The Decreasing-Term Anachronism

Abstract

The standard Irish mortgage-protection product is a decreasing-term assurance whose sum-assured schedule is fixed at policy inception to a notional 6% amortisation curve. The schedule is contractual — set on the day the policy issues — and it does not adjust when the underlying mortgage subsequently behaves differently. This paper asks a constructive question: in a contemporary Irish mortgage environment of low actual rates, short fixed-rate periods and routine principal events, what would a robust schedule design look like, and what would it cost the manufacturer to deliver? We approach the question by identifying three failure modes of the legacy product — a design failure (the schedule mechanic itself), a distribution failure documented elsewhere in this series (the bank-channel premium), and a capital question (whether modernisation prices fairly under the Solvency II Standard Formula). We then quantify the design failure at the household level and across the principal-dwelling-house stock, evaluate three modernisation candidates already in commercial use in comparable markets (re-rated annuity-falling, event-stepped, balance-linked), and compute the Solvency II capital implications under the Standard Formula.

The representative borrower is a 35-year-old non-smoking male with a 30-year, €322,000 mortgage at 4% against a 6% notional schedule. In steady state the household carries a mean over-insurance of €14,332 across the term, peaking at €22,292 in month 223. On a deterministic full-stock basis the aggregate annual premium dead-weight is approximately €12.34m; on a duration-weighted decomposition of the Bank of Ireland Asset-Covered-Securities pool — used as a published-evidence proxy for the in-force PDH cohort distribution — the more conservative central estimate is approximately €4.07m per year. A stochastic extension on the EIOPA EUR risk-free term structure at 30 April 2026, 50,000 paths over 360 months, seed 20260529, with bootstrap-Irish rate dynamics, produces a household over-insurance figure of approximately €26 per household per year (median, 5–95% interval €6–€44), against which three parametric Vasicek calibrations (C1 / C2 / C3) plus a C1 + MS2 spread-switching sensitivity bracket the headline between €14 and €24 per household per year.

Under the four common Irish-mortgage disturbances — fixed-rate roll-off, top-up advance, term extension, lump-sum overpayment — the schedule remains unmoved and the policy can fall structurally below the outstanding balance, with peak under-insurance of approximately €80,000 on a five-year term extension. We acknowledge, candidly, that the over-insurance gap also functions as a prudential cushion: at year five under a +200 bps upward rate spike, the cushion absorbs the resulting balance trajectory without ever leaving the policy under-cover, even as the mean cushion shrinks by approximately two-thirds. Two of the three modernisation candidates — the event-stepped (M2) and balance-linked (M3) designs — substantially close both the steady-state and the disturbance gaps. The third (M1, re-rated at issue) is a partial response: it closes the steady-state gap but trades it for a disturbance under-insurance gap.

The Solvency II Standard Formula mortality capital position is design-by-design: at baseline (D0) and in three of four disturbance scenarios, M1, M2 and M3 each reduce SCR_mort relative to the legacy product; in the term-extension disturbance specifically, M2 raises SCR_mort by approximately 27% relative to a legacy schedule that has, in that scenario, structurally collapsed below the outstanding balance. We compute the break-even expense loading per policy per year required to fund M1, M2 and M3 against the loss of premium income on the closed over-insurance gap and find it modest: approximately €10.61 per policy per year for M1, €13.01 for M2 and €11.87 for M3, equivalently approximately 2.7%, 3.3% and 3.0% of base premium respectively, against an implicit premium loading of approximately €1.2336 per €1,000 sum-assured per year (equivalently, approximately €397 per policy per year on the representative profile). The recommendation is that Irish life offices evaluate M2 as the proportionate near-term commercial response and treat M3 as the longer-run product-architecture target, within the existing product taxonomy and the European Insurance Distribution Directive's Product Oversight and Governance regime.

Keywords: mortgage protection insurance; decreasing-term assurance; product design; sum at risk; prudential cushion; Solvency II; Standard Formula; mortality stress; bancassurance; Ireland; Consumer Credit Act 1995; EIOPA; risk-free rate; Vasicek; product oversight and governance; protection gap.

JEL classification: G22; G28; G52; D14; D18.

Cite as: Milmo-Penny, D. (2026). The Decreasing-Term Anachronism — Schedule Mismatch in Irish Mortgage Protection and Three Modernisation Candidates. Mylife.ie Working Paper MWP-2026-03. SMP Financial Ltd, Dublin.

1. Introduction

Mortgage protection assurance is the life-insurance product most widely held by Irish households, because it is embedded in the statutory architecture of the residential mortgage. Section 126 of the Consumer Credit Act 1995 obliges a mortgage lender to ensure that a life-assurance policy is in place on a housing loan, subject to a small set of carve-outs, and the standard market response is a thirty-year decreasing-term assurance whose initial sum assured matches the loan and whose subsequent sum-assured schedule is calculated, at policy inception, on a notional amortisation curve.¹ The notional curve in near-universal Irish use is a six-per-cent amortisation. That convention is long-standing — confirmed at the level of primary policy schedules and key-features documents of the five principal Irish life offices.² It is the contractual basis on which the cover schedule is drawn, and it is independent of the interest rate the borrower actually pays on the mortgage.

This paper asks a constructive, pro-industry question. The legacy product has worked in its present form for decades; it does what its contractual specification says it should do; and the insurers who sell it have priced it on an actuarially honest basis under the contractual schedule. In a contemporary Irish mortgage environment, however — actual rates materially below the notional, short fixed-rate periods, routine principal events on a thirty-year horizon — what would a robust schedule design look like, and what would it cost the manufacturer to deliver? We treat this not as a regulatory critique but as a product-modernisation question: a design conversation that an Irish life office, an actuarial pricing function, or a manufacturer's product-oversight committee could constructively engage with under the existing supervisory framework.

The question is timely on several counts. The Irish mortgage book against which the legacy schedule was originally calibrated bore little resemblance to today's. Through the last three decades of the twentieth century the dominant Irish mortgage was variable-rate, on nominal mortgage rates that ran well into double digits for much of the period, with principal events confined largely to redemption or default; the six-per-cent notional schedule sat below prevailing rates and the resulting schedule tracked the actual amortising balance more closely than it does today. The contemporary Irish book is structurally different. The Central Bank of Ireland reports a weighted-average new-business PDH rate of three point five two per cent at end-March 2026,³ with approximately ninety per cent of new agreements on fixed-rate terms of one to ten years that systematically roll off within a thirty-year policy term.⁴ Top-up advances are a routine Banking and Payments Federation Ireland category;⁵ term extensions and arrears restructurings are an established Central Bank forbearance instrument;⁶ and lump-sum overpayments occur within published lender contractual caps.⁷ Each of these disturbances changes the outstanding balance B(t) but leaves the contractual schedule S(t) unmoved. The European Insurance Distribution Directive's product-oversight-and-governance regime (Article 25) imposes an ongoing-monitoring obligation on manufacturers to confirm that the product continues to meet target-market needs;⁸ on its face, that obligation invites the very modernisation conversation we attempt here.

We are explicit about what this paper is not. It is not an argument that the legacy product is in breach of section 126 — the schedule is contractually clear and the premium is actuarially honest under it. It is not an argument that insurers acting on the legacy schedule have acted in bad faith. It is not a regulatory call for action on the schedule mechanic — none of the matters discussed below requires a new regulatory mandate; they sit within the existing Insurance Distribution Directive, the existing Consumer Protection Code, and the existing Solvency II Standard Formula. It is not, finally, an argument for indiscriminate policy replacement: existing cover remains valid cover, and the modernisation argument is squarely about new business and the prospective book.

The principal findings are these. The representative Irish mortgage-protection policyholder — male, age thirty-five, non-smoker, €322,000 initial sum assured, thirty-year term, six-per-cent notional schedule against an actual four-per-cent mortgage rate — carries a mean over-insurance of €14,332 across the policy term and pays a small but non-trivial dead-weight premium on that excess cover for thirty years. The aggregate annual dead-weight across the Irish principal-dwelling-house stock, computed deterministically at the four-per-cent actual rate, is approximately €12.34m on a single-vintage uniform-cohort basis; the more conservative duration-weighted decomposition, anchored on the publicly disclosed Bank of Ireland Asset-Covered-Securities mortgage pool as a proxy for the in-force cohort distribution, gives approximately €4.07m per year. On a stochastic basis the headline household harm is approximately €25.94 per household per year (median, 5–95% interval €6.07–€44.35). Under the under-insurance disturbances, the largest single-policy exposures are €80,450 of peak under-cover on a five-year term extension and €38,363 of peak under-cover on a €50,000 top-up advance.

