The 1976 IMF crisis was a severe currency crisis which led the United Kingdom to borrow $3.9 billion from the IMF, at the time the largest amount ever requested. Capie argues that the crisis ‘stands out as a low point in British economic life’.Footnote ¹ It has often been presented as a humiliation, with the British government having to bow to the IMF’s conditions to receive this loan.Footnote ²

The focus here will not be on this political crisis and the dealings with the IMF. Rather, I will focus on discussing the causes of this significant crisis. How did it start? Was it triggered by British policymakers themselves, as some have argued? Or was it accidentally triggered by them, after trying to manage exchange rates down, as they had rarely done in the past?

In this chapter I use intervention data and new archival evidence to get a picture of the Bank of England operations during the crisis. I focus on the Bank of England, and not on the political drama unfolding following the initial currency crisis. On the political crisis, many authors have done a brilliant job describing the issue, with four full monographs on the subject.Footnote ³

Despite this large body of literature, there are still unanswered questions. The most important being whether the Bank started the crisis at least somewhat voluntarily. I argue that the crisis was not ‘natural’, so to speak, and triggered by the Bank of England at the request of the Chancellor. Other accounts have shown this based on archival material. Here I use quantitative evidence to show that the dollar purchase on the first day of the crisis was different from any dollar purchase in the past. Statistically speaking, it is unlikely that this crisis was just an accident. The Bank of England wanted to devalue the pound. If the initial devaluation was a policy success, the Bank rapidly lost control of the market and sterling fell further than was desired.

During this crisis, the Treasury frequently instructed the Bank on exchange rate management goals. Instructions were sometimes unclear or contradictory. Needham even goes to say that the Bank was ‘kept on a tight leash by the Treasury in its currency market operations’.Footnote ⁴ This was not the case in other crises. In 1967, for example, the instructions were clear and never changing: the Bank was simply asked to do its best to prevent the pound from dropping in value. As we have seen in previous pages, the Bank had discretion to do so. Here, because the pound was now floating and no longer fixed, things changed. The Chancellor and the Treasury were more involved in managing the pound on a day-to-day basis.

Another contribution of this chapter is to show that the Central Bank of Nigeria had nothing to do with the 1976 IMF crisis. In many (but not all) previous accounts, it was argued that the Nigerian sterling balances were at the root of the crisis. This was indeed a rumour circulating in newspapers at the time. But evidence from the Bank’s archives shows that these were nothing more than rumours.

Was the IMF Crisis ‘Natural’ or ‘Manmade’?

Sterling devaluations described in the previous pages of this book were always either officially announced or happened without the intervention of the Bank of England, which was fighting depreciation. Here the story is a bit different. At the start, there was a view at the Treasury that sterling was overvalued. The Bank broadly agreed with this view. A public devaluation was not an option, as it would upset public sterling balance holders.Footnote ⁵ As was the case before, a devaluation would also not protect sterling from future devaluations. If there was a devaluation, it would have to look like a natural market depreciation. The Bank nudged sterling into depreciation. But that little push scared markets, unused to such a dismissive attitude from the Bank regarding its currency. This triggered a proper currency crisis. And before long, the Bank was trying to reverse a fall it had itself initiated.

Part of the literature on the topic seems to think the manoeuvre was an accident, even though a depreciation was wanted.Footnote ⁶ Or that it could not have been on instruction from the Treasury, as the Chancellor and the Prime Minister were against it.Footnote ⁷ Most accounts seem to agree that the Treasury and Bank thought the pound was overvalued, but also that the event on 4 March was an accident. The Permanent Secretary of the Treasury, Douglas Wass, met with the Governor of the Bank on 1 March. Wass explained that the Treasury had decided on a depreciation of sterling.Footnote ⁸

Burk and Cairncross argue that one of the reasons the fall in the pound could not have been initiated by the Bank was that Wilson had been against it.Footnote ⁹ The fall, they argue, was not ‘knowingly initiated’ by either the Treasury or the Bank.Footnote ¹⁰ Wilson was planning to retire and did not need a currency crisis. But as Needham points out, Chancellor Denis Healey was hiding exchange operations from Wilson and the Cabinet. And the Bank did indeed initiate the fall, as we will see. Needham writes that ‘Healey also appears to have been playing a dangerous game with his cabinet colleagues, assuring the Prime Minister that the pound would be held at $2’.Footnote ¹¹ $2 was a symbolic threshold that was broken shortly after the Bank covertly devalued the pound.

