Socius NewsOn 16 September 1992, the Bank of England raised its lending rate from 10% to 12% before lunch, announced 15% for the next morning, and suspended sterling from the European Exchange Rate Mechanism before dinner. The 15% never took effect. By the close two days later the pound had lost about 6% against the Deutsche Mark. Black Wednesday 1992 is the day a fixed exchange rate met a market that had already priced its end.
Black Wednesday 1992 is the name given to 16 September 1992, when the United Kingdom left the ERM after two interest rate rises in one day failed to hold sterling above its agreed floor of about DM 2.778. It cost the Treasury an estimated £3.3bn and ended a two-year experiment with a pegged pound. This is the eighth episode of our three centuries of markets series, and it picks up where 1971 left off: with a currency that a government promised to hold, and could not.
Why did the UK join the ERM at DM 2.95?
Sterling joined the exchange rate mechanism on Monday 8 October 1990. The Bank of England's own bulletin records the terms: a bilateral central rate of £1 = DM 2.95, inside the wider 6% band. The band put the floor at about DM 2.778. Below that level, the Bank was obliged to buy pounds with foreign currency until the price came back.
The logic was inflation. Tying the pound to the Mark imported the Bundesbank's credibility. The price was that UK rates had to follow German ones, whatever the UK economy needed. In 1992 that meant a recession at home and a central bank that could not cut.
The entry was contested from the start. A 2022 review by David Marsh for OMFIF, republished by NIESR, notes that the 1990 decision was arranged without public debate and that the Bank of England opposed cutting rates on the day of entry. A peg that your own central bank doubts is a peg the market will test.
What went wrong between June and September 1992?
The pressure built in three steps, each dated in the Bank of England's bulletins.
First, politics. On 2 June 1992 Danish voters narrowly rejected the Maastricht Treaty. The Bank's Q3 bulletin calls that vote a turning point in the summer's currency turmoil. Sterling, above DM 2.94 in mid-May, lost about 1.5 pfennigs on the Danish result, slipped to the bottom of the ERM grid in mid-June and ended the quarter at DM 2.8992. The margin of UK over German short rates was only around 75 basis points. There was little premium left for holding pounds.
Second, Germany. On 16 July 1992 the Bundesbank raised its discount rate by 75 basis points to 8.75%. Reunification deficits were feeding German inflation, and with capital controls gone, money flowed to the Mark. For the UK, with base rates at 10% since 5 May, that meant a recession with no room to ease.
Third, the second referendum. France was due to vote on Maastricht on 20 September. On 25 August a poll showed, for the first time, a majority against. Traders needed only to know that the peg had a date on it.
Over the four months to September, the Bank's Q4 bulletin records, European central banks sold Deutsche Marks equivalent to more than $160bn. On 3 September the UK announced an ECU 10bn foreign currency borrowing programme, to be sold for sterling. On 14 September the Bundesbank cut its Lombard rate to 9.5% and its discount rate to 8.25%. Markets had expected more. Over the weekend of 12 and 13 September the lira was devalued by about 7%. Sterling was next in line.
What happened on Black Wednesday 1992, hour by hour?
The Bank of England's own account is precise.
8 October 1990
Sterling joins the ERM
Central rate £1 = DM 2.95, 6% band. The floor is about DM 2.778.
2 June 1992
Denmark votes no
Danish voters reject Maastricht. Sterling slides to the bottom of the ERM grid within two weeks.
16 July 1992
Bundesbank to 8.75%
German discount rate up 75 basis points. ERM members must hold rates up to stay in the grid.
12 to 13 September 1992
Lira devalued by about 7%
The first ERM realignment of the crisis. Attention moves to sterling.
16 September, 8.00 am
Sterling at DM 2.7794
The pound opens on its ERM floor. Official sales of sterling are heavy.
16 September, about 11.00 am
Rate up from 10% to 12%
The Bank raises its minimum lending rate. Clearing banks follow. Sterling does not lift off the floor.
16 September, 2.15 pm
15% announced for the next day
A second rise is announced, effective 17 September. No clearing bank ever moves to 15%.
16 September, after 7.30 pm
Sterling suspended from the ERM
The 15% is rescinded. The EC Monetary Committee endorses the decision that night.
17 September, 9.30 am
Rate back to 10%
The lira's ERM obligations are suspended the same day. Sterling closes 18 September at DM 2.6100.
30 September 1992
New low near DM 2.50
Sterling closes at DM 2.5095, about 10% below the former floor.
Three details matter more than the headline.
The first rate rise did not work. At 8.00 am sterling was at DM 2.7794, on its floor. The move to 12% around 11.00 am did not lift it off. When a 200 basis point rise produces no price response, the price is no longer about interest rates.
The second rise was never real. The 15% announced at 2.15 pm was due the next day. It was withdrawn the same evening, and no clearing bank ever moved to that level. The announcement was a message, not a policy, and the market read it as one.
The reserves ran out of road. The Bank's Q4 bulletin records that official sales of sterling were heavy before the suspension and that sterling counterpart transfers for foreign exchange transactions in the third quarter totalled £13.1bn. Marsh's review states that the government ended 16 September with negative foreign exchange reserves. The defence was spent.
How much did Black Wednesday 1992 cost?
