Special data series
Special data series are the market-wide reference series Fincanva keeps alongside instrument prices — interest rates, an inflation series and broad valuation indicators — used as inputs to a simulation or as comparisons, never as holdings: you cannot buy an inflation rate. Six exist today, and each has a defined behaviour when a date falls outside its data.
Also seen as: reference series, macro series, special symbols
How does Fincanva handle it?
- Six special data series exist today: risk-free rate, margin-loan rate, inflation, Shiller PE, Buffett Indicator, and Buffett Deviation.
- The risk-free rate is a live market series matched to your backtest's own dates, not a constant — see risk-free rate.
- The risk-free fallback is a flat 2% a year; the margin-loan and inflation fallbacks are fixed values that are not published.
- Special data series are references rather than holdings: they are not tradable instruments a strategy buys, so they never appear as a position in a backtest.
- They refresh alongside instrument prices, so a series' last data point moves forward as new data lands — see data freshness and frontier.
What does it look like in practice?
Suppose a backtest window runs 120 months and its first 12 fall before the risk-free series begins. The risk-free rate for that window is then a month-weighted blend: 2% a year for those 12 months and the series' own average for the other 108. If the series averaged 4% over its part, the window's rate is (12 × 2% + 108 × 4%) ÷ 120 = 3.8%, and the Sharpe ratio for that window subtracts 3.8%. Those 12 months are held to a 2% bar whatever short-term rates actually were then, and the longer the window, the smaller the share of it they can move.
Which special data series does Fincanva keep?
| Series | What it measures | Where it is used |
|---|---|---|
| Risk-free rate | The 3-month US Treasury Bill secondary-market rate, published by FRED as DTB3 | The near-riskless baseline that risk-adjusted metrics subtract — see risk-free rate |
| Margin-loan rate | A short-term reference interest rate — today the same 3-month US Treasury Bill series as the risk-free rate | The base rate that the borrowing markups sit on top of, and the rate idle cash is credited from — see interest-rate markups |
| Inflation | A long-history monthly US inflation series | Inflation-adjusted figures and inflation-based comparisons |
| Shiller PE | The cyclically adjusted price-to-earnings ratio of the US market: price divided by the average of the last ten years of inflation-adjusted earnings | A market-wide valuation reference |
| Buffett Indicator | Total US market capitalisation divided by US GDP | A market-wide valuation reference |
| Buffett Deviation | A proprietary Fincanva variant of the Buffett Indicator | A market-wide valuation reference |
The Shiller PE is taken exactly as its public reference series publishes it, so it matches the widely quoted figure; the Buffett Indicator is computed inside Fincanva as the standard ratio, from market-capitalisation and GDP data; the Buffett Deviation's construction is proprietary and not published.
What happens when a date falls outside a series?
A date outside a series' coverage — before its first data point or after its last — falls back to a fixed assumed value rather than failing the run. The rule is the same for the three rate and inflation series; what differs is the value and how it enters your results.
| Series | What stands in outside the series |
|---|---|
| Risk-free rate | A flat 2% a year for each month of the backtest window outside the series, blended into the window's period average — see risk-free rate |
| Margin-loan rate | A fixed assumed rate on each date outside the series. Its value is not published. |
| Inflation | A fixed assumed rate on each month outside the series. Its value is not published. |
The fallback covers dates outside a series, not a missing series. These series are loaded before any backtest runs, and if one of them could not be loaded no backtest would run at all.
The start-of-series case applies to dates before a series begins, as when a simulation start year is earlier than the series' first date. The end-of-series case applies when a series ends before your backtest window does, because it is published less often or later than prices; the dates past its end use the fallback.
Used in 8 pages
UPDATED 2026-10-07