For May 2026 (provisional), Destatis reports German inflation at 2.6% and at 2.7% — same month, same office, two index methods. This brief decomposes why two official German inflation numbers exist at all.
Both numbers are Destatis's, for the same month. The VPI (2.6%) includes imputed rent for owner-occupied housing and the broadcasting fee; the HVPI / HICP (2.7%) excludes both. The 0.1-pp gap is RAOSCAFF arithmetic (2.7 − 2.6). Both are correct under their own definitions — and the wedge is variable: it was 0 pp a month earlier.
For May 2026 (provisional), Destatis publishes two consumer-price numbers for Germany and releases them together. The national VPI read +2.6% YoY; the harmonised HVPI — the basis Eurostat aggregates as Germany's HICP — read +2.7% YoY. Same month, same statistical office, a 0.1-percentage-point wedge (RAOSCAFF arithmetic: 2.7 − 2.6).
The honest caveat first: this month's spread is small. But it is not noise and it is not an error. The two indices cover different baskets — the VPI carries owner-occupied imputed rent and the broadcasting fee; the HVPI leaves both out — so a 0.1-pp gap on identical source prices is the arithmetic consequence of two scopes, not a mistake in either index.
This brief decomposes why two official German inflation numbers can differ at all, and when the gap widens. It ranks neither index and recommends no security. Both numbers are correct under their own definitions; the question is which basket each one measures.
Destatis builds both indices from the same source prices, then sorts them into two baskets. The VPI keeps owner-occupied imputed rent and the broadcasting fee; the HVPI excludes them. The observed gap is variable — 0.0 pp in April 2026 (final), 0.1 pp in May 2026 (provisional).
Three axes, no more. A two-method gap; a structural driver that creates it; and a variable amplifier that opens and closes it month to month.
The core (the gap). For May 2026 (provisional), VPI +2.6% versus HVPI +2.7% = a 0.1-pp wedge (RAOSCAFF arithmetic). Eurostat independently carries the 2.7% line as Germany's HICP into the euro-area flash (3.2%, May) — not the 2.6% line.
The why (the driver). The baskets differ chiefly on owner-occupied housing: the VPI includes imputed rent via the net-cold-rent index (Mietäquivalenzansatz); the HVPI does not — Destatis states it is “bisher nicht berücksichtigt.” A secondary scope gap: the German HVPI has excluded the broadcasting fee (Rundfunkbeitrag) since 2020. Both items sit in the rent-heavy housing/services category — the seam where the gap is created.
The when (the amplifier). Because the baskets differ on a slow-moving housing/services component, the monthly gap reflects where inflation is concentrated. In April 2026 (final) both indices read +2.9% — a 0-pp wedge (RAOSCAFF arithmetic: 2.9 − 2.9) — with energy at +10.1%; in May 2026 (provisional) energy stood at +6.6% (a ~3.5-pp energy change, RAOSCAFF arithmetic: 10.1 − 6.6) and the wedge was 0.1 pp (RAOSCAFF arithmetic: 2.7 − 2.6). The structure is fixed; the monthly size is not.
| Axis | What it measures | The figures |
|---|---|---|
| The core · the gap | VPI vs HVPI, same month, same office | 0.1 pp — May 2026 prov. (2.7 − 2.6); the 2.7 line feeds the euro-area flash of 3.2% |
| The why · the driver | owner-occupied imputed rent (+ broadcasting fee since 2020) | in the VPI basket, out of the HVPI basket — a housing/services seam |
| The when · the amplifier | goods/energy moving against housing/services | wedge 0 pp in Apr (energy +10.1%) → 0.1 pp in May (energy +6.6%); ~3.5-pp energy swing |
The same May release shows energy +6.6%, services +3.1%, food +0.4%, and core (ex food & energy) +2.5% YoY. The wedge is the consequence of two correct baskets, sized by the month's inflation mix — not a data error in either index. The 0.1-pp and 0-pp wedges and the ~3.5-pp energy swing are RAOSCAFF arithmetic; May 2026 figures sourced to Destatis press release No. 182 (29 May 2026); April 2026 HICP 2.9% independently confirmed by Eurostat (20 May 2026 release).
If you read one thing: Germany's “inflation rate” is two baskets, not one rate — read which basket before reading the number.
“German inflation in May 2026” is two facts, not one: 2.7% is the EU-comparable HVPI that travels into the euro-area number (3.2%, May flash); 2.6% is the national VPI that carries owner-occupied housing. The observed wedge: 0 pp in April 2026 (final) and 0.1 pp in May 2026 (provisional) — variable, not a fixed offset. Read them as two baskets, not one rate. Mirror Brief DE-01 ranks neither index, names no policy actor, and recommends no security; both numbers are correct under their own definitions, and the wedge is the arithmetic of two scopes — not an error in either. The 0.1-pp and 0-pp wedges and the ~3.5-pp energy swing are RAOSCAFF arithmetic from Destatis's published figures.
Mirror format — RAOSCAFF anchors on the publisher's own filed figures and decomposes the number they print. DE-01 is a two-method brief: Destatis press release No. 182 (29 May 2026) supplies both the national VPI (+2.6%) and the harmonised HVPI (+2.7%) for May 2026 (provisional), plus the energy, services, food, and core components; Destatis's HVPI methodology page supplies the scope difference (owner-occupied imputed rent in the VPI, out of the HVPI; the broadcasting fee out of the German HVPI since 2020). Eurostat's 20 May 2026 release confirms Germany HICP +2.9% for April 2026 (final) — this is the primary source for the April figure; Eurostat's 2 June 2026 flash independently confirms the May HVPI 2.7% line. FACTS.md is the source-of-truth file; every figure here traces to it. The derived figures — the 0.1-pp and 0-pp wedges and the ~3.5-pp energy swing — are RAOSCAFF arithmetic from the published percentages, and are labelled as derived.
Source standards and limitations: every figure traces to a named, dated official document. May 2026 figures are provisional (final due 12 June 2026); April 2026 figures are final; each is labelled accordingly and is YoY against the same month a year earlier. The exact per-mille imputed-rent weight in the VPI Wägungsschema lives in a downloadable GENESIS table, not a primary HTML page, and is therefore stated qualitatively rather than as a fetched number — never invented. Editorial position: predict-not-recommend; defamation-disciplined — the brief critiques the two index definitions only, names neither Destatis nor Eurostat as wrong, slow, or inferior, and neither index is the “better” gauge; politically neutral — no view on monetary or fiscal policy, no party, government, or central bank as an actor, and no causal market claim; metric-identity held — the national VPI and the harmonised HVPI/HICP measure different baskets and are never conflated; magnitude-not-direction — this is a measurement decomposition, with no security, buy, sell, hold, or valuation view.
All six sources are official Destatis / Eurostat publications, fetched live 2026-06-04; the Destatis press release No. 182 and HVPI methodology page were read directly as the primary anchors. Full citations and the verified figures behind every number in this brief — including the arithmetic register — are listed in the companion FACTS.md, § H.