Bank of Japan · JPY · Pass-through 40% over 9 months
Average historical spread mortgage − policy rate: 1.45pp. Current spread: 1.65pp. Above the long-run average — banks are charging a wider risk premium than usual.
Policy rate vs. typical fixed mortgage rate; the shaded area is the spread.
Where the typical fixed mortgage rate ends up if the futures-implied policy path holds and the historical spread reverts to its long-run mean.
The four tiles at the top show the live policy rate (set by Bank of Japan), the interbank rate (TONA), and the typical fixed and variable mortgage rates available to a household in Japan. The gap between the policy rate and the mortgage rate is the spread — what the lender adds on top to cover funding, credit risk and profit.
The first chart shows that spread over the last five years. When the shaded area widens, banks are charging more on top of the policy rate, usually because the long bond market has moved or because banks are pricing in extra risk. When it narrows, competition or central bank bond-buying is squeezing margins. The second chart — the implied 12-month forward path — takes the current futures market's bet on where the policy rate is heading, applies the historical spread, and shows where your mortgage rate would land if both relationships hold. It is not a forecast: it is what current market pricing already implies.
The mortgage-minus-policy spread observed for Japan (1.65pp current vs 1.45pp historical mean) decomposes into four primary drivers. First, the funding curve: jurisdictions whose lenders fund predominantly via covered bonds (Germany, Denmark, France, Sweden) inherit the swap-plus-covered-spread basis, which moved from 5-15 bp pre-2022 to 25-50 bp during the ECB's APP/PEPP unwind. Lenders funded via deposit franchise (UK, Australia) anchor more to short-rate transmission and deposit beta. US lenders sell loans into agency MBS pools, so the spread is sensitive to the primary-secondary MBS basis and to Fed SOMA reinvestment policy.
Second, prepayment optionality and convexity: products without economic prepayment penalty (US 30Y, Danish callable bonds) trade at OAS rather than nominal spread; OAS widening during rate volatility regimes (VIX-Treasury MOVE comovement) bleeds straight into the borrower rate. Penalty-protected European products (German Festzins under §489 BGB, French indemnité de remboursement anticipé) carry minimal optionality premium. Third, lender duration mismatch: if the dominant local product is short-fixed (UK 2/5Y) the lender's asset-liability gap is small and the spread is stable; if long-fixed (US 30Y, German 10Y) lenders rely on swap and MBS markets to hedge duration, and spread widens when those hedge markets stress. Fourth, regulatory caps and capital treatment: France's taux d'usure, prudential LTV/DTI floors (Switzerland, Australia, Canada), and Basel III risk-weight differentiation across LTV buckets all alter the marginal cost of lending and feed back into quoted rates with lags of one to three quarters.
| Month | Implied Policy Rate | Projected Flat 35 (35-Year Fixed) | Spread |
|---|---|---|---|
| 2026-05 | 0.30% | 1.94% | +1.64pp |
| 2026-06 | 0.30% | 1.93% | +1.63pp |
| 2026-07 | 0.30% | 1.93% | +1.62pp |
| 2026-08 | 0.30% | 1.92% | +1.62pp |
| 2026-09 | 0.30% | 1.91% | +1.61pp |
| 2026-10 | 0.30% | 1.90% | +1.60pp |
| 2026-11 | 0.30% | 1.90% | +1.59pp |
| 2026-12 | 0.30% | 1.89% | +1.59pp |
| 2027-01 | 0.30% | 1.88% | +1.58pp |
| 2027-02 | 0.30% | 1.88% | +1.58pp |
| 2027-03 | 0.30% | 1.87% | +1.57pp |
| 2027-04 | 0.30% | 1.87% | +1.57pp |
Japan is the outlier of the developed world. The Bank of Japan sets the uncollateralised overnight call rate target (the policy rate). It anchors TONA (Tokyo Overnight Average Rate), the post-TIBOR risk-free benchmark.
But the transmission to the typical Japanese mortgage is uniquely slow and partial. The reason: variable mortgages — which account for over 70% of new originations — are not priced off TONA or the policy rate directly. They are priced off the short-term prime rate (短期プライムレート), which the largest commercial banks set themselves and historically adjust only in response to large policy moves. The short-term prime stayed at 1.475% from January 2009 through early 2024 — a 15-year stretch in which the BoJ moved its policy rate multiple times without the prime budging.
The result: pass-through from the BoJ policy rate to the typical variable mortgage is roughly 0.40 over a 9-month lag — by far the lowest of the seven banks tracked here. Even after the BoJ exited negative rates in 2024, the average variable mortgage moved only ~25 bp in the following twelve months.
The snapshot table shows the typical variable mortgage (変動金利), the Flat 35 35-year fixed average, TONA, and the BoJ policy rate. Variable rates remain extraordinarily low — under 1% for many borrowers — which is the single largest reason Japan’s mortgage debt service ratio is the lowest in the OECD.
Because the short-term prime rate is sticky, the spread between variable mortgages and the BoJ policy rate has actually been negatively correlated with the policy rate in some periods: when the BoJ raised rates in 2024-2025, the spread narrowed before banks slowly repassed the move. The historical chart shows this distinctive pattern.
The Flat 35 fixed rate, by contrast, prices directly off 10-year JGB yields plus a fixed margin, and so moves more in line with market expectations than with current policy.
Modelling Japan requires acknowledging the short-term prime rate stickiness. Our forward model applies a low pass-through (0.40) and 9-month lag, which approximates the historical relationship. The implication: even if markets price further BoJ tightening, the typical variable mortgage rate is projected to move only modestly over the next year. This is in sharp contrast to Australia or the UK where the same-magnitude policy change would translate almost 1-for-1.
The major megabanks (MUFG, Mizuho, SMBC) plus the regional banks dominate origination, with non-bank lenders like Aruhi specialising in Flat 35.
See the Bank of Japan page for BoJ meeting probabilities.