Liquidation preferences

Conversion indifference point

The exit value at which a non-participating preferred holder is indifferent between taking the preference and converting to common. Below it they take the preference; above it they convert.

Field	Value
Formula	E* = x*I / p
Worked	1x on 10,000,000 at 20 percent: E* = 10,000,000 / 0.20 = 50,000,000
Interpretation	Between I and E* the preferred is protected and the common is impaired

• The dead zone between the invested amount and the indifference point is where founder and employee equity is worth materially less than the headline ownership percentage implies.
• Raising at a high valuation with a high multiple widens that dead zone. A 2x preference doubles E* and doubles the range of outcomes in which common receives little.

Where a participation cap stops binding

A capped participating holder reverts to converting when straight conversion pays more than the cap. The cap therefore only binds over a finite band of exit values.

Field	Value
Formula	Cap binds while cap*I > p*E, i.e. for E < cap*I/p

• With a 3x cap on 10,000,000 at 20 percent, the cap ceases to bind above E = 150,000,000, where conversion pays 30,000,000.
• In the worked table above, the capped column tracks full participation until 200,000,000 and then reverts to the conversion value.

Stacked versus pari passu preferences

When multiple preferred series exist, the seniority rule determines the order of payment in a proceeds shortfall.

Field	Value
Pari passu	All series rank equally and share any shortfall pro rata to their preference amounts
Stacked (senior)	Later series are paid in full before earlier series receive anything
Tiered	Groups of series rank together, with seniority between groups

• Stacking matters only when proceeds are insufficient to cover all preferences. Above that level the distinction is irrelevant.
• Because it only bites in bad outcomes, stacking is frequently conceded in negotiation and then determines the entire result in the outcome that actually occurs.

Also described at: Wikipedia · Wikidata · NVCA model legal documents

The algebra of a multi-series stack

With more than one preferred series the payout is no longer a single expression. Each series has its own preference amount, and the seniority rule decides whether a shortfall is shared or absorbed in order. Both rules reduce to one line of arithmetic.

Field	Value
Formula	Pari passu: R_j = x_j*I_j * min(1, E/T) where T = sum of x_j*I_j. Seniority-ordered: R_j = min(x_j*I_j, max(0, E - S_j)) where S_j is the sum of preference amounts senior to j
Worked	Stack 2,000,000 / 8,000,000 / 20,000,000 all at 1x, so T = 30,000,000. At E = 18,000,000 the pari passu ratio is 0.60: Seed 1,200,000, Series A 4,800,000, Series B 12,000,000
Worked, seniority	Same stack at E = 24,000,000: B takes min(20,000,000, 24,000,000) = 20,000,000; A takes min(8,000,000, 4,000,000) = 4,000,000; Seed takes min(2,000,000, max(0, -4,000,000)) = 0
Check	Both rules distribute exactly min(E, T); above T the two are identical

• The seniority rule is a single word in the charter and it decides the entire result in a shortfall. Model both before agreeing to either.
• Pari passu shares pro rata to preference amounts, not to invested capital. A series carrying a multiple therefore recovers a larger fraction of its money than a 1x series ranking alongside it, without any seniority.
• A stack built round by round accumulates preference faster than it accumulates valuation. Track T against the exit values actually being discussed, not against the last post-money.

Also described at: Wikipedia · Wikidata

When seniority wipes a series out completely

Under seniority-ordered payment a series receives nothing at all once proceeds fail to cover the preferences ranking ahead of it. Under pari passu the same series always receives something as long as proceeds are positive. The threshold is exact and easy to compute.

Field	Value
Formula	Series j receives zero under seniority whenever E <= S_j, the aggregate preference senior to it. Under pari passu it receives x_j*I_j * E/T for any E > 0
Worked	Seed is junior to 8,000,000 + 20,000,000 = 28,000,000, so Seed receives nothing under seniority for any E <= 28,000,000
Same point, pari passu	At E = 28,000,000 Seed receives 2,000,000 * 28,000,000/30,000,000 = 1,866,667

• The seed investor who agreed to junior ranking has an all-or-nothing claim: full recovery above 30,000,000 and nothing below 28,000,000, with a 2,000,000 band in between.
• This is why later investors ask for seniority and earlier investors often grant it: the earlier investor is comparing outcomes that already look poor, while the later investor is buying protection in the region where its money is actually at risk.
• A pay-to-play or recapitalisation that converts the junior series to common produces the same result as seniority, by a different route. Check whether the charter already achieves what the new term asks for.

The three regimes of a capped participating series

A capped participating preference behaves as three different instruments above the point at which it has recovered its money: it participates in the residual, then plateaus at the cap with proceeds that do not move at all, then converts to common. Both boundaries are closed-form.

