Product Specification#
Status: Draft, 2026-08-29 (sources assembled); citations re-verified against the primary documents 2026-08-30.
Scope note. This is a standardized composite specification assembled for reference liability cash-flow modeling of a German indexgebundene Rentenversicherung — the Indexpolice: a deferred private annuity of Schicht 3 whose accumulated capital sits in the insurer’s Sicherungsvermögen (the ring-fenced general-account cover pool) under a guarantee, and whose annually declared Überschuss (surplus) is not credited as interest but spent as an option budget buying a one-year participation in a share index. It does not describe any single insurer’s contract.
Facts carrying a source tag — [S#] (primary product documents: AVB, Produktinformationsblatt,
Basisinformationsblatt (PRIIP-KID), Verbraucherinformation, Standmitteilung) and [R#]
(product-specific regulatory and actuarial references), both numbered per _research/indexpolice.md
and resolved in sources.md (same directory; numbering frozen, never renumbered), and [REG-R#] (the
cross-product reference library references/regulatory-and-actuarial-references.md, whose own
R-numbering is distinct and also frozen) — name the instrument the claim should be checked against.
Values marked std are standardizations introduced for the reference implementation; each
std table row carries a numbered footnote giving the rationale and, where one could be assessed,
the plausible market band. Claims that no retrieved document corroborates are flagged unverified.
Read this before reading any number below. delib was drafted with direct HTTP egress
blocked and, for this product, with the session’s WebSearch budget already exhausted: no AVB, no
Basisinformationsblatt, no statutory text and no index rulebook was opened, so the first draft
rested on the authoring model’s own knowledge of German insurance law and practice, disciplined by
std and unverified tags. That policy has since been lifted and this product’s citations
re-verified against the primary documents. Of the 38 entries in sources.md, 32 now read
Retrieved: yes, one reads partly, and five read no: a Produktinformationsblatt class
that does not exist for a Schicht 3 contract [S3], the annual parameter notice insurers send to
policyholders and never publish [S5], Finanztest behind its paywall [S13], the comparison portals
behind a bot wall [S15], and the rating houses behind their subscription tools R21. A
Retrieved: yes entry means the document was opened and the passage it rests on read; anything else
is a pointer rather than a certificate, and each entry says which it is.
The re-verification moved this document. Two carrier Bedingungswerke were read in full [S2] [S7], so the Indexbeteiligung clause set below is quoted rather than reconstructed; the death-benefit rule was cited to the wrong carrier series and is corrected (footnote 23); and rows that read “not established” now carry carrier figures. But the commercial envelope stays std throughout — no entry-age band and no minimum premium is published by any of the three carriers, so the model points remain construction. What was never in dispute is the mechanics — the financing identity between declared surplus and option budget, the sum-of-capped-monthly-returns payoff with uncapped negative months, the annual floor at zero, the permanent lock-in and the annual election — and this document puts its weight there. Three carrier products are named — Allianz Zukunftsrente IndexSelect [S2], R+V-PrivatRente IndexInvest [S7] and Stuttgarter index-safe [S8] — all three now established from the carriers’ own documents, and no fourth is added.
Product overview and market role#
An Indexpolice is an aufgeschobene Rentenversicherung: a deferred annuity on a single life, with an Aufschubphase running from inception to Rentenbeginn and a Rentenphase paying a lifelong Leibrente thereafter [S1] [S9]. Premium, reserve, death benefit before Rentenbeginn, Rückkaufswert, Beitragsfreistellung, Rentenfaktor, Kapitalwahlrecht and Rentengarantiezeit are the chassis of a klassische Rentenversicherung and are documented for that product. The delta is one clause set: how the annually declared Überschuss is applied. Three facts decide everything else, and the first two are the ones secondary descriptions usually get wrong.
1. The capital is in the general account, not in a fund. The accumulated Deckungskapital sits in the Sicherungsvermögen REG-R7, exactly as a classic annuity’s does: no Anlagestock, no unit account, no policy-level asset allocation. The policyholder owns a claim on the insurer measured in euros, not a number of units; the reserve rolls forward by a recursion, not by a unit price; and the Rückkaufswert is a reserve, not a Zeitwert of units R2 REG-R28. What the index does is define a payoff, not an investment — the policyholder is never invested in the index at any moment, and the insurer buys the option package that hedges the payoff it has itself written R9 REG-R7.
A terminological trap follows. In regulatory and accounting vocabulary, “Lebensversicherungen, bei denen das Anlagerisiko vom Versicherungsnehmer getragen wird” — the balance-sheet class containing fondsgebundene and indexgebundene life insurance — means contracts where the policyholder bears the investment risk. An Indexpolice of the kind described here does not belong there, and is booked and reserved as a conventional profit-participating contract, sitting in the Solvency II line insurance with profit participation R15, unverified as to the line-of-business numbering, which no retrieved document states. The reading itself is no longer a reading. § 125 Abs. 5 VAG requires an Anlagestock only “soweit Lebensversicherungsverträge Versicherungsleistungen … direkt an einen Aktienindex oder andere Bezugswerte binden”, and § 124 Abs. 2 VAG scopes the whole class by whether “das Anlagerisiko vom Versicherungsnehmer getragen wird” R15. A payoff financed out of declared surplus, floored at zero and payable in euros from the cover pool is neither. Both retrieved AVB say the capital is in the Sicherungsvermögen in terms — Allianz’s KID: “Die Kapitalanlage erfolgt während der gesamten Versicherungsdauer vollständig durch das Versicherungsunternehmen im Sicherungsvermögen” [S4]; R+V: “Ihr Policenwert ist Teil des Sicherungsvermögens der R+V Lebensversicherung AG” [S7]. The cross-product reference library still records the question as open REG-R7; this specification does not. delib therefore uses Indexpolice / Indexbeteiligung for the product and reserves indexgebunden for its regulatory sense.
2. The index participation is a form of Überschussverwendung, with no independent statutory footing. § 153 VVG gives the policyholder a right to participate in the surplus and in the Bewertungsreserven unless participation is excluded, and requires allocation by a verursachungsorientiertes Verfahren or another comparable appropriate method R1 REG-R24. What the policyholder is legally entitled to is a share of surplus; the AVB then say how that share is applied, and this product’s AVB say it is applied by buying a bounded index-linked payoff for one year. The Wahlrecht is therefore an Überschussverwendungswahlrecht, and the Indexbeteiligung stands or falls on the contract clause. That is the correct legal characterisation, and all three carrier documents now state it in their own words — Allianz: “Die Überschussbeteiligung eines Jahres kann zum einen für die Indexpartizipation … verwendet werden” [S4]; R+V finances the participation “mit den … jährlichen Überschussanteilen sowie mit der jeweiligen jährlichen Mindestbeteiligung an den Bewertungsreserven” [S7] § 3 Ziffer 9; Stuttgarter: “Es werden 100 % der laufenden Überschüsse für die Indexbeteiligung verwendet” [S11]. Both AVB add a component this specification did not carry: the budget is the declared surplus plus the year’s minimum share of the Bewertungsreserven, and Allianz’s is net of Verwaltungskosten. The subsection numbering is confirmed against the canonical text of § 153 VVG, and one correction follows from it: Abs. 4 makes the Bewertungsreserven half-share of Abs. 3 Satz 2 fall due at the end of the Ansparphase, not at termination, for a Rentenversicherung R1.
3. The option budget is the declared surplus, and nothing more. The insurer earns a return on the Sicherungsvermögen; the MindZV forces a minimum share of each result source to the policyholders — 90 % of the Kapitalanlageergebnis after the Rechnungszinsen are deducted, 90 % of the Risikoergebnis, 50 % of the übrige Ergebnis R8 REG-R18; the insurer declares an Überschussanteilsatz out of that; and a contract in the index arm has that declared amount spent on options instead of credited as interest. An Indexpolice therefore does not have a larger risk budget than a Klassik contract of the same vintage under the statutory minimum — the MindZV floor is the same instrument-wide floor. That is the most under-appreciated fact about the product and it belongs on the first page. But the declared rate is a different matter, and the market data contradict the stronger form of the claim: Assekurata’s 2026 survey puts the average declared laufender Überschusszins on Indexpolicen at 3,07 % against 2,62 % for classic private annuities and 2,65 % for Neue Klassik R20. Index tariffs sit in their own Bestandsgruppen and Überschussverbände ([S7] § 13 Ziffer 3), and the declaration for them is currently higher. The identity that holds without qualification is the narrower one: whatever is declared to this contract is spent once — on options or as interest, never both.
Why the product exists. The Höchstrechnungszins fell from 4,00 % (1994–2000) through 3,25 %, 2,75 %, 2,25 %, 1,75 %, 1,25 % and 0,90 % to 0,25 % for 2022–2024, and rose to 1,00 % from 1 January 2025 — the first increase in about thirty years, made by the Sechste Verordnung zur Änderung von Verordnungen nach dem Versicherungsaufsichtsgesetz of 19 July 2024, BGBl. 2024 I Nr. 250 R7 R18 REG-R14 REG-R15. At 0,25 % the guaranteed component of a conventional annuity’s return is negligible and the discretionary component is the whole story. An Indexpolice takes that same discretionary component and, instead of crediting it as a modest certain amount, converts it into a bounded lottery on an index. The product is a direct commercial response to a near-zero guaranteed rate, and the 2025 rise makes the sichere Verzinsung arm relatively more attractive again; whether observed elections have shifted is unverified R7. Qualitatively, and not in doubt: the family emerged in the second half of the 2000s, grew through the low-interest decade as the guaranteed component shrank towards nothing R7, became a standard offering across the large and mid-sized carriers, and was one of the main vehicles of the Neue Klassik generation [S6] REG-R53.
