sources & method
Where every number on this site comes from, the conventions used, and what was left out on purpose.
live datalive series
| Series | Source | Published | Latest | This build |
|---|---|---|---|---|
| SOFR | NY Fed markets API | each business day, for the prior day | 16 sep 2026 | live |
| SOFR 30-day average | NY Fed markets API | each business day | 17 sep 2026 | live |
| SOFR 90-day average | NY Fed markets API | each business day | 17 sep 2026 | live |
| HIBOR 1M (daily) | HKMA open API, daily interbank liquidity | each business day | 17 sep 2026 | live |
| HIBOR overnight (daily) | HKMA open API, daily interbank liquidity | each business day | 17 sep 2026 | live |
| Aggregate balance | HKMA open API, daily interbank liquidity | each business day | 17 sep 2026 | live |
| HIBOR by tenor | HKMA open API, monthly statistical bulletin | daily figures, about two weeks behind | 31 aug 2026 | live |
| USD/HKD | ECB reference rates via Frankfurter | each TARGET business day | 17 sep 2026 | live |
live = fetched for this build · cached = the source failed, last good copy used.
conventions
| Item | Convention used |
|---|---|
| USD money market | ACT/360 |
| HKD money market | ACT/365 |
| SOFR over a period | compounded in arrears; the NY Fed's 30/90-day averages as the published proxy |
| HIBOR − SOFR | one-month HIBOR minus overnight SOFR, on dates both are published. Different tenors, so a guide rather than an arbitrage. |
| FX forward | covered interest parity from published fixings; excludes the cross-currency basis |
| Carry tables | simple annual interest, so the argument stays visible |
illustrative assumptions
Not observed anywhere free. Kept identical across pages so the examples agree with each other.
| Assumption | Value |
|---|---|
| Long swap spread | 45bp |
| Short swap spread | 30bp |
| Hard-to-borrow fee | 5% |
| Dividend yield | 2% |
left out on purpose
Stock borrow fees and dealer swap spreads. These are the numbers a financing desk argues about most, and no free source publishes them. Securities-lending data comes from paid vendors. Rather than show a made-up number, the log explains how they're built and labels every assumption.
Credit, commodity and crypto financing pages. An earlier version of this site charted them. They were accurate but didn't help explain a stock position, so they're gone.
Trade-level DTCC data. It's used locally for the lifecycle finding, but not republished here.