Field	Value
Formula	Participation reaches the cap at E_cap = I*(1 + (cap - 1)/p). Conversion overtakes the cap at E_conv = cap*I/p. Between them proceeds are constant at cap*I
Worked	I = 10,000,000, cap = 3, p = 0.20: E_cap = 10,000,000*(1 + 2/0.20) = 110,000,000; E_conv = 3*10,000,000/0.20 = 150,000,000
Check at E_cap	I + p*(E - I) = 10,000,000 + 0.20*100,000,000 = 30,000,000 = cap*I
Check at E_conv	p*E = 0.20*150,000,000 = 30,000,000 = cap*I
Plateau width	E_conv - E_cap = I*(1 - p)/p = 10,000,000*0.80/0.20 = 40,000,000 of exit value over which the holder's proceeds do not change

• Inside the plateau the holder is economically indifferent to price. A 40,000,000 improvement in the sale price is worth nothing to it, which removes its incentive to push the buyer and can quietly misalign it from common in a live negotiation.
• The cap is usually presented as a founder-friendly limit on participation. It is, but it also creates a range where the holder stops caring about the outcome. Check where that range sits relative to the exits being modelled.
• The plateau width simplifies to I*(1 - p)/p and does not depend on the cap at all. At I = 10,000,000 and p = 0.20 a 2x cap, a 3x cap and a 4x cap all produce a 40,000,000 flat band; the cap only decides where the band starts. Ownership is what narrows it - at p = 0.50 the same band is 10,000,000 wide.

Full participation at 1x versus a higher non-participating multiple

These are the two standard ways to give a preferred holder more than its money back, and they cross exactly once. Below the crossover the multiple pays more; above it participation pays more.

Field	Value
Formula	1x full participation equals an N x non-participating preference at E_eq = I*(1 + (N - 1)/p)
Worked	I = 10,000,000, p = 0.20, N = 2: E_eq = 10,000,000*(1 + 1/0.20) = 60,000,000
Below the crossover	At E = 40,000,000: participating pays 10,000,000 + 0.20*30,000,000 = 16,000,000; 2x non-participating pays max(20,000,000, 8,000,000) = 20,000,000
At the crossover	At E = 60,000,000 both pay 20,000,000
Above the crossover	At E = 100,000,000: participating pays 28,000,000; 2x non-participating pays max(20,000,000, 20,000,000) = 20,000,000

• An investor asking for participation and an investor asking for a 2x are not asking for the same thing, and which is worse for common depends entirely on the exit range. Compute E_eq before conceding either.
• A higher multiple is bounded: it can never pay more than N*I unless conversion is better. Participation is unbounded. Founders trading a 2x away for participation are trading a known cost for an open-ended one.
• The comparison is only valid at a fixed p. Any change to the fully diluted denominator moves E_eq, so run it on the post-round cap table, not the pre-round one.

Conversion elections across series are interdependent

Each series elects preference or conversion to maximise its own proceeds, but a converting series joins the residual pool and changes what every other converting series receives. The elections must be solved together, not one at a time.

Field	Value
Formula	A converting series j receives s_j / (s_common + sum of s_k over converting k) * (E - sum of preference amounts of non-converting series)
Setup	Common 5,000,000 shares; Seed 1,000,000 shares carrying 2x on 2,000,000 (preference 4,000,000); Series A 4,000,000 shares carrying 1x on 8,000,000. E = 25,000,000
Worked	Seed takes its 4,000,000 preference (conversion would pay 1,000,000/10,000,000 * 25,000,000 = 2,500,000). Series A converts: 4,000,000/9,000,000 * 21,000,000 = 9,333,333, against an 8,000,000 preference
Result	Seed 4,000,000; Series A 9,333,333; common 5,000,000/9,000,000 * 21,000,000 = 11,666,667. Total 25,000,000

• Solving series by series in charter order gives the wrong answer, because an early series' election changes the residual that a later series is comparing against. Iterate until no series wants to switch.
• The equilibrium is a mixed one here: the series with the multiple takes cash, the series with the larger ownership converts. Mixed outcomes are the normal case, not the exception.
• A waterfall model that hard-codes each series' election rather than solving it will be wrong precisely in the exit range where the answer matters. Test it by sweeping E and checking that every series' proceeds curve is non-decreasing.

Management carve-out and how it reorders the waterfall

A carve-out reserves a slice of proceeds for management ahead of the preferred stack. Because it is paid before the most senior preference, it inverts the payment order the charter otherwise establishes.