2. Institutional setting

2.1 The statutory obligation

Mortgage protection assurance is required, on Irish housing loans, by section 126 of the Consumer Credit Act 1995. The substantive obligation runs to the lender, not the borrower. The operative wording of section 126(1) requires a mortgage lender to arrange, through an insurer or an insurance intermediary, a life-assurance policy providing, in the event of the death of a borrower before a housing loan has been repaid, for payment of a sum equal to the amount of the principal estimated by the lender to be outstanding in the year in which the death occurs on the basis that payments have been made in accordance with the mortgage, such sum to be employed in repayment of the principal.¹ Subsection 126(2) provides for a small set of carve-outs — the property is not the borrower's principal residence, the borrower is over fifty at approval, the borrower belongs to a class for which cover is unobtainable except at a significantly higher premium, or the borrower has otherwise arranged sufficient life assurance — and subsection 126(5) expressly provides that excess proceeds, over the outstanding mortgage at the date of death, are payable to the surviving borrower or the deceased borrower's estate rather than retained by the lender.⁹

Two features of the statutory language matter for this paper. First, the reference point is the actual mortgage and the actual outstanding principal — not a notional amortisation curve drawn at inception against a market-convention rate. The legislative intent is balance-tracking against the loan as it stands. Second, the Competition and Consumer Protection Commission has confirmed and re-confirmed that the borrower is not obliged to take the policy offered by the lender: the borrower may purchase from the lender, from an insurance company directly, or through an intermediary.¹⁰ The decreasing-term schedule is therefore a market convention, not a statutory requirement: section 126 demands that the death benefit cover the outstanding principal, but it does not prescribe the mathematical mechanic by which the sum assured tracks the principal across the policy term. We also note — and return to the point in Section 8.6 — that the statutory provision for excess proceeds to flow to the estate means that a steady-state over-insurance position is not, in welfare terms, simply dead-weight on the household; a portion of it is a transfer that benefits the deceased borrower's estate on the insured event.

2.2 The Irish principal-dwelling-house book

At end-June 2025, the Central Bank of Ireland reported 698,335 private residential mortgage accounts secured on principal dwelling houses, with an aggregate outstanding balance of approximately €106 billion.¹¹ The Central Statistics Office's Household Finance and Consumption Survey 2023 reported that thirty per cent of Irish households held a mortgage on their main residence, with a median outstanding balance of €117,900.¹² The Banking and Payments Federation Ireland publishes the quarterly mortgage drawdowns report; the Q4 2025 release records average first-time-buyer PDH drawdown value of approximately €322,000 and an average first-time-buyer age of thirty-five.⁵ Those two numbers are the representative-borrower anchor used in this paper.

The composition of the new-business book matters for the disturbance analysis. The Central Bank's monthly Retail Interest Rates release shows that roughly ninety per cent of new mortgage agreements in recent quarters have been on fixed-rate terms,³ and the Governor's blog has confirmed that approximately one in ten PDH mortgages sees a fixed-rate period expire each year.⁴ Irish fixed-rate periods are short by international comparison — typically one to ten years against a thirty-year amortisation — so a single mortgage almost certainly transits multiple rate regimes within a single policy term. The BPFI top-up category is a separate, well-defined line in the quarterly drawdowns release: the Q4 2025 release records 3,280 top-up drawdowns nationally.⁵ The Central Bank's mortgage arrears statistics, and Deputy Governor Sibley's July 2021 statement, document that over 95,000 PDH accounts — approximately thirteen per cent of the book — were assessed as facing a shortfall at end of term, with term extension a routine forbearance instrument.⁶ Lender published terms show overpayment caps ranging from €5,000 per year (Allied Irish Banks) to ten per cent of the normal monthly repayment or €65 — whichever is greater — without breakage fee on fixed-rate accounts at Bank of Ireland (variable-rate accounts have no cap) to twenty per cent per year at ICS Mortgages and certain other lenders.⁷

These are not edge cases. Each of the four disturbance categories used in Section 5 — fixed-rate roll-off, top-up advance, term extension, lump-sum overpayment — is a regular Irish-mortgage event with a published prevalence. The decreasing-term schedule, as currently engineered, is unmoved by any of them.

2.3 The six-per-cent notional convention

The Irish decreasing-term schedule is constructed at policy inception as the amortisation curve of a level-payment loan of the initial sum assured at a notional interest rate, over the policy term. The market standard for that notional rate is six per cent. Formally, for an initial sum assured S(0), term n months, and notional monthly rate i = 0.06/12, the contractual schedule at month t is

S(t) = S(0) · ((1+i)^n − (1+i)^t) / ((1+i)^n − 1)

and the schedule is locked at issue. The premium is computed, on the insurer's pricing basis, against the integral of the mortality-discounted sum at risk across this fixed S(t) curve. Neither the schedule nor the premium adjusts subsequently for the actual mortgage rate, the actual outstanding balance, or any disturbance to the loan.

The six-per-cent figure is a historical artefact. It reflects the long-run average nominal mortgage-rate environment under which Irish decreasing-term assurance crystallised as a product class — a period in which Irish mortgage rates of six to fifteen per cent were ordinary, and in which fixing the schedule against a mid-range nominal rate gave a reasonable expected match to the typical loan's actual amortisation profile across a long term. The convention was reasonable when it was set; it is not, by itself, evidence of bad faith. The six-per-cent convention is confirmed in the current public policy-schedule and key-features documentation of all five principal Irish life offices — Irish Life, Royal London Ireland, Zurich Life Ireland, New Ireland Assurance, and Aviva Life and Pensions Ireland.²

The downstream consequence — and the analytical focus of this paper — is that S(t), the contractual schedule, and B(t), the actual outstanding balance, are now two different objects. In a steady-state Irish mortgage at four per cent actual against a six-per-cent notional, the schedule sits above the balance for almost the entire term, peaking at a divergence of approximately €22,000 mid-policy. Under any of the four disturbance categories, the schedule and balance can diverge in the opposite direction, with the schedule falling below the balance and the policy structurally under-insured. The two-sided mismatch — chronic over-insurance, episodic under-insurance — is the central object of analysis in Sections 5 and 6.

3. Three failure modes

This paper sits at the intersection of three independent but cumulatively-incident failure modes of the legacy Irish mortgage-protection product. The framing is constructive: each failure mode is a candidate for product modernisation rather than for regulatory action, and the modernisation candidates evaluated in Section 5 are already in commercial use in markets comparable to Ireland's.

3.1 The three failure modes

(a) The design failure — the schedule mechanic. The legacy schedule fixes the sum-assured curve at policy inception, against a notional six-per-cent rate, and holds it fixed across the policy term. In steady state the schedule sits above the balance (chronic over-insurance), and under principal events the schedule can fall below the balance (episodic under-insurance). The schedule mechanic was a reasonable engineering shortcut in the environment in which the decreasing-term product class crystallised — a variable-rate book on high nominal rates, with principal events confined largely to redemption or default. It is the focus of this paper because, in the contemporary low-rate, short-fixed-period environment, the legacy mechanic generates a systematic two-sided mismatch with the legislative intent of section 126.

(b) The distribution failure — the bank-channel premium. Documented in The Bank Premium (MWP-2026-02), a structural premium attaches to lender-channel mortgage protection in Ireland, with whole-of-market quotation suggesting typical price gaps of twenty to thirty per cent against the cheapest broker-quoted equivalent for an otherwise-identical risk.¹³ This is a distribution-side failure: it arises from tied bancassurance and from the absence of routine consumer comparison at the point of mortgage drawdown. It is independent of the schedule mechanic but compounds the household-level impact. The present paper does not re-litigate that finding; we cite it once and proceed.

(c) The capital question — does modernisation price fairly? This is not strictly a failure of the legacy product so much as an open question about modernisation: if Irish life offices are to invest in a modernised schedule (M2 or M3), the Solvency II Standard Formula's mortality capital charge must price modernisation fairly. As Section 6.4 shows, the Standard Formula does price modernisation roughly correctly — at baseline and in three of four disturbance scenarios, modernisation reduces SCR_mort; in the term-extension scenario alone it raises the charge by approximately twenty-seven per cent — and that asymmetry is itself diagnostic of the underlying design failure.

For each failure mode the regulatory frame is the existing European supervisory architecture — the Insurance Distribution Directive's product-oversight-and-governance regime (Article 25), the Mortgage Credit Directive's borrower-choice provisions (Article 12), the Solvency II Standard Formula's life-mortality sub-module — and not a new regulatory mandate.