Market size. There is no industry figure to quote: no published statistic isolates the German index-participation segment. The GDV’s Die deutsche Lebensversicherung in Zahlen 2024 splits the in-force book at 31.12.2023 into Renten- und Pensionsversicherungen 61,8 %, Kapitalversicherungen (klassisch) 15,7 %, Invaliditätsversicherungen 9,2 % and Risikoversicherungen 6,5 % — with no index line, and the word “Index” occurring nowhere in the publication except its own table-of-contents index R19. These contracts are counted within conventional annuity business because that is what they are R15. The only counts available are a single carrier’s own, and they are press statements rather than statistics: Allianz reported 400.000 IndexSelect contracts in October 2016 [S12] and “über 500.000” in May 2019 [S16]. The frame it sits in: German life premium income (life insurers, Pensionskassen and Pensionsfonds, GDV basis) was +2,8 % to 94,6 Mrd. € in 2024 — laufende Beiträge 66,3 Mrd. €, roughly flat, Einmalbeitragsgeschäft about +10 % to 28 Mrd. € — with the contract count −1,4 % to 80,3 Mio.; on the BaFin basis, life-segment verdiente Bruttobeiträge were 90,4 Mrd. € REG-R53. The two measure different populations on different bases and are never combined. The relevant market rate is the declared one. Assekurata’s 24th Marktstudie (March 2026) gives, for 2026, 2,62 % laufende Verzinsung on classic private annuities (3,23 % including Schlussüberschüsse), 2,65 % on Neue Klassik (3,32 %), 3,07 % on Indexpolicen — “etwa dem Vorjahresniveau” — and 2,49 % on guaranteed fund policies R20; the 2026 averages recorded from other surveys in the sibling delib files (2,6–2,7 %, 2,87 %, 2,54 %) measure different panels and are not combined with these REG-R53. One carrier publishes its own figure directly: Stuttgarter’s sichere Verzinsung for all Indexstichtage from 1.2.2026 to 31.1.2027 is 2,16 % [S8]. That declared rate is the option budget R8.
The same index module is written on four chassis — Schicht 3 private annuity (this document), Basisrente, Riester and Direktversicherung in bAV (outside delib). The wrapper changes the guarantee requirement R12 REG-R43, the tax treatment R13 R14 REG-R41 REG-R45 and the accessibility of the capital — and not the index mechanics.
Representative specification#
The representative design is a composite and remains one: it is not a transcription of either carrier wording, because the two retrieved wordings differ from each other on the payoff itself. The GDV publishes Musterbedingungen for the deferred-annuity chassis but no model wording for an index-participation module — its life catalogue runs to eleven model conditions and nine Muster-Standmitteilungen, and not one of them is an index module [S1]. That is the structural reason the clause set varies more across insurers here than for any other delib product, and the variation is now demonstrated rather than asserted: Allianz caps each month and multiplies the capped sum by a Partizipationssatz [S2]; R+V applies a Beteiligungsquote to the point-to-point year return of a house index and has no cap at all [S7]; Stuttgarter is a quota design too [S8]. Every representative choice below is argued against the plausible band in Variations across insurers, and the choices that a retrieved AVB now settles say so in their Basis column.
Product identity and issue rules#
Parameter |
Representative value |
Basis |
|---|---|---|
Design type |
Single-life aufgeschobene Rentenversicherung, Schicht 3, profit-participating, general account; Neue Klassik guarantee architecture |
[S1] [S6] R15 |
Where the capital sits |
Sicherungsvermögen; no Anlagestock, no units, no policy-level asset allocation |
|
Überschussverwendung |
Indexbeteiligung or sichere Verzinsung, at the policyholder’s annual election |
|
Lives basis |
Single life; no joint-life form in this family. Sex may not be a rating factor — unisex since 21 December 2012 |
[S1]; REG-R34 |
Eintrittsalter |
25 to 55 |
std (2) |
Rentenbeginn age |
62 to 70; 67 representative |
std (2) |
Aufschubdauer |
12 to 40 years; 12 is the tax minimum |
|
Age basis |
Age last birthday at inception |
std (3) |
Underwriting |
Light or absent — a short declaration or none, the sum at risk before Rentenbeginn being small |
unverified; std (4) |
Garantieniveau (Beitragsgarantie) |
90 % of the Beitragssumme, due at Rentenbeginn |
std (5) |
Guaranteed rate (Rechnungszins) |
1,00 % — the Höchstrechnungszins for 2025 and 2026 |
|
Anchor model cell |
Eintrittsalter 40, Rentenbeginn 67, Jahresbeitrag 2 400,00 €, annual mode, Cap design, 90 % Beitragsgarantie, full index election in every year |
std (6) |
Footnotes to std rows:
No index AVB was obtained [S2] [S7] [S8], and the GDV publishes no model clause for the module [S1]. The clause set in Contractual mechanics below is reconstructed from the mechanics the research file establishes, is attributed to no carrier, and is labelled a composite wherever used.
No Produktinformationsblatt was located [S3] [S11] [S15], so no entry-age band, minimum premium, term band or maximum sum was established for any carrier. The envelope is an uncontroversial std construction: 25–55 is the mid-career segment this product is sold into, 67 is the German statutory retirement age, and the 12-year minimum is the tax threshold of § 20 Abs. 1 Nr. 6 EStG R14 REG-R45 rather than a product limit.
The German market’s own Eintrittsalter convention is frequently the calendar year of inception less the year of birth, stepping on 1 January rather than on the birthday. delib runs age last birthday across all ten products, the registry fixing one age basis for the library; with the attained age stepping on the policy anniversary the two differ by at most one year, and mortality here is a timing rather than an amount assumption.
No underwriting rule of any carrier was established; the reasoning is structural. The Aufschubphase death benefit is a return of capital rather than a sum at risk, so the Risikoüberschuss is small and § 161 VVG (Selbsttötung, three years) is close to inoperative R6 REG-R26.
Carriers offer a choice of Garantieniveau — because every euro of guarantee not promised is a euro that can back risk assets, and therefore a larger option budget. Three levels are now observed in retrieved documents: 90 % for Allianz Zukunftsrente IndexSelect (“Sie haben Anspruch darauf, mindestens 90 % Ihres Kapitals zurückzuerhalten”) and 80 % for IndexSelect Plus, whose Chancenturbo is precisely what the released ten points buy [S4]; 90 % in R+V’s AVB (“Zum vereinbarten Rentenbeginn entspricht der Policenwert mindestens 90 % der Summe der gezahlten Beiträge”) [S7] § 1 Ziffer 2; and 85 % for the Stuttgarter Basisrente variant [S11]. delib’s 90 % is the modal observed level. That the menu once ran down to 60 % is the research file’s recollection and stays unverified. The wrapper decides the floor: a Riester variant must guarantee that at least “die eingezahlten Altersvorsorgebeiträge” are available at the start of the payout phase under the AltZertG Beitragserhaltungszusage — up to 20 % of the Gesamtbeiträge disregarded where they buy biometric cover — and so has the smallest option budget of the four R12 REG-R43.
Eintrittsalter 40 with Rentenbeginn 67 gives a 27-year Aufschubdauer, long enough for the ratchet to compound visibly and short enough to print in one table. The 2 400,00 € Jahresbeitrag is the research file’s 200,00 € per month, taken on the annual mode so the anchor is free of the Ratenzahlungszuschlag and the loading is exercised by the fractionated model points instead.
The index participation module#
This is the product, and it is the one table in this document where the mechanics are firm and every level is std.
Parameter |
Representative value |
Basis |
|---|---|---|
Indexjahr |
Twelve months, aligned in the model to the policy year |
mechanic [S2] [S5]; alignment std (11) |
Observation |
The index level is read at thirteen Beobachtungstage — one at the start and one per month — and month |
mechanic firm; convention std (11) |
Payoff design |
Cap: each month’s return capped above at |
mechanic firm |
Monthly Cap |
3,00 % per month |
std (12) |
Partizipationsquote variant |
|
mechanic firm; level std (13) |
Partizipationsquote |
60 % on an equity price index; 100 % on a low-volatility house multi-asset index |
std (13) |
Base of the participation |
The accumulated capital at the start of the Indexjahr, before that year’s premium |
std (14) |
Höchststandsicherung |
Whatever is credited is permanently added to the guaranteed capital and enters the base of every later Indexjahr |
mechanic firm |
Declared surplus rate |
2,50 % per year of |
std (15) |
Wahlrecht |
Annual election between the two arms, exercisable without the insurer’s consent, without medical evidence and without charge; election as a fraction |
mechanic firm; |
Mindest-Cap |
None |
std (17) |
Minimum option budget |
None — if no surplus is declared the Indexbeteiligung buys nothing and the year credits zero whatever the index does |
|
Mid-year exit |
No index credit in the year of exit — death, surrender or annuitisation inside an Indexjahr forfeits that year’s payoff |
std (18) |
Index participation in the Rentenphase |
None — the Wahlrecht lapses at Rentenbeginn |
std (19) |
Underlying |
Parameterised by an explicit monthly-return path with a stated volatility, not by a named index |
std (20) |
Ersatzindex |
The insurer may substitute a comparable index on notice if the index ceases to be published, is materially restructured, or ceases to be available on terms on which the hedge can be bought |
mechanic firm; procedure std (21) |
The level read is a closing level, not an average: monthly movements are “die prozentuale Veränderung des Index zwischen 2 Bewertungsstichtagen, die wir Ihnen jährlich mitteilen” [S2] Ziffer 3.3 Absatz 2 a), and R+V’s Bewertungsstichtag is “der letzte Börsentag eines Versicherungsjahres in Frankfurt am Main” [S7] § 3 Ziffer 3. An averaging (Asian) reading would lower the effective volatility and buy a higher Cap out of the same budget; neither carrier averages, so delib’s closing-level convention is the market’s. The alignment is a genuine variation. For R+V the Indexjahr is the Versicherungsjahr [S7] § 3 Ziffer 3; for Allianz it need not be — Ziffer 3.5 contemplates that “der Beginn des →Indexjahres nicht mit dem Beginn eines →Versicherungsjahres übereinstimmt” — and Stuttgarter runs a common calendar window, its published quota applying to “alle Indexstichtage vom 1.2.2026 bis 31.1.2027” [S8]. delib’s alignment with the policy year is R+V’s rule and a std simplification against the other two.