Field	Value
Formula	Proceeds available to the stack = E - carve-out. Each series then receives x_j*I_j * min(1, (E - carve-out)/T)
Worked	E = 18,000,000 with a carve-out of 10 percent of E = 1,800,000. Available to the stack = 16,200,000 against T = 30,000,000, a ratio of 0.54
Result	Seed 1,080,000; Series A 4,320,000; Series B 10,800,000; management 1,800,000
Without the carve-out	Seed 1,200,000; Series A 4,800,000; Series B 12,000,000

• A carve-out defined as a percentage of E is a very different instrument from one defined as a percentage of proceeds above the aggregate preference. The first pays in every outcome; the second pays only in outcomes where common was already getting paid. Read which one the plan says.
• Carve-outs are usually documented as a cash bonus plan rather than as equity, so they never appear on the cap table. A waterfall model built from the cap table alone will silently omit the most senior claim in the structure.
• The carve-out exists because a preference overhang leaves management with no reason to run a sale. Investors approve it for the same reason they resist it: it is a transfer from them to the people who have to execute the transaction.

Preference overhang and the first dollar to common

With non-participating preferred, common receives nothing until proceeds clear the entire aggregate preference plus anything senior to it. That break-even, not the last post-money valuation, is the number that determines whether employee equity is worth anything.

Field	Value
Formula	With no carve-out, common's break-even is E = T. With a carve-out set at a fraction c of E, it is E = T/(1 - c)
Worked	T = 30,000,000 and no carve-out: common receives its first dollar above E = 30,000,000
Worked, with a carve-out	T = 30,000,000 with c = 0.10: E = 30,000,000/0.90 = 33,333,333. Check: 0.10*33,333,333 = 3,333,333 to management, leaving exactly 30,000,000 for the stack

• T is a sum of historical cheques and is therefore known exactly. The break-even is one of the few numbers in venture financing that involves no estimate at all, and it is rarely the number quoted to employees.
• Adding a participating series does not move the break-even, but it does reduce common's share of every dollar above it. The break-even and the slope are separate questions.
• If T has grown past the exit values the company can realistically reach, no amount of further operating progress makes common valuable. That is the condition a recapitalisation exists to reset.

Pari passu with unequal multiples

Pari passu ranking is often described as equal treatment. It is equal ranking of preference amounts, which is not the same thing: a series carrying a multiple recovers a larger fraction of its invested capital than a 1x series ranking alongside it.

Field	Value
Formula	R_j = x_j*I_j * E/T for E < T, so the recovery multiple on invested capital is R_j/I_j = x_j * E/T
Setup	Seed 2,000,000 at 2x (preference 4,000,000); Series A 8,000,000 at 1x; Series B 20,000,000 at 1x. T = 32,000,000
Worked	At E = 16,000,000 the ratio is 0.50: Seed 2,000,000, Series A 4,000,000, Series B 10,000,000
Recovery on capital	Seed 2,000,000/2,000,000 = 1.00x; Series B 10,000,000/20,000,000 = 0.50x

• A multiple buys effective priority without asking for seniority, and it does so in a term that is negotiated as an economic point rather than a control point. The later investor conceding pari passu ranking to an earlier 2x has conceded more than the word suggests.
• The arithmetic generalises: within a pari passu group, relative recovery is set entirely by the ratio of preference amounts. Invested capital never enters the calculation.
• When a stack contains mixed multiples, check T against the aggregate of actual cheques written. A 32,000,000 preference on 30,000,000 of capital is a 6.7 percent invisible increase in the overhang.

Also described at: Wikipedia · Wikidata

Where each series flips election in a four-series stack

In a single-series cap table the conversion indifference point is x*I/p. In a stack it is not, because the preferences of every series that has not converted come off the top first and the residual is shared by a smaller group. Solved jointly, every flip point moves higher or stays where it is - here by 47,166,667 for Seed, 32,000,000 for Series A, 5,000,000 for Series B and nothing for Series C, which is the most senior and therefore has nothing ranking ahead of it.

Field	Value
Formula	Series j converts when s_j/(s_common + sum of s_k over converting k) * (E - sum of preference amounts of non-converting series) > x_j*I_j
Setup	Seed 2,000,000, Series A 8,000,000, Series B 20,000,000, Series C 30,000,000, all 1x non-participating. p = 9.600, 16.000, 16.000, 20.000 percent; common and pool 38.400 percent
Solved alone	Seed 20,833,333, Series A 50,000,000, Series B 125,000,000, Series C 150,000,000
Solved jointly	Seed 68,000,000, Series A 82,000,000, Series B 130,000,000, Series C 150,000,000
Understatement	Seed 47,166,667, Series A 32,000,000, Series B 5,000,000, Series C 0
Worked, the Seed flip	Seed converting alone shares the residual with common and pool: 9.600/(38.400 + 9.600) = 0.20000. Setting 0.20000*(E - 58,000,000) = 2,000,000 gives E = 68,000,000
Worked, the Series A flip	With Seed already converted, A's residual share is 16.000/(38.400 + 9.600 + 16.000) = 0.25000. Setting 0.25000*(E - 50,000,000) = 8,000,000 gives E = 82,000,000
Check at the top	At E = 150,000,000 the first three have converted and Series C is exactly indifferent: 0.20000 * 150,000,000 = 30,000,000 = its preference