3.2 European Union — EIOPA and the credit-protection-insurance review

The European Insurance and Occupational Pensions Authority's thematic review of credit-protection insurance, published on 28 September 2022 with a parallel Article 9(3) Warning on 4 October 2022, is the most directly relevant piece of European supervisory work.¹⁴ The review examined 174 insurance undertakings and 145 banks across the European Economic Area, covering data from 2018 to 2020, and concluded that mortgage credit-protection insurance, as then sold via banks, was structurally unprofitable for the consumer: the average claims ratio across the sample was approximately twenty-six per cent of gross written premium, with the remaining seventy-four cents of each euro of premium consumed by insurer costs, distribution commissions, and profit. For single-premium credit-protection products the position was worse — sixty-five per cent of the insurers in the sample reported claims ratios below twenty per cent.

The review identified two structural design problems that bear directly on the schedule-mismatch question. The first is that the dominant distribution mechanic — group policies under which the bank is policyholder and the consumer is an insured life — constrains the insurer's ability to tailor cover to the individual borrower's circumstances. The second is that the standardised group product does not travel with the consumer: a consumer that purchased a group policy may not be able to maintain it on remortgaging and would have to purchase a new one. That re-start of cover, at older entry age and potentially higher premium, is itself a manifestation of the schedule-mismatch problem.

EIOPA issued a public Warning to manufacturers and distributors in parallel with the review, under Article 9(3) of the EIOPA Regulation, calling for fair-value products, for action on conflicts of interest arising from high commissions, and for an end to tied and pressure-sales practices.¹⁵ A follow-up summary of supervisory activities was published in April 2025.¹⁶

The legislative framework behind EIOPA's expectations is twofold. The Mortgage Credit Directive (2014/17/EU) provides the tying-and-bundling architecture: Article 12(1) prohibits pure tying, while Article 12(4) permits creditors to require the borrower to hold an insurance policy provided the borrower is free to choose an alternative insurer offering an equivalent level of guarantee; Recital 25 explicitly preserves the borrower's right to choose his own insurance provider.¹⁷ The Insurance Distribution Directive (2016/97) provides the demands-and-needs and product-oversight architecture. Article 20 requires that any insurance product proposed to a customer be consistent with that customer's demands and needs; Article 25 establishes the product-oversight-and-governance (POG) regime, requiring manufacturers to operate a product approval process that specifies an identified target market, assesses the risks to that target market, and monitors on an ongoing basis whether the product continues to meet the target market's needs.⁸ The Central Bank of Ireland's domestic implementation via the Consumer Protection Code, the Insurance Distribution Regulations 2018, and the September 2022 Differential Pricing Review applies the same architecture in the Irish jurisdiction.¹⁸

3.3 France — ACPR and assurance emprunteur

France operates the largest borrower-insurance market in the European Union, with annual premium volume of approximately €7.6 billion, and the longest record of supervisory scrutiny of the product class. The market has historically been dominated by bancassureurs — banks distributing captive group policies — who held approximately eighty-one per cent of stock in force as of 2024. The Autorité de contrôle prudentiel et de résolution has, in successive supervisory exercises, identified the same structural problems EIOPA later flagged at European level: low claims ratios, high commissions, and elevated claim-refusal rates.¹⁹

The ACPR's most detailed published data are from Vice-Président Jean-Paul Faugère's presentation at the Matinée de la protection de la clientèle on 14 March 2025.²⁰ The analysis covered six banking groups and their alternative-provider counterparts. Sinistres/primes ratios — the French regulatory analogue of the loss ratio — varied substantially: group contracts averaged thirty-nine per cent (range twenty-one to sixty); contrats défensifs averaged twenty-one per cent (range six to forty-seven); and external contracts averaged thirty-two per cent. Claim-refusal rates were elevated, with group contracts averaging a refusal rate of forty-one per cent. The ACPR's stated position is that a sustained S/P ratio below thirty per cent is presumptively problematic without management explanation.

The French legislative response to bancassurance captivity has been a sequence of laws progressively widening borrower substitution rights: the Loi Hamon (2014), the Amendement Bourquin (2017), and ultimately the Loi Lemoine (Law no. 2022-270, enacted 28 February 2022), which introduced free substitution of assurance emprunteur at any time during the mortgage term, and which abolished the medical questionnaire for loan amounts under €200,000 per insured life and maturing before age sixty.²¹ While the Loi Lemoine addresses distribution competition rather than the schedule-mismatch problem directly, it presupposes that a switching borrower can obtain a policy with a genuinely equivalent level of guarantee, which in turn requires that the sum assured at the switch date accurately represents the outstanding capital. The Association pour la promotion de la concurrence en assurance des emprunteurs (APCADE) has documented in its 2025 Baromètre that thirty-eight per cent of substitution requests still exceed the legal ten-day deadline, with some delays exceeding two months.²²

3.4 United Kingdom — FCA reviews and the Term & Health Watch data

The United Kingdom is the closest comparator to Ireland on language, regulatory architecture, life-office practice, and product taxonomy. The Financial Conduct Authority has not, to date, published a thematic review of decreasing-term mortgage protection as such. The broader record is, however, directly relevant. Thematic Review TR19/2 (general-insurance distribution chain, April 2019) found that significant commission or other payments to firms in the distribution chain significantly increased the cost of the product to the end user, with the FCA stating that its findings were relevant to all GI products.²³ Thematic Review TR24/2 (product oversight and governance in general insurance and pure protection, August 2024) found, across twenty-eight manufacturers and thirty-nine distributors covering ten product lines including pure protection, that many firms lacked effective product governance frameworks and were not fully meeting the requirements under PROD 4.²⁴ The ongoing Pure Protection Market Study MS24/1, launched in March 2025 with an interim report published in January 2026, examines term assurance alongside critical illness, income protection and whole-of-life, and identifies lower claims ratios for some protection products than others as one of three priority improvement areas.²⁵

The market-structure paper accompanying the study, MS24/1.3 (December 2025), reports that in 2024 mortgage-related term contributed £188 million in annual new premiums with top-five-firm concentration of eighty-four per cent, compared with £318 million for non-mortgage-related term with seventy-three-per-cent concentration.²⁶ The supervisory work to date does not examine whether the standard fixed-schedule decreasing-term product actually tracks actual outstanding balances. That gap is the gap this paper attempts to characterise quantitatively.

The most granular published dataset on level-versus-decreasing-term mix in the UK market is the Swiss Re Term & Health Watch 2025 report, published in May 2025.²⁷ Of 1,421,512 new term-assurance policies sold in the UK in 2024 (including those with critical-illness cover): 644,478 (forty-five per cent) were level-term without critical illness; 317,149 (twenty-two per cent) were decreasing-term without critical illness; 266,179 (nineteen per cent) were level-term with critical illness; and 140,097 (ten per cent) were decreasing-term with critical illness. The 2024 data showed decreasing-term without critical illness growing by three point six per cent year-on-year while level-term without critical illness declined by two point one per cent. The frequently-stated narrative that the UK has moved decisively to level-term cover for mortgage protection is therefore overstated by Swiss Re's data: decreasing-term remains a meaningful share of new mortgage-linked business, accounting for approximately thirty-two per cent of total new term sales by volume in 2024.

3.5 Canada — FCAC retail-banking review

The Financial Consumer Agency of Canada's Domestic Bank Retail Sales Practices Review, published 20 March 2018, examined the retail sales practices of Canada's six largest banks, with creditor insurance — including bank-channel mortgage life insurance — identified as one of the four highest-risk product and distribution categories.²⁸ Three findings are directly relevant. First, the FCAC characterised bank creditor insurance as a "sold" product rather than a "bought" product: consumers rarely initiate purchase, and banks rely on staff to cross-sell at the point of mortgage approval, with penetration targets — typically thirty per cent attachment — incentivising sales irrespective of consumer need. Second, the FCAC documented post-claim underwriting as a structural feature: underwriting at purchase is performed by assessing a handful of broadly worded yes-or-no questions, with the insurer conducting medical assessment only on death-claim submission, so that a pre-existing condition or application inaccuracy can lead to claim decline. Third, bank mortgage life insurance offered no portability. While the post-claim-underwriting concern is structurally distinct from the schedule-mismatch concern, the two compound.

3.6 United States — the shift to level-term mortgage cover

Decreasing-term mortgage life insurance was a significant individual-life product category in the United States in the 1980s and 1990s, but the dominant market direction over the past two decades has been toward level-term insurance for mortgage protection. The shift reflects consumer preference (for a policy that can be assigned to any beneficiary, provides a fixed benefit regardless of remaining mortgage balance, and is portable across lenders) and insurer commercial preference (for a standardised product amenable to systematic underwriting and reinsurance).²⁹ LIMRA reports that term life insurance represented approximately nineteen per cent of total United States individual life insurance new premium in 2024 (approximately three billion dollars in new annual premium).³⁰ The US precedent demonstrates that the market can resolve the schedule-mismatch problem at the level of product taxonomy by selecting a level benefit; it does not, in itself, vindicate the alternative resolution (a more accurate decreasing schedule) that this paper analyses.

