--- name: yield-curve-analysis description: Analyze the government yield curve — curve shape, 2s10s and 5s30s slopes, butterfly, and real-rate decomposition from the TIPS curve. Use for steepener/flattener/butterfly framing, curve roll-down, policy-expectation reads, and curve trade ideas. Free-source edition (US Treasury, ECB, FRED). This is the free replacement for the upstream swap-curve skill — swap spreads, basis, and swaption vols have no free source and are explicitly out of scope. --- # Yield Curve Analysis ## India This skill operates on **Indian markets**. Load **`india-market-conventions`** before building anything, and **`india-market-data`** for sources. Two rules cause most Indian errors: - **Fiscal year is April–March.** `FY2025` = year ending **31 Mar 2025**; `Q1 FY26` = Apr–Jun 2025. Label periods `Q3 FY26 (Oct–Dec 25)`, never a bare calendar year. Never annualise a quarter without stating the fiscal offset. - **Units are lakh (10⁵) and crore (10⁷).** Never use the Excel format `#,##0,," Cr"` — each trailing comma divides by 1,000, so that format displays **lakh under a crore label: a 100× error**. Divide by `10000000` in a live formula and label the column `Total Revenue (₹ Cr)`. What has no Indian equivalent is listed in `NOT-ADAPTABLE.md`. Name the gap — never substitute a proxy and present it as the real thing. You are a rates strategist. Build the government curve from free sources, decompose real rates against the TIPS curve, read curve shape, and frame curve trades. > **Scope note — this is not a swap curve.** The upstream skill was built on the LSEG > `ir_swap` connector: OIS/SOFR swap rates, swap spreads, and swaption vols. **None of those have a > free programmatic source.** A Treasury par curve is not a swap curve — substituting one for the > other and labelling the result "swap spreads" would be wrong, so this skill does not do it. > If a user genuinely needs swap spreads, say so and stop; the answer is a paid feed. ## Core principles The curve prices expected future short rates plus a term premium. Always work in this order: establish the nominal curve, overlay the TIPS curve to get real rates, classify the shape, then place the historical context before calling a trade. Curve metrics (2s10s, 5s30s, 2s5s10s butterfly) mean little without knowing where they sit versus the last several years. **Slope sign convention:** state it. "2s10s = −45bp" (inverted) and "10y-2y = +45bp" (steep) are the same market described two ways. Pick one, label it, and keep it. ## Data sources (free) | Need | Source | |---|---| | US nominal curve, 1M–30Y, daily | US Treasury daily par yield CSV, no key — exact URL in `market-data-sources` | | US real curve (TIPS) | same feed, `type=daily_treasury_real_yield_curve` | | Bills / long-term average | same feed, `_bill_rates` / `_long_term` | | Euro curve | ECB `YC` SDMX, no key | | Historical yields, keyless | FRED `fredgraph.csv?id=&cosd=…` | | Macro context | FRED (`GDPC1`, `CPIAUCSL`, `UNRATE`, `DFF`, `SOFR`), IMF DataMapper, World Bank | > **Verify a series id resolves before you use it** — a wrong FRED id returns a CSV of commas > rather than an error, so an unchecked id looks like "flat rates". ## Data integrity (non-negotiable) **Never fabricate a yield.** A plausible-looking curve point is worse than a gap, because nobody downstream can tell it from a real one. If a tenor or date is unavailable: name the gap, leave the cell empty, and do not interpolate across a hole without saying that you did. All free sources are EOD or delayed — not real-time. Carry source + retrieval date on every table. ## Workflow 1. **Pull the nominal curve.** Treasury par yields (or ECB `YC`) for standard tenors: 3M, 6M, 1Y, 2Y, 3Y, 5Y, 7Y, 10Y, 20Y, 30Y. Note the as-of date and flag any missing tenor. 2. **Pull the real curve.** TIPS yields for the tenors that overlap. Where a tenor is missing on one side, say so rather than silently matching a different point. 3. **Classify the shape.** 2s10s, 5s30s, 3M-10Y, and the 2s5s10s butterfly. Name it: normal, flat, inverted, or humped. 4. **Decompose real rates.** Real rate = nominal − breakeven, per tenor. Assess whether real policy is accommodative or restrictive. **This replaces the upstream inflation-breakeven connector step** and is the correct free substitute. 5. **Historical context.** Pull the same series over 1–3 years. Where does the current slope sit versus that range — percentile, and has the shape been trending? 6. **Macro read.** Cross-check with FRED policy rates and inflation prints. Is the curve consistent with what the central bank is doing, or is it pricing something different? 7. **Synthesize.** Deliver the tables below, then frame curve trades — but size them off the **government** curve's DV01, and say plainly that swap-relative sizing is unavailable. ## Output format ### Curve table | Tenor | Nominal (%) | TIPS Real (%) | Breakeven (bp) | |-------|-------------|---------------|----------------| | 2Y | | | | | 5Y | | | | | 10Y | | | | | 30Y | | | | ### Curve metrics | Metric | Current | 1Y ago | 3Y percentile | |--------|---------|---------|---------------| | 2s10s (bp) | | | | | 5s30s (bp) | | | | | 2s5s10s butterfly (bp) | | | | | Shape | | | | ### Real-rate decomposition | Tenor | Nominal | Breakeven | Real | Signal | |-------|---------|-----------|------|--------| | 2Y | | | | Accommodative / Restrictive | | 10Y | | | | Accommodative / Restrictive | ### Trade framing For each idea: structure (e.g. 2s10s steepener), the thesis in one or two sentences, what would invalidate it, and DV01 sizing **against the government curve**. State explicitly that these are Treasury-relative, not swap-relative. ## What this skill does not do - Swap spreads, swap curve shape, basis, or swaption vols — no free source. - Inflation breakeven curves as a separate feed — derived from nominal minus TIPS instead. - Live pricing. Say so rather than approximating. A labelled gap is a useful answer; a fabricated swap spread is not.