The depreciation of sterling after the Brexit referendum is often viewed as an exogenous shock to the economy caused by the ‘leave’ vote. In this comment on an article in the January Newsletter Giles Keating suggests the truth is more complex.

Dear Sir,

The estimates for the costs of Brexit, succinctly summarised in the January edition of this Newsletter, are essentially estimates of the impact of sterling’s 2016 depreciation, which caused a squeeze on consumer real incomes and spending that appears to have been much larger than any stimulus to net trade or corporate expenditure. As the article points out, any counterfactual discussion is uncertain, but a key issue that has seen little discussion is the question of how far the depreciation was caused by BoE policy, and to the extent that it was, whether that was an endogenous move within a policy reaction function, or an exogenous policy shock that influenced the economy independently of Brexit.

The depreciation (using daily averages of the TWI) can be divided broadly into three stages. The first was a fall of around 9 per cent that occurred within a week of the referendum result, followed by a 2 per cent recovery by June 29th for a net fall of some 7 per cent, following the pattern of overshooting and partial/full recovery often seen in market reaction to news. Up to this point, the Bank’s signalling had focussed on financial stability, with the Governor’s speech on the day of the results stressing bank capital strength and announcing emergency liquidity lines, but making only the barest hint at monetary policy changes. This is the last data point before a further speech by the Governor on 30th June, which explicitly signalled the likelihood of monetary ease, with media interpretations focused on a rate cut.

This clearly triggered the second stage of depreciation, with the pound falling immediately through key chart points and then declining further in the period up to and beyond the policy meeting on 4th August, at which the Bank went much further than cutting rates, also introducing a new round of QE and signalling likely further easing. By 15th August, the pound had fallen 5.5 per cent from the level just before the Governor’s end-June speech.

In a third stage, after a partial recovery to early September that failed to regain the key level seen before the 30th June, the pound trended down to new lows in mid-October. This period saw newly hawkish language from the Fed that caused market implied probabilities of a US December rate hike to rise to around 60 per cent, with the ECB and BoJ protecting their currencies by subtly less dovish rhetoric. The BoE did not do this and would have faced a credibility problem had it done so, because it had shifted so decisively and so recently to a forward bias toward easing. This left the pound vulnerable.

There was then a modest recovery but in the 18 months after the referendum, the pound never regained the level seen just before the Governor’s 30th June speech, and was centred around the mid-August low reached soon after the BoE’s big policy ease.

In summary, the last available data point prior to the key June 30th speech showed a post-Brexit depreciation of 7 per cent with the pound on a rising trend; after that speech, the pound averaged some 5.5 per cent lower than that level, for at least 18 months. While we will never be sure, this provides strong circumstantial evidence that at least 5.5 percentage points of the depreciation — approaching half the total — are attributable to the BoE’s policy shift.

The Bank’s decision to ease so substantially seems to have been based on three core elements: first, a view that Brexit implied a lower equlibrium real exchange rate; second, a belief that decisive action was needed to bolster confidence; and third, the sharp drop in confidence shown by business surveys published after the result and before the August meeting. The first of these is still an area of debate, while the second raises the question whether it was monetary stability (as emphasised in the Governor’s first speech) or monetary policy that were crucial for confidence. On the third, there have been historic occasions when survey responses initially weakened sharply in response to unexpected events, only to rapidly rebound. This is exactly what happened on this occasion, but since the BoE did not wait for a second round of surveys, it is very difficult to determine whether the rebound would have happened anyway, or was a response to the policy easing.

Since Brexit is a one-off event, there are major research challenges in analysing whether the Bank acted endogenously within a given reaction function, or exogenously by changing that function, But this difficulty should surely not prevent a best attempt.

Yours faithfully

Giles Keating
Werthstein and St Catherine's College, Oxford
Beethoven Str 43, Zurich 8002, Switzerland