Key takeaways
- Seven policy rates run from 0% at the Swiss National Bank to 4.35% at the Reserve Bank of Australia, a spread of 4.35 percentage points in August 2026.
- Three of the seven raised rates in 2026. The RBA added 0.75 percentage points across three meetings; the ECB and the Bank of Japan each moved once, both on 17 June 2026.
- Four have not changed since 2025. The Fed's last move was 11 December 2025, the Bank of England's 18 December 2025, the Bank of Canada's 30 October 2025.
- The Bank of England's own table holds 258 changes, from 11.25% in January 1975 to 3.75% in December 2025, with a high of 17.00% and a low of 0.10%.
- The 2021 to 2023 tightening lifted Bank Rate by 5.15 percentage points and the top of the Fed's range by 5.25 points, but the ECB's deposit rate by 4.50.
Central bank interest rate history, dated to the day each rate took effect
Where does each policy rate sit right now, and which banks have moved this year? Here's the table.
What follows is a central bank interest rate history for seven institutions, with every figure taken from the setting bank's own published series. As of 20 August 2026 they run from 0% in Switzerland to 4.35% in Australia. That's a spread of 4.35 percentage points between the highest and the lowest. Three of the seven have raised rates during 2026. The other four haven't moved since 2025.
| Central bank | What it sets | Rate | In effect from |
|---|---|---|---|
| Reserve Bank of Australia | Cash rate target | 4.35% | 6 May 2026 |
| Bank of England | Bank Rate | 3.75% | 18 December 2025 |
| Federal Reserve | Federal funds target range | 3.50% to 3.75% | 11 December 2025 |
| European Central Bank | Deposit facility rate | 2.25% | 17 June 2026 |
| Bank of Canada | Target for the overnight rate | 2.25% | 30 October 2025 |
| Bank of Japan | Uncollateralised overnight call rate | around 1.0% | 17 June 2026 |
| Swiss National Bank | SNB policy rate | 0% | 20 June 2025 |
The "in effect from" column is the point of the table. A policy rate quoted without a date isn't a fact about anything; it's a fact about the day someone happened to look. Two of these seven changed within the last ten weeks, and one hasn't changed in over a year.
Three banks moved in 2026 and four sat still, on the same shock
The Reserve Bank of Australia did the most. It ended 2025 with a cash rate target of 3.60%, then raised at three consecutive meetings: to 3.85% on 4 February 2026, to 4.10% on 18 March, and to 4.35% on 6 May. That's 0.75 percentage points in a quarter of a year. It then held at 4.35% in June and again in August.
The European Central Bank raised its deposit facility rate to 2.25% with effect from 17 June 2026, from the 2.00% that had stood since June 2025. On the same day the Bank of Japan's new guideline took effect. Its Policy Board had voted 7 to 1 the day before to encourage the uncollateralised overnight call rate, the rate Japanese banks charge each other for money overnight, "to remain at around 1.0 percent". It held there on 31 July, by 8 votes to 1, with the dissenting member wanting 1.25%.
The other four did nothing. The Bank of England kept Bank Rate at 3.75% at its meeting ending 29 July 2026, by 6 votes to 3, with three members preferring 4%. The Federal Reserve's own table of target-range changes records no move in 2026 at all; its most recent entry is still 11 December 2025. The Bank of Canada has held its overnight rate target at 2.25% since 30 October 2025. The Swiss National Bank left its policy rate at 0% on 18 June 2026.
What makes the split interesting is that the banks are describing the same shock. The SNB's June 2026 assessment says inflation "has risen in recent months as a result of higher energy prices", and notes that "key interest rates have been raised in the euro area, while they have remained unchanged in the US". The Bank of England's July minutes record that "financial conditions had tightened materially compared with prior to the conflict". One energy shock, seven starting points, and no common answer.
The 2021 to 2023 tightening was not the same size in every currency
Every one of these series has the same shape over the last six years: a floor, a violent climb, then a partial retreat. The heights differ more than the shape does.
| Central bank | Floor | Peak | Rise |
|---|---|---|---|
| Bank of England | 0.10%, 19 March 2020 | 5.25%, 3 August 2023 | 5.15 points |
| Federal Reserve (top of range) | 0.25%, 16 March 2020 | 5.50%, 27 July 2023 | 5.25 points |
| Bank of Canada | 0.25%, January 2022 | 5.00%, 13 July 2023 | 4.75 points |
| ECB (deposit facility) | -0.50%, 18 September 2019 | 4.00%, 20 September 2023 | 4.50 points |
| Swiss National Bank | -0.75%, to 16 June 2022 | 1.75%, 23 June 2023 | 2.50 points |
The Bank of England got there in 14 consecutive increases, starting from 0.25% on 16 December 2021 and finishing at 5.25% on 3 August 2023. The Fed needed 11, starting later, on 17 March 2022, and moving in bigger steps. The ECB's climb looks smaller only because it began below zero: its deposit facility rate had been negative since 11 June 2014, reached -0.50% in September 2019, and did not return to 0.00% until 27 July 2022. Eight years of negative deposit rates ended in a single afternoon.