3.7 Reinsurer published positions

The published research output of the major life reinsurers bearing directly on the schedule-mechanic question is limited. RGA has published a case study describing an RGAX partnership to build a prototype for integrating protection into a digital mortgage journey, observing that the market does not have an adequate insurance solution for customers who engage with the mortgage journey digitally.³¹ Swiss Re has confirmed in sigma 5/2024 that demand for protection products is generally less responsive to changes in interest rates and that, looking ahead, demand for risk protection will be driven by cyclical factors such as improving mortgage markets.³² Searches of the published output of Munich Re, Hannover Re, and SCOR did not identify a research note that specifically addresses the product-design question of decreasing-term schedule accuracy or mortgage-balance linkage. We record this gap candidly.

3.8 Academic literature

Peer-reviewed academic work directly on decreasing-term mortgage protection product design is sparse. Colquitt, Fier, Hoyt, and Liebenberg (2012) provide the most directly relevant empirical study, testing for adverse selection in the US credit-life market; their result is that adverse selection arises from the absence of underwriting, not from consumers outsmarting insurers.³³ Bernheim, Forni, Gokhale, and Kotlikoff (2003) quantify the mismatch between life-insurance holdings and household financial vulnerabilities using the US Health and Retirement Study; their framework — comparing the insured sum with the actual financial liability at each age — is directly analogous to the schedule-tracking problem.³⁴ Zhang, Barnard, Jumanca, Francis, and Hussain (2018) examine how regulatory intervention can improve customer outcomes in life products, including through product design, disclosure and ongoing monitoring.³⁵ Finkelstein, Einav, and Cullen (2010) provide the canonical welfare analysis of insurance-market distortions arising from imperfect pricing.³⁶ The Society of Actuaries in Ireland's Irish Insured Lives Mortality Investigation (2019) is the most recent published Irish actuarial work bearing on the pricing of the product class.³⁷

3.9 What the international evidence does and does not establish

The international and academic record establishes four propositions. First, the schedule-mismatch problem is a recognised — though under-articulated — feature of the credit-life and mortgage-protection product class across multiple jurisdictions. Second, the supervisory direction of travel is toward greater accountability for product design under the EU IDD Article 25 POG regime, the UK FCA PROD 4 sourcebook, and the equivalent Irish Consumer Protection Code architecture. Third, the modernisation candidates evaluated in Section 5 are not novel: re-rated schedules at issue (M1) are routine UK commercial practice; event-stepped schedules (M2) sit conceptually close to the French assurance emprunteur architecture under the Loi Lemoine; balance-linked schedules (M3) sit closest to the US level-term taxonomy. Fourth, no jurisdiction has yet quantified the schedule-mismatch cost at the level of granularity attempted here.

4. Data

This section specifies the data used in the deterministic and stochastic modelling. We use only first-party authoritative sources: EIOPA, the Central Bank of Ireland, the Banking and Payments Federation Ireland, the Central Statistics Office, EUR-Lex / Commission Delegated Regulations, the Society of Actuaries in Ireland, the Institute and Faculty of Actuaries / CMI Bureau, first-party Irish life-office policy schedules and key-features documents, and first-party lender Terms and Conditions pages. No commentary site, broker blog, forum, or aggregator source is used in the calibration.

4.1 Representative borrower

The representative borrower is identical to that used in The Bank Premium (MWP-2026-02), to permit direct comparability across the series.

Table 1 — Representative borrower parameters

Two reconciliation footnotes. First, the Central Bank's PDH stock figure used as the macro denominator — 697,964 — differs slightly from the 698,335 used in The Switching Gap; the difference reflects different vintages of the Residential Mortgage Arrears and Repossessions dataset and is immaterial.³⁸ Second, the six-per-cent notional rate is the contractual basis on which the cover schedule is computed at inception, and is independent of the borrower's actual mortgage rate.

4.2 Mortality basis

The assured-lives mortality curve is constructed as a three-layer multiplicative composition: a population base, a population-to-insured-lives adjustment, and a select-period adjustment.

Table 2 — Mortality stack

The IILMI Actual/Expected ratio is retained against ILT17. The CMI TMN00 table is used because it is built specifically for temporary-assurance experience and remains the canonical Irish-market reference in the absence of an SAI-published insured-lives select table. Beyond age seventy-five the curve is linearly extrapolated using the local ratio q_x / q_{x−1}; the representative policy never reaches that range during the thirty-year horizon.

4.3 Discount basis

All present-value computations use EIOPA's monthly Risk-Free Rate publication, file EIOPA_RFR_20260430.zip, reference date 30 April 2026, published 6 May 2026, EUR curve, no Volatility Adjustment.⁴¹ Headline curve points are 1-year 2.68%, 10-year 3.00%, 30-year 3.19%, with the Ultimate Forward Rate at 3.30%, the Last Liquid Point at twenty years, the Smith-Wilson α at 0.069243, and the Credit Risk Adjustment at ten basis points. The technical documentation is EIOPA-BoS-25-471 (October 2025).⁴²

4.4 Capital basis

Capital is computed on the Solvency II Standard Formula, life mortality sub-module, per Commission Delegated Regulation (EU) 2015/35 Article 137: the mortality stress is an instantaneous permanent increase of 15% in the mortality rates used for the calculation of technical provisions, applied per Article 83's scenario-based methodology and Article 105(3)(a)'s life mortality risk specification.⁴³ For a pure term assurance the premium leg cancels in the SCR difference, so the mortality SCR reduces to

SCR_mort = max(0, BEL_shock − BEL_0) = 0.15 · PV[q_x(t) · S(t)]

where S(t) is the sum at risk. The Cost-of-Capital rate is 6.0% under the current Standard Formula and falls to 4.75% from 30 January 2027 under Commission Delegated Regulation (EU) 2026/269.⁴⁴ Risk margin is explicitly omitted — a standard product-modernisation simplification — because the comparative interest is in BEL and SCR movement between designs.

4.5 Disturbance prevalences

Section 5 specifies four disturbance scenarios. The aggregate harm calculations in Section 6.7 weight the per-household disturbance integrals by published Irish prevalences.

Table 3 — Disturbance prevalences (Irish-data calibration)

The calibration is deliberately conservative on the under-insurance prevalences D2 and D3, and uses the highest-prevalence assumption available for the over-insurance disturbance D1.

4.6 Macro denominator

The Irish PDH stock used to convert household-level harms to aggregate harms is 697,964 mortgages — the Central Bank of Ireland's reported figure at the closest available vintage to the valuation date.³⁸ The deterministic full-stock aggregate is computed both on a uniform-cohort multiplier and on a duration-weighted decomposition anchored on the Bank of Ireland Asset-Covered-Securities mortgage pool at 30 June 2025.⁴⁵

5. Methodology

The deterministic core is straightforward: a representative policy is simulated month-by-month for 360 months under a fixed contractual schedule against the actual outstanding balance, the gap is integrated, and the integrals are scaled to the PDH stock. Four disturbance scenarios perturb the actual mortgage trajectory. Three modernisation candidates redefine the contractual schedule. A present-value layer recasts all flows on the EIOPA risk-free curve. The stochastic extension, detailed in Annex A, replaces the fixed actual rate with a Monte Carlo distribution of plausible rate paths.

5.1 Schedule, balance, gap

The contractual schedule is S(t) = S(0) · ((1+i)^n − (1+i)^t) / ((1+i)^n − 1) with S(0) = €322,000, n = 360 months, i = 0.06/12. The actual outstanding balance is the amortisation of the same initial principal at the actual monthly rate j = r/12: B(t) = S(0) · ((1+j)^n − (1+j)^t) / ((1+j)^n − 1). The base case is j corresponding to four per cent annual; sensitivities are run at three, three-and-a-half, four-and-a-half, and five per cent. The gap function is G(t) = S(t) − B(t), with sign convention: G > 0 is over-insurance, G < 0 is under-insurance. The integral metrics are the mean gap, the peak gap (and its month), the mean share, the nominal integral, and the present-value integral on the EIOPA discount factor.

5.2 Disturbance scenarios and the prudential-cushion role

Four disturbance scenarios are defined, each of which redefines B(t) from the disturbance month forward while S(t) is unchanged. D0 · Baseline: no disturbance. D1 · Fixed-rate roll-off: at policy year three the actual rate transits from 3.5% (fixed) to 4.75% (variable); an over-insurance disturbance; prevalence 95%. D2 · Top-up advance: at policy year seven the borrower takes a €50,000 top-up; an under-insurance disturbance; prevalence 13.1%. D3 · Term extension: at policy year five the borrower extends from twenty-five years remaining to thirty; the largest single-policy under-insurance disturbance; prevalence 8%. D4 · Lump-sum overpayment: at policy year five a €20,000 overpayment; an over-insurance disturbance; prevalence 18%.

