Why a waterfall exists
In a typical syndication, limited partners (LPs) supply most of the equity but are passive; the general partner (GP, or sponsor) supplies a sliver of equity, then sources and operates the deal. A flat “split profits by ownership” rule would pay the sponsor almost nothing for the work and reward them the same whether the deal outperforms or underperforms. The waterfall fixes both problems: it pays LPs first to a target return, then lets the sponsor earn a disproportionate share, the promote, for delivering above that bar. It aligns incentives.
The tiers, in order
Cash flows down the tiers like water, each one filling before the next gets a drop:
- 1. Return of capital. LPs (and any preferred equity) get their invested dollars back before anyone splits a profit. Some structures return capital at the very end; others blend it with the preferred return.
- 2. Preferred return (the “pref”). LPs earn a target annual return — commonly 7–9% — on their capital before the sponsor shares in profit. An 8% pref means LPs receive 8% a year (often accruing and compounding if cash flow can’t cover it) off the top.
- 3. The promote / carried interest. Above the pref, the sponsor starts taking a disproportionate cut. A “70/30 over an 8% pref” means once LPs have their 8%, further profits split 70% to LPs and 30% to the GP, even though the GP may have contributed only 10% of the equity.
- 4. Additional promote tiers. Many deals add hurdles: 70/30 over an 8% IRR, then 60/40 over a 14% IRR, then 50/50 over 18%. Each tier hands the sponsor a bigger slice as performance climbs.
Two flavors of hurdle: a preferred return hurdle is met when LPs have received a set percentage on capital; an IRR hurdle is met when their time-weighted return crosses a threshold. IRR hurdles are stricter: they account for when cash came back, so early distributions count more.
A worked distribution
Say a deal returns $14.2M at exit after a five-year hold, against $5.66M of LP equity and $0.63M of GP equity, with an 8% pref and a 70/30 promote over an 8% IRR, then 60/40 over 14%. The hold matters as much as the splits: the pref compounds, so the same proceeds distributed a year later leave less for the promote tiers.
Exit distribution — five-year hold, no catch-up
The sponsor put in ~10% of the equity but, through the promote, earns roughly $1.73M of the $7.91M profit, about 22%. That premium is the reward for sourcing and running the deal, and it only materializes because the LPs cleared their pref first. Every figure above is what our free calculator returns for those inputs; punch them in and it reproduces every one of them, shown as separate LP and GP columns rather than this combined table.
To try your own numbers, use our free equity waterfall calculator: it takes the equity, preferred return, and promote tiers and shows the LP/GP split, IRR, and multiple, with no login. It has a catch-up toggle too, which is the next section.
The GP catch-up
The tiers above assume profit moves straight from the preferred return into the promote split. Many agreements insert one more band first, the catch-up, and it is the single clause most likely to change the answer without changing a single tier percentage.
The logic: once LPs have their pref, the sponsor argues it should hold its full carry on all the profit distributed so far, the preferred return included. So the catch-up band pays the sponsor most or all of each dollar until that is true. With a full (100%) catch-up to a 30% carry, the band closes when the sponsor’s promote equals 30% of everything paid above return of capital.
Take the same deal as above ($5.66M of LP equity and $0.63M of GP equity, a five-year hold, an 8% preferred return, 70/30 then 60/40 over a 14% IRR hurdle, and $14.2M of exit proceeds) and switch the catch-up on. These are the numbers our free calculator returns; you can reproduce every one of them:
Full catch-up to a 30% carry
The formula for a full catch-up is just that ratio: carry ÷ (1 − carry) × the LP pref. At a 30% carry it is 30/70, about 0.43× the preferred return. A partial catch-up splits the band instead of handing it all over (80/20 and 50/50 are common compromises when LPs resist a 100% band), which keeps LP cash flowing but makes the band run longer to reach the same carry, because the LP’s slice of each band dollar also raises the base.
