly and carry them period to period:
- Contributed capital (CC): Total capital called, adjusted only as the LPA permits.
- Unreturned capital (URC): CC less capital treated as returned under the LPA. Recycling can make this tricky; define it, don’t improvise it.
- Preferred return accrual (PRA): Pref owed to LPs net of pref already paid, using the correct rate, compounding (or not), and day count.
- GP carry paid (GPCash): Cash the GP actually received as carry, net of givebacks.
- GP carry allocated (GPAlloc): Allocations to the GP capital account when the LPA books carry without cash.
- Escrow/reserve (Escrow): Cash withheld from the GP to secure future true-ups.
- NAV/unrealised value (NAV): Needed when crystallisation is tied to fair value.
Impact tag: if these balances are wrong, every output is wrong-IRR optics, clawback sizing, and close certainty. For practical context on how waterfall tiers work in funds, see the American vs European waterfall breakdown.
Start from a baseline waterfall, then show what changes
A baseline European waterfall often runs like this. First, pay LPs until URC is zero. Second, pay LPs until PRA is zero. Third, pay the GP catch-up until the GP’s share of profits reaches the target promote. Finally, split remaining proceeds at the promote rate.
Crystallisation and resets modify the definition of “proceeds,” the timing of allocations, or the starting balances for the next period. You cannot shortcut by taking a promote percentage times a mark and calling it carry. The catch-up tier and any caps matter, because the LPA uses them to decide who gets the next dollar.
Crystallization often uses a “deemed distribution and recontribution”
Many LPAs implement crystallisation by pretending the fund made a distribution and then immediately took some of it back as a recontribution. That structure is helpful because it forces the crystallisation through the same waterfall you already model.
If crystallisation occurs on a valuation date using NAV, do it in four clean steps:
- Compute deemed value (DV): This is usually fair value of a defined pool, net of liabilities and reserves that the LPA treats as reducing distributable proceeds.
- Run the waterfall on DV: Treat DV as if it were distributed proceeds and calculate tiers, including catch-up, exactly as drafted.
- Separate cash from allocation: The GP may get cash, a book allocation, or both. Escrow and holdbacks move cash timing and GP liquidity, even when economics are unchanged.
- Reset the future base (only if stated): Some agreements reset URC and PRA as if DV had been distributed and recontributed. Others lock carry entitlement but keep original balances running.
If you cannot point to the clause that resets URC or PRA, assume they do not reset. People love to “assume” a reset. Auditors do not.
Crystallization tied to a real transaction: continuation funds and partial sales
When crystallisation is tied to a continuation transaction, the anchor is the transfer price and what counts as proceeds. The modelling job is to translate the legal form into the economic base used by the waterfall and then reconcile to actual cash movement.
- Gross vs net proceeds: Does the carry base come after transaction costs, debt repayment, and reserves, or before? One word in “Distributable Proceeds” definitions can move the GP’s check and the LP’s multiple.
- Sale vs in-kind distribution: Some structures treat the old fund as selling; others treat it as distributing assets (or interests) in kind, with LPs electing to roll. The economics can be similar, but capital accounts, tax reporting, and timing of allocations can differ. For a checklist mindset, compare with an in-kind distributions checklist.
- Cash vs non-cash consideration: Rolled value can satisfy waterfall tiers if the LPA treats it as distributed then reinvested. That can trigger carry even when cash does not leave the ecosystem.
Impact tag: if you mix cash and non-cash, you will misstate GP liquidity and LP distribution profiles-two things investment committees watch closely.
Promote resets: switch regimes, don’t hack the spreadsheet
A reset clause changes the incentive engine after an event date. The clean approach is to model two regimes and switch on the reset date. On that date, freeze the pre-reset state variables, then define the post-reset rules for pref, catch-up, split, and measurement period.
On the reset date, lock the items that should never change silently and document them:
- Freeze balances: Print URC, PRA, GPAlloc, GPCash, and Escrow immediately before the reset.
- Apply new rules: Input the new pref rate, compounding method, accrual base, catch-up, promote split, and start date.
- Maintain audit trail: Tie each switch to a clause reference so reviewers can trace numbers back to the agreement.
Common reset designs (and what they imply)
- Hard reset: Treat everything after the reset as a fresh start with a new URC and new pref base. Prior carry may be settled, locked, or left subject to clawback.
- Soft reset: Change the promote split on future profits while leaving URC and PRA running under the old terms. It sounds fair until you try to administer it.
- Two-pocket approach: Keep separate pockets with separate waterfalls-pre-reset and post-reset-and allocate assets and proceeds by agreed attribution.
A small numerical anchor (because words can drift)
Assume a European waterfall: 8% annual simple pref; 100% catch-up to the GP until the GP reaches 20% of profits above capital and pref; then an 80/20 split. At the crystallisation date: URC is 100 and PRA is 10. DV (NAV net of liabilities) is 150.
