RaoscaffResearch
Prediction Series · Lock · Issue P-03
Prediction Series · P-03

India's June inflation, locked before the print.

Headline CPI for June 2026, locked at 4.6% YoY with a full probability distribution — before MoSPI publishes (~July 13). Our edge: the market is under-weighting the full-month pass-through of May's fuel hikes, the June LPG hike, and a tomato spike, stacked on a deeply favourable year-ago base.

Type · Prediction Lock · single-print numerical Locked · 2026-06-20 · before the MoSPI release Resolves · ~2026-07-13 · MoSPI June-2026 CPI, base 2024=100 Scored · headline-bucket Brier + point error vs the first-print provisional figure
Headline CPI · June 2026 · our locked point
4.6%
YoY · first-print provisional · base 2024=100

Above the ~4.5% Street consensus (ICRA, India Ratings). The distribution skews up — P(4.6–4.8%)=28% is the mode. CFPI locked at 5.6%. We lock the literal MoSPI first-print figure and will score it honestly.

— 1 · The Locked Forecast

Headline 4.6% YoY — with the full distribution, not just a point.

We lock the headline at 4.6% YoY and CFPI (food) at 5.6%, with the full probability distribution below. The mode sits in the 4.6–4.8% bucket — above the ~4.5% Street consensus — reflecting a rising five-month trend meeting a low base.

Headline CPI June-2026 · locked probability distribution (first-print, base 2024=100)
Outcome bucketLocked probability
< 4.2%8%
4.2 – 4.4%14%
4.4 – 4.6%26%
4.6 – 4.8%  (modal)28%
4.8 – 5.0%16%
≥ 5.0%8%
— 2 · The Data

Five months rising, a low base, and a full-month fuel pass-through.

May 2026 printed 3.93% — the fifth straight monthly rise (Jan 2.75 → May 3.93). The June print absorbs three upside forces the May index only partly captured: the cumulative May 15–26 OMC fuel hikes (~Rs7.50/L, fully effective for all of June), the June 7 LPG cylinder hike (Rs29, ~+6.3 bps), and an accelerating vegetable spike (wholesale tomato +54.7% MoM as of June 10, AgMarkNet). All of this lands on June 2025's deeply favourable low base (CPI then 2.10%, food in deflation).

Sources: MoSPI/PIB (May 2026, base 2024=100); AgMarkNet via CommodityOnline/FreshPlaza (June 2026 mandi prices); DiscoveryAlert / IndMoney / Business-Standard (May fuel hikes); HDFC Bank (LPG impact).

— 3 · The Variant View

Above the ~4.5% consensus — by design.

Consensus (ICRA, India Ratings) sees ~4.5%. Our +0.1 comes from sizing the full-month fuel pass-through and the tomato spike precisely, on a low base — the consensus is reading the headline narrative, not the base-effect collision. We are confident in the call and explicit that it is a lock placed before the print; it will be scored against the published number, not a vibe.

— 4 · Methodology

Single-print numerical: component build to the headline.

The P-01 format: a component build (food / fuel / core) to a single headline number, with a probability distribution for calibration. Core is modelled near ~3.7–3.9% (sticky, modest services pass-through). We lock the MoSPI first-print provisional figure on the 2024=100 base — revisions are explicitly excluded.

— 5 · Pre-Committed Post-Mortem

How we'd be wrong — named in advance.

Below 4.4% if the back-cast June-2025 base is less favourable than estimated, or vegetable prices reverse on an early monsoon recovery. Above 5.0% if fuel pass-through is sharper or a second food shock lands late-June. A bucket miss if the first-print methodology surprises. We pre-commit these branches so resolution cannot be spun.

This prediction is locked in writing — and will be scored against the published number.

RAOSCAFF locks P-03 on 2026-06-20, before MoSPI publishes the June-2026 CPI. The commit timestamp is the cryptographic lock. We will score the locked distribution by Brier against the first-print provisional headline figure (base 2024=100) — not against a revision, not against a directional vibe.

Locked
2026-06-20 (commit timestamp on origin/main)
Resolves
~2026-07-13 — MoSPI/NSO June-2026 CPI first-print release (July 12 is a Sunday; resolution = the actual first-print date)
Source
MoSPI / NSO, base 2024=100, All-India Combined headline YoY (one decimal)
Scored by
headline-bucket Brier + point error; CFPI scored separately

If the print lands outside our distribution's mass, we will say so plainly — the lock is what makes the score checkable.