Guaranteed single-premium endowment through bancassurance

Implied guarantee cost. ₹1,150,000 / ₹600,000 = 1.9167 over 15 years, i.e. 4.34% p.a. compounded. Deduct upfront commission of 3.2% (₹19,200) and per-policy expenses of ₹40,000, leaving ₹540,800 invested; the required return becomes 5.03% p.a. At the higher ₹1,180,000 guarantee it rises to 5.20%. So the incremental volume offer costs about 17 bp of required yield before any capital effect.

Net earned rate. 7.25% gross less tax of roughly 12.5% on gains, 0.20% investment management and 0.10% for default/illiquidity gives about 6.0–6.2%. That leaves a spread of only 1.0–1.2% to cover capital, the cost of the guarantee and mismatch risk - thin for a 15-year irrevocable promise.

Duration gap. Assets at 9 years against liabilities at 11.5 years mean a 100 bp fall in yields raises liabilities by roughly 12% and assets by only about 9.5%, a loss of about 2.5% of the fund. Maturing assets are reinvested at lower yields exactly when the guarantee bites.

Ranking. (i) Earned/reinvestment rate, because the spread is small and the gap is long; (ii) lapse and surrender behaviour, since a policy lapsing after the 3.2% commission is immediately loss-making, and economic lapse rises when competitors pay more; (iii) expense and commission levels. A 9.8% IRR against a 13% hurdle implies roughly a quarter of the expected value is missing. Raising the guarantee worsens both spread and capital strain, so volume alone cannot close the gap.

The economics can be repaired only by changing at least two levers at once; a single tweak leaves the IRR near 10%.

• Commission: move from 3.2% upfront to about 2.0%, with a trail of 0.25% p.a. This alone recovers roughly 25–30 bp of required yield and aligns the bank with persistency.
• Guarantee and term: reduce the maturity guarantee to ₹1.08–1.10 million, or shorten the term to 12 years, so the required return drops towards 4.5% and the duration gap narrows.
• Product structure: consider a participating product with a lower guaranteed floor plus non-guaranteed bonuses, or a unit-linked plan with a maturity floor funded by an explicit charge. Both transfer some investment risk and preserve solvency capital.
• Surrender terms: use a disclosed surrender-charge scale so that early exits do not crystallise the upfront commission loss, subject to the regulator's minimum guaranteed surrender value rules.
• Asset strategy: match duration with 15-year government securities and state development loans, and use receive-fixed swaps for the residual gap; rebuild the profit test with the matched portfolio rather than the current blended yield.

Evidence before launch: a stochastic profit test at 5th/25th/50th/75th/95th percentiles of yields; dynamic lapse testing; the cost of the guarantee under a 200 bp fall; risk-adjusted return on capital rather than IRR alone; and a signed product-filing opinion confirming compliance with the regulator's product and pricing norms. The volume commitment should be stress-tested, not treated as a benefit.

Volume does not fix unit economics. A ₹500 crore single-premium tranche at a product earning 9.8% against a 13% hurdle destroys value in proportion to size. If required capital is, say, 6% of premium, the tranche absorbs about ₹30 crore of capital at a time when the ratio is already stretched; if the product is sold at a loss, the capital is consumed faster.

Concentration and conduct. One bank channel would dominate new business, creating a single point of distribution failure and a mis-selling risk if the guarantee is marketed as a fixed deposit. It also makes the product politically difficult to withdraw; the insurer should confirm the regulator's position on withdrawal and on servicing existing policies before accepting a large commitment.

Appropriate lapse assumption. A flat 8–10% is not defensible. Dynamic lapse should be modelled as a function of the ratio of the competitor/prevailing rate to the product's implicit rate:

• When yields fall below the guarantee, the policy becomes more valuable to hold, so surrenders should fall - but liquidity-driven surrenders remain.
• When yields rise above the guarantee, rational policyholders surrender, and the insurer is left with the policies that are expensive to guarantee - adverse selection.
• Early-duration lapse should be modelled separately because of the 3.2% upfront commission; a 10% lapse in year 1 is far more damaging than 10% in year 10.

The bank's own persistency data, not industry averages, should calibrate the base rate, and the model should be stress-tested with a 20% relative increase in surrenders when the guarantee is out of the money.