Two cap and quota levels are now established, but not a market distribution. Allianz’s own worked example runs at a Cap of 3,2 % with a Partizipationssatz of 75,00 %, both “exemplarisch gewählt” [S2] [S5]; the 2018 litigation records the cap then in force as 3,3 % [S14]. What is still missing is a panel — a year’s levels across named carriers side by side, which only a rating house publishes R21 and none was reachable. The band quoted throughout — 1,5 % to 5,0 % per month, typically 2,5 % to 4,0 % — remains the research file’s assessment and is unverified; 3,00 % is its midpoint and sits just below Allianz’s own illustration. The Cap is not a free parameter, and the AVB says so: it is set annually “auf der Grundlage von Angeboten mehrerer Finanzinstitute” and depends on the surplus, the Bewertungsreserven Sockelbetrag and market factors “wie der Volatilität und der Dividendenrendite des jeweiligen Index” [S2] Ziffer 3.3 Absatz 2 b) — so there is exactly one Cap at which the capped-sum payoff costs the budget, and the technical notes publish that consistency check.
One quota is published and one illustrated. Stuttgarter’s current Partizipationsquote on its house multi-asset index is 70 %, with the Index-Turbo options at 120 % and 172 % [S8]; Allianz illustrates 75,00 % on the EURO STOXX 50 [S2]. delib’s 60 % on an equity price index is below both, and its 100 % on the house path is above the one published house-index figure; both remain std, and 50–80 % / 80–120 % remains the research file’s assessment, unverified as a band. The Cap design is the base because it is the design the product’s reputation and its criticism both rest on; the Quote is a switchable variant — and it is R+V’s and Stuttgarter’s actual design, not a hypothetical one.
Settled, and delib’s reading is the carriers’. “Bezugsgröße für die →Indexpartizipation ist der →Policenwert zu Beginn des →Indexjahres” [S2] Ziffer 3.3 Absatz 2 e), which expressly excludes that year’s premiums and Zuzahlungen; R+V’s Bezugsgröße is likewise “der Wert, der ab Beginn des Versicherungsjahres … das gesamte Versicherungsjahr vorhanden ist. Dabei werden weitere Beiträge und Zuzahlungen während des Versicherungsjahres nicht berücksichtigt” [S7] § 3 Ziffer 2. It is the whole capital at the year start, before that year’s premium — not a sub-account and not the accumulated Überschussguthaben. The alternative readings are withdrawn, and this is no longer a named model risk.
The declared rate is the option budget R8, and delib’s 2,50 % is now on the low side of the evidence rather than at its midpoint: Assekurata’s 2026 index-segment average is 3,07 % R20 and Stuttgarter’s own published sichere Verzinsung is 2,16 % [S8]. The value is a shipped input and is not changed in a provenance pass;
model.mdrecords the comparison. It is exogenous in the reference implementation: the feedback from the Garantieniveau through the asset mix to the declared rate is real, is the whole design logic of Neue Klassik, and is not modeled. Both AVB also add a component delib does not model — the year’s minimum share of the Bewertungsreserven is part of the budget alongside the declared surplus [S2] Ziffer 3.3 Absatz 1, [S7] § 3 Ziffer 9.A split election is permitted, in 25-percent steps: “Die Aufteilung kann in 25-Prozentschritten erfolgen, wobei die Summe 100 Prozent ergeben muss” [S2] Ziffer 3.1. delib’s continuous fraction
wis therefore a std relaxation of a discrete menu, with all-or-nothing the special casew ∈ {0, 1}. The notice period is 7 days before the Indexstichtag at both carriers ([S2] Ziffer 3.1, [S7] § 2 Ziffer 3) — and the far more consequential question is answered in delib’s favour: the Cap is announced before the election deadline, Allianz notifying the Caps and the Partizipationssatz “spätestens 3 Wochen vor dem Indexstichtag”.These are different promises: a Mindest-Cap bounds the Cap given a budget, a minimum budget bounds the budget — and a Mindest-Cap is worthless in a year in which no surplus is declared R1 R8. Neither appears in either retrieved AVB, and delib assumes neither. Both AVB instead carry the opposite provision: index participation is excluded for a year in which the Policenwert does not exceed the Deckungsrückstellung required for the guarantee ([S2] Ziffer 3.5, [S7] § 2 Ziffer 1), so the budget is not merely unguaranteed but contractually switched off when the guarantee binds.
Settled, and delib’s std is the carriers’. The participation is credited only “zu Beginn des folgenden →Indexjahres” ([S2] Ziffer 3.3 Absatz 1, [S7] § 3 Ziffer 5), R+V’s Bezugsgröße is by definition the value present for the whole Versicherungsjahr [S7] § 3 Ziffer 2, and on surrender Allianz adds only a pro-rata Schlussüberschussanteil and Sockelbetrag [S2] Ziffer 9.2 Absatz 4 — no pro-rata index credit and no refund of the unspent budget. The behavioural consequence stands: the product rewards surrendering just after an Indexjahr end and penalises surrendering just before one. A model stepping annually puts every exit at a year end and so implicitly assumes the favourable convention; the monthly grid dates the forfeiture instead, though the payoff is still not pro-rated and the surrender rate is still unconditional.
Confirmed for Allianz: the participation runs “vor Beginn der Rentenzahlung” only [S2] Ziffer 3.3, so the Wahlrecht lapses at Rentenbeginn and payout-phase surplus is applied to the annuity in payment. Whether any other carrier offers index participation in the payout phase is still unestablished.
Two German house multi-asset indices are now named in this file — the Solactive Multi Anlage Stabil Index (SOMAS), built for R+V by Solactive [S7], and the Stuttgarter M-A-X Multi-Asset Index, alongside a Stuttgarter Grüne Zukunft Index [S8] — and the equity underlyings are the EURO STOXX 50 and the S&P 500, the latter with a Währungsfaktor applied to the year return [S2]. No volatility target and no index-level fee is published for either house index, which is why the model still parameterises the underlying by an explicit table of monthly returns with a stated drift and volatility, shipping an equity case, a low-volatility house case and an all-zero case, and names no index in any shipped file.
Substitution requires no unabhängiger Treuhänder. Allianz may replace an index “mit Wirkung zu Beginn des nächsten →Indexjahres” on material changes it is not responsible for, and if it cannot replace one may exclude the participation for subsequent Indexjahre [S2] Ziffer 3.7; R+V replaces the index “zum nächsten Indexstichtag”, choosing one that “dem zu ersetzenden Index weitestgehend entspricht”, at no cost to the policyholder [S7] § 3 Ziffer 11. Neither grants a Sonderkündigungsrecht; R+V instead gives the policyholder a choice whether to continue with the new index. Both also carry a suspension clause delib does not model: where no suitable capital-market instrument can be bought, the participation is suspended and the budget goes to the Verzinsung arm [S7] § 3 Ziffer 10. The legal frame is under Contractual mechanics.
