Financial Reporting
Cumulative Dividends on Preferred Stock: The 8% Nobody Pays That Still Comes Out of Common
An 8% cumulative dividend never costs cash, but it grows the liquidation preference, shrinks common stock at exit, and now has a FASB measurement rule.
By 409.AI Team - 2026-09-23
# Cumulative Dividends on Preferred Stock: The 8% Nobody Pays That Still Comes Out of Common
The term sheet says the Series B carries an 8% cumulative dividend. You ask about it, and the answer sounds reassuring: nobody ever writes that check. That part is true. Venture-backed companies almost never pay a dividend in cash, and most charters don't let them.
What the answer leaves out is that the dividend still gets paid. It gets paid at the exit, out of the money that would otherwise have reached common stock, and by then it has been compounding for five years.
Here's what the term does, what it costs in real numbers, and why the accounting for it changed this spring.
Cumulative, non-cumulative, and paid in kind
Most venture preferred carries a non-cumulative dividend. The model language reads that dividends "will be paid on the Series A Preferred on an as-converted basis when, as, and if paid on the Common Stock," which [Morrison Foerster](https://www.mofo.com/resources/insights/230628-common-provisions-in-venture-capital-term-sheets-dividends) describes as a provision that stops a company from paying cash out to common holders without returning value to its venture investors. It costs nothing until a dividend is declared, and in a startup that never happens. It's a guardrail, not a claim.
A cumulative dividend is a different instrument wearing the same word. It accrues on a schedule whether or not the board declares anything, and it becomes payable at a liquidation, a redemption, or whenever the board finally does declare. Morrison Foerster's example runs at 5%: a share with a $1.00 issue price carries an accrued dividend of "$0.05/share at the end of the first year, $0.10 at the end of the second year." Nothing moves through the bank account. The claim just grows.
The third variant is the paid-in-kind dividend, where the company settles the accrual by issuing more preferred shares rather than cash. [BDO's summary](https://arch.bdo.com/PIK-Dividends-on-Equity-Classified-Preferred-Stock-Initially-Measurement) puts the definition plainly: a PIK dividend is satisfied "either by delivering to the holder additional preferred stock with the same terms as the original preferred stock or by increasing the value of the original preferred stock." Same economics as an accrual, plus a cap table effect that shows up further down.
There's also a legal reason nobody pays these in cash. Under [Section 170 of the Delaware General Corporation Law](https://delcode.delaware.gov/title8/c001/sc05/index.html), a board can declare dividends only out of surplus, or out of net profits for the current or preceding fiscal year. A company burning venture capital has neither. So the accrual sits there, growing, until an event forces a settlement.
Cumulative dividends stay uncommon in priced venture rounds. Fenwick's [Q1 2026 Venture Beacon](https://www.fenwick.com/insights/publications/q1-2026-venture-beacon-key-vc-market-trends) reports that investor-friendly terms including participating preferred, greater-than-1x preferences and cumulative dividends "remained rare." They cluster where the balance of power shifts: structured late-stage rounds, extensions, insider bridges, and the [pay-to-play recapitalizations](https://409.ai/articles/pay-to-play-recapitalization-409a-cram-down-cap-table) that show up when a round doesn't come together on clean terms.
What 8% costs at an exit
Take a company with 30 million fully diluted shares: 12 million of common and options, 8 million of Series A bought for $10 million, and 10 million of Series B bought for $20 million at $2.00 a share. Both preferences are 1x non-participating. Only the Series B carries the 8% cumulative dividend, compounding annually, and the company sells five years after that round closes.
Five years of compounding turns the $20 million preference into $29.39 million. The accrued dividend is $9.39 million, and not a dollar of it was ever recorded as cash out the door.
Sell the company for $75 million and the damage looks modest. Without the accrual, the Series B would convert to common, because 33.3% of $75 million is $25 million and that beats a $20 million preference. With the accrual, conversion is worth $25 million against a preference of $29.39 million, so the Series B takes the preference instead. That leaves $45.61 million. The Series A converts, and common and options split 60% of what's left: $27.37 million, or $2.28 a share against the $2.50 they'd have seen otherwise.
Now sell for $40 million, which is the outcome this term is actually written for. Without the accrual, the Series B takes $20 million and the Series A takes $10 million, leaving $10 million for 12 million common shares, or about $0.83 each. With the accrual, the Series B takes $29.39 million and the Series A still takes its $10 million, which leaves $610,000. Common lands at roughly $0.05 a share.
One clause, no cash, and the common stock loses 94% of its proceeds. This is the same mechanic as a [liquidation preference stack](https://409.ai/articles/liquidation-preferences-waterfall-common-stock-exit), except the stack keeps growing on a calendar.
Compounding is the whole negotiation
Run the same deal with a simple 8% accrual instead of a compounding one and the arithmetic changes more than it looks like it should. Simple accrual adds $8 million over five years rather than $9.39 million, so the Series B preference is $28 million. At the $40 million exit, common gets $2 million instead of $610,000, more than triple the outcome, from a single word in the charter.