What it costs the LPs is not small. Run that same deal with the catch-up switched on and the sponsor’s promote goes from $1.73M to $2.42M, while the LPs’ total falls from $11.84M to $11.15M, a $0.68M transfer, and the LP’s IRR drops from 15.9% to 14.5%. Not one tier percentage changed. That is why a catch-up is negotiated rather than assumed: it lives in the operating agreement, and it is worth reading out of the document instead of inferring it from the splits.
Preferred equity vs the preferred return
These two get used interchangeably and they are not the same thing. A preferred return is a hurdle inside the common equity: the LPs earn their 8% before the sponsor shares in profit, but they are still common equity and still absorb the loss if the deal disappoints. Preferred equity is a separate layer of the capital stack sitting between the senior loan and the common equity, with its own negotiated rate, paid before any common equity sees a dollar.
In practice a preferred-equity tranche is paid in two pieces: a current-pay rate distributed in cash as the property produces it, and an accruing carry (PIK) that compounds on the balance and is settled when the deal exits. Only what survives that layer flows into return of capital, the LP pref, and the promote tiers, so a preferred-equity piece quietly re-orders everything below it.
What about deals with two LPs?
Most syndications treat every common LP as one class. The tier percentages apply to the class, and each investor takes its pro-rata share of whatever the class receives, so the waterfall arithmetic is identical whether there are two investors behind it or two hundred. Adding investors does not add tiers.
When people describe a “two LP” structure they usually mean something else: one investor negotiated priority over the other. That is almost always a preferred-equity tranche ahead of the common LPs (the layer above) rather than a second common class with its own hurdles. Modeling it as pref equity plus one LP common class is both simpler and closer to how the cash actually moves.
Where this stops working: a genuinely bespoke agreement with several common classes, each with its own pref rate and its own promote schedule, is not a shape Nivora models. It handles one preferred-equity layer plus one LP common class, and a true multi-class agreement still needs a side calculation.
The terms you’ll see
- LP / GP. Limited partner (passive investor) and general partner (sponsor/operator).
- Pref. The preferred return: the LP’s hurdle before the GP shares in profit.
- Promote / carry. The GP’s disproportionate share of profits above a hurdle.
- Catch-up. A clause that lets the GP “catch up” to its target promote split once the LP pref is met, before the normal tier split resumes (worked through above).
- Preferred equity. A layer that sits between debt and common equity, paid before LP common, often with its own fixed return and an accruing (PIK) portion paid at exit. Not the same as a preferred return.
- Clawback. A protection requiring the GP to return promote if early distributions overpaid them relative to the final result.
Why the waterfall is the hardest part to model
A waterfall is recursive: the IRR hurdles depend on the timing of distributions, but the distributions depend on which hurdle has been cleared — so the split in month 50 depends on every dollar paid in months 1 through 49. Build it by hand in a spreadsheet and a single broken reference quietly misallocates thousands of dollars between LP and GP, and the error compounds across tiers. It’s the line item most likely to be wrong in a manually built model.
This is where Nivora earns its keep: the full waterfall (preferred equity with current-pay and accrued carry, LP pref, an optional GP catch-up band, and multi-tier promote with IRR hurdles) runs on the monthly cash flow, with the exit cascade reconciled to that cash flow by construction, so the distributions always sum to what the deal actually produced. The catch-up ships off, because assuming one is how a model quietly overpays the sponsor; you turn it on to match the agreement. The lender export strips sponsor economics automatically. Watch it pay out tier by tier on the live sample deal.
Frequently asked questions
How does a GP catch-up work?
A catch-up is a band of profit that sits after the LP preferred return and before the normal promote tiers, in which the sponsor takes most or all of each dollar until it has earned its target carry on the profit distributed so far. With a full (100%) catch-up to a 30% carry, the sponsor receives every dollar in the band until its promote equals 30% of the LP pref plus the band itself, which works out to 30/70 × the LP pref. On a deal paying a $2.66M preferred return, the band is about $1.14M, and once it closes the sponsor holds exactly 30% of all profit paid above return of capital. The normal tier split then resumes on whatever is left.
What is the formula for a GP catch-up?
