Housecoin

Start with the arithmetic, because it is not in dispute. A house in the United States costs $410,700 at the median; in 2000 it cost $165,300. Indexed properly, by repeat sales of the same homes rather than by whatever mix happened to close that quarter, American housing is worth 3.37 times its 2000 value. Consumer prices in general are worth 1.97 times theirs. Real median household income is up sixteen per cent.1 The roof over your head appreciated at nearly twice the rate of everything else you buy, and wages did not move.

Rates finished the job. At the January 2021 low a thirty-year fixed cost 2.65%; it now costs 6.66%. On the median house with a fifth down that is a payment near $2,110 a month before tax, insurance or a single repair, and roughly $431,000 of interest on a $329,000 loan.2 Homeownership has fallen to 65.0% while unemployment sits at 4.1%, which is the uncomfortable part: this is not a recession pricing people out, it is an ordinary economy in which the ordinary asset has moved beyond the ordinary wage.

The textbook response to an asset outrunning you is to own some of it, and that door is what closed: a house is indivisible, costs a fifth of its price to enter, and cannot be bought in the size a normal saver has. So the saver holds cash, which is the one position guaranteed to lose, for exactly the reason the house keeps winning.

This paper does not solve housing and does not track house prices; anyone selling a memecoin as an inflation hedge is selling a story. What it does is narrower. A coin produces one real thing, trading fees, and in almost every case those fees leave. Here they are converted into the closest instrument a chain can hold to the risk-free rate and handed to the holders.

Behind the price of an ordinary memecoin there is nothing: the float is the float, the chart is the product, and when volume stops there is no asset anywhere in the arrangement that a holder has a claim on. Fees are the exception and also the proof, because every coin on every launchpad earns them and in almost every case the creator claims them and spends them somewhere the chart never sees.

$HOUSE differs in one respect. The fees do not leave: every trade pays 0.3%, the crank claims it, buys USDY with it, and sends that USDY to the wallets holding the coin in proportion to what they hold. There is no staking, no lock-up, no claim button and no governance. Holding is the entire qualification, and the payment arrives in the wallet without anyone signing anything.

What gets paid matters as much as the fact of paying. A distribution in the coin itself is a share issue: it hands holders more of the thing they already hold and dilutes the price by roughly what it pays. A distribution in a plain dollar is honest and inert. A distribution in a note against three-month Treasuries is honest and keeps working: it arrives, it sits in the recipient’s wallet, and it earns there whether or not they ever look at it again.

One consequence should be said in the open, because it is unusual enough to sound like a mistake. A holder who never trades still gets paid; a trader who buys and sells within a block pays the fee and receives nothing. The coin pays the patient out of the impatient, which is the opposite of how this industry usually runs.

$HOUSE lives on a pump.fun bonding curve. One instruction, create_v2, mints a fixed supply into the curve account under Token-2022 and discards the mint authority in the same transaction it uses it. The mint carries an empty extension set: no transfer hook, no transfer fee, no permanent delegate, no freeze authority. That is deliberate, and unlike a promise it is verifiable in one RPC call. There is no allocation, no vesting and no team wallet, because there is nothing left to allocate.

Two properties of that program matter here, and both belong to code we did not write and cannot amend. The creator’s share of the trading fee is a flat 0.3% of quote volume, charged to buyer and seller alike, and it does not scale with market capitalisation or stop at graduation. And coin_creator is written once at creation, with no instruction to reassign it, so the destination of every fee this coin will ever earn was decided in the transaction that created it and is beyond the reach of the person who created it.

The asset being distributed is Ondo US Dollar Yield: a token backed by short-term US Treasuries and bank demand deposits, whose interest accrues into its price rather than into a rebasing balance. That is the whole reason for choosing it over a plain stablecoin. A dividend paid in a stablecoin stops being useful the moment it lands; a dividend paid in USDY keeps earning in the recipient’s wallet, at whatever the short end of the curve is paying, without them doing anything at all. It trades near $1.14 against USDC’s $1.00, and that gap is accrued yield rather than a depeg.3

It is also, unglamorously, the right shape for the job: a classic SPL mint with six decimals and about three million dollars of routable liquidity, filling one SOL at 0.0045% price impact. A distribution is a few hundred ordinary token transfers, and an exotic mint with transfer hooks or a permanent delegate would make each of them a different and more expensive problem.

The only money entering this system is the creator leg of the pump.fun trading fee. If V is cumulative quote volume, the revenue is

and that is the entire monetary base. No emission, no inflation, no treasury sale, no second round, and no promise of a rate: what is distributed is what was earned, and if nothing is earned nothing is distributed. Equation (1) does not depend on price, only on turnover, so a coin trading sideways on constant volume pays at exactly the rate one trading upward does.

A distribution does not run on a clock. Time is not what the crank is waiting for, money is, and paying out dust costs more than the dust is worth. A block closes when the vault holds enough for the run to be worth its own cost.

Two costs set the threshold. The first is execution: below roughly 0.020 SOL the swap into USDY loses more to fees and slippage than it delivers. The second is rent, and it is the one that governs. A holder who has never held USDY has no account for it, and somebody must pay about 0.00204 SOL to open one; that somebody is the vault. So a distribution to a thousand new recipients costs roughly two SOL in account rent before a cent reaches anyone, which is why payouts below a floor are not made at all and why the floor is a published parameter rather than a secret.4 The rent is paid once per holder: the second distribution to the same wallet costs nothing.