Run DV through tiers. First, return capital: 100 to LP, leaving DV of 50. Next, pay pref: 10 to LP, leaving DV of 40. Then split profit: total profit above capital and pref is 40, so the GP’s 20% target is 8 and the LP’s target is 32. Under a standard catch-up design, the GP receives 8 out of the remaining 40 and the LP receives 32, though the path depends on the exact catch-up clause.
So the implied carry at crystallisation is 8. If the clause is allocation-only, GPAlloc increases by 8 and GPCash stays at zero. If the GP is paid in cash with a 30% escrow, GPCash increases by 5.6 and Escrow increases by 2.4.
Now the question that matters is what happens next. If the LPA treats DV as deemed distributed and recontributed, URC may go to zero and then restart, which restarts pref accrual on the post-event base. If the LPA only locks the carry amount, PRA may keep accruing on remaining URC, and the GP’s 8 stays exposed to clawback if later outcomes reduce total profits. Same day-one carry number. Different long-term risk. As a rule of thumb, if the clause changes your “base,” it changes your downside.
Where the truth is written: your document map
These mechanics rarely live in one paragraph. Read the whole stack, then model.
Start with the LPA: definitions of Distributable Proceeds, Capital Contributions, Returned Capital, Preferred Return, Catch-Up, Carried Interest, and Clawback. The crystallisation trigger, deemed distribution language, valuation references, and true-up mechanics often sit in a schedule or amendment. If your team needs consistent term naming, keep a shared glossary (for example, a fund terms glossary).
Then read side letters. They can add governance rights, valuation constraints, MFN elections, and reporting requirements. Some LPs prohibit NAV-based crystallisation without independent valuation.
If there is a continuation transaction, read the sale/contribution agreement and rollover election materials. Confirm whether rollovers are treated as deemed distributions and recontributions. That single choice drives whether carry triggers without cash.
Finally, read the valuation policy and any escrow or guarantee documents. If a subscription line or NAV facility exists, check lender covenants; restricted payment tests and collateral definitions can force timing changes.
Impact tag: document conflicts can delay consent, and a delayed crystallisation date can move pref accrual and marks-real dollars, not theory.
Costs, accounting, tax, and compliance can quietly move economics
Crystallisation and resets bring extra costs: third-party valuation work, conflicts counsel, fairness processes, administrator effort, audit procedures, and transaction costs. Unless the LPA clearly excludes an expense from the carry base, treat it as a reduction to the value that flows through the waterfall. That conservative assumption avoids over-crediting carry and later explaining why.
Accounting matters because allocation is not cash. A NAV-based allocation can shift capital accounts today. If marks reverse later, the LPA may require reversing allocations or may rely on clawback. Your model must specify which rule applies, because it changes who bears mark risk and when. For a related modelling mindset, the same discipline applies in NAV financing, where covenant timing can change cash reality without changing headline valuation.
Tax is jurisdiction-specific, but your model should flag timing risk. A booked allocation without cash can create taxable income in some settings, or at least create reporting complexity. Also watch any structure that makes carry look like a guaranteed payment. The model shouldn’t give tax advice, but it should force tax counsel to sign off on the assumptions.
Compliance adds timing risk. Continuation transactions bring conflicts disclosures, LPAC processes, and eligibility checks. A quarter’s delay can change pref accrual and valuation inputs. Build a timing sensitivity. You do not need ten scenarios; you need one honest one that shows the effect of time.
Diligence and “kill tests” that catch most model failures
Before trusting outputs, run a few tests that catch most failures. These checks work because they attack the areas where crystallisation and resets usually break spreadsheet logic: tier ordering, state resets, and cash versus allocation reconciliation.
- Waterfall invariance: With no crystallisation and no reset, your model should replicate historical distributions and carry to the cent.
- Deemed distribution consistency: Setting DV to zero should change nothing. Setting DV to a small amount should produce marginal allocations consistent with the next tier.
- Clawback stress: Assume a material post-crystallisation write-down and compute final entitled carry versus carry paid, net of escrow.
- Regime switch audit trail: On the reset date, print state variables before and after and show exactly what changed, with the clause reference.
- Cash vs allocation reconciliation: GPAlloc, GPCash, and Escrow must reconcile over time.
Implementation notes and closeout discipline
Good execution follows the contract, not the other way around. Lock a term summary that mirrors LPA definitions. Draft amendments and escrow terms. Pre-clear valuation approach with auditors when possible. Build the model with the administrator’s capital account logic, then reconcile to pro forma partner statements. Obtain LPAC consent or broader LP approvals as required, and assume timing can slip.
At the end, treat records like you expect a dispute. Archive the index, versions, Q&A, user access records, and full audit logs. Hash the final package so you can prove what existed when. Apply retention schedules, then obtain vendor deletion and a destruction certificate when permitted. If there is a legal hold, it overrides deletion-every time.
Closing Thoughts
Promote crystallization and promote resets are not “special calculations” you tack onto a waterfall at the end. They are contract-driven state changes that can reshape the carry base, the preferred return ledger, and clawback risk, especially in continuation deals and fund restructurings. If you model the balances, document the switches, and reconcile cash versus allocation, you will get outputs that match the LPA-and avoid unpleasant surprises at close.