When Rates Rise

Policy does not hit households in one step. Central bank decisions reprice markets; banks reprice loans; the monthly payment moves; what remains after essentials and debt service is the buffer.

Hold a hypothetical household at €4,000 monthly income and €2,000 essentials. Before a rate step-up: €1,000 debt service and €1,000 buffer (25% of income). Nothing here is a survey respondent — the point is the accounting identity.

If your monitoring stack stops at the policy rate, you are watching the first node and pricing risk as if the last two did not exist. The rate only becomes risk when it reprices the payment.

Lift debt service on that household to €1,600 and the buffer falls to €400. The payment ate the residual.

Now give three hypothetical households the same coupon shock. The high-buffer name still has capacity. The low-buffer name is already against the wall. A parallel rate move is not a parallel risk move.

Remember that picture. When we zoom to a whole book, the question becomes: which starting rooms turn the same shock into a problem account?

Zoom from three names to a population. A seeded illustrative book of 2,000 amortising mortgages (seed 42): about 35% floating by unpaid balance, the rest fixed. Thin buffer means residual cash-flow under 6% of income — a teaching cut calibrated to rhyme with published DSTI-stress intensification, not a regulatory definition. Fixed coupons hold; floaters reprice by +300bp and payments are recomputed.

Watch what happens to the thin-buffer share of unpaid balance. It rises from 27.8% → 33.3%. Newly thin accounts appear. Most of the field is still fine — that is the point. The story is not “the whole book broke.” The story is who crossed, and whether that crossing is random.

Do not cut on the thin share alone yet. Thin can include fixed-rate names that were already stressed. The next act asks a stricter question.

Keep only the balances that can still reprice and sit under the thin-buffer line: floating ∩ thin.

Watch the panel filter the field. What remains is not “everyone who felt the shock.” It is the sleeve where mechanism and exposure meet. In this calibrated book that sleeve is 21.0% of unpaid principal.

Of the accounts that newly crossed into thin, 90.4% were already in the weakest pre-shock buffer tercile. Stress landed on names that were short of room before the move.

One more check: hold the generator fixed and vary only floating mix. At +300bp, new thin balance runs 3.4% / 5.5% / 7.9% as floating share rises from 20% to 50%. Mix rivals the shock. The dial you argued about in Act 1 is upstream. The cut lives in the sleeve.

Pause the model. Does Europe after 2022 rhyme with payment-up / buffer-down / stress-up?

Published readings: CES housing costs about +10.2% (Jul 2022–Jan 2024) vs HICP +5.5%; mortgagors about +12%. WP 3053 simulations: median DSTI +6 pp; share above 40% DSTI 26% → 33%. Lower-income expected late mortgage payments ~30% in early 2024 — an early signal, not defaults.

Counterpoint you need: household debt-to-income still fell 92.8% → 87.0%. Aggregate leverage eased while pockets tightened.

So the mechanism is not “Europe blew up.” It is the same lesson as the book: averages can improve while a sleeve deteriorates. That is why the cut is a filter, not a headline rate.

You have already seen the sleeve. The panel restates the book as two parts: the majority that is not the first-loss corridor for a floater shock, and the floating × thin slice you filtered to.

That is where the watchlist starts — not on a parallel shift across every balance, and not on the overnight print alone.

Tag reset dates. Refresh buffer or DSTI at account level. Treat floating × thin as the first-loss corridor for rate scenarios. Prefer names that were already weak before the move. Use euro-area aggregates as context.

Illustrative model plus published evidence — not live-portfolio advice, not a default forecast. Methodology and sources follow.