Part surrender vs segment surrender: same cash, different tax

Most investment bonds are written as a cluster of identical mini-policies — segments — commonly 100 or 1,000 of them. That structure gives every withdrawal two possible legal routes. A part surrender takes the cash proportionately across all segments and is taxed against the cumulative 5% allowance (IPTM3540): withdraw more than the allowance and the excess is a chargeable gain, regardless of performance. A segment surrenderfully surrenders K whole segments (IPTM3520): each is a final event on a separate policy, so the gain is each segment's share of actual growth — value minus premiums.

The same £30,000, both ways

The bond is worth less than was paid in, so each segment's final-event gain is nil — the segment route raises the cash with no chargeable event gain at all. The part-surrender route manufactures a £15,000 gain because it is measured against the allowance, not the investment. Same client, same bond, same need; the £2,400 difference is purely the mechanism.

IPTM3520 · IPTM3540 — figures engine-computed, 2026/27 config

Which route wins is case-specific

Segments do not always win. On a bond standing at a large gain, a segment surrender crystallises real growth now, while a withdrawal inside the unused 5% pool defers tax entirely — there the part surrender is the quiet route. The variables are the bond's gain position, the unused allowance, the client's marginal band this year, and how many complete years have run (more years, bigger slice relief). The point is not that one route is better; it is that the routes differ, and the difference is knowable before the instruction is sent.

The common error: the default form

Provider withdrawal forms commonly default to a part surrender across all segments, and a request phrased as “please send £30,000” will be processed that way. The chargeable event certificate arrives months later, after the tax year has crystallised. The election is made by the shape of the instruction, before payment — so the comparison has to happen at the point of request, in writing. The chargeable event gain calculator prices all three routes (full surrender, part surrender, K segments) on your figures and flags the cheapest.

A hybrid is often optimal in practice: whole segments for the bulk + a part surrender within the remaining 5% pool for the balance.