The steady-state over-insurance also functions as a prudential cushion: if actual rates rise sharply mid-term, the balance trajectory steepens and the cushion absorbs the move. Section 6.5 quantifies this with a +200 bps upward rate spike at year five, showing the mean cushion shrinks from €14,332 to €5,032 but the policy never falls under-cover. Whether the loss of the cushion is welfare-positive is a question of household risk preference that this paper does not resolve.

5.3 Modernisation candidates

M1 · Re-rated annuity-falling. The schedule is recomputed at inception using the actual rate, rather than the six-per-cent notional. It closes the steady-state gap by construction but, because the schedule remains a curve and not a balance, it does not adjust to subsequent disturbances. M1 is the design most easily implemented inside legacy actuarial systems.

M2 · Event-stepped. The schedule is computed at inception against the actual rate, and re-set at each contractually-defined event: rate change, top-up, term extension, material lump-sum overpayment. It captures the disturbance cases without requiring continuous balance reconciliation; it is conceptually equivalent to the French assurance emprunteur substitution architecture under the Loi Lemoine.

M3 · Balance-linked. The schedule is the outstanding mortgage balance, reconciled to lender data on a periodic basis (annual in the base case). It is the conceptually clean design — the policy benefit tracks the legislative intent of section 126 — but it requires reconciliation infrastructure between insurer and lender that does not currently exist at industry scale in Ireland.

The implementation overhead for M2 and M3 is not modelled in this paper; the capital cost is quantified in Section 6.4, and the per-policy commercial budget within which any implementation cost would need to be amortised is computed in Section 8.2.

5.4 Aggregate harm constructions — HARM A and HARM B

HARM A — Premium dead-weight (over-insurance). For each rate sensitivity, the mean excess SA across the term is multiplied by the per-euro-SA premium (€0.1028 per €1,000 SA per month) to give a monthly per-household dead-weight, scaled to annual and thirty-year-cumulative metrics. The full-stock construction is reported in two parallel decompositions: a uniform-cohort multiplier and a duration-weighted decomposition anchored on the Bank of Ireland ACS pool.⁴⁵

HARM B — Coverage gap (under-insurance). The under-insurance integral ∫ max(0, B(t) − S(t)) dt is computed per scenario per design, expressed in €·years, and weighted by the disturbance prevalence.

5.5 Present-value layer

All nominal €·year integrals are recast on the EIOPA EUR risk-free term structure at 30 April 2026, no Volatility Adjustment. The PV ratio is typically approximately sixty per cent of nominal for front-loaded over-insurance integrals; for under-insurance integrals it ranges from approximately seventy per cent of nominal (D2) to approximately fifty per cent (D3, back-loaded).

5.6 Stochastic extension — EIOPA-anchored rate paths

Annex A documents a stochastic extension in which the single actual rate is replaced by a Monte Carlo distribution of plausible rate paths: 50,000 paths, 360 months each, seed 20260529, with the EIOPA EUR risk-free curve at 30 April 2026 used for discounting only. The bootstrap-Irish primary dynamics are anchored on the Central Bank of Ireland's Retail Interest Rates series (last observation 3.50%, March 2026 modal 3.49%): historical monthly rate changes are sampled with replacement and applied to the Irish spot rate. Three Vasicek calibration lenses are used as diagnostic comparators, driven by a short-rate process dr = κ(θ − r) dt + σ dW: C1 (Data-MLE, κ = 0.063, empirical θ), C2 (Reinsurance convention, κ = 0.20, θ = 2.50%), and C3 (Near-RW, κ = 0.02). A Markov-switching sensitivity (C1 + MS2) layers a two-state regime on the Irish-EUR spread. The Irish translation is floored at zero per cent; there is no driver floor. The bootstrap is the headline construction because it admits the full empirical distribution of realised Irish rate changes without imposing a parametric mean-reverting structure that would damp the extreme over-insurance tails.

6. Results

6.1 Household-level — steady-state over-insurance

Under the base-case actual rate of four per cent against the six-per-cent notional schedule, the representative single policy carries a mean excess sum assured of €14,332 and a peak of €22,292 at month 223 (year 18 and 7 months), a mean share of excess over schedule of 9.4 per cent, a nominal integral of 431,154 €·years and a present-value integral of 264,636 €·years. Table 4 reports the over-insurance summary across the rate sensitivities; the central slope is approximately €224,000 of nominal integral per 100 basis points near the four-per-cent centre.

Table 4 — Over-insurance under rate sensitivities (single representative policy, 30-year term)

The steady-state over-insurance is monotonically decreasing in the actual mortgage rate, and approximately linear in the rate spread: each 100 bps change in the actual rate moves the mean excess sum assured by approximately €7,400 near the four-per-cent centre.

6.2 Household-level — disturbance scenarios

Table 5 — Under-insurance integral (€·years) by scenario × design (single representative policy)

The largest single-policy under-insurance exposures are in D3 (term extension, €80,450 peak under-cover) and D2 (top-up advance, €38,363 peak under-cover). Table 5 reveals the central design finding: the minimal modernisation, M1, creates a substantial disturbance-side under-insurance liability that does not exist in the legacy product, because the M1 schedule is drawn against the actual rate at issue and is therefore tighter than the legacy schedule. M1 closes the steady-state over-insurance gap fully but trades it for a disturbance under-insurance gap in three of the four disturbance scenarios. M2 closes all four disturbance gaps fully; M3 closes them to within rounding (the small D2 residual is attributable to the annual reconciliation lag).

6.3 Household-level — modernisation outcomes (over-insurance side)

Table 6 — Over-insurance integral (€·years) by scenario × design (single representative policy, 4% actual)

All three modernisations substantially close the steady-state over-insurance gap (D0): M1 and M2 fully by construction; M3 leaves a small residual of approximately €26,770 €·years per year (roughly six per cent of the legacy D0 figure), a reduction of approximately ninety-four per cent rather than full closure. Under D4, M1 carries a residual over-insurance integral of 291,191 €·years because the M1 schedule does not adjust downward when the principal is paid down; M2 and M3, both event-aware, close it.

6.4 Household-level — Solvency II Standard Formula mortality SCR

Table 7 — Solvency II Standard Formula mortality charge by scenario × design (single representative policy, €)

At baseline (D0) and in three of four disturbance scenarios — D1, D2 and D4 — modernisation candidates M1, M2 and M3 each reduce SCR_mort relative to the legacy product, by approximately ten per cent (D0) to fifteen per cent (D4). The exception is D3 (term extension): the legacy product's SCR falls to €475 because the cover has collapsed below the outstanding balance; restoring cover under M2 or M3 raises the SCR back to approximately €603–€607, a twenty-seven per cent increase. That twenty-seven per cent is the price of product integrity in the D3 scenario specifically — the additional capital required to back a cover schedule that actually does what section 126 intends.

6.5 Prudential-cushion stress test — Legacy schedule under +200 bps at year 5

To quantify the prudential-cushion role we hold the 6%-notional schedule unchanged and apply a +200 bps upward shock to the actual mortgage rate at policy year five, stepping the actual rate from 4% to 6% with re-amortisation over the remaining twenty-five years. The headline results: monthly payment €1,537.28 → €1,876.47 (an increase of approximately twenty-two per cent); mean cushion €14,332 → €5,032 (a reduction of approximately sixty-five per cent); peak over-insurance €22,292 at month 223 → €8,394 at month 60; and a cushion deficit of zero across all 360 months — the cushion never falls below the post-shock balance. The cushion shrinks by approximately two-thirds but does not exhaust. The question Section 8.6 returns to is whether the cushion is the most efficient way to deliver the rare-stress insurance.

6.6 Aggregate — HARM A (premium dead-weight, deterministic)

We report two parallel aggregate constructions: a single-vintage uniform-cohort multiplier across the entire PDH denominator, and a duration-weighted decomposition anchored on the Bank of Ireland ACS pool at 30 June 2025 (weighted-average remaining term 20.71 years on an aggregate balance of €10.82 billion across 73,622 loans across seven remaining-term buckets).⁴⁵⁴⁶

Table 8 — HARM A: premium dead-weight on excess SA (4% actual base case unless stated)

The duration-weighted aggregate is approximately one-third of the uniform-multiplier construction; both are valid constructions answering different questions. The conservative duration-weighted figure of approximately €4.07 million per year is the more defensible central estimate of today's system-level annual dead-weight; the €12.34 million figure is the more relevant comparator for a new-business cohort. We lead with the duration-weighted figure as the central estimate and report both transparently. (Removing the published 30–35-year remaining-term bucket — which sits outside this paper's thirty-year frame — lowers the duration-weighted headline to approximately €3.70m, so the BoI ACS proxy is best read as a €3.7m–€4.07m range.)