• The naive figure x*I/p is not a conservative estimate, it is wrong in the founder-unfriendly direction. It tells the Seed investor it is protected only up to 20,833,333 when in fact it holds a cash claim all the way to 68,000,000, and it tells common that the preferred starts sharing far earlier than it does.
• The flip points are ordered, and the order is not the order of the naive points. Which series converts first is decided by the ratio of its preference to its ownership relative to every other series, not by vintage or by cheque size.
• Each flip point is a kink in the proceeds curve for every other holder. A model that samples exit values on a round grid can miss all four kinks and still look smooth. Sample at the flip points themselves.
• The practical use is negotiation timing: between 60,000,000 and 68,000,000 the Seed investor is a cash claimant with no interest in price, and above 82,000,000 both Seed and Series A are equity holders aligned with common. Those are different counterparties in the same room.

A stack with mixed multiples and mixed participation

Replacing the clean Series C with a 1.5x participating preference capped at 2x changes three things at once: the aggregate preference, the residual sharing group, and every other series' flip point. The arithmetic is unchanged; only the inputs move.

Field	Value
Formula	T = sum of x_j*I_j. A participating series joins the residual pool while still holding its preference, so the residual denominator includes its shares whether it has converted or not
Setup	Seed, A and B unchanged at 1x non-participating. Series C 30,000,000 at 1.5x participating, capped at 2x invested
Aggregate preference	2,000,000 + 8,000,000 + 20,000,000 + 45,000,000 = 75,000,000, against 60,000,000 for the clean stack
Flip points	Seed 87,166,667, Series A 107,000,000, Series B 160,000,000, Series C 300,000,000
Worked at E = 150,000,000	Series C 60,000,000 (capped), Seed 10,500,000, Series A 17,500,000, Series B 20,000,000, common and pool 42,000,000
Same exit, clean stack	Series C 30,000,000, Seed 14,400,000, Series A 24,000,000, Series B 24,000,000, common and pool 57,600,000
Cost to common and pool	57,600,000 less 42,000,000 = 15,600,000 at this single exit value

• One structured series moves the whole stack. Series B's flip point rises from 130,000,000 to 160,000,000 and Series A's from 82,000,000 to 107,000,000 purely because Series C is taking more off the top and then sharing the remainder. The earlier investors pay for the later investor's structure before common does.
• A participating series is in the residual pool in every state of the world, so it can never be squeezed out the way a non-participating series can. That is what makes participation the more valuable term at the same headline multiple.
• When comparing a structured proposal with a clean one, compute the cost at three exit values, not one. Here the 15,600,000 at 150,000,000 becomes 0 below 75,000,000 - where common gets nothing under either structure - and shrinks again far above 300,000,000 once the cap forces conversion.

The cap plateau widens when the series sits in a stack

For a single capped participating series the flat band over which its proceeds do not move is I*(1 - p)/p. Inside a stack the band is wider, because the senior preferences of the other series delay the point at which participation reaches the cap while the conversion crossover stays where it was.

Field	Value
Formula	Participation reaches the cap where x*I + p_resid*(E - sum of other non-converting preferences) = cap*I. Conversion overtakes the cap at E = cap*I/p
Setup	Series C 30,000,000 at 1.5x participating, capped at 2x, p = 20.000 percent, in the four-series stack
Cap reached	With Seed and Series A converted the residual group is 84.000 percent of the ledger and C's share of it is 20.000/84.000 = 0.238095. 45,000,000 + 0.238095*(E - 65,000,000) = 60,000,000 gives E = 128,000,000
Conversion overtakes the cap	0.20000*E = 60,000,000 gives E = 300,000,000
Plateau width	300,000,000 - 128,000,000 = 172,000,000 of exit value over which Series C receives exactly 60,000,000
Same series standing alone	I*(1 - p)/p = 30,000,000*0.80/0.20 = 120,000,000

• The plateau is 52,000,000 wider inside the stack than it would be alone. Over that whole range the largest and most recent investor is economically indifferent to the sale price, and it is usually the investor with the board seats and the consent rights.
• The width is not a drafting artefact that can be negotiated away by moving the cap: the cap only decides where the band starts. Narrowing the band requires changing p or removing participation.
• Check where the band sits relative to the exits actually being discussed. A plateau that starts above every credible outcome costs common nothing; one that starts inside the likely range removes the pricing incentive of the holder who controls the process.

Escrow and holdbacks - two allocation conventions that differ by millions

An escrow defers part of the consideration. The waterfall can be run on cumulative proceeds as each tranche is released, or the escrow can be allocated in the same proportions as the closing distribution. Both appear in real allocation schedules and they are not close to each other.