For a full catch-up, the band equals the target carry divided by the LP share, times the preferred return already paid: catch-up = (carry ÷ (1 − carry)) × LP pref. At a 30% carry that is 30/70 × the pref, or 0.4286 × the pref. The check is simply that the sponsor’s promote divided by total profit distributed above return of capital equals the target carry exactly once the band closes. For a partial catch-up (where the sponsor takes, say, 80% of the band rather than 100%), the band is larger, because the LP’s share of each band dollar raises the base the carry is measured against.
What is the difference between a full and a partial catch-up?
A full catch-up sends 100% of the band to the sponsor, so the target carry is reached in the fewest dollars. A partial catch-up (commonly 50/50 or 80/20 during the band) splits those dollars with the LPs, so the LPs keep receiving cash through the catch-up and the band has to run longer to reach the same carry. Partial catch-ups are gentler on LP cash flow and are frequently the negotiated compromise when LPs resist a 100% band. If the sponsor’s band share is at or below the target carry, the target can never be reached and the band simply runs at that split.
Does every waterfall have a catch-up?
No. A catch-up is a negotiated clause, and many syndication waterfalls go straight from the preferred return into the promote tiers. Its presence materially changes the split: on the same deal, turning a full 30% catch-up on can move well over a million dollars from the LPs to the sponsor without changing a single tier percentage. Because it is optional, it is worth confirming against the operating agreement rather than assuming either way. That is why Nivora ships the catch-up off by default and requires the underwriter to enable it deliberately.
What is the difference between preferred equity and a preferred return?
A preferred return is a hurdle inside the common equity: the LPs earn a target rate — commonly 7–9% — before the sponsor shares in profit, but they are still equity and still take the loss if the deal underperforms. Preferred equity is a separate layer of the capital stack that sits between the senior debt and the common equity, with its own negotiated rate, and it gets paid before any common equity sees a dollar. The words are close enough to be confusing; the position in the stack is what actually differs.
How does preferred equity get paid in the waterfall?
Preferred equity is typically paid in two pieces: a current-pay rate distributed in cash as the property produces it, and an accruing or PIK portion that compounds on the balance and is settled when the deal exits. It sits ahead of the LP common equity, so the preferred-equity balance and its accrued carry are cleared first, and only what remains flows into return of capital, the LP preferred return, and the promote tiers. Nivora models both pieces: a cash current-pay rate and an accruing carry, with simple or compounded accrual.
How does an equity waterfall work with multiple LPs?
Most syndications treat all common LPs as one class: the tier percentages apply to the class as a whole, and each investor receives its pro-rata share of whatever the class receives, so the waterfall arithmetic is unchanged no matter how many investors sit behind it. Structures described as “two LP” deals usually mean something different: one investor negotiated a higher priority, which in practice is a preferred-equity tranche ahead of the common LPs rather than a second common class. Nivora models that shape directly: one preferred-equity layer plus one LP common class. It does not model arbitrary per-investor classes with separate hurdles, so genuinely bespoke multi-class agreements still need a side calculation.
Do LPs get their capital back before the sponsor gets its promote?
In the standard structure, yes: return of capital and the preferred return come before any promote. What varies is whose capital is returned first: an LP-first waterfall returns the LPs’ invested dollars ahead of the sponsor’s co-investment, while a pari-passu structure returns LP and GP capital pro-rata at the same time. The distinction rarely matters when the deal performs and matters a great deal when it does not, so it is worth reading out of the operating agreement rather than assuming. Nivora supports both orderings as an explicit setting.
What is a clawback in a real estate waterfall?
A clawback requires the sponsor to hand back promote it already received if the final result does not justify it. It exists because promote is often paid during the hold (from operating cash flow or a refinance) against hurdles measured over the whole life of the deal. A strong early year followed by a weak exit can leave the sponsor having collected more than the agreed carry on the deal’s actual performance, and the clawback trues that up at the end.
See it computed live.
Every concept on this page is a number Nivora computes live: the real engine is running on a fictional 104-unit sample deal right now, no login required.