Benefit provisions#
Parameter |
Representative value |
Basis |
|---|---|---|
Benefit at Rentenbeginn |
The accumulated capital, floored at the greater of the Beitragsgarantie and the guaranteed capital including every locked-in credit |
|
Conversion |
Lifelong Leibrente at the greater of the guaranteed Rentenfaktor fixed at issue and the insurer’s current factor at Rentenbeginn |
chassis fact, two carrier documents in the sibling research; level std (22) |
Guaranteed Rentenfaktor |
25,00 € per month per 10 000 € of capital at Rentenbeginn 67 |
std (22) |
Kapitalwahlrecht |
Lump sum instead of the annuity, applied for at latest one month before Rentenbeginn; once exercised the Rentenbeginn may no longer be deferred |
[S7] § 1 Ziffer 8; the window is that carrier’s |
Death benefit in the Aufschubphase |
The accumulated capital excluding the running Indexjahr, floored at 50 % of the Beitragssumme |
[S7] § 1 Ziffer 5; floor std (23) |
Selbsttötung |
No liability on a death cover within three years of conclusion or reinstatement; the Rückkaufswert is then owed |
|
Schlussüberschussanteil / Bewertungsreserven |
Half of the Bewertungsreserven determined at the end of the Ansparphase — § 153 Abs. 4 VVG makes that the relevant date for a Rentenversicherung, not the end of the contract — subject to the Sicherungsbedarf restriction |
Taken over from the sibling klassische Rentenversicherung std — and now within 3 % of a published figure for an index tariff: Stuttgarter’s Muster-Produktinformationsblatt for BasisRente index-safe discloses a guaranteed Rentenfaktor of 25,74 € per 10.000 € on a 100 €-a-month, age-37-to-67 model case [S11]. The max-of-two rule is confirmed in a retrieved AVB rather than inherited: “Ergibt sich bei Rentenbeginn auf der Grundlage der Sterbetafel und des Rechnungszinses, die wir für den Neuzugang von vergleichbaren sofort beginnenden Rentenversicherungen verwenden, eine höhere Rente …, dann wird die höhere Rente garantiert” [S7] § 1 Ziffer 4 — and Allianz’s Ziffer 1 adds the Treuhänder leg for the case where no comparable annuity is on sale. R+V also discloses the bases of its guaranteed factor: a Rechnungszins of 0,1 % p. a. and “eine auf der DAV-Sterbetafel 2004 R basierende unternehmenseigene vom Geschlecht unabhängige Sterbetafel” [S7] § 1 Ziffer 3 — well below the 1,00 % Höchstrechnungszins that applies to the accumulation guarantee, which is why a Rentenfaktor cannot be read off
guar_rate. The base run sets the current factor equal to the guaranteed one, so the max-of-two rule is exercised by a test rather than by the base path. A Rentenfaktor is the arithmetic image of an annuity table plus a guaranteed rate, and the market-standard table is DAV 2004 R, a Generationentafel in attained age and calendar year REG-R49 — the property of the Deutsche Aktuarvereinigung, not public and not redistributed here. delib ships std proxies and states what a replacement must preserve.The citation for this row was wrong and is corrected. The retrieved Zurich Verbraucherinformation — the [S9] chassis — provides the opposite default: “Ist keine der folgenden Erweiterungsmöglichkeiten … eingeschlossen, so erlischt im Falle des Todes der versicherten Person die Versicherung, ohne dass eine Leistung fällig wird”, and where cover is agreed the standard form is Beitragsrückgewähr, a return of premiums [S9]. The return-of-accumulated-capital form delib models is R+V’s: “Stirbt die versicherte Person vor Rentenbeginn, wird der Policenwert, mindestens jedoch 90 % der Summe der gezahlten Beiträge fällig” [S7] § 1 Ziffer 5, and Allianz’s KID shows the same amount in the death and survival scenarios [S4]. Both shapes are in the market, and delib models one of them. The 50 % floor is a std representative choice whose statutory basis is narrower than stated: EStG § 20 Abs. 1 Nr. 6 Satz 6 Buchst. a is written for a “Kapitallebensversicherungsvertrag”, not for the Rentenversicherung mit Kapitalwahlrecht that Satz 1 also covers, so reading its 50 % across to this product is an inference and stays unverified R14 REG-R45. The commencement date is exact: § 52 Abs. 28 Satz 8 applies Satz 6 to contracts concluded after 31 March 2009. R+V’s own floor is 90 % of premiums, comfortably above 50 %. The floor is on in the base run and off on one model point, keeping the plain return-of-capital form testable.
The Bewertungsreserven leg is path- and balance-sheet-dependent in a way a gross liability cash-flow model cannot reproduce, and the Sicherungsbedarf test REG-R9 REG-R18 has for most of the last decade reduced the payable half to zero on high-guarantee portfolios. delib models the declared laufende surplus explicitly and excludes these two components, saying so.
Charges#
Nothing about the charge structure is special. Two levels are now established from carrier disclosures and delib’s acquisition charge turns out to be the market’s; the rest stay std. The comparators, both for index tariffs: Allianz Zukunftsrente IndexSelect, Einstiegskosten “2,5% der kumulierten Anlagen” plus 1,5 % of the annual payment from year 6, Verwaltungsgebühren 3,5 % of the payment a year plus 1,0 % of the value a year, Transaktionskosten 0,1 %, total 1,6 % a year over 30 years [S4]; Stuttgarter BasisRente index-safe, Abschluss- und Vertriebskosten 2,50 % of premiums, Verwaltungskosten 9,00 % of premiums plus 0,04 % of the accumulated capital monthly, Effektivkosten 1,80 Prozentpunkte [S11]. Both sit well inside the “over four percent” level at which BaFin says an appropriate customer benefit “erscheint zweifelhaft” R17.
Parameter |
Representative value |
Basis |
|---|---|---|
Abschluss- und Vertriebskosten |
2,5 % of the Beitragssumme, financed by Zillmerung, against a DeckRV § 4 Höchstzillmersatz of 25 ‰ (“Der Zillmersatz darf 25 Promille der Summe aller Prämien nicht überschreiten”), cut from 40 ‰ on 1 January 2015 |
matches both retrieved carrier disclosures [S4] [S11]; ceiling R7 REG-R16 REG-R20 |
Acquisition-cost spread |
Over the first five years |
|
Verwaltungskosten, premium-based |
3 % of each gross premium |
std (26) |
Verwaltungskosten, reserve-based |
0,25 % of the Deckungskapital per year |
std (26) |
Stückkosten |
36,00 € per policy per year, inflating at 1,5 % |
std (26) |
Stornoabzug |
2 % of the Deckungskapital, subject to the § 169 Abs. 3 floor |
std (27) |
Option dealing cost and spread |
Inside the Cap, not a charge line — a wider spread simply produces a lower Cap |
structural (28) |
House-index level fee and volatility-target drag |
Inside the index, and therefore inside neither the Cap nor the disclosed costs |
structural (28) |
Dividend yield of a price index |
Not a charge at all, but a permanent give-up of the same order — of the order of 3 % a year on euro-area equity |
unverified; structural (28) |
Effektivkosten |
Required to be disclosed as the Minderung der Wertentwicklung to the start of the payout phase; a validation target, not a model input |
Two different rules with two different functions, and delib keeps them apart: the DeckRV governs what the insurer may reserve (the Höchstzillmersatz, 25 ‰) REG-R16, while § 169 Abs. 3 VVG governs what it must pay — at least the Deckungskapital obtained by spreading acquisition and distribution costs evenly over the first five contract years REG-R28. delib’s charge profile uses the five-year spread, so the floor is satisfied by construction; it is nevertheless computed and applied, so a user who shortens the spread sees it bite.
Inherited std from the sibling delib endowment and classic-annuity products, and below both retrieved comparators:
β = 3 %of premium sits between Allianz’s 3,5 % of the annual payment and is far below Stuttgarter’s 9,00 % of premiums, whileγ = 0,25 %of the reserve is a quarter of Allianz’s 1,0 % of value and about half Stuttgarter’s 0,04 % monthly (≈ 0,48 % a year) [S4] [S11]. Neither carrier’s structure is delib’s — Allianz and Stuttgarter both charge on premium and on value, as delib does, but at higher levels, and Allianz adds a second entry charge of 1,5 % of the payment from year 6. The values are shipped inputs and are not changed in a provenance pass; the effect on the reported Effektivkosten is a live sensitivity intechnical-notes.md. The wider frame: the 2024 Verwaltungskostenquote was 2,4 % on one measurement and 2,19 % on another, spread from under 2 % to over 4 % REG-R53, and BaFin makes cost a supervisory focus R16 R17 REG-R35.§ 169 Abs. 5 VVG permits a deduction only if it is agreed, quantified and appropriate — “nur berechtigt, wenn er vereinbart, beziffert und angemessen ist” — with a deduction for unredeemed acquisition costs expressly ineffective R2 REG-R28. Both retrieved AVB apply one and both put the amount outside the AVB, which is how the beziffert requirement is met in practice: R+V states it “in EUR” in the Verbraucherinformationen [S7] § 11 Ziffer 2, Allianz in the Versicherungsinformationen [S2] Ziffer 9.2 Absatz 2 — so no published deduction level was obtained for any index tariff, and delib’s flat 2 % remains std inside a 0–20 % band. Allianz’s AVB does disclose the structure: the deduction falls away on surrender in the last year of the Aufschubdauer, and in the last seven years where the insured is at least 55 and the contract at least ten years old — a taper delib does not model. Both AVB also carry the § 169 Abs. 6 power to reduce the surrender value for one year at a time.
These three are the index-specific give-ups, and none appears in any charge table, so the disclosed Effektivkosten understate the economic give-up relative to holding the index by an amount disclosed nowhere — a structural fact about the product class, not a claim about any carrier, and the most substantive fair-criticism point in this specification.