In volume terms, a steady-state block is 12 SOL of trading against the coin. Not a figure anyone is asked to believe in, just the arithmetic of 0.3% against the threshold. If the market is quiet, distributions are far apart. If nobody trades, none happen, and nothing is lost but time.

Five steps, in order, each signed by the same key: the wallet the coin was launched from, which pump.fun recorded as its creator and which is therefore the only key that can claim anything at all.

1. Read. The unclaimed creator balance is read across both pump programs, the bonding curve and the AMM, because the revenue moves from one to the other at graduation. Below 0.003 SOL nothing is claimed, so a quiet minute costs one RPC call rather than a wasted transaction.
2. Claim. The vault signs collect_coin_creator_fee and the rewards arrive as SOL. After graduation the AMM pays in wrapped SOL, which is unwrapped in the same step.
3. Buy. The whole deployable balance, less a gas reserve, buys USDY through an aggregator route, bounded at 3% slippage. There is no buyback leg: a buyback is one side of a pairing, and money spent on it is money that does not reach a holder.
4. Snapshot. Every token account holding $HOUSE is read, summed by owner, and filtered. §8 is what that means in practice.
5. Pay. Transfers go out in batches of six, each batch its own transaction, opening the recipient’s account where one is missing. A batch that fails is recorded and skipped rather than stopping the run, and the amount it was carrying stays in the vault for the next distribution.

A distribution is only as honest as the list it pays, so the list is read from the chain at the moment of paying rather than from any record of ours. Three details decide who is on it, and each of them changes who gets money.

Balances are held by accounts, not people. One wallet can hold the same mint in several token accounts, so accounts are summed by owner before any share is computed. Paying per account rather than per owner would pay someone twice for splitting their own balance.

Some of the largest holders are not holders. A liquidity pool’s vault holds the coin on behalf of whoever trades against it; paying a dividend into that vault pays the pool, which distributes to nobody and quietly leaks the dividend to arbitrage. Pool authorities are excluded by address, as are burn addresses and the vault itself.

Below a floor, a payment costs more than it delivers. Holders whose pro-rata share is under the floor are not paid, and their share stays in the vault and rolls into the next distribution rather than being redistributed by a tie-break nobody can audit. The same is true of what integer division leaves over. Nothing is destroyed and nothing is kept: it is simply distributed later, when it is worth the transaction.

The snapshot is taken fresh every time and nobody accrues a claim between them. A holder who sells before a distribution is not in it; a holder who buys before one is. That is the simplest rule that cannot be gamed by anything except holding the coin, which is the behaviour the mechanism is trying to pay for.

Write b for a holder’s balance, S for the eligible supply after exclusions, and R for the reserve bought in this block. The payment is

computed in integer base units, so the sum of every payment plus the remainder is exactly R and can be checked against the transactions on chain. Cumulative across blocks, the total any holder has received is a non-decreasing quantity that price action cannot reverse: what has been paid has left this protocol’s control and sits in their wallet.

pump.fun fixes the fee destination and the aggregator executes the swap. Neither is ours. The cycle in §7 is: a keeper runs it on a 60-second loop and holds the key to the launch wallet.

That key cannot mint, because create_v2 discards the mint authority at creation. It cannot redirect the fees, because coin_creator is immutable. It cannot freeze or claw back a transfer, because the mint carries no extension that would allow it. What it can do is stop: nothing on chain compels a distribution, and the same key that pays the holders could simply keep the money instead.

Say that plainly rather than in a footnote. Every distribution that has happened is on chain and is listed above; every distribution that has not happened yet is a choice. Until the vault is owned by a program whose only instruction is to pay out, what stands between the fees and the holders is somebody’s intention. Price the operator, not the arithmetic.

Notes

1. Housing and rate figures are published series, read on the day of writing: median sales price of houses sold, $410,700, Q2 2026 (FRED MSPUS, against $165,300 in Q1 2000); Case-Shiller US national index 336.663, June 2026, on a January 2000 base of 100 (CSUSHPINSA); CPI-U 332.813 against 169.300 (CPIAUCSL); real median household income $83,730 in 2024 against $71,790 in 2000, in 2024 dollars (MEHOINUSA672N); thirty-year fixed 6.66% on 27 August 2026 against the 2.65% low of 7 January 2021 (MORTGAGE30US); homeownership 65.0% (RHORUSQ156N); unemployment 4.1% (UNRATE). United States only.
2. The payment figures are arithmetic on those series: $410,700 at 20% down is a $328,560 loan, which at 6.66% over 360 months amortises to about $2,110 a month of principal and interest, roughly $760,000 in total, of which about $431,000 is interest. Taxes, insurance and maintenance are excluded, which flatters the number rather than the argument.
3. The reserve is Ondo US Dollar Yield, mint A1KLoBrKBde8Ty9qtNQUtq3C2ortoC3u7twggz7sEto6, classic SPL, 6 decimals, roughly $3.0M routable, one SOL filling at 0.0045% price impact. Its mint and freeze authorities are both live; §4 is the disclosure, not this note.
4. Rent for a token account is a chain parameter, about 0.00204 SOL, and it is read from the chain rather than hard-coded. The payout floor and the batch size are both published parameters, imported by this page from the same module the keeper reads.
5. pump.fun charges a total trading fee of which the creator leg is one part; every figure here quotes the creator leg only, the 0.3% of volume this protocol actually receives, never the total the trade pays. The bonding curve program is 6EF8rrecthR5Dkzon8Nwu78hRvfCKubJ14M5uBEwF6P.
6. Every parameter quoted in the prose is imported from the module the keeper reads, so the document cannot drift from the process it describes.