Switzerland's is the odd one. The SNB's policy rate sat at -0.75% until 16 June 2022, went to -0.25% the next day, and cleared zero on 23 September 2022 with a single 0.75 point move. It peaked at 1.75% in June 2023, roughly a third of the height Bank Rate reached, and it's already back to 0%.
The retreats have been slower than the climbs, and uneven. The top of the Fed's range is 1.75 points below its 2023 peak after six cuts, which gives back about a third of what it added. The Bank of Canada's rate is 2.75 points below its own peak, which is rather more than half of its climb.
Bank Rate has been 17% and it has been 0.10%
The Bank of England publishes the longest of these series, and it's worth reading for scale. It runs from 20 January 1975, when Bank Rate stood at 11.25%, to 18 December 2025, and it records 258 separate changes across those 50 years.
The high is 17.00%, set on 15 November 1979. The low is 0.10%, set on 19 March 2020 and held for 21 months. That's a range of 16.90 percentage points inside one published table, for one country, in the working lifetime of people still holding mortgages.
Anyone assembling a central bank interest rate history from official sources runs into a practical limit here. The Federal Reserve's published archive of target-rate changes starts on 13 July 1990, at 8.00%, and its current table picks up again in 2003. So "the full Bank Rate history" and "the full federal funds rate history" cover different spans of time. That isn't a data quality problem. It's a reminder that these tables are records of what each institution chose to publish, not a natural measurement of the world.
Why the seven numbers in this central bank interest rate history aren't comparable
Here's the strongest objection to a table like the one above, and it's a good one: those seven numbers are not the same kind of number, so lining them up in a column implies a precision that isn't there.
Start with the instruments. The Fed doesn't set a rate at all; it sets a 25 basis point wide range, so any single figure for the US is a choice you made. The ECB sets three rates, and the one you quote changes the answer: since 18 September 2024 the spread between its main refinancing rate and its deposit facility rate has been 15 basis points, so the ECB is simultaneously at 2.25% and 2.40%. The Bank of Japan sets neither, strictly. It publishes a guideline for where it wants a market rate to sit, "around 1.0 percent", while its basic loan rate sits at 1.25%. The SNB pays its policy rate on sight deposits only up to a threshold, with a 0.25 point discount above it.
Then there's the more serious problem: these are nominal rates, and nothing in the table tells you what they mean in real terms. The Bank of Japan's own June 2026 statement is blunt about it. Real interest rates in Japan, it says, "have been negative, mainly in the short- to medium-term zone", even at a 1.0% policy rate, because the price stability target is 2 percent. Switzerland's 0% looks like the loosest setting in the table until you notice Swiss inflation was 0.6% in May 2026, up from 0.1% in February. A 0% nominal rate against 0.6% inflation is a different animal from the same 0% against the 3.5% Switzerland recorded in August 2022.
Finally, dates. The Bank of Japan decided on 16 June 2026 and the guideline took effect on 17 June. The SNB decided on 19 June 2025 and the rate applied from 20 June. The published tables label this differently too: the Bank of England's column is headed "Date Changed", the Federal Reserve's simply "Date". Two tables built honestly from the same sources can disagree by a day, and any comparison finer than that is noise.
What the series can't tell you is what you pay or receive
A policy rate is the price of overnight money between banks and the central bank. It isn't a mortgage rate, a savings rate or a bond yield, and the gap between them moves on its own. The Bank of England's July 2026 minutes make the point directly: financing costs faced by households and firms had risen because financial conditions tightened, while Bank Rate itself sat at 3.75% and didn't move at all.
That gap is where most of the money actually is. The fixed vs variable mortgage record shows how far the same policy path can push two borrowers apart, and the arithmetic of bond duration starts from a curve, not from a single overnight rate. A gilt or Treasury holding responds to the whole term structure, so the duration arithmetic takes Bank Rate's 5.15 point journey between March 2020 and August 2023 as an input to the repricing, not as the repricing itself. Real yields are the version of these numbers that a long-dated portfolio actually discounts by.
The other thing the table can't do is tell you where any of these rates goes next. The Fed dot plot exists precisely because the level today says so little about the level in a year.
The limitations, and what would change this table
Three limitations are worth stating plainly. First, this is a snapshot with a date on it, refreshed quarterly, and three of the seven rates changed during 2026; the dates in the table matter more than the levels. Second, the Bank of England's 258 changes are one country's record over 50 years, and a 50-year sample of one institution is not a forecast of anything. Third, the levels are nominal, and the real rate that matters to a saver or a borrower depends on an inflation rate the table doesn't carry.
There's also a conditionality built into the sources themselves that's easy to miss. The SNB's own inflation forecast, in its words, "is based on the assumption that the SNB policy rate is 0% over the entire forecast horizon". That is a conditional projection, not a commitment, and every central bank in this table publishes something similar.
What would move the numbers? The vote counts are the honest place to look, because they're published and they're specific. Three of the Bank of England's nine members voted for 4% on 29 July 2026. One of the Bank of Japan's nine wanted 1.25% on 31 July. The RBA has changed its cash rate target at three of its 2026 meetings and held at the other two. Those splits don't predict anything, but they do tell you which of these seven rates is being argued over inside the building and which is not, and a table that lists only levels hides exactly that.