Field	Value
Formula	Cumulative: each holder's share of a release equals its allocation at (closing proceeds + release) less its allocation at closing proceeds. Pro rata to closing: each holder's share equals release * its closing allocation / closing proceeds
Setup	E = 90,000,000 with a 9,000,000 escrow, so 81,000,000 is distributed at closing. All-1x four-series stack
At closing, 81,000,000	Seed 4,600,000, Series A 8,000,000, Series B 20,000,000, Series C 30,000,000, common and pool 18,400,000
Cumulative at 90,000,000	Seed 6,000,000, Series A 10,000,000, Series B 20,000,000, Series C 30,000,000, common and pool 24,000,000
Escrow allocated cumulatively	Seed 1,400,000, Series A 2,000,000, Series B 0, Series C 0, common and pool 5,600,000
Escrow allocated pro rata to closing	Seed 511,111, Series A 888,889, Series B 2,222,222, Series C 3,333,333, common and pool 2,044,444
Difference to common and pool	3,555,556 - the cumulative convention pays common 2.74 times as much

• The cumulative convention is the economically correct one: the escrow is part of the merger consideration and the charter allocates total consideration, not each instalment separately. The pro rata shortcut silently pays the preferences twice, once out of the closing tranche and again out of the escrow.
• The error only appears when the closing distribution and the total sit on opposite sides of a flip point. Here 81,000,000 is below Series A's 82,000,000 flip and 90,000,000 is above it, so the shortcut denies Series A and common the entire benefit of A's conversion.
• Ask for the allocation schedule to state which convention it uses, and ask for the escrow line specifically. It is a single sentence in an exhibit and it is worth more to common than most of the terms negotiated in the term sheet.
• A holdback for a working capital true-up behaves the same way, with the added feature that it can be reduced to nothing. Model the escrow at full release and at zero release; the two waterfalls are not proportional.

Source: Delaware General Corporation Law section 251 (merger consideration and its allocation)

Transaction expenses come off the top, so common pays most of them

Banker fees, legal fees and other transaction expenses are paid before the waterfall runs. Every holder in the residual group bears them in proportion to its share of the residual, not its share of the company - and above the aggregate preference the residual group is much smaller than the company.

Field	Value
Formula	With the residual shared among a group holding fraction f of the ledger, common's share of each expense dollar is p_common/f, not p_common
Setup	E = 90,000,000, all-1x four-series stack, transaction expenses of 3,000,000
Without expenses	Seed 6,000,000, Series A 10,000,000, common and pool 24,000,000
With expenses, 87,000,000 reaches the stack	Seed 5,550,000, Series A 9,250,000, common and pool 22,200,000
Common's share of the expense	1,800,000 of 3,000,000 = 60.000 percent, against an ownership of 38.400 percent
Why	At this exit the residual group is Seed, Series A, common and pool, holding 38.400 + 9.600 + 16.000 = 64.000 percent. Common's share of the residual is 38.400/64.000 = 0.60000

• The multiplier is p_common divided by the residual group's total, and it rises as more of the stack sits outside the residual. Below the aggregate preference the multiplier is zero because common is receiving nothing; between the preference and the last flip point it is at its highest.
• This is why a fee negotiated as a percentage of proceeds is a transfer from common specifically. A 1 percent success fee on a 90,000,000 sale costs common 0.60 percent of the whole enterprise, not 0.38 percent.
• Expenses paid by the company before closing rather than out of the consideration produce the same result by a different route, because they reduce the price the buyer will pay. Where they sit in the documents changes the accounting and not the incidence.

A percentage carve-out and a fixed carve-out pool are different instruments

A carve-out sized as a percentage of proceeds grows with the exit; one sized as a fixed pool does not. They cross exactly once, and on either side of the crossing the two structures transfer value in opposite directions between management and everyone else.

Field	Value
Formula	Percentage plan: carve-out = c*E. Fixed plan: carve-out = K. The two are equal at E = K/c, and the fixed plan is larger below it
Setup	c = 0.10 against a fixed pool K = 6,000,000, so the crossing is at E = 60,000,000
At E = 20,000,000	Percentage pays management 2,000,000 and the fixed pool pays 6,000,000. Common receives nothing either way; the preferred absorb the whole difference
At E = 20,000,000, Series C proceeds	Percentage 9,000,000, fixed 7,000,000
At E = 150,000,000	Percentage pays 15,000,000 and the fixed pool pays 6,000,000
At E = 150,000,000, common and pool	Percentage 50,400,000, fixed 54,720,000 - a difference of 4,320,000

• Management prefers the fixed pool in bad outcomes and the percentage in good ones, which is exactly backwards from the incentive the plan is meant to create. A plan that pays 6,000,000 on a 20,000,000 sale rewards the outcome nobody wanted.
• The structure that aligns is a percentage of proceeds above a threshold, because it pays nothing in the outcomes where the sale destroys value and scales in the outcomes where the buyer had to be persuaded upward.
• Whichever plan is used, it is documented as a bonus plan and not on the cap table, so it will be missing from any waterfall built from the stock ledger. Ask for the plan document by name.