Termination and values#
Parameter |
Representative value |
Basis |
|---|---|---|
Rückkaufswert |
The Deckungskapital on the calculation bases of the premium calculation, floored by the five-year-spread Mindestrückkaufswert, less the Stornoabzug |
|
Locked-in credits |
Inside the Rückkaufswert — they are guaranteed capital by then, not a contingent entitlement, and § 169 Abs. 7 VVG requires already-allocated Überschussanteile to be paid on top of the Abs. 3 amount in any event |
R2; mechanic firm |
The running Indexjahr |
Not inside it — the payoff exists only at the year end, and on surrender only a pro-rata Schlussüberschussanteil and Sockelbetrag are added |
[S2] Ziffer 9.2, [S7] § 3 Ziffer 5 (18) |
Beitragsfreistellung |
Conversion to a paid-up contract at any time for the end of the current insurance period, provided the agreed Mindestversicherungsleistung is reached (below it the Rückkaufswert is paid instead), on the § 169 Abs. 3–5 value; the index participation continues on the capital and the Wahlrecht survives, § 165 Abs. 3 Satz 2 leaving Überschussbeteiligung claims untouched |
|
Premium-default conversion |
The insurer’s termination converts to prämienfrei automatically |
|
Widerruf / Kündigung |
30 days’ Widerruf for life insurance; Kündigung at any time for the end of the current insurance period where laufende Prämien are payable |
|
Expiry |
There is none — the Aufschubphase ends at Rentenbeginn with a benefit, not with a lapse |
mechanic firm |
German lapse is a three-way decrement — surrender, Beitragsfreistellung and premium-default conversion — and the last two keep the policy in force with a reduced benefit and a continuing expense loading REG-R28. The reference implementation models surrender only, because a paid-up population’s per-policy account diverges from the premium-paying one from the moment of conversion and tracking it needs a conversion-cohort ledger. The technical notes say so, state what the paid-up path would do, and expose a shortened Beitragszahlungsdauer as a model-point column, so the deterministic form of the same effect is exercised and tested.
Contractual mechanics#
Each subsection states the operative rule, in this document’s own words, and says what it does to the
projection. Where a retrieved AVB states a rule in terms, it is now quoted exactly and attributed
to the carrier and the edition — Allianz Zukunftsrente IndexSelect (Plus) E25, E---A0025Z0 (014) 12/2025 [S2], and R+V-IndexInvest-Rentenversicherung IL55, Stand 01.07.2025 [S7]. A quotation is
evidence about that carrier; nothing below generalises a wording to the market on the strength of
one of them.
The Überschuss as an option budget — the financing identity#
Each year the insurer declares an Überschussanteilsatz out of the surplus its results and the MindZV
permit R8 REG-R18. In the sichere Verzinsung arm that rate is credited to the Deckungskapital as
interest, on top of the guaranteed Rechnungszins. In the Indexbeteiligung arm the same amount is
not credited — it is spent, becoming the Optionsbudget with which the insurer buys, for the
coming Indexjahr, the option package replicating the promised payoff. With G the participating
capital at the start of the Indexjahr and b the declared rate:
option budget = b × G
price of the promised payoff on G = b × G ← the Cap (or the Quote) is set to make this hold
The Cap is not a marketing parameter. It is the solution of a pricing equation — which is why caps move from year to year with no change in the contract, and both retrieved AVB say so in terms. Allianz: “Den jeweiligen →Cap eines Index legen wir jährlich zu Beginn des →Indexjahres auf der Grundlage von Angeboten mehrerer Finanzinstitute neu fest”, the level depending on the year’s Überschussanteile, the Bewertungsreserven Sockelbetrag and “Faktoren des Kapitalmarkts wie der Volatilität und der Dividendenrendite des jeweiligen Index” [S2] Ziffer 3.3 Absatz 2 b). R+V, on the Beteiligungsquote: “Je niedriger der Preis der Kapitalmarktinstrumente und je höher die Überschussbeteiligung einschließlich der Mindestbeteiligung an den Bewertungsreserven sind, umso höher ist die Beteiligungsquote” [S7] § 3 Ziffer 4. One refinement the AVB force on the identity above: the budget is the declared surplus plus the year’s minimum share of the Bewertungsreserven, and at Allianz net of Verwaltungskosten; delib models the declared surplus alone. Priced risk-neutrally, the index arm is worth exactly what the safe arm is worth, the whole difference being the equity risk premium earned on the option package’s delta, less dealing costs. The product is a redistribution of one year’s surplus across states of the world, not extra return. And the budget can be zero — “Im ungünstigsten Fall kann die Überschussbeteiligung Ihres Vertrags der Höhe nach null sein” [S2] Ziffer 2.1 — in which case the year credits nothing whatever the index does R1 R8. Both AVB go further and switch the participation off entirely for a year in which the Policenwert does not exceed the Deckungsrückstellung required for the guarantee ([S2] Ziffer 3.5, [S7] § 2 Ziffer 1).
The annual Wahlrecht#
The policyholder elects, once a year and for the coming Indexjahr only, between Indexbeteiligung and sichere Verzinsung. The election is a contractual right, exercisable without the insurer’s consent, without medical evidence and without charge; doing nothing leaves the policyholder in the arm they were in.
Arm |
The year’s surplus is |
Outcome |
|---|---|---|
Sichere Verzinsung |
credited to the Deckungskapital as interest |
certain, positive, immediately guaranteed |
Indexbeteiligung |
spent on the index option package |
zero in a bad year; a multiple of the surplus in a good one; never negative |
The choice is informed, and the AVB settles it. The insurer fixes the Cap on market conditions shortly before the Indexjahr starts and the policyholder must elect before it starts; whether the Cap is announced before the election deadline decides whether the choice is informed or blind. Allianz notifies the indices, “die Höhe der →Caps der jeweiligen Indizes”, the Partizipationssatz, the year’s surplus net of Verwaltungskosten and the Bewertungsreserven Sockelbetrag “spätestens 3 Wochen vor dem Indexstichtag”, and the election “muss uns spätestens 7 Tage vor dem nächsten →Indexstichtag vorliegen” [S2] Ziffer 3.1; R+V likewise informs the policyholder “jeweils rechtzeitig vor Beginn eines Versicherungsjahres” and takes the election up to 7 days before it [S7] § 2. delib’s assumption that the Cap is known at election time is the carriers’ rule. Allianz’s default on silence is not simply “stay where you were”: the previous split rolls over only if index participation was at least 50 %, and otherwise the contract is put to 50 % index participation [S2] Ziffer 3.2 — a nudge delib does not model. The Wahlrecht attaches to the capital, so it survives Beitragsfreistellung R3 and persists to Rentenbeginn, ceasing there ([S2] Ziffer 3.3: “vor Beginn der Rentenzahlung”). In delib’s assumption taxonomy the election is a behavioural assumption, not a contractual or an insurer-discretionary one, and is exposed as a per-year path.
The Indexjahr — the sum of capped monthly returns#
This is the single most important and most misunderstood feature of the product, and it is now quotable. Allianz’s AVB defines the maßgebliche Jahresrendite as follows [S2] Ziffer 3.3 Absatz 2 a):
“Sie bestimmt sich dadurch, dass die negativen monatlichen Wertentwicklungen und die mit dem jeweiligen →Cap (siehe Absatz b)) des gewählten Index gedeckelten positiven, monatlichen Wertentwicklungen am Ende eines →Indexjahres aufsummiert werden. Die monatlichen Wertentwicklungen entsprechen dabei der prozentualen Veränderung des Index zwischen 2 Bewertungsstichtagen, die wir Ihnen jährlich mitteilen. Ergibt sich nach der Aufsummierung eine negative jährliche Summe, setzen wir diese auf null.”
Negative months in full, positive months capped, the twelve summed, the sum floored once at zero:
that is the formula below, clause for clause. The Indexjahr is divided into twelve monthly
observation periods. For each month m the index level is read at the two Beobachtungstage bounding
the month, and
r(m) = I(m) / I(m−1) − 1
x(m) = min( r(m), C ) ← capped above at C; NOT floored below
S = Σ over m = 1…12 of x(m) ← summed, NOT compounded
Indexrendite = max( S, 0 ) ← the floor is on the YEAR, not on the month
Indexgutschrift = max( S, 0 ) × G
The three features that define the payoff and must never be separated:
Upside is capped monthly. A month in which the index rises 8 % contributes
C, not 8 %.Downside is not capped at all. A month in which the index falls 8 % contributes the whole −8 %. There is no floor on
x(m), only onS.The twelve are summed, not compounded. Summation is close to compounding for small numbers but is not the same, and the contractual formula is a sum.
Why the asymmetry is the whole story. The payoff is a capped cliquet: the policyholder is long
the index’s monthly returns, short a strip of twelve monthly calls struck at C, with an annual floor.
Truncating each month’s right tail while leaving its left tail intact removes far more expected return
than the cap level suggests. At a monthly standard deviation of 5 % — about 17 % annualised, ordinary
for a broad European equity index — a 3 % cap gives away roughly one percentage point of expected
return per month, twelve times a year, against an expected monthly return well under 1 %. The
technical notes do that arithmetic: at those parameters the expected value of a capped month is
negative, and the product’s positive expectation rests entirely on the annual floor.
The trap. The max(S, 0) floor operates on the sum, not on each month, so it is not true that
a year with more up-months than down-months credits something. It is perfectly ordinary for a year in
which the index finished higher to credit zero — the research file’s constructed Example B is exactly
that case, the index rising 6,44 % and the credit being 0,00 € — and it is a required test.
And the insurer publishes the same trap, worked, on its own product page. Allianz sets out two Indexjahre on the EURO STOXX 50 at an exemplary Cap of 3,2 % and Partizipationssatz of 75,00 % [S2] [S5]. In 2020/2021 the twelve monthly movements ran +18,06 % (capped to 3,20 %), +2,26 %, −2,52 %, +4,45 % (3,20 %), +7,78 % (3,20 %), +1,42 %, +1,63 %, +0,61 %, +0,62 %, +2,62 %, −3,53 % and +5,00 % (3,20 %) — negative months passing through in full — summing to 15,90 %, which at 75 % gave an Indexpartizipation of 11,92 %, against a point-to-point index gain of 43,69 %. In 2021/2022 the same arithmetic summed to −26,96 % and the maßgebliche Jahresrendite was 0 %. Allianz’s own footnote makes delib’s arithmetic point for it: “Die Wertentwicklung des EURO STOXX 50® ergibt sich aus der Differenz der Kurse zu Beginn und zum Ende des Betrachtungszeitraumes, nicht aus der Summe der monatlichen Wertentwicklungen.”
