Blog · · EN
The quiet steepener of July 2026
Four Swiss banks, the same steep curve, the same week — and a ranking that tips for the first time in years.
In mid-July 2026, at least four Swiss banks updated their fixed-rate mortgage quotes: Bank Cler, Migros Bank, Zug Cantonal Bank, and Lucerne Cantonal Bank. All within a few days, all stamped "as at 14.07.2026" on the rate page.
When one bank shifts its curve, that's a bank-specific adjustment. When four banks do it simultaneously, it's a market event. When the curves additionally show the same steep shape — with the short end sitting well below the long end — the swap market has spoken, not the SNB, and your rule of thumb "longer duration = better deal" deserves fresh scrutiny.
The observation
LUKB gives us the cleanest before-and-after comparison, because bankproducts.db holds both the April and the July curve:
| Tenor | 2026-04-07 | 2026-07-14 | Delta |
|---|---|---|---|
| 2 years | 1.62 % | 1.53 % | −9 bp |
| 3 years | 1.68 % | 1.66 % | −2 bp |
| 4 years | 1.74 % | 1.75 % | +1 bp |
| 5 years | 1.79 % | 1.82 % | +3 bp |
| 6 years | 1.84 % | 1.88 % | +4 bp |
| 7 years | 1.89 % | 1.94 % | +5 bp |
| 8 years | 1.93 % | 2.01 % | +8 bp |
| 9 years | 1.98 % | 2.07 % | +9 bp |
| 10 years | 2.03 % | 2.13 % | +10 bp |
That's not a parallel shift — it's a steepener. The 2y–10y spread widened from 41 bp to 60 bp in three months.
The same steep shape appears at all four banks that refreshed in July:
| Bank | 2y | 5y | 10y | 2y–10y spread |
|---|---|---|---|---|
| Migros Bank | 1.40 % | 1.66 % | 2.00 % | 60 bp |
| ZGKB | 1.50 % | 1.72 % | 2.01 % | 51 bp |
| LUKB | 1.53 % | 1.82 % | 2.13 % | 60 bp |
| Cler | 1.49 % | 1.74 % | 2.05 % | 56 bp |
Absolute levels differ (Migros is around 20 bp cheaper than LUKB across the entire curve) — but the steepness is remarkably homogeneous.
Two important caveats to the interpretation:
- Only for LUKB and Cler do we have a solid Q2 baseline in
bankproducts.db. The steepener is directly proven there via a delta comparison. - For Migros and ZGKB, we had no multi-tenor curve in the DB in Q2 — we now see Q3, but can't formally quantify the shape change. The shape (steep) is consistent with a market spread that has widened.
- At Cler, two sub-products break the pattern:
nachhaltigkeitshypothek(sustainability mortgage) andsupercard_hypothek(loyalty mortgage) cut their 10y end (−16 bp and −11 bp respectively). The steepener holds for the standard fixed mortgage — the marketing sub-products are apparently being repriced with different intent.
What the steepener is not
It doesn't come from the SNB. The policy rate has stood unchanged at 0 % since 19 June 2025. The SNB meeting of 18 June 2026 didn't change that.
It comes from the swap market. LUKB documents this itself in its rate forecast of 10 July 2026: the 10-year swap sat at 0.64 % on 9 July 2026, with a 12-month forecast of 0.80 % — a 16 bp move at the long end. The bank curves are moving in exactly the same direction, only faster: they're pricing the expected steepness in today, not a year from now.
Why this matters to borrowers
The standard recommendation to mortgage customers has been the same for years: "the longer the fixation, the better the hedge — and the small premium for security is worth paying."
In a flat curve, that logic works. At a 41 bp spread over 8 tenors, the security premium costs ~5 bp per year of tenor. At 60 bp, it costs ~7.5 bp. Still not much — but the relative attractiveness of the 2-year has risen: it now sits about 60 bp below the 10-year, compared to 41 bp in April.
A borrower who signed a 5-year LUKB fixed mortgage in April at 1.79 %, and who could roll into a 2-year follow-on today at 1.53 %, saves roughly 26 bp per year over the next two years — provided the refinancing after those two years doesn't land at 3 %.
That's exactly the risk. LUKB's swap forecast (12 months, +16 bp at the long end) implies a further-steepening curve short term, but not dramatically higher levels. Which means: the 2-year is today's best price at the short end, but possibly the worst time to sign a two-year fixation, if you're going to roll into an expected mildly-higher rate environment in 24 months.
The rankings tip
On bankproducts.info we run four quarterly rankings for fixed mortgages:
best-festhypothek-2ybest-festhypothek-5ybest-festhypothek-10y
Before July, the 10y rankings were the crowd favourite — borrowers clicked the longest tenor because it "sold security". After July, the relative attractiveness looks like this:
- Migros Bank 2y at 1.40 % — among the cheapest standard fixed-mortgage rates in our dataset for this tenor (only LLB CH goes lower at 1.36 %; below that are only specialised eco, sustainability, and loyalty products such as ZKB Umweltdarlehen at 0.72 % or Cler Nachhaltigkeitshypothek at 1.24 %)
- LUKB 10y at 2.13 % — 60 bp more expensive than its own 2-year, and 73 bp more expensive than the cheapest provider (Migros) at 2 years
If you renew your mortgage today and pick the 10-year purely because "safer is better," you pay the full steepener premium. If you pick the 2-year, you buy the steepest discount we've seen in our coverage so far (Q2 and Q3 2026) — with refinancing risk in 24 months.
The choice is ultimately the same as always (rate lock vs. rate risk), but the price differential is now at a level that makes the "longer is always better" heuristic honestly testable.
What you can't see on a single bank's website
LUKB documents its own rate development transparently. But LUKB documents only LUKB. That four independent banks publish the same steep curve shape within a single week — and that the two with a clean Q2 baseline (LUKB, Cler standard fixed) even show the same steepener move — is only visible if you place several curves side by side. That's exactly what bankproducts.db is for.
If you want to slice the dataset or the rankings yourself: the SQLite DB and JSON snapshots are open on github.com/gbits-io/bankproducts_info, and the Q3 snapshot tables (fee_snapshot, interest_rate_snapshot) make a quarter-over-quarter comparison possible in a single SQL query.