6.7 Aggregate — HARM B (coverage gap, deterministic)

Table 9 — HARM B: coverage gap (under-insurance, system level)

The aggregate HARM B figure of 62 billion €·years is the cumulative under-cover exposure across the entire PDH stock under the legacy product, weighted by Irish-data-calibrated prevalences. It is fully closed by modernisation candidates M2 and M3.

6.8 Stochastic-rate headline

Table 10 — Stochastic HARM A, bootstrap-Irish headline with three Vasicek calibration lenses and the C1 + MS2 sensitivity (50,000 paths, 360 months each, seed 20260529)

On bootstrap-Irish dynamics the median household-level over-insurance harm is approximately €25.94 per household per year, with a 5–95% interval of €6.07–€44.35; the corresponding thirty-year present value across the Irish PDH stock is €345 million (interval €100m–€587m). The three parametric Vasicek calibrations plus the C1 + MS2 sensitivity bracket the headline between €13.76 and €23.60 per household per year — confirming that the headline is not an artefact of the rate-model choice. The stochastic headline on the uniform-cohort comparator is approximately forty-seven per cent higher than the deterministic uniform-cohort point estimate, consistent with the convexity of the over-insurance function in the actual rate.

7. Discussion

(i) The six-per-cent notional as an amortisation-shape shortcut. The schedule is the amortisation curve of a level-payment loan against a fixed notional rate. The choice of six per cent is not a balance-tracking choice; it is an amortisation-shape choice. The legislative intent of section 126 is balance-tracking; the schedule mechanic delivers an approximation that depended, for its accuracy, on the actual mortgage rate being close to the notional. In the historic Irish nominal-rate environment that condition was more or less met; in the 2020s it is not.

(ii) Two-sided mismatch as the mathematical consequence. A fixed S(t) is by construction a poor tracker of a stochastic B(t). The over-insurance side dominates in steady state because the notional rate sits above the actual rate; the under-insurance side dominates under disturbance because principal events move B and leave S unchanged. The two failures are the two consequences, with opposite signs, of the same root cause.

(iii) The prudential-cushion role and the welfare-ambiguity it creates. The over-insurance gap has three conceptually distinct components: a steady-state component (dead-weight under the central rate scenario), a rate-state-contingent insurance component (non-zero in rising-rate states), and an estate-benefit component (the statutory surplus on death). The welfare-relevant dead-weight is therefore smaller than the gross HARM A figure suggests. The aggressive reading is that the cushion is an inefficient way to deliver those services. We record the trade-off transparently rather than resolving it.

(iv) Why M1 alone is insufficient. M1 closes the steady-state gap but still draws a curve fixed at issue, tighter than the legacy schedule and with materially less slack to absorb disturbances. Under D1, D2 and D3 the M1 under-insurance integral is therefore materially larger than under the legacy product. M1 trades steady-state over-insurance for disturbance under-insurance in three of the four scenarios; it is a different — and in three scenarios materially worse — failure profile.

(v) Why M2 closes events without explicit balance linkage. Because the events themselves are the disturbances, re-setting the schedule at each event reconstitutes it on the new loan parameters precisely when those parameters move. The empirical magnitude of between-event drifts is small in the Irish context: M2's residual disturbance under-insurance integrals are zero across all four scenarios. M2 is operationally equivalent to balance linkage without requiring the data-exchange infrastructure that M3 demands.

(vi) Why M3 requires reconciliation infrastructure. M3 sets S(t) ≡ B(t) at each reconciliation point, which requires a data-exchange channel between insurer and lender. Such channels are routine in some markets but are not currently established at industry scale in Ireland. The marginal payoff of M3 over M2 in the Irish environment is the residual D2 reconciliation lag and the elimination of the materiality-threshold judgement on event-stepping.

(vii) Capital signature — system-level versus insurer-level reading. Read at the insurer level, the D3 result looks like a +27% capital cost of modernisation. Read at the system level, the legacy SCR_mort under D3 is low because the cover has collapsed; that low figure is an artefact of the under-insurance failure rather than a measure of legacy capital efficiency. Across the entire scenario set the picture is not "modernisation is more capital-intensive than legacy" but "modernisation prices cover correctly across all scenarios, and the legacy product under-prices capital specifically in the disturbance scenario where its cover has failed."

(viii) How international precedent maps onto the Irish modernisation question. The UK level-term shift is a taxonomic response rather than a schedule-mechanic response; the French Loi Lemoine substitution regime is structurally close to M2; the EIOPA POG architecture bears on all three modernisations equally; the Canadian FCAC post-claim-underwriting findings are orthogonal but complementary in concern. The cumulative read is that the schedule-mechanic question is the missing piece in the Irish product-design discussion.

8. A spectrum of responses

The framing is deliberately a spectrum of responses rather than a single recommendation. Each sits within the existing legal and supervisory framework; none requires a new regulatory mandate.

8.1 The response spectrum

Response 1 — Disclosure-only. Provide, alongside the premium quote, a single-page illustration of the schedule against an indicative balance trajectory under the actual rate at issue, with a sensitivity to a rate-roll-off scenario. Computationally trivial; substantively addresses the demands-and-needs assessment under IDD Article 20. It makes the existing mismatch visible to the consumer at the moment of purchase.

Response 2 — M1 re-rate, no event mechanism. Re-rate the schedule at issue against the actual rate. The minimal product change; the steady-state over-insurance is closed by construction. The principal trade-off is the disturbance under-insurance gap. For the typical Irish book with 13.1-per-cent top-up prevalence and eight-per-cent term-extension prevalence, M1 is a partial response.

Response 3 — M2 event-stepped. Re-rate at issue and re-set the schedule at each contractually-defined event. M2 closes the disturbance gaps fully. The operational uplift over M1 is modest in principle but requires the policy-administration system to recognise the event, capture the new loan parameters, and trigger the recalculation. M2 is the proportionate near-term response.

Response 4 — M3 balance-linked. Set the schedule equal to the outstanding mortgage balance at each periodic reconciliation. The conceptually cleanest design; it requires industry data-exchange infrastructure that does not currently exist at scale in Ireland. M3 is the longer-run product-architecture target.

Response 5 — Product taxonomy switch to level-term. Abandon the decreasing-term mechanic entirely and price the cover as level-term. The trade-off is that level-term breaks the decreasing-term economics of pricing against a diminishing sum at risk; the premium would be materially higher.

A reasonable manufacturer-level pathway might be: Response 1 in the next six months at near-zero cost; Response 3 (M2) over a twelve-to-eighteen-month roadmap; Response 4 (M3) as a longer-term target once industry-scale lender-data infrastructure is in place.

8.2 Break-even expense loading

The manufacturer's premium income on a modernised policy is lower than on the legacy policy by exactly the dead-weight premium on the closed over-insurance gap. The break-even expense loadings — the per-policy-per-year revenue uplift required to leave the manufacturer indifferent between the legacy and the modernised product on a single-vintage basis — are M1: €10.61, M2: €13.01, and M3: €11.87 per policy per year. Against the implicit per-policy premium of approximately €397 per year (€33.10 monthly), these correspond to approximately 2.7%, 3.3% and 3.0% of base premium respectively. The share-of-premium framing is invariant to the premium-per-mille assumption. The modernisation conversation is therefore not about an order-of-magnitude commercial impact; it is about whether a manufacturer can deliver M2 or M3 within an expense uplift of roughly €11–€13 per policy per year.

8.3 Product oversight under IDD Article 25 / Consumer Protection Code

A schedule that diverges materially and persistently from the outstanding mortgage balance, both above it in steady state and below it under common Irish-mortgage events, is on its face a candidate for manufacturer-led monitoring review under the IDD Article 25 product-oversight-and-governance regime and the Central Bank of Ireland's domestic implementation.⁸¹⁸ The point is not that the legacy product is in breach; it is that the ongoing-monitoring obligation is the natural framework within which a manufacturer would consider the modernisation candidates this paper analyses.

8.4 Solvency II Standard Formula — capital as optionality, not imposition

The proper framing of the Section 6.4 result is capital optionality. A manufacturer that modernises to M2 takes on the D3-scenario-specific capital uplift in exchange for the D0/D1/D2/D4 capital relief, the closure of the under-insurance disturbance liabilities, and the per-policy commercial budget computed in Section 8.2. The Standard Formula is not imposing modernisation; it is offering a fair price for it. The publicly disclosed solvency basis at each of the principal Irish life writers of mortgage protection is the Standard Formula on the life-mortality sub-module.⁴⁸

8.5 Interaction with MWP-2026-01 (demand) and MWP-2026-02 (supply)

The three papers in this series address three independent mechanisms of welfare-relevant cost. The Switching Gap (MWP-2026-01) identifies the demand-side problem: Irish households rarely review their cover.⁴⁷ The Bank Premium (MWP-2026-02) identifies the supply-side problem: a structural premium in lender-channel mortgage protection.¹³ The Decreasing-Term Anachronism identifies the product-design problem. The three problems are independent in mechanism but cumulative in incidence; the remedies are correspondingly independent and can be progressed in parallel, by different actors, without dependency.