One structural difference from delib’s model. Allianz applies both a monthly Cap and a
Partizipationssatz to the capped sum — “Die →Indexpartizipation ermitteln wir, indem wir die
maßgebliche Jahresrendite … mit dem →Partizipationssatz … multiplizieren” [S2] Ziffer 3.3 Absatz 2.
delib’s cap payoff form has no participation rate (its w is the election share, not a
Partizipationssatz) and its quote form has no cap, so an Allianz-shaped tariff is not directly
representable in the reference implementation. That is a model matter, recorded here and in
model.md, and deliberately not acted on in a provenance pass.
The floor and the Höchststandsicherung#
An Indexjahr can never end below zero: max(S, 0) is contractual and universal in this family, and it
is the feature the product is sold on. The floor is what makes it a life-insurance product rather
than a bet, and it is genuine — the worst imaginable Indexjahr credits zero and leaves the capital
untouched. Whatever is credited is locked in: at the end of the Indexjahr the Indexgutschrift
is added to the capital and becomes part of the guaranteed capital, no longer at risk in any later
year, earning the guaranteed Rechnungszins thereafter like any other part of the Deckungskapital,
and entering the base G of every subsequent Indexjahr. That is the Höchststandsicherung, and
it is what makes the year-by-year floor add up to a path-independent guarantee.
Two consequences. The ratchet is not free: each year’s option package is a fresh strip on a larger base whenever the previous year credited something — it finances itself automatically, because the surplus is declared as a rate on that same larger base, which is why the financing identity is written in rates. And the guarantee is a floor on the path, not only on the maturity value: under a plain maturity guarantee the insurer can recover a bad year with a good one, whereas here every credited amount is permanent, so the guarantee’s cost rises with every good year. A within-year Höchststandsicherung, locking in the highest level reached inside the year, is a different and rarer feature; no German carrier is established as offering it and delib implements the annual lock-in only.
The Partizipationsquote variant#
Instead of capping each month, the contract credits a fixed fraction q of the year’s index
movement, floored at zero: Indexrendite = max( q × ( I(12)/I(0) − 1 ), 0 ). There is no monthly cap
and no monthly asymmetry — a down-month is not penalised relative to an up-month because only the
year’s net movement matters — and the whole of the give-up is in q. The two designs are not
equivalent and they fail differently. The Cap design gives away the large monthly moves and is hurt
by volatility even when the year ends well; the Quote design gives away a constant fraction in every
state. On the research file’s Example A the Cap variant credited more (8,90 % against 60 % of a
compounded 13,4548 %); on Example B it credited nothing while the Quote variant credited 60 % of
6,4402 % — the cleanest possible demonstration that the two are not interchangeable.
The Cap-Festlegung — who sets it, when, and on what#
The Cap is fixed by the insurer, for one Indexjahr at a time, before that Indexjahr begins, and is then binding for its whole length, not adjustable during the year. The determination is a pricing calculation rather than a discretion in substance, and the directions of movement follow from that:
If this rises |
the Cap |
because |
|---|---|---|
the declared surplus rate (the option budget) |
rises |
more money buys more upside |
the index’s implied volatility |
falls |
monthly caps are strips of options, and volatility makes them dearer |
the index’s dividend yield |
falls |
options are written on the price index; a higher dividend yield lowers the forward |
the risk-free rate |
rises, indirectly |
it raises the investment return and hence the surplus available |
The legal frame, and the distinction this product turns on. The Cap-Festlegung is a unilateral determination by the insurer of a term deciding the policyholder’s return for the coming year, and is therefore reviewable under § 315 BGB for billiges Ermessen: “so ist die getroffene Bestimmung für den anderen Teil nur verbindlich, wenn sie der Billigkeit entspricht. Entspricht sie nicht der Billigkeit, so wird die Bestimmung durch Urteil getroffen” (§ 315 Abs. 3 BGB) R22. It is not an adjustment under § 163 VVG (Prämien- und Leistungsänderung), which lets the insurer reset the premium where “sich der Leistungsbedarf nicht nur vorübergehend und nicht voraussehbar gegenüber den Rechnungsgrundlagen der vereinbarten Prämie geändert hat”, on an unabhängiger Treuhänder’s confirmation; nor under § 164 VVG (Bedingungsanpassung), which needs the clause to have been declared ineffective “durch höchstrichterliche Entscheidung oder durch bestandskräftigen Verwaltungsakt” and involves no trustee at all R4 REG-R27. Keeping the three apart is the most important legal distinction in this product and no delib document may blur it: redetermining the Cap exercises a discretion the contract confers, while replacing an ineffective clause changes the contract. Substituting the index is a third thing again — both retrieved AVB reserve it to the insurer on material change, with no trustee and no Sonderkündigungsrecht ([S2] Ziffer 3.7, [S7] § 3 Ziffer 11). No decided German case on the Cap-Festlegung is known, so the § 315 framing, doctrinally sound, is untested. German litigation over an Indexpolice does exist — the Verbraucherzentrale Hamburg sued Allianz over the IndexSelect web advertising and won at first instance (LG München I, 23.03.2018, Az. 37 O 12326/17) before the OLG München dismissed the claim on 04.04.2019 with no Revision admitted [S14] [S16] — but that was a UWG case about how the participation was described, not a review of a cap determination.
The underlying index, and the move to house indices#
The classic underlying is the EURO STOXX 50 — confirmed as Allianz’s, alongside the S&P 500, whose non-euro quotation brings in a Währungsfaktor applied to the year return [S2] — and two of its properties drive the economics. It is quoted and used as a Kursindex — a price index, dividends excluded — and options are written on the price index, so the euro-area dividend yield, of the order of 3 % a year unverified, never reaches the policyholder in any state of the world: a permanent structural give-up on top of the cap, invisible to a purchaser comparing the product to “the index”. And it is volatile, of the order of 18–22 % annualised unverified, which makes the monthly cap strip expensive and forces the Cap down.
From the mid-2010s a substantial part of the market replaced it with bespoke multi-asset indices, whose common features are: multi-asset composition, so volatility is structurally lower than an equity index’s; volatility targeting, a rule scaling exposure to hold realised volatility at a target often around 5 % unverified — the decisive engineering step, because at a 5 % target the option package costs a fraction of what it costs on a 20 %-volatility index, so the same budget buys a participation rate near or above 100 %; an excess-return construction with an embedded fee of the order of 0,5–1,5 % a year unverified, which reduces the return without appearing in any cost disclosure; and a short live history behind a long backtest. The honest summary: the shift moved the give-up from somewhere the purchaser can see — a 55 % participation rate, a 3 % cap — to somewhere they cannot. Headline numbers improved; expected outcomes did not necessarily improve with them, because the financing identity still binds.
Two such indices are now named from carrier documents, which is the first evidence in this specification that the pattern above is real and not a recollection. R+V’s underlying is the Solactive Multi Anlage Stabil Index (SOMAS), developed for the tariff by R+V and Solactive and described by the carrier as combining equities, bonds and gold with a “Stabilitätsmechanismus” [S7]; Stuttgarter’s are the M-A-X Multi-Asset Index and a Grüne Zukunft Index, the M-A-X described as investing “in mehreren Anlageklassen, um eine kontinuierliche Wertentwicklung zu erzielen” [S8]. Both are quota designs, and Stuttgarter’s published quota of 70 % on its house index is far below the “near or above 100 %” the volatility-target argument predicts, which is a caution against pushing that argument too far. No volatility target, no index-level fee and no excess-return construction is published for either, so the 5 % target and the 0,5–1,5 % embedded fee remain unverified and nothing about their level is asserted. No index is named in any shipped delib input file; the model parameterises the underlying by an explicit monthly-return path.
The guarantee at Rentenbeginn — Neue Klassik#
What the contract promises is a garantiertes Kapital zu Rentenbeginn, expressed as a percentage
of the premiums paid — the Beitragsgarantie — plus every index credit locked in along the way,
and a guaranteed Rentenfaktor converting that capital into an annuity. It is not a guaranteed
annual interest rate on the reserve. That is the defining feature of Neue Klassik and the reason
index products are grouped under that label [S6]: by owing the guarantee only at one future date rather
than at every balance date, the insurer can hold a materially riskier asset mix behind it and generate
the surplus that becomes the option budget. A model that reserves an Indexpolice as though it
guaranteed the Rechnungszins every year overstates the guarantee. The effective guarantee is
max( Beitragsgarantie on the premiums paid , guaranteed capital including all locked-in credits ),
with the second term dominating after a few good years. A projection must carry both, and a test must
assert that the guaranteed capital is monotone non-decreasing.