The number usually quoted is £3.3bn. Its source is an internal Treasury minute titled "The cost of Black Wednesday reconsidered", released under the Freedom of Information Act on 9 February 2005. Its subtitle states that "the most sensible assumptions give a loss of £3.3bn measured at February 1994". Central Banking reported the same papers that day and gave the Treasury's best estimate as $4.9bn, or £3.3bn: the value the reserves would have had with no intervention, against their actual value at later dates.
Two caveats belong next to that figure. It is an accounting loss on reserves, not a cash payment. And it depends on the date chosen to value what was left. Marsh notes that the Treasury had already given £3.3bn in December 1993, with caveats. The documented number is the one in the Treasury's own file.
DM 2.95
ERM central rate
Set at entry, 8 October 1990, 6% band
10% to 12%
First rate rise
About 11.00 am, 16 September 1992
15%
Announced, never applied
2.15 pm, rescinded the same evening
£3.3bn
Treasury loss estimate
Measured at February 1994, released 2005
£13.1bn
Q3 1992 FX outflow
Sterling counterpart transfers, Bank of England
On the other side, Marsh reports that one hedge fund manager was amassing about $10bn against sterling as the Treasury arranged its ECU 10bn loan. The profit figure attached to that name in most retellings never appeared in a document we can cite, so it stays out of this article. The position size makes the point. One private balance sheet matched the government's emergency borrowing.
Why did the UK leave the ERM instead of devaluing?
Spain devalued the peseta by 5% on 16 September and stayed in. Why did sterling leave?
Part of the answer is timing. A realignment needs the other members to agree, and that takes a weekend. Sterling hit its floor on a Wednesday morning with the French vote four days away.
Part of the answer is credibility. Marsh's fourth lesson is the one traders recognise: do not make promises you cannot keep. The Chancellor had pledged to do "whatever is necessary" to stay in. The 15% announcement was the last form that pledge could take. Once it failed, a devaluation inside the mechanism would have reopened the same test at a new floor. Suspension ended the test.
And part of the answer is what the UK needed. The Bank's 1993 annual report puts it in one line: UK base rates were cut by 4% between September and January. That is 10% to 6% in four months. No country inside the grid could have done that in late 1992.
What happened to the pound and to rates after 16 September?
The pound kept falling for two weeks and then found a level. The Bank's figures: DM 2.7784 at the close on 16 September, DM 2.6100 on 18 September, about 6% below the former floor, and DM 2.5095 on 30 September.
Sterling against the Deutsche Mark, 1990 to 1992
ERM central rate DM 2.95, floor about DM 2.778
Source: Bank of England Quarterly Bulletin 1990 Q4, 1992 Q3 and 1992 Q4 (Table A)
Rates went the other way. The minimum lending rate went back to 10% at 9.30 am on 17 September. On 22 September the Bank lent to the discount houses at 9%, a 1 point cut. Further cuts in October, November and January took base rates to 6%, the 4 point reduction the Bank's annual report describes.
Then came the part that outlived the crisis. In October 1992 the Chancellor set an explicit inflation target, initially a range of 1% to 4%. Five years later, in a lecture at the LSE on 29 October 1997, Mervyn King reported the results. UK inflation had averaged 5.2% a year in the decade before the target and 2.8% a year in the five years after it. The variance of inflation had fallen from 2.21 to 0.09. GDP growth over the same five years averaged 2.9%. King called it one of the most successful episodes of post-war UK economic performance.
The 1971 Nixon shock ended a system from the centre. Black Wednesday ended a membership from the edge. The mechanics were the same: a fixed price, a bank obliged to defend it, a market that could sell faster than the bank could buy. The difference is speed. In 1971 the decision took a weekend. In 1992 it took eleven and a half hours. In 1987, portfolio insurance showed how a rule everyone follows can sell into its own crash. In 1992, a peg everyone doubts broke in one session, and the two rate rises meant to defend it became the clearest evidence that it was over.
What your journal would have shown
Take an illustrative demo account on 16 September 1992. Illustrative demo data, not a record of any real trader.
The demo trader is long sterling against the Mark at DM 2.7810 from the 15 September close, stop just under the ERM floor at DM 2.7750, risking 1R, on the thesis that the Bank must defend the floor. At 11.00 am the rate rise arrives and the price does not move. A second unit goes on at DM 2.7790. At 2.15 pm the 15% announcement arrives and the price still does not move. A third unit goes on at DM 2.7785. The stop is never hit inside the ERM, because the Bank is buying at the floor. When the ERM stops at 7.30 pm, so does the floor. The demo trader waits for a bounce that does not come and closes all three units at the 18 September close of DM 2.6100. Against the original 60-point stop, each unit shows about -28R.
A read-only journal would not have stopped any of that. It is not built to. What it would have shown, before the next peg trade, is the shape. Three entries in one direction in one session, each placed after news that was supposed to help and did not. Risk that was 1R on paper and about 28R in the fill. In Socius Trades that day is one bar on the drawdown chart, every trade is measured in R rather than in pounds, and the add-on pattern shows under results by hour. You can see the day you bet on a promise. You can see the hour the promise broke. Plans start with a Free tier and the paid plans are listed on the pricing page.
We look. We never touch.
Trading involves risk of loss. Socius Trades is an analytics tool, not investment advice.