Mastering the Flow: A Deep Dive into the Waterfall Financial Model
Waterfall financial model: who gets paid first on a $100M exit
Two funds can earn the same return and still send investors different checks. The difference sits in the waterfall, the contract schedule that decides who gets paid first, who gets paid most, and what must happen before the manager earns a dollar of profit.
A waterfall financial model turns that schedule into arithmetic. Cash fills one tier completely before any of it spills into the next. The order of the tiers matters more than the headline split, because order decides when an investor sees money and who carries the risk if later deals disappoint.
This guide walks through every tier with real numbers. It also shows where the math usually goes wrong, and it covers both places waterfalls appear: investment funds and project finance.
Two kinds of waterfall model

Finance teams build waterfalls in two settings. The mechanics differ, so it helps to separate them early.
An equity waterfall governs private equity, venture capital, and real estate funds. It splits profit between the general partner (GP), who manages the fund, and the limited partners (LPs), who supply the capital. Hurdles such as the preferred return and the internal rate of return (IRR) drive every tier.
A cash flow waterfall governs project finance and infrastructure deals. It sets the order in which project revenue pays operating costs, taxes, lenders, reserve accounts, and finally equity holders. Debt gets paid before equity sees a cent.
Many people call both a cash cascade. The word fits, since money fills one tier completely before it spills over.
The four tiers of an equity waterfall
A standard fund waterfall has four tiers. The same logic runs through any financial model for a venture capital or private equity fund, and each tier must be coded to match the limited partnership agreement (LPA).
| Tier | What happens | Why it exists |
| 1. Return of capital | LPs receive 100% of distributions until they recover their contributed capital | Principal comes back before anyone discusses profit |
| 2. Preferred return | LPs keep receiving 100% until they earn the agreed annual return, often 8% | Sets a minimum return before the GP shares in profit |
| 3. GP catch-up | The GP receives most or all of the next distributions until its share of total profit equals the carry percentage | Brings the GP up to the agreed split, usually 20% |
| 4. Carried interest split | Everything left splits by the agreed ratio, commonly 80/20 between LPs and the GP | Pays the GP for performance above the hurdle |
These figures are market habits, not rules. In Proskauer’s 2026 buyout fundraising reports, a 20% carry appeared in 77% of surveyed North American funds and an 8% preferred return in 89% of them. Every LPA still sets its own terms, so the model must follow the document.
A worked example: $100M exit
Assume a fund sells a portfolio company and distributes $100 million. LPs contributed $80 million. The preferred return is 8%, simplified to one year. Carry is 20% with a 100% catch-up.
| Step | Amount | What it covers |
| Return of capital to LPs | $80.0M | LPs recover their full contribution |
| Preferred return to LPs | $6.4M | 8% on the $80M |
| GP catch-up | $1.6M | GP receives 100% until it holds 20% of profit so far |
| Remaining 80/20 split | $12.0M | $9.6M to LPs, $2.4M to the GP |
LPs receive $96.0 million in total. The GP receives $4.0 million, which is exactly 20% of the $20 million profit.
The catch-up is where models break. A common shortcut takes 20% of the preferred return, which gives $1.28 million. That leaves the GP short of its agreed share. The correct method grosses the pref up first. Divide $6.4 million by 0.80 to get $8.0 million of total profit through tier 3, then take 20% of it. The answer is $1.6 million, and the profit split lands on 80/20.
How catch-up terms change the result
Not every LPA uses a 100% catch-up. Some set 50%, and some have none. The term matters most when the exit is modest.
Take the same $80 million of LP capital and 8% pref, but sell for $90 million. Profit is $10 million, so a full 20% carry would be $2.0 million.
| Catch-up term | GP carry | LPs in total |
| 100% catch-up | $2.00M | $88.00M |
| 50% catch-up | $1.80M | $88.20M |
| None (hard hurdle) | $0.72M | $89.28M |
With no catch-up, the GP takes 20% only of the profit above the pref. A 50% catch-up shares that tier equally, so it needs more cash to finish. At the $100 million exit above, the 100% and 50% versions both end at $4.0 million, because the profit is large enough to complete the catch-up either way.
Preferred return: simple or compounded
An 8% pref sounds like one number, but it can mean two things. A simple pref accrues 8% of contributed capital each year. A compounded pref accrues 8% on the growing balance.
| Years held | Simple 8% on $80M | Compounded 8% on $80M | Gap |
| 3 | $19.2M | $20.8M | $1.6M |
| 5 | $32.0M | $37.5M | $5.5M |
| 10 | $64.0M | $92.7M | $28.7M |
Compounding raises the bar the GP must clear before carry starts, so it favors LPs. Also check whether the hurdle applies to all contributed capital or only to capital in realized deals. A pitch deck that says “8% pref” rarely says which.