Death before Rentenbeginn#
The Todesfallleistung delib models is the return of the accumulated capital rather than a sum at risk, so the Risikoüberschuss is small, underwriting is light, and § 161 VVG is close to inoperative R6. That is one of two shapes in the market, and the retrieved documents show both. R+V’s index tariff pays “der Policenwert, mindestens jedoch 90 % der Summe der gezahlten Beiträge” [S7] § 1 Ziffer 5, and Allianz’s KID shows the same amount in the death and survival scenarios [S4] — the shape delib models. But the conventional Zurich chassis this specification once cited for it provides the opposite default: no death benefit at all unless an extension is agreed, and where one is, a Beitragsrückgewähr returning premiums rather than capital [S9]. The representative design floors the benefit at 50 % of the Beitragssumme, for the tax reason at footnote 23 R14 REG-R45. Death mid-Indexjahr attracts no credit in the year of exit, as for surrender — R+V computing the Policenwert “zum Ende des Monats, in dem der Todestag … liegt” with no index element for the incomplete year [S7] § 1 Ziffer 5.
Rückkaufswert and Beitragsfreistellung#
Surrender delivers the Rückkaufswert under § 169 VVG: the Deckungskapital computed by recognised actuarial rules on the calculation bases of the premium calculation, floored at the value obtained by spreading acquisition and distribution costs evenly over the first five contract years, less a Stornoabzug effective only if agreed, quantified and appropriate R2 REG-R28. Locked-in index credits are inside the reserve and therefore inside the surrender value — they are guaranteed capital by then. The running Indexjahr is not: a surrender in month 7 forfeits that year’s payoff. Beitragsfreistellung under § 165 VVG leaves the capital in place, continues the index participation on it, preserves the Wahlrecht, and gives a reduced guaranteed benefit on the same § 169 value R3.
Rentenbeginn — Rentenfaktor and Kapitalwahlrecht#
At Rentenbeginn the capital is converted at monthly annuity = capital / 10 000 × Rentenfaktor, and
the applied factor is the maximum of the guaranteed factor fixed at issue and the insurer’s current
factor at Rentenbeginn — a guarantee with upside. The index mechanic ends here: the capital is
fixed, the Wahlrecht lapses, and payout-phase surplus is applied to the annuity in payment. The
Kapitalwahlrecht takes the capital as a lump sum instead. The annuity itself is out of scope for this
model and belongs to products/sofortrente/.
Riders and options#
In scope (modeled or parameterized). The annual Wahlrecht, as a per-year election fraction w,
with four shipped paths — always index, always safe, a constant half-and-half split, and a switch from
the index arm to the safe arm mid-term. The payoff design, Cap or Partizipationsquote, as a
model-point column, so the two can be compared on an identical index path. The choice of underlying,
as a model-point key into an external monthly-return table, with an equity case, a low-volatility house
case and an all-zero case. The Kapitalwahlrecht, deciding whether the terminal capital is reported
as a lump sum or as the annuity it buys. The Stornoabzug and the 50 % death-benefit floor, as
model-point switches. And the shortened Beitragszahlungsdauer, the deterministic form of
Beitragsfreistellung.
Out of scope, and said so. Beitragsfreistellung as a stochastic decrement (footnote 29);
Dynamik (footnote 10); Zuzahlungen; the Rentengarantiezeit and every other payout-phase
feature, which belong to products/sofortrente/; Hinterbliebenenrente and Beitragsrückgewähr;
a Berufsunfähigkeits-Zusatzversicherung, a rider on this chassis in the market and a stand-alone
product in delib (products/berufsunfaehigkeit/); the Schlussüberschussanteil and the
Bewertungsreserven share (footnote 24); and a within-year Höchststandsicherung.
Variations across insurers#
This table is now a comparison rather than a record of what could not be compared. Two carrier AVB were retrieved in full — Allianz’s Zukunftsrente IndexSelect (Plus) E25, edition 12/2025 [S2], and R+V’s IndexInvest-Rentenversicherung IL55, Stand 01.07.2025 [S7] — and Stuttgarter’s published product documents and current parameters were read, though its AVB is not published and its row is correspondingly thin [S8] [S11]. Three product names are given, all now established; none may add a fourth, and a cell reading “not established” means exactly that.
Feature |
Allianz [S2] [S4] [S5] |
R+V [S7] |
Die Stuttgarter [S8] [S11] |
Anyone else |
delib std |
|---|---|---|---|---|---|
Index AVB obtained; product name |
yes; Zukunftsrente IndexSelect (Plus) E25 |
yes; PrivatRente IndexInvest, tariff IL55 |
no (AVB unpublished); index-safe |
no |
composite [S1] |
Payoff design (Cap / Quote / both) |
monthly Cap and a Partizipationssatz on the capped sum |
Beteiligungsquote on the year return; no cap |
quota on the year return; no cap |
not established |
Cap; Quote as a switchable variant |
Cap / quota level, current |
Cap 3,2 % and Partizipationssatz 75,00 %, both illustrative |
not published |
Partizipationsquote 70 % (Turbo 120 % / 172 %), 1.2.2026–31.1.2027 |
not established |
3,00 % monthly; |
Mindest-Cap guaranteed |
none in the AVB |
none in the AVB |
not established |
not established |
none |
Underlying index |
EURO STOXX 50, S&P 500 (with a Währungsfaktor) |
SOMAS (Solactive Multi Anlage Stabil Index) |
M-A-X Multi-Asset, Grüne Zukunft |
not established |
generic, by volatility |
Wahlrecht notice period |
election 7 days before the Indexstichtag; splits in 25 % steps |
election 7 days before the Versicherungsjahrestag |
annual |
not established |
annual, at the year end |
Cap announced before the election deadline |
yes — parameters notified ≥ 3 weeks before |
yes — “rechtzeitig vor Beginn” |
quota published in advance |
not established |
assumed yes — correct |
Base |
the Policenwert at the start of the Indexjahr, excluding that year’s premiums |
the Policenwert present the whole year, excluding that year’s premiums |
not stated (the budget is 100 % of laufende Überschüsse) |
not established |
the whole capital at the year start |
Garantieniveau menu |
90 %; 80 % for IndexSelect Plus |
90 % |
85 % (BasisRente variant) |
not established |
90 % |
Mid-year exit treatment |
no index credit; pro-rata Schlussüberschuss and Sockelbetrag only |
no index credit (credited at the next year’s start) |
not established |
not established |
no credit |
Effektivkosten / total cost |
1,6 % a year over 30 years; entry 2,5 % of cumulative payments |
not retrieved |
1,80 points; entry 2,50 % of premiums |
not established |
std, above |
Indexjahr aligned with the policy year |
not necessarily |
yes |
no — a common 1.2.–31.1. window |
not established |
aligned |
Parameter bands, restated. Two rows are now placed against real carrier figures; the rest remain unverified assessments and are the reason this specification still carries many std rows.
Parameter |
Band |
Who sits where |
|---|---|---|
Monthly Cap |
1,5 % – 5,0 %, typically 2,5 % – 4,0 % unverified |
Allianz illustrates 3,2 %; 3,3 % recorded in 2018 [S14]; no panel |
Partizipationsquote |
50 % – 80 % on an equity price index; 80 % – 120 % on a house index unverified |
Allianz illustrates 75,00 % on equity; Stuttgarter publishes 70 % on a house index — below the band |
Garantieniveau |
80 % / 85 % / 90 % observed; 60 % and 100 % unverified |
Allianz 90 % / 80 %, R+V 90 %, Stuttgarter 85 %; 100 % statutory for Riester R12 REG-R43 |
Declared surplus rate, 2026 |
3,07 % index segment, 2,62 % Klassik, 2,65 % Neue Klassik |
Assekurata survey averages R20; Stuttgarter publishes 2,16 % [S8] |
Höchstrechnungszins by cohort |
0,25 % (2022–2024) – 4,00 %; 1,00 % from 2025, recommended again for 2027 |
|
Index volatility (annualised) |
15 % – 22 % equity; 5 % – 8 % house index unverified |
no target published for either named house index |
Verwaltungskostenquote 2024; Ratenzahlungszuschlag |
under 2 % to over 4 %, average 2,19–2,4 %; 2 % / 3 % / 5 % |
REG-R53; convention unverified |
Stornoabzug |
0 % – 20 % of the Deckungskapital |
both carriers apply one and neither publishes its level — it is quantified per contract R2 |
Stornoquote, market-wide |
2,56 % (2023), 2,51 % (2022), by count, all Hauptversicherungen |
no index-specific rate exists R19 |
What does not vary, and what turned out to. Three things are firm across all three carriers, and they are the reason a composite is possible at all: the surplus finances the participation rather than sitting beside it; the year’s outcome is floored at zero, never negative; and what is credited is locked in permanently. The fourth item this section previously listed as invariant — that monthly returns are capped above and not below — is not invariant: it is Allianz’s design, and R+V and Stuttgarter have no monthly cap at all. delib ships both payoff forms, so the model spans the variation; the specification’s choice of the Cap form as representative is a std choice about which design to lead with, not a finding about the market.