8.6 What the paper does not argue — including the estate-benefit acknowledgement

The paper does not argue that the legacy product is in breach of section 126; that insurers have acted in bad faith; that policies should be replaced indiscriminately; or for a regulatory intervention on the schedule mechanic. We make the estate-benefit acknowledgement explicitly: section 126(5) provides that proceeds exceeding the amount due to the lender are payable to the surviving borrower or the estate. The steady-state over-insurance is therefore not pure dead-weight; a fraction functions as a small additional death-benefit transfer to the estate. We have not netted that estate benefit into the headline HARM A figure, on the grounds that its expected present value is small and that a household wanting estate benefit has more direct instruments. The headline HARM A figure is, in that strict sense, an upper bound on the welfare-relevant dead-weight.

9. Limitations

Single representative borrower. All household-level results are computed on a single representative borrower; the aggregate harm figures are appropriately read as central estimates rather than population means. Prevalence calibration uncertainty. The disturbance prevalences are authoritative point estimates subject to revision and vintage; the directional consequence for the headline is small. Mortality basis vintage. The mortality stack uses ILT17, IILMI (2019) and TMN00 (2005); the comparative SCR result (the twenty-seven per cent under D3) is robust to plausible select-table updates because it is driven by the schedule-balance gap rather than the underlying mortality. Discount curve forward bias. The EIOPA curve is a market-implied forward curve at a single observation date; the stochastic extension addresses rate-path uncertainty conditional on the forward curve, not mis-specification of the curve itself. Modernisation implementation cost is not modelled. The capital cost is quantified; the implementation overhead is acknowledged as an additional, market-level cost. Expense loading is an upper bound on welfare-relevant dead-weight. HARM A does not net out the prudential cushion or the statutory estate-benefit component. Standard Formula capital basis is the only basis considered. Internal Model treatments would be insurer-specific and out of scope. Stochastic floor choice. The asymmetric floor (no driver floor, zero on translation) is mildly conservative on the lower tail; the Vasicek lenses provide a robustness check. Bootstrap as headline, parametric as diagnostic. We lead with the bootstrap because it reproduces the empirical distribution of realised rate changes without embedding an undocumented mean-reversion judgement; an opposite reading (the Vasicek-central C1 figure of €23.60) is defensible. International evidence completeness. Several reinsurer and academic sources could not be confirmed and are marked candidly. Series scope. The analysis is confined to Irish principal-dwelling-house mortgage protection.

10. Conclusion

The question this paper has asked is constructive: in a contemporary Irish mortgage environment of low actual rates, short fixed-rate periods and routine principal events, what would a robust mortgage-protection schedule design look like, and what would it cost the manufacturer to deliver? The answer is encouraging for the industry. The legacy product is not broken in any catastrophic sense — the schedule is contractually clear, the premium is actuarially honest under the schedule, and the failures it exhibits are systematic, well-characterised, and tractable. The modernisation candidates close the failures, are already in commercial use in markets comparable to Ireland's, and price fairly under the Solvency II Standard Formula.

The systematic two-sided mismatch is real. In the steady-state Irish mortgage at four per cent actual against the six-per-cent notional, the schedule sits structurally above the outstanding balance, with mean over-insurance of €14,332 across the term and peak over-insurance of €22,292 mid-policy. Across the principal-dwelling-house stock the deterministic annual premium dead-weight, computed on the duration-weighted Bank of Ireland ACS cohort proxy, is approximately €4.07 million per year; on the uniform-cohort multiplier construction the same figure is approximately €12.34 million per year. On stochastic rate dynamics the household figure is approximately €26 per household per year median, with a thirty-year present value across the stock of €345 million. Under the four common Irish-mortgage disturbances the schedule remains unmoved and the policy can fall structurally below the outstanding balance, with peak under-insurance of approximately €80,000 on a five-year term extension.

We acknowledge transparently that the steady-state over-insurance also functions as a prudential cushion: under a +200 bps upward rate spike at policy year five, the mean cushion shrinks by approximately two-thirds but the policy never falls under-cover. The headline HARM A figure is therefore an upper bound on the welfare-relevant dead-weight.

The break-even expense loading required to fund M2 (approximately €13.01 per policy per year) and M3 (approximately €11.87) is modest against the implicit per-policy premium of approximately €397 per year — approximately 3.3 per cent (M2) and 3.0 per cent (M3) of base premium. At baseline and in three of four disturbance scenarios, modernisation reduces SCR_mort by approximately ten to fifteen per cent; under the term-extension disturbance specifically, M2 raises SCR_mort by approximately twenty-seven per cent — the cost of restoring cover that has structurally collapsed below the outstanding balance.

The recommendation is that Irish life offices evaluate Response 1 (disclosure) as an immediate near-zero-cost intervention; Response 3 (M2 event-stepped) as the proportionate near-term commercial modernisation; and Response 4 (M3 balance-linked) as the longer-run product-architecture target once industry-scale lender-data infrastructure is in place. The implementation can proceed within the existing product taxonomy, does not require regulatory intervention, and aligns with the supervisory direction of travel under the European Insurance Distribution Directive's product-oversight-and-governance regime and the Central Bank of Ireland's domestic implementation. Schedule modernisation is the third piece, alongside the demand-side review-trigger remedy proposed in The Switching Gap and the supply-side channel-test remedy proposed in The Bank Premium, of a coherent technical response to the structural mortgage-protection problems documented across the Mylife.ie working-paper series.