Callaghan, the incoming resident of 10 Downing Street, was also not aware of the exchange rate operations at the Bank. He later thought that it was orchestrated:

While Callaghan is right that the manoeuvre eventually got out of hand, it is not clear when. The Bank surely did not want to trigger the IMF crisis, but, I will argue, it did want to trigger a sharp fall on 4 March.

Needham argues that the idea from the Treasury and the Chancellor was that the Bank would cream off the market. The practice, which was often used, involved ‘selling sterling on a rising market to replenish the reserves’.Footnote ¹³ This would prevent sterling from appreciating too much. On this occasion, however, the creaming off was much stronger than usual. The devaluation was therefore manufactured by the Bank of England, and not an accident.

The Bank had constantly been accumulating reserves since September 1975, building up to $2,253 million by the day of the crisis. But what is really striking is the size of the purchase on 4 March. The Bank bought, on one day, just about 10 per cent of what it had taken six months to accumulate. The purchase of dollar on 4 March was fifteen times bigger than the average purchase over the previous six months. Figure 13.1 shows the distribution of the Bank’s operations for one year up to September 1975. Only one operation was a purchase of more than $100 million. The distribution shows that the operations of 4 March stand out by several orders of magnitude.

Figure 13.1. Number of interventions by amount, September 1975–September 1976

Source: Dealers’ reports (C8).

Another way to understand the exceptional nature of that fateful day is to compare it to all the operations the Bank did after the Second World War. Comparing intervention data over time can be problematic. Inflation plays a role. Because of inflation, interventions are generally larger in nominal terms as time passes. But for transparency, I compare 4 March with two samples. One sample includes potentially ‘larger’, later operations. And the other includes only potentially ‘smaller’, earlier operations. Looking at all the interventions from 1952 to 1992, 3 March is 3.5 standard deviations from the mean. It is the ninety-ninth largest purchase out of 9,586 intervention operations. These include purchases and sales. When looking at the same figures from 1952 to 1976, it is the twelfth largest dollar purchase from a total of 4,678 intervention operations. As many of these operations occurred in 1971 around the fall of Bretton Woods, reducing the sample to 1972 to 1976 makes it the second largest dollar purchase over that four-year period.

These numbers show that the operation was exceptional in nature. The purchase was unlikely to have been an accident by the dealers. They wanted sterling to depreciate on that given day. This is corroborated by the archival evidence presented by Needham that we will dive into later on.Footnote ¹⁴

Unfounded Rumours

The literature has argued that the fall of sterling on 4 March was due in part to sales from the Central Bank of Nigeria. Looking more closely at archival evidence, this is untrue. Nigeria was an OPEC member, and official sterling balance holder. It could at any point ask the Bank to convert sterling balances into dollars. But in early March 1976, it did not.

The idea possibly started with Burk and Cairncross. They quoted The Economist, which wrote that the market participants ‘feared that the Bank was either selling pounds under instructions from an overseas central bank, such as Nigeria, or that it was carrying out deliberate government policy to make sterling cheaper’.Footnote ¹⁵ But Burk and Cairncross do not take a stance on the article in The Economist and just present it as is. Most of the later literature up to the present seems to have gone with this interpretation, with only a few exceptions.Footnote ¹⁶

The narrative goes that the Nigerian Central Bank had sold sterling to the Bank of England. To rebuild reserves which had gone to Nigeria, the Bank of England had purchased dollars on the market. According to the narrative, this was in ‘accordance with normal practice’.Footnote ¹⁷ But this makes little sense. The Bank constantly tried to build reserves when market conditions allowed. It did not link its purchases to previous operations for third-party central banks. Reserves could always be rebuilt later, while a currency crisis could never be undone with a few operations. The two things had different time horizons. The focus was always on exchange rate management first, reserve building second.

Let us illustrate this with an extreme example. Imagine that one day before Black Wednesday, the Bank had sold dollars for sterling to the Nigerian central bank. On Black Wednesday, the Bank’s main concern would not be to rebuild its reserves according to the previous day’s sale, it would have been to defend the pound. This is to say, reserve management considerations came second to exchange rate considerations. Yes, the Bank wanted to hold as much reserves as possible, but never at the cost of losing control of the exchange rate. Reserves are just a tool to defend the currency, not the other way around. Needham goes further. He argues that this would go against an implicit pledge by dealers: ‘This was interpreted by currency dealers as the Bank selling sterling on a falling market, something it was pledged never to do.’Footnote ¹⁸ Here I show that the Bank wanted to devalue the currency.

