A working engine · built before we bid
Total return is the one number neither Alameda nor its peers reports, the metric California's CDIAC calls "the most ideal" for judging a public pool. We did not propose to build the capability to produce it. We built it, and tested it. It ingests the standard exports your SymPro and custody systems already generate, strips out your own cash flows, and returns an auditable, benchmark-ready total return. On day one of data access, it runs on your files.
What it ingests
Nothing exotic. Every input below is a report your systems already produce, SymPro publishes 50+ reports to Excel/CSV and auto-interfaces with your custodian, pricing service and general ledger. The engine reads them as uploaded files.
Per CUSIP: par, book / amortized cost, market value, accrued interest, coupon, maturity, call schedule, rating.
SymPro Portfolio / Investment Status report.
Buys, sells, calls, paydowns, maturities, coupons, with dates, price, principal, realized gain/loss.
SymPro Activity report.
The County's own deposits and disbursements, the property-tax apportionment cycle, dated, so they can be stripped from performance.
GL / Auditor-Controller apportionment.
Interest received, accrued-interest change, amortization / accretion.
SymPro Earnings report.
The custody bank's holdings and market values, for an independent re-price and reconciliation.
Custody statement / data feed.
How it works
Read the five series; aggregate holdings to a portfolio market value at each period boundary.
Remove the County's own contributions/withdrawals, CDIAC's rule, so apportionments aren't counted as performance.
Compute a CDIAC-standard time-weighted total return, chain-linked month over month.
Split income vs. price; attribute by security through calls, paydowns and maturities.
Tie holdings to custodial marks, to the dollar, and benchmark against a duration-matched index.
Why an engine, not a spreadsheet. At the granularity a gold-standard total return needs, every security valued at each cash-flow date, ideally daily, across six years, the dataset runs into the hundreds of thousands of rows, beyond what a spreadsheet can practically compute or audit. The math runs in this engine; the downloadable Excel workbook is the auditable output, not the compute platform. It is also why the standing report belongs in the County's own database-backed SymPro system, where, in fact, total return is already an available report.
Proof, not promise
Run on a full set of production-format files, the engine reproduces the six-year total return we had already reconstructed from your public reports, to the penny, and it shows precisely why the cash-flow data matters.
That 16-point gap is the apportionment cycle being mistaken for performance, the clearest reason an auditable total return requires the cash-flow series, and exactly what this engine handles.
Recast vs. production
Everything across this portal, the total-return reconstruction, the security-level recast, the EIP backtest, was built from your public monthly reports alone and validated three independent ways. That is ~95% of the picture. The engine converts it to 100%, exact, auditable, to the dollar, the moment your security-level and transaction files are uploaded.
Income from book yield, price from the pool-level NAV ratio; reconstructed from 75 monthly reports. Proves the method and shows the shape, at zero cost to the County.
Actual income and security-level marks; flows stripped at each date; every bond tracked through its lifecycle; reconciled to custodial and benchmarked. The metric, made permanent, and yours.
The ~95% recast we built before the bid, and the production shell that turns it into an exact, auditable result the day your data arrives. Both are downloadable, right here.
All 401 securities repriced from the County's public reports, the 75-month total-return series, and the EIP classification, built entirely from public data and validated three independent ways. This is what an evaluator can hold in hand before any award.
⬇ Preliminary recast, 3.2 MBThe real 31,129-row holdings history plus the auditable, flow-stripped time-weighted total return. The production shell, exact to the dollar the day your live SymPro and custodial files replace the modeled series.
⬇ Production shell, 3.0 MBRun it yourself · AI-driven
This is what the historical recast actually produces. Load the County's data set, the real 31,129-row holdings history plus the activity, cash-flow, income and custodial series, or upload your own. Our AI-driven engine analyzes and classifies the data, organizes it into the five series, strips the County's own cash flows, and computes a CDIAC-standard time-weighted total return.
Drag & drop your five exports onto the boxes below (or click to browse), or load the demonstration set.
Drag the five files onto the boxes above, or load the demonstration set, to run the engine.
Where the month-to-month price change comes from: each security's month-end market value in the holdings file (A), sourced from the custodian / pricing service and reconciled to the custodial marks (E) within 0.03%, no separate pricing feed needed for this high-grade book.
Engine output · historical total return (2020–2026)
Eleven tabs, including a live, formula-driven Calculation (TWRR) tab you can audit cell by cell, and a Portfolio MV tab that sums each month's value live from the 31,129 holdings and ties to the engine, so every number traces in the spreadsheet itself, plus the full Transactions, Cash Flows, Income & Custodial data. The total return is computed by a deterministic engine (reproducible code, not a language model); the workbook reproduces that math in auditable Excel formulas. Built from the County's real 31,129-row holdings with the supporting series modeled for the demonstration; your live SymPro / custodial files replace the modeled series in the engagement.
Upload-ready
You're engaging us to run it on your data and hand you the result, and the capability, for good. The engine is built and tested; it needs only your file uploads.