The pref is a priority of payment, not a guaranteed return. If the fund never earns enough to fill that tier, the GP earns no carry. LPs are not owed the shortfall.
Real estate waterfalls: IRR hurdles and the promote
Real estate sponsors call carry the promote. Any real estate financial model for a syndication or fund needs this logic, and many deals replace the single 80/20 split with tiers tied to IRR.
A common illustrative structure pays 100% to LPs until an 8% IRR. It then pays 70/30 up to a 12% IRR, 60/40 up to a 15% IRR, and 50/50 above that.
Say LPs invest $1,000,000 and the deal returns $3,000,000 in one exit after five years.
| Tier | LPs must reach | Split (LP/GP) | Distributions in tier |
| 1 | 8% IRR | 100/0 | $1,469K |
| 2 | 12% IRR | 70/30 | $419K |
| 3 | 15% IRR | 60/40 | $415K |
| 4 | Above 15% | 50/50 | $697K |
LPs end with about $2.36 million and the GP with about $0.64 million. The deal IRR is 24.6%, but the LP net IRR is 18.7%. That gap is the cost of the promote.
An IRR hurdle is an annualized rate, not 8% of capital each year. So timing matters, and early distributions lower the dollars needed in later tiers.
Sponsors usually invest their own money too. The GP’s share of distributions above the hurdle then equals the promote plus its pro rata share of what remains. With a 20% promote and a 10% stake, that is 20% + 10% × 80% = 28%. Fund managers who co-invest alongside LPs get the same adjustment.
American vs European waterfalls
The two structures differ in when the GP collects carry. A European (whole-fund) waterfall pays no carry until LPs have received all contributed capital and the preferred return across the entire fund. An American (deal-by-deal) waterfall lets the GP collect carry on each profitable exit as that deal clears its own tiers.
The labels describe habits more than geography. In Proskauer’s 2026 reports, 85% of surveyed European buyout funds used a whole-fund waterfall and 12% a hybrid. In North America, 64% used deal-by-deal and 36% whole-fund.
| Feature | European (whole-fund) | American (deal-by-deal) |
| Carry measured on | The whole fund, cumulatively | Each exited deal |
| First carry paid | After all contributions and the pref are returned | On the first exit that clears its own tiers |
| Main LP protection | The order of payment itself | Clawback, escrow, guarantees |
| Share of surveyed buyout funds | 85% in Europe, 36% in North America | 64% in North America |
Hybrid waterfalls blend the two. The LPA sets fund-level thresholds that must be met before deal-level carry is released, along with any escrow or later true-up. Emerging managers often push for deal-by-deal terms because carry arrives sooner. The Institutional Limited Partners Association (ILPA) publishes model agreements in both forms.
Clawbacks: a paragraph or a protection
A clawback forces the GP to return carry it was paid but did not earn once final results are known. It matters most under deal-by-deal waterfalls.
Here is a simplified case. A $100 million fund makes two $50 million investments with 20% carry. Deal A exits for $90 million, a $40 million profit, so the GP collects $8 million of carry. Deal B then goes to zero.
Across the whole fund, LPs put in $100 million and got back $90 million. The GP earned nothing, so the full $8 million is due back. If 30% of carry sat in escrow, $2.4 million comes from escrow and the GP or its principals must fund the other $5.6 million. This example ignores the pref, fees, and taxes.
Terms to check in the LPA:
- Testing: at fund end only, or also at interim dates
- Cap: repayment limited to the lesser of excess carry or total carry received
- Tax: whether the GP repays on an after-tax basis
- Security: escrow, guarantees, or letters of credit
- Certification: whether an auditor signs off on the calculation
Practice varies widely. In Proskauer’s 2026 European sample, 60% of funds used no escrow, and those that did most often escrowed 50% of carry. ILPA’s model agreements bracket 30%. So “the fund has a clawback” says little on its own. The useful questions are when it is tested, how it is capped, and who stands behind it.
Cash flow waterfalls in project finance
Project finance runs on debt seniority, not profit sharing. The exact order depends on the facility agreement, but a typical sequence looks like this.
| Order | Payment | Purpose |
| 1 | Operating costs | Keeps the asset running |
| 2 | Taxes | Pays amounts due before lenders |
| 3 | Senior debt service | Interest and principal to the most senior lenders |
| 4 | Reserve accounts | Funds the debt service reserve account (DSRA) and major maintenance reserve |
| 5 | Subordinated debt | Repays junior lenders |
| 6 | Equity distributions | Remaining cash, released only if coverage tests pass |
Lenders track the debt service coverage ratio (DSCR), which equals cash flow available for debt service (CFADS) divided by debt service. Covenant minimums typically run from 1.2x to 1.5x, depending on project risk. A lock-up trigger sits above the default level. When DSCR falls below it, distributions to equity stop and cash stays in the project.