The PIK version adds a second effect. Settle that 8% in shares and the Series B's 10 million shares become about 14.69 million after five years. Fully diluted count goes to 34.69 million, so the Series B's as-converted stake climbs from 33.3% to 42.4% while common's falls from 40% to 34.6%. The accrual dilutes on the way up as well as subordinating on the way down.
What it does to your 409A
The strike price your employees pay comes out of an allocation model, usually an option pricing model that treats the equity as a series of call options struck at the points where each class starts sharing. If you've read how [OPM and PWERM allocate value](https://409.ai/articles/409a-allocation-methods-opm-pwerm-backsolve), the accrual does something simple to the inputs: it moves the breakpoints.
In the example above, common begins participating only after the preferences are satisfied. At signing that threshold sits at $30 million. Five years later, with the accrued dividend included, it sits at $39.39 million. Push that first breakpoint up by $9.39 million and hold every other input constant, and less value lands on common. Part of the reason a [409A comes in below the post-money](https://409.ai/articles/why-is-your-409a-valuation-lower-than-post-money-valuation) is exactly this, and an accruing dividend widens the gap every quarter.
Two practical points follow. First, the accrual grows between valuations, so two 409As on the same cap table and the same enterprise value won't produce the same strike price. Second, your valuation provider needs the certificate of incorporation, not the term sheet. The charter is what states whether the dividend is cumulative, whether it compounds, at what rate, on what base, and whether it's payable on a conversion or only on a redemption. Those details decide the number, and they're the ones that get redlined late and quietly. When you commission a [409A valuation](https://409.ai/products/409a), send the current charter and a schedule of accrued dividends by series as of the valuation date.
ASU 2026-01 settles how the accrual gets measured
Until this year, U.S. GAAP said nothing about how an issuer should initially measure a PIK dividend on equity-classified preferred stock. Some companies used the stated rate in the agreement. Others used the fair value of the additional shares they issued. Two companies with identical terms could report different numbers.
In April 2026 the FASB closed that gap with [ASU 2026-01, Equity (Topic 505)](https://storage.fasb.org/ASU%202026-01.pdf). In-scope PIK dividends are now measured using the rate stated in the preferred stock agreement, applied to the liquidation value. KPMG's [summary](https://kpmg.com/us/en/frv/reference-library/2026/fasb-issues-final-asu-pik-dividends-equity-classified-preferred-stock.html) uses the obvious illustration: an 8% PIK rate on a $1,000 liquidation value produces an $80 dividend, full stop, with no fair value exercise attached.
The scope covers equity-classified preferred stock, convertible or not, including preferred classified as temporary equity, which is where most redeemable venture preferred sits. It excludes preferred classified as a liability, dividends settled in a different equity instrument such as common shares, and nonmonetary transactions. Check the classification before you assume the rule reaches your preferred, because that gating question is the same liability-versus-equity analysis that decides how [a SAFE lands on the balance sheet](https://409.ai/articles/safe-liability-vs-equity-classification-asc-480-815-40).
The amendments take effect for annual periods beginning after December 15, 2026, and the interim periods within them, for all entities. Calendar-year companies are looking at fiscal 2027. Early adoption is permitted in any period whose financial statements haven't yet been issued or made available for issuance, and adopters can choose prospective application or a modified retrospective approach with a cumulative-effect adjustment.
For a private company with no public reporting obligation, the visible effect is contained: the carrying amount of preferred and the equity rollforward. One group should care more. ASC 260 requires earnings per share from entities whose common stock trades publicly and from those that have filed, or are in the process of filing, with a regulatory agency to sell securities in a public market. A company drafting an S-1 has to present EPS for prior periods, and PIK dividends reduce income available to common shareholders in every one of them. Sorting out the measurement before an auditor does is cheaper than sorting it out during a public offering, which is the same lesson as [the first audit of a 409A](https://409.ai/articles/first-audit-409a-valuation-au-c-540-assumptions).
What to push on before you sign
If an investor wants a cumulative dividend, the rate is rarely the most valuable thing to argue about. Compounding versus simple accrual moved common's proceeds by more than 3x in the example above. Whether the dividend is payable on conversion or only on a redemption decides whether it touches a normal exit at all. A cap on the accrual period, say five years, puts a ceiling on a claim that otherwise grows for as long as the company takes to sell. Each of those is a smaller ask than deleting the term, and each is worth more than shaving the rate from 8% to 6%.
The reason this one slips through is that it never announces itself. It doesn't change the post-money headline, doesn't hit the bank account, doesn't appear on a standard cap table summary, and doesn't cost anything in the years when the company is doing well. It shows up once, in the waterfall, in the deal that gets done at a number nobody planned for. If your charter has an accruing dividend in it, put the running balance on the board deck next to the cap table, and reprice it in your head every time someone quotes a valuation. And if a round goes sideways and that accrual has to be dealt with, the cleanup options look a lot like [a down round with underwater options](https://409.ai/articles/down-round-409a-underwater-options-repricing): possible, but far more expensive than the conversation you could have had at signing.