The evidence that there were no sales from Nigeria is clear in the archives. First there is no record in the dealers’ report of a sale of dollars for sterling on 4 March 1976.Footnote ¹⁹ But there are on other days. For example, on 13 September 1977, the dealers’ report reads ‘Nigeria –20’. This means that the Bank sold $20 million against sterling from the Central Bank of Nigeria. Such operations are common with holders of sterling balances. Just a day following the start of the 1976 IMF crisis, one such operation happened. The Bank sold $5 million against sterling from the Reserve Bank of New Zealand. This was again recorded in the reports as ‘New Zealand –5’.Footnote ²⁰

The other piece of evidence comes from a special report by the Bank’s dealers. The report aimed to take stock of the ongoing crisis. It was written a few days after the crisis started. It reads:

Other authors have already argued that Nigerian sales played no role, but the reference seems to have stuck in the literature. In 1997, Hamon quoted a Cabinet meeting where the Chancellor said that: ‘The Bank had not intervened to bring the rate down and it wasn’t true that Nigeria was selling sterling.’Footnote ²² This we now know is untrue. The Bank did intervene. The Chancellor, however, conceded that the ‘Treasury had allowed for a fall of this size and there was nothing to worry about’.Footnote ²³ The Chancellor would have denied intentional intervention in any case. It is likely that he wanted the Bank to devalue the pound, but he did not want the Cabinet to know his intention. This is in line with what Needham argues.Footnote ²⁴ In the meeting, the Chancellor also affirmed that the Nigerian rumours were unfounded. As shown, on this last point he probably told the truth. The market slide was not a result of Nigerian sales.

Capie also bases his narrative on the dealer reports and notices that there were no sales from Nigeria. He argues it was only a ‘City theory’ following a ‘political skirmish with Britain’.Footnote ²⁵ It was true though that the Nigerian Central Bank reduced its position from £1.8 billion in early 1975 to £800 million around the time of the crisis.Footnote ²⁶ These operations happened over a longer time span. They are not the direct cause of the March 1976 crisis. Other sterling holders were also liquidating their sterling holdings. Nigeria was not specific in that regard. But no holder did it on 4 March, when our crisis was triggered.

Rate Cut and Further Fall

The following day, 5 March, the Bank lowered the Minimum Lending Rate (MLR). The rate cut made the pound drop below the symbolic threshold of $2 (Figure 13.2). The cut in itself probably did not surprise the market. Interest rates were on a downward trend. What surprised markets was that it happened the day after a large reserve purchase. This certainly helped the pound drop. It was an endorsement of the fall in the exchange rate.

Figure 13.2. Prime lending rate differential of main UK and US banks. Lower differentials were likely to put more pressure on sterling (average monthly rates, for the 15th of each month).

Source: Refinitiv, references UKBANKR and USBANKR.

The dealers were fighting a sterling depreciation following the rate cut. They reported:

Instead of encouraging the rate to fall further, the Bank started selling considerable amounts of its reserves to keep the pound from dropping too fast. They needed to look as if they were stopping the fall. Not doing so could have made things worse. Figure 13.3 shows that right after their aggressive operation to make the pound drop (the positive vertical bar pointed with an arrow), the Bank intervened to keep it from dropping. See the negative values in intervention figures following 4 March. The goal was to avoid suspicion. Needham quotes the notes from a meeting between the Bank and the Chancellor. Healey said that his ‘instinct was not to intervene further’ to make the pound drop as ‘the essence of a slide was to conceal the authorities’ hand as long as possible’.Footnote ²⁸ This policy is visible in Figure 13.3. The Bank spent large amounts of resources to stop the pound from falling over the next few days. It spent $1.35 billion in two weeks, which represented 20 per cent of its reserves.

Figure 13.3. Intervention and exchange rate

Sources: Dealers’ reports for the intervention figure and Accominotti et al., ‘Currency Regimes’, for the exchange rate.

The next Monday (8 March), pressure was still high. Sterling dropped by a further 5 per cent in the first hours of trading. It was ‘the biggest drop since the float of sterling in June 1972’.Footnote ²⁹ The dealers wrote: ‘Sterling suffered its largest ever fall in a single day.’Footnote ³⁰ They continued: ‘The severest pressure occurred in the first three hours, when widespread selling – though with no evidence of any single large sellers – propelled sterling down through 32.4% [decrease since 1971] when the morning effective rate was published … to an extreme of 33.8%.’ Intervention to defend the pound did little to help. The Bank’s dealers wrote: ‘[B]ut as the dealers [on the currency market] came to appreciate that the authorities were prepared for stronger resistance at lower levels, sterling bottomed out at 1.9287.’ Finally the Bank was gaining some control over the monster it had itself unleashed. ‘Market regained a degree of poise’ they wrote on Tuesday, 9 March,Footnote ³¹ but the respite was short-lived, and on Wednesday they reported that ‘the calm proved deceptive’.Footnote ³² The Bank spent the rest of the year trying to make the currency appreciate again. From 4 March until November, the Bank of England spent a net total of $5.5 billion on the foreign exchange market.

