One relative-change function, three applications — and a broker's blend taken apart. The exercise hands you both vintages; the diff is yours.

The diff

static double RateChange(double y2026, double y2027) => (y2027 - y2026) / y2026;

• Layer 1: (0.078 − 0.075) / 0.075 = +4.00%
• Layer 2: (0.034 − 0.032) / 0.032 = +6.25%
• Subject premium: (127.5m − 120m) / 120m = +6.25%
• Premium @100%: (1,056,000 − 1,005,000) / 1,005,000 = +5.07%

Is "circa +5%" honest?

As a blend, yes — the premium total moves +5.07%. As a description, no: the increase is not uniform. Layer 2 — the remote 15m xs 10m — is repricing half again as fast as Layer 1. That usually encodes a view: the market (or this cedent's own claims, see CL-2023-001's development in Part 2) has made large losses look larger, moving the tail rate. A reply to Cartwright Re worth sending: "We read +4.0% on Layer 1 and +6.25% on Layer 2 against expiring — please confirm, and share what drives the differential."

Live and green

Implement RateChange as (new − old)/old, run it over both Alpina MTPL XL layers and the subject premium, and show the battery green: +4.00%, +6.25%, +6.25%, with the premium blend +5.07% called out underneath.

← Back to the exercise · Lesson 1 — The submission · All solutions