Public Cash MoneyCS — THE EMPIRICAL LAB.
Backtests, a Registered Validation Protocol, a Payments Panel, and an Engine You Can Drive Yourself
Most monetary proposals ask to be believed. This page is where the Citizens Standard asks to be checked. Everything below is reproducible: the data are public, the code is published, and each package runs end to end with one command.
Neo-Solon · citizensstandard.org · 2026
1. The Interactive Engine
The issuance engine of Chapter 2 — every channel, every Mode, every dial — implemented as a single self-contained page running on sixty-five years of US data. Set the citizenship deposit, the growth split, the transactional share, the debt-band target; watch the price level, the floors, and the debt path respond. The engine is not an illustration of the papers; it is the same arithmetic the papers verify, executed live. A separate verification note documents that the engine’s outputs match the published results to the dollar.
2. The Horse Race: Does the Transactional Aggregate Actually Predict Inflation?
The framework’s load-bearing empirical claim — goods prices are set by transaction-active money, not total money — was tested under a pre-registered protocol: constructions, benchmarks, sample splits, and the falsification condition all fixed before estimation, and published with its own kill switch (“if, out of sample and across all constructions, the transactional aggregate adds no goods-inflation information over M2, the decomposition claim fails”).
Results of the first execution: in high-money-growth regimes, the transaction-active aggregate carries four to five times the goods-inflation information of simple-sum M2 (next-12-month CPI R² 0.19 vs 0.04), statistically displaces M2 in a joint regression (M2’s coefficient falls to zero), and beats it out of sample in that regime. Reported with equal prominence: the advantage is regime-conditional — it vanishes in quiet-inflation samples — and out of sample the transactional aggregates beat M2 but only tie a naive inflation-persistence baseline. Money beats broad money; nobody beats persistence. The falsification condition stands untriggered, and unclaimed victories stay unclaimed.
3. Three Independent Constructions, One Band
The transactional share is triangulated, not assumed: a composition construction (from Federal Reserve component data), a user-cost construction (Divisia), and a payment-flow construction (from Nacha/RTP/Payments-Study volumes). The first two converge tightly (growth correlation 0.82, level correlation 0.99). The third — executed on an assembled payments panel with per-row source attribution — brackets the share between a narrow-anchor lower bound and the registered band of 0.46–0.57 rather than pinning a point, with both ends of the bracket corroborated by independently measured account-turnover data. The construction constrains the share; it does not yet settle it, and the page says so because the distinction matters.
4. The 2020–22 Episode, in Payments Data
The sharpest single exhibit in the lab. From February 2020 to March 2022, M2 grew +40.6 percent; consumer prices followed with a lag, peaking at 9.0 percent. The two-circuit model makes a specific prediction about the sequence: the new money should appear as balances before it appears as transactions. The quarterly ACH payments panel shows exactly that — payments value growing +4 to +22 percent year over year through 2020–21 while balance aggregates grew +8 to +35 percent, with correlations near zero in the surge window. The money parked first, and priced later. The decomposition’s premise, observed directly in the payments system.
5. Backtests and Stress
The engine’s sixty-five-year backtest reproduces the framework’s headline accumulation results and their honest tails: the flagship floor of ≈$413K real at the general-equilibrium return of 4.26 percent (band 3.30–5.03, derivation in closed form); the sequence-of-returns tail, where a crash on the eve of retirement halves a cohort’s floor ($209,942 → $102,114); and the lost-decade result, where ten years of zero growth costs a mid-career cohort only ~4 percent — the floor’s enemy is a badly timed crash, not a slow economy. The crisis companion runs the architecture through depression, stagflation, crash, pandemic, and secular-stagnation scenarios; the transition companion reproduces the 102 → 84 → 58 percent debt path year by year.
6. Reproduce Everything
Every chapter’s numbers trace to a replication package: monetary data from FRED, payments data from Nacha and the Federal Reserve Payments Study (per-row provenance recorded), wealth data from the Survey of Consumer Finances, all code published, each package executable with a single command, and a claim-backing index mapping every published figure to the script that generates it. If a number on this site does not reproduce, that is a defect in the site or the papers — report it, and it will be fixed or retracted.
Papers, data, and code: citizensstandard.org · Replication archive: github.com/Neo-Solon/Citizens-Standard