A few days later on 10 March, Second Permanent Secretary of the Treasury D. J. Mitchell wrote to the Chancellor. The sterling slide the Bank had manufactured was still secret. The letter reads: ‘while some suspect that the recent lurch was a deliberate act, no one can prove it … we have just had a new demonstration of how much an unplanned burst of pressure can achieve’.Footnote ³³ Burk and Cairncross cite articles in The Sunday Times by Fay and Young. The articles had hinted at the possibility of a Bank manipulation. Another article in The Economist (quoted earlier) hesitated between voluntary devaluation and Nigerian sales as a cause of the drop.Footnote ³⁴ But no one knew for sure. The Treasury was celebrating this victory while the Bank was more concerned. There was a rapid fall in the currency.

The Cat Is Out of the Bag

From 4 March onward, for the rest of the year, the Bank was busy defending the pound. The depreciation had worked too well. The Bank spent two-thirds of trading days trying to make the pound appreciate again. During only a third of the days was the Bank trying to make sterling depreciate or accumulating reserves (Figure 13.4).Footnote ³⁵ Looking at this more formally, I run a Bai–Perron break test on the Bank’s foreign exchange interventions for the year 1976. This is a way to obtain a periodisation of the data in an objective fashion. The break test works without human decision and is based on the average values of each period. The results in Table 13.1 highlight a period ‘before 4 March’. This pre-crisis period shows average daily interventions of around $24 million purchases to try to depreciate sterling. The pound was still on the rise. And then there is a period ‘after 4 March’. This second period shows interventions to try to appreciate the currency at around $67 million per day.Footnote ³⁶ The Bank went from being able to buy reserves, to being forced to sell reserves. The intervention pattern is completely reversed after the event of 4 March.

Figure 13.4. Percentage of days in a month with Bank of England sale and purchase operations (thirty-day rolling average)

Another way to understand what happened to the Bank is to analyse its behaviour on the market. Figure 13.4 shows dollar sales (meant to support sterling) and dollar purchases (meant to either devalue sterling or simply build up reserves). The crossing point in the graph was reached in March 1976, after the Bank’s undercover devaluation. From this point on, the Bank spent more days defending the currency than trying to devalue it. And this was in large part its own doing. Its 4 March operation had reversed the sterling bull run, or at least it reversed the trend more quickly than without its action.

After 4 March, the Bank completely lost control of the market. The $232 million purchase led sterling to fall. But the $1.3 billion then spent in the following eight days to potentially reverse the trend had little or no effect (see Figure 13.3). Part of it was the fact that the devaluation operation surprised the market. But also a central bank is more credible when spending domestic currency (which it has in infinite supply) than spending precious limited foreign exchange reserves.

On 23 April, the currency again was under pressure and interest rates had to be increased. The MLR was raised by 1.5 points to 10.5 per cent. It is likely that this had a positive impact on intervention, as the Bai–Perron breaks in Table 13.1 show. Interventions after the end of April 1976 were on average lower than before that date. Monetary authorities had managed to somewhat reassure markets that they no longer wanted sterling to fall. The table also shows that the IMF loan finally managed to reverse the reserve loss tendency. On average, after the IMF loan request in September 1976, the Bank was again able to build up reserves.

On 21 May, the Bank again had to raise the MLR rate to 11.5 per cent. Figure 13.5 shows that from the secret devaluation on 4 March, the Bank started to constantly bleed reserves. The small devaluation initiated by the Bank of England transformed into a full-blown currency crisis. This does not account for a counterfactual. What would have happened had the Bank not tried to devalue sterling discreetly on 4 March? The timing shows that 4 March marked a complete reversal of trends. It led the Bank to constantly lose reserves, eventually having to turn to the IMF. In fact, at the peak of the crisis, on 1 November 1976, the Bank had lost $5,539 million to defend sterling on the market. This reserve loss of $5.5 billion more than helps explain the need for the $3.9 billion financing that the IMF loan provided.

Figure 13.5. Bank of England cumulative reserve accumulation, September 1975–December 1976. September 1975 is used as a benchmark at the value of 0.

Source: Dealers’ reports (C8), author’s calculation of a cumulative sum of intervention figures.