A carve-out on proceeds above the preference costs common much less

The same headline percentage applied to proceeds above the aggregate preference rather than to total proceeds pays nothing in a shortfall and less in every outcome. It is a one-line drafting change with a large arithmetic consequence, and it is the version that actually rewards clearing the stack.

Field	Value
Formula	Threshold plan: carve-out = c*max(0, E - T). Total plan: carve-out = c*E. The difference is c*min(E, T), which is c*T for any exit above the preference
Setup	c = 0.10, T = 60,000,000, all-1x four-series stack
At E = 40,000,000	Threshold plan pays 0; total plan pays 4,000,000, taken entirely from the preferred
At E = 90,000,000	Threshold plan pays 3,000,000 against 9,000,000; common and pool receive 22,200,000 against 18,400,000
At E = 150,000,000	Threshold plan pays 9,000,000 against 15,000,000; common and pool receive 53,280,000 against 50,400,000
The general difference	c*T = 0.10*60,000,000 = 6,000,000 of consideration, in every outcome above the preference

• Read which base the plan uses before arguing about the percentage. Ten percent of proceeds above the preference and ten percent of proceeds are the same sentence with a six-million-dollar difference on this cap table.
• The threshold version is harder to sell to management precisely because it pays nothing in the outcome management is most worried about. That is the argument, not a drafting oversight: the reason to grant a carve-out at all is to create a reason to run a sale that clears the stack.
• A hybrid - a small fixed floor plus a percentage above the preference - is common and is easy to model as the sum of the two plans above.

A carve-out against a participating stack leaks back to the preferred

A carve-out is paid before the waterfall, so it reduces the residual as well as the preference payments. Where the senior series participates in the residual, part of every carve-out dollar is taken from that series and part of it is handed straight back, which changes who is really funding the plan.

Field	Value
Formula	With a participating series holding residual share p_r, the carve-out reduces its proceeds by p_r*carve-out and reduces common's by p_common/f * carve-out, where f is the residual group's total
Setup	Series C at 1.5x participating capped at 2x, E = 150,000,000, carve-out at 10 percent of E = 15,000,000
Without the carve-out	Series C 60,000,000, Seed 10,500,000, Series A 17,500,000, Series B 20,000,000, common and pool 42,000,000
With the carve-out	Series C 60,000,000, Seed 8,250,000, Series A 13,750,000, Series B 20,000,000, common and pool 33,000,000
Cost to common and pool	9,000,000 of the 15,000,000, or 60.000 percent
Cost to Series C	0 - Series C is at its cap in both cases, so it funds none of it

• A series sitting on its participation cap is immune to the carve-out: its proceeds are fixed at cap*I regardless of what is taken off the top. The plan is therefore funded entirely by the series below the cap and by common.
• That is the opposite of the usual framing, in which the carve-out is described as a concession by the most senior investor. Check whether that investor is inside or outside its cap at the exit values being discussed before crediting the concession.
• The same immunity applies to any non-converting series whose preference is fully covered. Only claimants in the residual pool fund a carve-out.

Accrued dividends across series of different vintages

A cumulative dividend accrues from each round's closing, so the oldest series has the largest proportional accrual and the newest has none. Applied across a stack the effect is to reweight the preference towards the earliest money, which is the opposite of the seniority the later rounds negotiated.

Field	Value
Formula	Preference of series j at the measurement date = x_j*I_j*(1 + d*t_j) for simple accrual, or x_j*I_j*(1 + d)^t_j if it compounds, with t_j measured from that series' closing
Setup	d = 0.08 cumulative. Years accrued: Seed 6, Series A 4, Series B 2, Series C 0
Simple accrual	Seed 2,960,000, Series A 10,560,000, Series B 23,200,000, Series C 30,000,000
Compounding	Seed 3,173,749, Series A 10,883,912, Series B 23,328,000, Series C 30,000,000
Aggregate preference	60,000,000 with no dividend, 66,720,000 simple, 67,385,660 compounding
Effect on the common break-even	The first dollar to common moves from 60,000,000 to 66,720,000, a 11.200 percent increase from a clause that was never declared
Seed's recovery multiple at the aggregate	Seed's preference has grown 48.000 percent against Series B's 16.000 percent

• Compounding costs 665,660 more than simple accrual here on a 60,000,000 stack, and the whole difference is a single word in four charters. It is worth reading rather than assuming, because the two are drafted almost identically.
• The reweighting matters in a shortfall under pari passu ranking, where recovery is proportional to preference amounts. The dividend hands the oldest series a larger share of a small outcome without any change to seniority.
• Accrual usually runs to the date of the distribution, not to the signing date, so a slow closing keeps enlarging the preference. On a stack this size each additional month adds roughly 400,000 to the amount ahead of common.
• A non-cumulative dividend accrues nothing unless declared and is economically nil on these facts. Conceding it is free; conceding a cumulative dividend is the same as agreeing to a rising preference multiple.

