The workings behind the Plan 2 threshold freeze analysis

With the freeze - the path every headline figure on this site runs on - the Plan 2 repayment threshold stays at £29,385 until April 2030, then rises with simulated inflation (RPI) each April. The income levels that set Plan 2's interest rate move with the threshold, so the freeze holds those still too.

Without the freeze, we assume the threshold would have kept rising with RPI each April from April 2027, and would have carried on that way after 2030. The two paths therefore never re-converge: the without-freeze threshold stays permanently higher.

That second path is our assumption of how the threshold would have moved absent the freeze, not an announcement anyone made. In practice the government sets the threshold year by year: the rises in April 2025 and April 2026 followed RPI, which is why RPI continuation is the counterfactual used here, but in earlier years the threshold was often raised in line with average earnings instead. If you read the without-freeze column as 'the world where the last announced uprating basis had simply continued', you are reading it as we built it.

Every figure comes from the same simulation engine as the student loan overpayment calculator, run outside the browser on the same statutory figures. Each scenario simulates 10,000 possible futures for inflation and salary, steps one year at a time, adds interest before taking that year's repayments, and runs to Plan 2's 30-year write-off.

Each future is priced twice - once under each threshold path - on identical inflation and salary draws. The difference between the two runs is caused by the threshold policy alone; nothing else varies. Plan 2's interest is capped at 6% for the 2026/27 year only, as announced, with the normal rule resuming afterwards.

Stages of the repayment journey are set by the year repayments started: early = 2025 (29 years left to write-off), mid = 2020 (24 left), late = 2016 (20 left - the earliest possible Plan 2 start). Salary means today's gross salary; by default it then grows with each future's inflation plus 2% a year, with year-to-year randomness.

The lifetime-extra table fixes the balance at £45,000 and varies salary (£25,000 / £50,000 / £75,000) and stage. The clearing-chance and overpayment-return grids are mid-journey, at salaries £50,000 to £125,000 and balances £30,000 / £50,000 / £70,000; the return grid evaluates £100 a month of voluntary overpayments. The boundary map searches salaries £25,000 to £125,000 against balances £10,000 to £80,000, in £5,000 steps, mid-journey.

Middle-of-the-road outcomes (medians) are computed on paired paths: each simulated future's difference between the two threshold runs is taken first, and the middle value (median) of those differences is what the tables show. That is not the same as subtracting one path's middle outcome from the other's, and it is the convention the calculator itself uses.

Today's money strips each future's own simulated inflation out of every cashflow, so amounts paid in different decades can be added honestly. Cash terms sums the pounds of the years they are paid, with no adjustment - the two are shown separately and are not comparable with each other.

Returns on overpayments follow the calculator's convention: the money paid in and the repayments it saved are lined up year by year on the same future, and the result is glossed as working out like interest of X% a year on the money paid in. A future where the loan was written off anyway counts as -100% - the whole overpayment was lost - and such futures are included, never hidden.

Every figure on both pages comes from one set of results, produced in a single run of the calculator's own engine. No number is typed in by hand, and nothing is recalculated in your browser: the pages are built from those results and then published as fixed text.

The run is seeded, so repeating it produces exactly the same figures - not similar ones, the same ones. Whenever the model or the parameters change, the dataset is regenerated and the pages rebuild from it, so the two can't drift apart.

The engine also has a check outside itself. Its year-by-year core is reconciled against the Department for Education's published student loan forecast method, and agrees to the penny wherever the two make the same assumptions. Where they differ - the postgraduate threshold basis, most notably - the difference is measured and written down rather than smoothed over.

The dataset behind these pages and the full reconciliation working are available on request. Email [email protected] and we'll send them over.

It cannot tell you what inflation, salaries or future governments will do - the futures here are simulated, and real life will not follow any of them exactly. It models one loan at a time, steps annually rather than monthly, and assumes announced policy runs as announced: no further freezes, no cap renewals, no early changes.

And it reports outcomes for example borrowers, not for you. For your own numbers, the calculator runs the same comparison on the details you enter.

To run it on your own loan, use the free student loan overpayment calculator - the Plan 2 threshold comparison card shows the freeze's effect on your inputs.