References

1. Consumer Credit Act 1995, section 126 (Ireland). Irish Statute Book. irishstatutebook.ie
2. Six-per-cent notional convention — primary policy-schedule and key-features documents of the five principal Irish life offices: Irish Life; Royal London Ireland; Zurich Life Ireland; New Ireland Assurance; Aviva Life and Pensions Ireland.
3. Central Bank of Ireland · Retail Interest Rates (monthly release). Weighted-average new mortgage agreement rate 3.52% at end-March 2026. centralbank.ie
4. Central Bank of Ireland · Governor's Blog, "The resilience of household borrowers", 8 February 2024. centralbank.ie
5. Banking and Payments Federation Ireland · BPFI Mortgage Drawdowns Report Q4 2025. Average first-time-buyer drawdown value, age, and top-up category data. bpfi.ie
6. Central Bank of Ireland · "More action is needed by lenders to resolve long-term mortgage arrears", Deputy Governor Ed Sibley, 13 July 2021. Over 95,000 (13%) of PDH mortgage accounts assessed as facing a shortfall at end of term. centralbank.ie
7. Irish lender overpayment caps (first-party Terms and Conditions, compiled): AIB (€5,000/yr on fixed-rate accounts); Bank of Ireland (10% of normal monthly repayment or €65, whichever is greater, without breakage fee on fixed-rate accounts; variable uncapped); Permanent TSB; Avant Money (10% p.a.); ICS Mortgages (20% p.a.); MoCo / Núa Money (10% p.a.). Aggregated reference: Park Financial, Mortgage Overpayments in Ireland 2026. parkfinancial.ie
8. Directive (EU) 2016/97 (Insurance Distribution Directive), Articles 20 and 25 (product oversight and governance). eur-lex.europa.eu
9. Consumer Credit Act 1995, sections 126 to 130 (Ireland), including the provisions on surplus proceeds payable to the borrower or the deceased borrower's estate. irishstatutebook.ie
10. Competition and Consumer Protection Commission (Ireland) · Mortgage Protection Insurance — consumer guidance on the borrower's right to purchase from a lender, insurer, or broker. ccpc.ie
11. Central Bank of Ireland · Residential Mortgage Arrears and Repossessions Statistics, Q2 2025. PDH stock 698,335 accounts at end-June 2025; aggregate outstanding €106 billion. centralbank.ie
12. Central Statistics Office (Ireland) · Household Finance and Consumption Survey 2023. Median outstanding mortgage balance €117,900; 30% of households with a mortgage on main residence. cso.ie
13. Mylife.ie · The Bank Premium (Working Paper MWP-2026-02). mylife.ie/research/the-bank-premium
14. European Insurance and Occupational Pensions Authority · Thematic Review on Credit Protection Insurance sold via Banks, 28 September 2022. Average claims ratio approximately 26% of GWP across 174 insurers and 145 banks. eiopa.europa.eu
15. EIOPA · Warning on Credit Protection Insurance products sold via banks, 4 October 2022 (Article 9(3) of the EIOPA Regulation). eiopa.europa.eu
16. EIOPA · Follow-up Summary of Supervisory Activities — Credit Protection Insurance, April 2025. eiopa.europa.eu
17. Directive 2014/17/EU (Mortgage Credit Directive), Articles 12 and 14; Recitals 25 and 42. eur-lex.europa.eu
18. Central Bank of Ireland · Consumer Protection Code; European Union (Insurance Distribution) Regulations 2018 (S.I. No. 229 of 2018); Differential Pricing in the Private Car and Home Insurance Markets — Final Report, July 2021. centralbank.ie
19. Autorité de contrôle prudentiel et de résolution (France) · Rapport annuel 2024 and successive Analyses et Synthèses on assurance emprunteur. acpr.banque-france.fr
20. ACPR (France) · Matinée de la protection de la clientèle, 14 March 2025; intervention of Vice-Président Jean-Paul Faugère on assurance emprunteur S/P ratios. acpr.banque-france.fr
21. Loi n° 2022-270 du 28 février 2022 pour un accès plus juste, plus simple et plus transparent au marché de l'assurance emprunteur ("Loi Lemoine"). legifrance.gouv.fr
22. Association pour la promotion de la concurrence en assurance des emprunteurs (APCADE) · Baromètre de l'assurance emprunteur 2025. apcade.fr
23. Financial Conduct Authority (UK) · Thematic Review TR19/2 — General insurance distribution chain, April 2019. fca.org.uk
24. Financial Conduct Authority (UK) · Thematic Review TR24/2 — Product oversight and governance in general insurance and pure protection, August 2024. fca.org.uk
25. Financial Conduct Authority (UK) · Pure Protection Market Study MS24/1 — Interim Report, January 2026. fca.org.uk
26. Financial Conduct Authority (UK) · MS24/1.3 — Pure protection market structure, December 2025. fca.org.uk
27. Swiss Re · Term & Health Watch 2025, May 2025. UK 2024 term-assurance sales breakdown across 1,421,512 new policies. swissre.com
28. Financial Consumer Agency of Canada · Domestic Bank Retail Sales Practices Review, 20 March 2018. canada.ca
29. LIMRA · US Life Insurance Industry Estimates, 2024 calendar-year aggregate. limra.com
30. LIMRA · US Individual Life Insurance Sales — Term Life Market, 2024 annual review. Term life approximately 19% of total individual life new premium. limra.com
31. RGA · "Life Design Sprints: How to Deliver Innovation in the Digital Insurance Customer Journey", RGA Knowledge Centre case study. rgare.com
32. Swiss Re Institute · sigma 5/2024 — World insurance: strengthening global resilience with a new lease of life. swissre.com
33. Colquitt, L. L., Fier, S. G., Hoyt, R. E., and Liebenberg, A. P. (2012). Adverse Selection in the Credit Life Insurance Market. SSRN Working Paper. papers.ssrn.com
34. Bernheim, B. D., Forni, L., Gokhale, J., and Kotlikoff, L. J. (2003). The Mismatch Between Life Insurance Holdings and Financial Vulnerabilities. American Economic Review, 93(1): 354–365. aeaweb.org
35. Zhang, J., Barnard, R., Jumanca, T., Francis, J., and Hussain, M. (2018). How might regulation improve customer experience and the operating model of life insurance products? British Actuarial Journal, 23: e21. cambridge.org
36. Finkelstein, A., Einav, L., and Cullen, M. R. (2010). Estimating Welfare in Insurance Markets Using Variation in Prices. Quarterly Journal of Economics, 125(3): 877–921. academic.oup.com
37. Society of Actuaries in Ireland, Demography Committee · Irish Insured Lives Mortality Investigation Report, February 2019. doras.dcu.ie
38. Central Bank of Ireland · Residential Mortgage Arrears and Repossessions Statistics — PDH stock 697,964 accounts at the closest published vintage to the valuation date (the variance against the 698,335 figure used in The Switching Gap reflects different release vintages and is immaterial). centralbank.ie
39. Central Statistics Office (Ireland) · Irish Life Tables No. 17, 2015–2017. Statistical Release, 16 December 2020. cso.ie
40. Institute and Faculty of Actuaries / Continuous Mortality Investigation Bureau · Working Paper 12 (2005), table TMN00 (Temporary Assurances, males non-smokers, five-year select). actuaries.org.uk
41. EIOPA · Monthly Risk-Free Rate publication, file EIOPA_RFR_20260430.zip, reference 30 April 2026, published 6 May 2026. EUR curve, no Volatility Adjustment. eiopa.europa.eu
42. EIOPA · Risk-Free Rate Technical Documentation, October 2025 (EIOPA-BoS-25-471). eiopa.europa.eu
43. Commission Delegated Regulation (EU) 2015/35 of 10 October 2014 (Solvency II Delegated Acts), Articles 83, 105(3)(a) and 137. eur-lex.europa.eu
44. Commission Delegated Regulation (EU) 2026/269 of 29 October 2025, OJ L, 18 February 2026, amending Delegated Regulation (EU) 2015/35 (Cost-of-Capital reducing from 6.0% to 4.75%); entry into force 10 March 2026; application 30 January 2027. eur-lex.europa.eu
45. Bank of Ireland Group plc · ACS Mortgage Pool Disclosure, 30 June 2025. Weighted-average remaining term 20.71 years; aggregate balance €10.82 billion; 73,622 loans across seven remaining-term buckets. investorrelations.bankofireland.com
46. The seven remaining-term buckets used for the duration-weighted decomposition are taken directly from Bank of Ireland's published ACS pool stratification table (reference 45). The balance-weighted weighted-average remaining term computed from the bucket midpoints reproduces the published 20.71 years to within six basis points.
47. Mylife.ie · The Switching Gap (Working Paper MWP-2026-01). mylife.ie/research/the-switching-gap
48. The publicly disclosed Solvency II basis at each of the principal Irish life writers of mortgage protection is the Standard Formula (Irish Life Assurance plc SFCR 2024; New Ireland Assurance Company plc SFCR 2024; Aviva Life & Pensions Ireland DAC SFCR 2024; Standard Life International DAC SFCR 2024). No public Irish life SFCR reports an approved Internal Model on the life-mortality sub-module.

About this paper

The Decreasing-Term Anachronism is Working Paper MWP-2026-03 in the Mylife.ie Working Paper Series. The paper characterises a product-design question — schedule mismatch between fixed decreasing-term policy schedules and actual mortgage balances under the contemporary Irish rate and event environment — and evaluates three modernisation candidates against the Solvency II Standard Formula. Earlier papers in the series (MWP-2026-01 The Switching Gap and MWP-2026-02 The Bank Premium) address independent dimensions of Irish mortgage-protection market structure. The present paper does not assume or require familiarity with them.

About the author

Donal Milmo-Penny, QFA FLIA, is the founder of Mylife.ie and Research Lead of the Mylife.ie Working Paper Series. He is a founding partner of SMP Financial Ltd, a financial services firm regulated by the Central Bank of Ireland (C42382). His public professional profile records twenty-five years of business experience in life assurance, pension and inheritance planning for Irish individuals, families, professionals and business owners. He has previously served as President / Chairman of the Professional Insurance Brokers Association (PIBA), as a Director of Brokers Ireland, and as a member and former Chair of Brokers Ireland's Legislation and Compliance Committee. His professional qualifications include the Qualified Financial Adviser (QFA) and Fellow of the Life Insurance Association (FLIA) designations.

Remuneration disclosure

The author is the owner of SMP Financial Ltd, whose commercial activities include the sale of life-assurance and mortgage-protection products to Irish consumers through the Mylife.ie trading channel. The firm has no direct commercial interest in any of the modernisation candidates (M1, M2, M3) evaluated in this paper, none of which is a Mylife.ie product. The author has no consulting, advisory, or remuneration relationship with any Irish life insurer, lender, or reinsurer named in this paper.

Use of AI

This paper was drafted with the assistance of an AI research-and-writing tool acting under the direct supervision of the author. All numerical results were independently verified against the locked model output documented in the companion workbook (MWP-03-Workbook.xlsx), and all source citations were verified at the URLs given in the References. The author retains full editorial and professional responsibility for the paper's content, conclusions, and any errors.

Copyright

© 2026 SMP Financial Ltd. Mylife.ie is a trading name of SMP Financial Ltd. All rights reserved. Material may be quoted in academic, regulatory, or professional contexts with attribution. SMP Financial Ltd is regulated by the Central Bank of Ireland, reference number C42382.