Recapitalisation by inserting a new senior series

A company can clear an overhang by converting the existing preferred to common, or it can leave the existing stack in place and put a new series senior to all of it. The second route is faster and does not need the old holders' economics to be renegotiated, and it makes the overhang worse rather than better.

Field	Value
Formula	After a senior round of R_new at 1x, T' = T + R_new and the common break-even moves from T to T'. The new series' ownership is R_new/POST and every existing holder is multiplied by PRE/POST
Setup	The four-series ledger, FD = 24,509,805, T = 60,000,000. A new Series D invests 20,000,000 at PRE = 30,000,000, senior to everything
Price and shares	Price = 30,000,000/24,509,805 = 1.2240; new shares 16,339,870; FD after = 40,849,675
Ownership after	Series D 40.000 percent, founders 19.584 percent, Seed 5.760 percent, Series A and B 9.600 percent each, Series C 12.000 percent
New aggregate preference	60,000,000 + 20,000,000 = 80,000,000, so the first dollar to common moves up 20,000,000
Worked at E = 100,000,000	Series D 24,418,605, Series C 30,000,000, Series B 20,000,000, Series A 8,000,000, Seed 3,516,279, common and pool 14,065,116

• A senior insertion is the cheapest financing to document and the most expensive to live with. It raises the preference by the full amount raised while diluting everyone, so common is worse off on both axes at once.
• Issuing a senior or pari passu security is on every standard list of protective provisions, so the existing series have to consent. In practice the consent is obtained by the same holders funding the new series, which is why this transaction is usually an inside round.
• The genuine reset is the other route: convert the whole stack to common and reduce the old cap table to a stated residual. That extinguishes the preference instead of adding to it, and it is the transaction a new management pool needs in order to be worth anything.
• Compare the two by computing the common break-even under each, not by comparing the headline valuations. Here the senior insertion sets it at 80,000,000 while a conversion recapitalisation sets it at the new money alone.

Asset sale, merger and initial public offering are three different payouts

The same enterprise value produces three different distributions depending on the legal form of the exit. A merger runs the charter waterfall on the consideration. An asset sale adds a corporate tax layer before anything is distributed. An initial public offering usually converts the whole preferred stack to common and pays no preference at all.

Field	Value
Formula	Merger: distribute E through the waterfall. Asset sale by a C corporation: distribute E - corporate tax on the gain, then run the waterfall. Qualifying public offering: all preferred converts, so each holder receives p*E with no preference and no election
Merger at 100,000,000	Seed 7,500,000, Series A 12,500,000, Series B 20,000,000, Series C 30,000,000, common and pool 30,000,000
Asset sale at 100,000,000	Tax basis 10,000,000, corporate rate 0.21: tax = 18,900,000, so 81,100,000 is distributable. Seed 4,620,000, common and pool 18,480,000
Common's share of the corporate tax	11,520,000 of 18,900,000 = 60.952 percent
Public offering at a 100,000,000 market value	Preferred converts: Seed 9,600,000, Series A 16,000,000, Series B 16,000,000, Series C 20,000,000, common and pool 38,400,000
What the conversion costs Series C	30,000,000 as a preference against 20,000,000 as common - the preference is worth 10,000,000 more at this value, which is why mandatory conversion is drafted with a price threshold

• The rate used above is the stated federal corporate rate; state tax, the character of the assets and the seller's attributes all move the real number. The structural point is that an asset sale interposes an entity-level tax that a stock sale or merger does not, and that common bears most of it.
• Buyers prefer asset purchases for liability and basis reasons and sellers prefer stock deals for exactly the tax reason above. The gap is negotiated as price, so the form of the deal is an economic term and not a legal detail.
• Mandatory conversion on a public offering is the one event that removes the entire preference at once. That is why its definition - a minimum price, a minimum size, sometimes a named exchange - is negotiated harder than the preference itself.
• A charter that treats an asset sale as a deemed liquidation but does not address a subsequent dissolution can strand proceeds at the company level. Read the deemed liquidation definition and the dissolution waterfall together.

Source: Delaware General Corporation Law section 271 (sale of substantially all assets) and section 251 (merger)

Deemed liquidation events

A liquidation preference is written to apply on a liquidation, dissolution or winding up. Almost no venture-backed company reaches an exit that way, so the charter defines a list of transactions that are treated as if they were a liquidation. That definition, not the preference itself, decides whether the preference applies to the transaction that actually happens.

