Fitted Instantaneous Forward Rate 1 Year Hence (THREEFF1) | FRED | St. Louis Fed
- Rows
- 9,519
- Columns
- 2
- Data range
- 1990-01-02 to 2026-06-26
Graph and download economic data for Fitted Instantaneous Forward Rate 1 Year Hence (THREEFF1) from 1990-01-02 to 2026-06-26 about 1-year, rate, and USA.
AI analysis
Analysis: Fitted Instantaneous Forward Rate 1 Year Hence (THREEFF1)
1. Dataset Overview and Value
This dataset, sourced directly from the St. Louis Fed's FRED API (fredgraph.csv?id=THREEFF1), represents a daily fitted instantaneous forward rate one year hence — essentially a model-derived estimate of what short-term interest rates are expected to be one year from each observation date, based on the Treasury yield curve. As a Federal Reserve publication, it carries strong methodological credibility and is refreshed on an ongoing basis (coverage extends through mid-2026, suggesting this includes forward-looking/projected periods or a very current pull). This makes it valuable for correlation studies involving monetary policy expectations, yield curve dynamics, and forward-looking rate sentiment — particularly useful as a leading indicator against realized economic outcomes like inflation, employment, or equity market performance.
2. Data Quality Observations
The dataset is nearly complete but not pristine. Of 9,519 total rows, there are 412 null cells (4.3%), all concentrated in the single value column, THREEFF1 — the observationdate join key has zero nulls and full distinct-value integrity (9,519 distinct dates matching row count, confirming no duplicate dates at the key level). A 4.3% null rate in the primary metric is moderate and likely reflects non-trading days, holidays, or periods where the yield-curve model couldn't converge (common in Fed fitted-rate series during market stress or data gaps). Duplicate-row recomputation is still pending (Phase D), so full deduplication confidence isn't yet established, but the distinct-date count is a good proxy for row uniqueness. Analysts should filter or interpolate the 412 nulls before running time-series or correlation models, since gaps could distort lag/lead relationships.
3. Key Column Distributions
The THREEFF1 column shows a mean of 3.28% and median of 3.44% — the mean sitting below the median suggests a mild left-leaning pull, though the skewness value of 0.206 indicates only slight positive skew overall, meaning the distribution is roughly symmetric with a modest right tail. The range is notable: from -0.0332% (reflecting the near-zero/negative rate environment of 2020-2021) to a high of 9.0519% (likely capturing early-1990s or inflation-shock-era rate expectations). The interquartile range (Q1=0.972%, Q3=5.004%) spans about 4 percentage points, reflecting the wide historical variation in rate regimes across 36+ years. With 8,372 distinct values out of 9,519 rows, the series is highly granular and continuous rather than discretized — ideal for fine-grained trend and volatility analysis.
4. Recommended Join Key
observationdate is the clear and only viable join key — it's complete (zero nulls), fully unique (distinct count equals row count), and spans a long, consistent daily frequency from 1990 to 2026. This makes it straightforward to align with other FRED time series or macroeconomic datasets using standard date-based joins (with attention to weekday/business-day alignment, since this appears to be a business-day series).
5. Recommended Pairings for Correlation Analysis
Strong candidates for correlation studies include: (a) other FRED interest rate series such as the Fed Funds Rate (FEDFUNDS), 10-Year Treasury Yield (DGS10), or the broader yield curve spread (T10Y2Y) to study term-structure relationships; (b) inflation indicators like CPI or PCE to test whether forward rate expectations predict realized inflation; (c) equity market indices (S&P 500) or credit spreads to examine risk-sentiment linkages; and (d) employment data (unemployment rate, nonfarm payrolls) to explore how forward rate expectations respond to labor market conditions. Given its forward-looking nature, this series pairs especially well with any dataset where testing "expectation vs. realization" lag structures is the analytical goal.
Columns
- observation_date (date)
- THREEFF1 (decimal)