Take a project with $9.0 million of annual debt service. If CFADS is $12.0 million, DSCR is 1.33x and $3.0 million flows down toward equity. If CFADS drops to $10.0 million, DSCR falls to 1.11x. Under an illustrative 1.15x lock-up, the $1.0 million left after debt service stays trapped, and equity gets nothing that period.
A DSRA protects lenders, but it delays equity distributions while the reserve builds. Some deals add a cash sweep, which sends surplus cash to prepay debt. Oak’s CNG and LNG energy case study shows how a project finance model ties debt service and capital structure to investor returns.
Building the waterfall in Excel
A waterfall should be easy to audit. Keep inputs on one sheet, build one row per tier, and put checks at the bottom. For a single exit with a 100% catch-up, the logic is short.
| Tier | Excel logic |
| 1. Return of capital | =MIN(Proceeds, LP_Capital) |
| 2. Preferred return | =MIN(Proceeds – T1, Pref_Due) |
| 3. GP catch-up | =MIN(Proceeds – T1 – T2, T2 * Carry / (1 – Carry)) |
| 4. Residual | =Proceeds – T1 – T2 – T3, then split by Carry |
For a partial catch-up rate, replace the tier 3 cap with Pref_Due * Carry / (CatchUp – Carry). Give the GP the catch-up rate of that tier and the LPs the rest.
Multi-period models compound the pref balance each period and test IRR hurdles with XIRR on dated cash flows. Oak’s guide to Excel formulas for financial models covers XIRR and the error checks worth adding.
Build these checks into every waterfall:
- LP plus GP distributions equal total proceeds
- The GP’s share of profit never exceeds the carry percentage
- No tier returns a negative amount
- The model reproduces a hand-calculated case, such as the $100 million example above (LPs $96.0M, GP $4.0M)
Metrics a waterfall model should report
| Metric | What it shows | Watch out for |
| IRR | Annualized return on invested capital | Sensitive to timing, and most hurdles are set on it |
| Equity multiple | Total distributions divided by capital invested | A 2.0x multiple says nothing about how long it took |
| DPI | Cash distributed divided by paid-in capital | Counts only realized cash, so it lags early on |
| TVPI | Distributions plus remaining value, divided by paid-in capital | Depends on how the fund marks unsold assets |
Report returns net of fees and carry, because LPs live on net numbers. Management fees come off the top for years before any tier fills, so a model that ignores them overstates what investors keep.
Questions to ask before signing an LPA
A waterfall’s fairness lives in details that are easy to skim in a long agreement. Ask these first:
- Is the waterfall European, American, or hybrid?
- What is the pref rate, and does it compound?
- Does the hurdle apply to all contributed capital or only realized deals?
- Is the catch-up 100%, partial, or absent?
- How is the clawback tested, capped, and secured?
- Do fees, expenses, and recycled capital count in the calculations?
Vague answers to any of these justify pushing back before capital moves.
Frequently Asked Questions
How is the GP catch-up calculated?
Divide the preferred return paid to LPs by one minus the carry percentage, then multiply by the carry percentage. With a $6.4 million pref and 20% carry, that is $6.4M ÷ 0.80 × 0.20 = $1.6M. This assumes a 100% catch-up, and a lower rate stretches the tier.
Is a European or American waterfall better for investors?
European waterfalls generally favor LPs, since no carry is paid until the whole fund clears its hurdle. American waterfalls pay the GP earlier and rely on clawbacks if later deals lose money. The strength of the clawback security often matters more than the label.
Does a clawback guarantee LPs get overpaid carry back?
No. Repayment depends on how often the clawback is tested, whether it is capped after tax, and whether escrow, guarantees, or letters of credit back it. In Proskauer’s 2026 European sample, 60% of funds used no escrow.
What is a cash flow waterfall in project finance?
It is the order in which project revenue pays operating costs, taxes, senior debt, reserve accounts, junior debt, and equity. Lenders test coverage with DSCR, and a lock-up blocks distributions to equity when coverage falls below the trigger.
Is the preferred return guaranteed?
No. It is a priority of payment, not a promise. If the fund never earns enough to fill that tier, the GP earns no carry and LPs absorb the shortfall.
How do you build a waterfall model in Excel?
Put inputs on one sheet and build one row per tier with MIN functions, so each tier takes only what remains. Add checks that LP plus GP equals total proceeds. Test the model against a hand-calculated case before loading real figures.
Conclusion
A waterfall is a contract translated into arithmetic. The risk rarely sits in the arithmetic itself. It sits in the translation: a catch-up rate misread, a pref that compounds in the LPA but not in the spreadsheet, a clawback test that exists only on paper. Teams that tie every tier to a clause and a check cell catch these errors before an investor does.
Oak Business Consultant builds investor-ready financial models with tier-by-tier checks for fund, real estate, and project finance waterfalls. Contact us to build a waterfall model your investors can trust.