Field	Value
Formula	No arithmetic. The definition is a list, and the preference applies if and only if the transaction is on it
Commonly on the list	A merger or consolidation in which the company's stockholders cease to hold a majority of the voting power of the surviving entity; a sale, lease or exclusive licence of all or substantially all assets; a sale of a majority of the outstanding stock
Commonly excluded	A reincorporation or holding-company reorganisation; a financing; a transaction with a wholly owned subsidiary
Usually separate	A qualifying public offering, which triggers mandatory conversion instead of a preference payment
Where it bites	An exclusive licence of the core technology can be a disposal of substantially all assets in substance while not looking like a sale in form

• The gap that matters is between a transaction that transfers the business and a transaction that is on the list. A sale of a subsidiary holding the main product, a licence granting all commercial rights in every field, or a sale of a bare majority of the stock can each move the business without meeting a narrowly drafted definition.
• Whether the definition is drafted as automatic payment or as an option for the preferred to demand payment changes who controls the timing. An option gives the holder a second decision after the price is known.
• The definition also drives the acceleration triggers in equity grants and often the drag-along. A change of control defined one way in the charter and another way in the option plan produces a transaction that pays the preference without accelerating any employee equity.
• Ask for the definition as drafted rather than the summary. It is the single clause on which the entire liquidation section of the corpus depends for having any application at all.

Source: NVCA model Amended and Restated Certificate of Incorporation (definition of Deemed Liquidation Event)

Also described at: Wikipedia · Wikidata · NVCA model certificate of incorporation

Strict seniority across four series - the zero thresholds

Under seniority-ordered payment each series has a single number that determines everything about its recovery in a shortfall: the aggregate preference ranking ahead of it. Below that number it receives nothing at all, and above it plus its own preference it is paid in full. The band between the two is narrow.

Field	Value
Formula	R_j = min(x_j*I_j, max(0, E - S_j)), where S_j is the sum of preference amounts senior to j. Series j receives nothing for E <= S_j and is paid in full for E >= S_j + x_j*I_j
Setup	Series C senior to B, B to A, A to Seed. All 1x, T = 60,000,000
Senior preference S_j	Seed 58,000,000; Series A 50,000,000; Series B 30,000,000; Series C 0
Full recovery from	Seed 60,000,000; Series A 58,000,000; Series B 50,000,000; Series C 30,000,000
Worked at E = 40,000,000	Series C 30,000,000; Series B 10,000,000; Series A 0; Seed 0
Same exit, pari passu	Series C 20,000,000; Series B 13,333,333; Series A 5,333,333; Seed 1,333,333
Seed's band	Nothing below 58,000,000 and 2,000,000 above 60,000,000 - a 2,000,000-wide transition over a 60,000,000 range

• Seniority converts every junior series into a binary claim. That is materially worse than it sounds, because the exit values at which seniority matters are exactly the ones the company is most likely to reach when it is being sold under pressure.
• The comparison to run is not stacked against pari passu at the last post-money, which is above T and where the two are identical. Run it at half of T and at three quarters of T.
• A four-series stack has three seniority decisions, taken in three separate negotiations, years apart, each time by an investor comparing outcomes that already look poor. That is how a stack ends up strictly ordered without anyone having agreed to order it.
• Tiered ranking - groups of series equal within the group and ordered between groups - is the usual landing point and is modelled by treating each tier as a single series with the tier's aggregate preference.

The first dollar to common, with every deduction added

The common break-even is the aggregate preference, adjusted for anything paid ahead of it. Each item is small and each is documented somewhere other than the charter, and together they move the number by half again.

Field	Value
Formula	Break-even = (T + accrued dividends + net debt + transaction expenses + fixed carve-out) / (1 - c), where c is any carve-out expressed as a fraction of proceeds
Aggregate preference T	2,000,000 + 8,000,000 + 20,000,000 + 30,000,000 = 60,000,000
Plus accrued dividends at d = 0.08 by vintage	66,720,000 - an increase of 6,720,000
Plus net debt of 6,000,000 and expenses of 3,000,000	75,720,000
Divided by (1 - c) for a 10 percent carve-out	84,133,333
Check	0.10 * 84,133,333 = 8,413,333 to management, leaving exactly 75,720,000 for debt, expenses and the stack
Against the last post-money	150,000,000. The break-even is 56.089 percent of it

• Every input here is a known historical number or a stated contract term. The break-even involves no forecast at all, which makes it the one number in a venture cap table that can be computed to the dollar and audited.
• It is also the number that is never quoted. Employees are told the last post-money valuation, which is a price paid for a senior instrument, and are left to infer that their common is worth the same fraction of it.
• Recompute it after every financing and after every year of accrual, not just at the point of a sale. The whole purpose of the figure is to know in advance which outcomes are worth pursuing.
• If the break-even has passed the exit values the business can credibly reach, further operating progress cannot make common valuable. That is the specific condition a recapitalisation exists to reset, and the break-even is how it is diagnosed.