Regulatory context#
Contract law — VVG. The hinge is § 153: the policyholder participates in the surplus and in the Bewertungsreserven unless participation is excluded, and such an exclusion can only be made for the whole of the profit participation; the insurer must allocate by a verursachungsorientiertes Verfahren or another comparable appropriate method; the Bewertungsreserven are recomputed annually and half of the amount determined is paid — and for a Rentenversicherung § 153 Abs. 4 makes the relevant date the end of the Ansparphase, not the end of the contract — subject to the LVRG’s Sicherungsbedarf override R1 REG-R24 REG-R9 REG-R20. §§ 165–170 supply the exit machinery R2 R3 REG-R28. § 163 (Prämien- und Leistungsänderung) permits the insurer to reset the premium where the Leistungsbedarf has changed unforeseeably and not merely temporarily against the premium’s calculation bases, the new premium being appropriate and necessary for permanent solvency and an independent trustee having confirmed both — the policyholder may take a reduced benefit instead — and § 164 (Bedingungsanpassung) permits a clause to be replaced only after it has been declared ineffective by a highest-court decision or a final administrative act, and without any trustee R4 REG-R27. Those are the two statutory channels through which this contract’s terms can be changed against the policyholder’s will, and neither of them is the annual Cap. § 161 excludes suicide within three years, the Rückkaufswert then being owed R6 REG-R26. § 155 requires an annual Standmitteilung stating the current status of the policyholder’s claims including profit participation and disclosing to what extent that participation is guaranteed — which is why a Standmitteilung specimen is a legitimate primary-source class here, and why the research file’s gap 4 (no completed Indexjahr with its twelve monthly movements was ever obtained) is its most frustrating absence [S10] R5 REG-R25.
§ 154 and the Modellrechnung. Where the insurer makes quantified statements about possible benefits beyond the guaranteed ones it must give a Modellrechnung on three interest rates, which § 2 Abs. 3 VVG-InfoV fixes as the Höchstrechnungszins × 1,67, that rate plus one point and that rate minus one point — at a 1,00 % Höchstrechnungszins, 1,67 % / 2,67 % / 0,67 % R5 REG-R25 REG-R31. Two things follow from reading § 154 itself. Its Abs. 1 Satz 2 exempts only contracts with benefits of the § 124 Abs. 2 Satz 2 VAG kind, so the duty does apply to an Indexpolice — one more consequence of the classification at R15. And a Modellrechnung for an Indexpolice is intrinsically awkward, because the interest assumption drives the option budget, which drives the Cap, which drives the payoff non-linearly. How German carriers discharge § 154 for this product is still not established: neither retrieved AVB reproduces one.
Prudential. § 124 VAG imposes the prudent-person standard with no quantitative investment limits since 1 January 2016 — permitting derivatives only “sofern diese zur Verringerung von Risiken oder zur Erleichterung einer effizienten Portfolioverwaltung beitragen” (Abs. 1 Nr. 5) — and § 125 Abs. 5 Nr. 4 ring-fences the Sicherungsvermögen, requiring a separate Anlagestock section only where life contracts “direkt an einen Aktienindex oder andere Bezugswerte binden” R9 R15 REG-R7. Buying index options to back an index-participation obligation is the paradigm of a derivative hedging a liability the insurer has itself written — liability and hedge matched by construction, month for month and cap for cap R9. § 139 VAG governs the surplus participation from the supervisory side and carries the Sicherungsbedarf rule REG-R9; §§ 140 and 145 govern the RfB REG-R10; the MindZV sets the 90 % / 90 % / 50 % minima on the three result sources — § 6 Abs. 1 requiring “90 Prozent der … anzurechnenden Kapitalerträge abzüglich der rechnungsmäßigen Zinsen”, § 7 90 % of the Risikoergebnis and § 8 50 % of the übrige Ergebnis, each floored at zero — so the guarantee is funded first and only the excess is shared R8 REG-R18. The DeckRV caps the technical rate and the Zillmersatz R7 REG-R14 REG-R16, and its § 5 Abs. 3 Zinszusatzreserve machinery sits behind the declared rate this product spends REG-R17. Above it all is Solvency II REG-R1 REG-R2 REG-R13.
Conduct and disclosure. An Indexpolice is a Versicherungsanlageprodukt and therefore a PRIIP: a three-page Basisinformationsblatt with a summary risk indicator, four performance scenarios and the cost tables is required R10 REG-R32. It is a Category 4 PRIIP: the DAV’s Ergebnisbericht records that Ziffer 7 of Anhang II RTS assigns to that category “Versicherungsanlageprodukte, deren Wertentwicklung teilweise von nicht am Markt beobachteten Faktoren abhängt” — here the discretionary surplus declaration — with the MRM to be determined under “einem anerkannten Branchen- oder Regulierungsstandard”, which that report supplies and aligns with the PIA standard used for certified products R10 R11 REG-R32. Category 4 permits the insurer’s own model for that component, which is why two Indexpolicen with similar mechanics can publish very different favourable scenarios; the DAV report is generic to Category 4 and says nothing specific about index mechanics. A Basisinformationsblatt for a German index product was retrieved [S4]: risk indicator 1 of 7 at 30 years, a moderate scenario of 2,1 % a year after costs against 3,7 % before, and total costs of 1,6 % a year. The Effektivkosten duty REG-R31 and BaFin’s Merkblatt 01/2023 (VA) on angemessener Kundennutzen R16 REG-R35 complete the frame. The Merkblatt does not name index products — it applies to kapitalbildende Lebensversicherungsprodukte as a class and makes the Effektivkosten the measure of cost — and BaFin’s Risiken im Fokus 2026 records that individual products reached Effektivkosten “über vier Prozent”, above which “erscheint ein angemessener Kundennutzen zweifelhaft” R17. Both retrieved index tariffs sit well below that. A design that credits zero in a substantial fraction of years while carrying a full acquisition-cost load is nevertheless exactly what a value-for-money regime exists to interrogate.
Taxation — context, not a cash flow; delib publishes gross cash flows and computes no tax. A Schicht 3 Leibrente is taxed only on its Ertragsanteil, a percentage fixed once and for all by the age reached at Rentenbeginn and read off the statutory table in § 22 Nr. 1 Satz 3 Buchst. a Doppelbuchst. bb Satz 4 EStG — 17 % at 67, 18 % at 65–66 and 22 % at 60–61 R13 REG-R41. A lump sum under the Kapitalwahlrecht is taxed on the Unterschiedsbetrag between the payment and the premiums paid; where the contract has run at least twelve years and the payment falls after the 62nd birthday, only half that difference is taxable and at the personal marginal rate rather than by final withholding. The 62 is a transitional rule, not the enacted text: § 20 Abs. 1 Nr. 6 Satz 2 EStG says the 60th birthday, and § 52 Abs. 28 Satz 7 substitutes the 62nd “für Vertragsabschlüsse nach dem 31. Dezember 2011” — so a pre-2012 contract keeps 60. The relief is subject, for contracts concluded after 31 March 2009 (§ 52 Abs. 28 Satz 8), to the 50 % Mindesttodesfallschutz condition — which § 20 Abs. 1 Nr. 6 Satz 6 Buchst. a states for a “Kapitallebensversicherungsvertrag”, so its application to a Rentenversicherung mit Kapitalwahlrecht is a reading and stays unverified R14 REG-R45. The index credits are not separately taxed: they are absorbed into the capital as credited, so there is no annual tax event, no Abgeltungsteuer on the year’s index gain and no Teilfreistellung under the Investmentsteuergesetz — the last because there is no fund. This tax deferral is one of the two genuine advantages over holding an index fund directly, the other being the guarantee. Exercising the Wahlrecht is not a change of contract and does not restart the twelve-year clock unverified. The duration-12 / age-62 double threshold is the strongest single driver of German surrender behaviour and shapes the lapse assumption in the technical notes REG-R45.
The criticism, stated fairly, because a specification that omits it is not a specification. Its home is the German consumer publishers and trade press — Finanztip [S12], Stiftung Warentest [S13], the Verbraucherzentralen [S14] and the trade titles [S16]. Two of the four were retrieved and are now quoted; Finanztest stays behind its paywall [S13] and no figure is taken from it. Finanztip’s assessment, from its own press release: returns “von mehr als 4 Prozent sind aber nur schwer zu erreichen”, with “nach Abzug aller Kosten Werte von 0,5 bis 2,5 Prozent” more likely, and “Verbraucher können oft nicht wirklich nachvollziehen, was sie da eigentlich kaufen” [S12]. The Verbraucherzentrale Hamburg’s, from the release announcing its first-instance win: the participation runs “nicht über die eingezahlten Beiträge, sondern ausschließlich über die … jährlich zu ermittelnde Überschussbeteiligung”, and the monthly measurement means the annual outcome can fall short of the index “selbst dann …, wenn der Cap in der Jahresbetrachtung gar nicht überschritten wird” [S14]. The first of those is this specification’s own characterisation, put by a consumer body; the second is Example B. Both are positions, and the litigation they came from was decided the other way on appeal [S16]. The cap’s effect on the expected credit is large and is not disclosed in a usable form: the purchaser is told the cap, is not told the volatility, and cannot do the calculation. Negative months are uncapped, which is genuinely counter-intuitive. Against a direct index investment the product loses on every axis but two — it gives up the dividends of a price index and the tail of every good month, and adds acquisition, administration and possibly index-level costs — but what it gives back is real: the capital cannot fall, credits lock in permanently, the guarantee is the insurer’s, and the accumulation is tax-deferred with a favourable exit. The Cap is redetermined annually at the insurer’s discretion, constrained in principle by § 315 BGB and by no decided case on that point R22; the move to house indices moved the give-up out of sight, and the two now named publish no volatility target or index fee [S7] [S8]; and complexity is itself a defect in a retail product — a point the LG München I accepted in 2018 and the OLG München rejected in 2019 [S14] [S16]. The counter-argument, fairly stated: the relevant benchmark for most purchasers is not an index fund but the sichere Verzinsung arm of the same contract, and against that the index arm has a higher expected value, cannot do worse than zero in any year, costs nothing extra, and can be abandoned at any anniversary. The reference implementation lets a reader run that comparison.