Forex Institutional Research: HSBC FX Brexit

Forex Institutional Research: HSBC FX Brexit

Key quotes from the HSBC FX Brexit report: 

Beware Signal Failure

We believe that sharp price moves in GBP-USD on the day of the UK referendum will be, at best, merely a noisy signal of the outcome or totally uninformative.

In “GBP: Cable’s magic formula” 15 June 2016 we described how the price of GBPUSD has been driven by the probability of a Leave vote in the referendum on 23 June. Given this fixation, it is quite likely that the FX market will be very skittish on the day of the vote. Furthermore, liquidity in GBP pairs will be lower than usual. Therefore, we may well see sudden, sharp price moves in GBP-USD at some point during normal trading hours on 23 June.

The natural inclination of the market will be to ascribe a superhuman ‘wisdom of crowds’ intelligence to such price moves and thus interpret the price action as giving a strong signal of the referendum result. This temptation is likely to be particularly strong in this referendum given the reports of unnamed financial institutions commissioning their own, private, exit polls1 . In this piece we ask the question: to what degree should we interpret the price action on the day as an indication of the likely referendum outcome? We conclude that the prudent approach is to resist the temptation to trade in response to any such price moves since the risk-reward tradeoff is likely to be very unappealing.

Exit polls In recent UK general elections the exit polls have proven surprisingly accurate. This was particularly notable at the most recent general election in which it transpired that the opinion polls in the run up to the election were very inaccurate whereas the exit poll was very accurate. This means that the market is primed to apply a significant weight to exit poll evidence in the UK.

Somewhat surprisingly, there is no public exit poll being conducted for the UK EU referendum. There have, however, been press reports that some hedge funds have commissioned private exit polls for the UK EU referendum. As a result, sharp moves in GBP-USD on the day of the referendum are likely to cause serious consternation as many in the market may think that the moves are related to the content of these private exit polls.

Proceed with caution

We would caution against reading too much into such price action for several reasons:

  • The fact that no institution has deemed it worthwhile to conduct a public exit poll despite the intense public interest in the referendum outcome raises serious questions about the statistical power of a referendum exit poll.
  • Even if the private exit poll methodology is robust enough that this first concern can be ignored it is unlikely that an exit poll result early in the day2 could be trusted. There are stark differences in voting intentions3 from different demographic groups. If there is a difference in the time of day at which different demographic groups tend to vote then this may skew the intraday exit polls significantly; however, it would not give an accurate indication of the likely result.
  • There are other possible explanations for sharp price moves in Cable on the day of the referendum. We discuss four options in the following section; only one of these would make the price action an informative signal of the referendum result.
  • Most likely explanations for the price action do not provide an attractive risk-reward profile.

Four options

We see four plausible reasons why there might be a sharp move in GBP-USD on 23 June:

1. The private exit polls have produced evidence which, whilst close, appears “statistically significant”.

2. The private exit polls have produced overwhelming evidence.

3. The price action has been driven by non-speculative flows.

4. The price action has been driven by speculators; however, these speculators have no extra information and are simply making a call on the referendum outcome.

Whilst all four of these options are plausible explanations, only option 2 indicates an attractive trade opportunity. Sadly, this is the least likely of the four options.

1. Private exit poll result hits “statistical significance”, but signal remains quite weak It is quite likely that the private exit polls will find themselves in the situation where at some point during the day there appears to be a statistically significant difference between the two outcomes, given the evidence which has been collected so far. This will place the speculators which have commissioned the polls in a tricky situation. The significance thresholds will be calculated based on the assumption that the polling so far has been a representative sample of the voting population. However, given the worries about different demographic groupings voting at different times of day the actual confidence in this poll result is inevitably going to be much lower than the statistical significance level applied.

Market interpretation: If you believe this explanation then the risk-reward profile is not very appealing. The actual result of the referendum remains highly uncertain in this situation

2. Private exit polls have produced overwhelming evidence Clearly this is the option which is most concerning for the market. If the private exit polls have produced a vastly overwhelming signal then the funds which conducted the poll would have a significant informational advantage over the market. On the balance of probability, we feel that this option is quite unlikely; however, if you believe that this explanation is true then the market interpretation is clear.

Market interpretation: This move is one which is likely to continue; this option has an appealing risk-reward profile.

3. Not driven by speculative flows The most innocuous explanation for the price action is that it is not driven by referendum speculation. Given the anxiety in the market on the day of the vote, liquidity is obviously going to be very thin. If there are people who are compelled to trade Cable on 23 June then the market impact of their trades will be much larger than usual. This is particularly true if the market misinterprets their price impact as being informative of the referendum outcome since then their flows will be significantly magnified by speculators trading along with the momentum.

Market interpretation: If you believe this option, then the price move is an anomaly which has been driven by thin liquidity. All else being equal, one would expect such an anomaly to correct itself. However, the looming referendum outcome clearly means that all else will not be equal so the risk-reward profile here is, once again, unappealing.

4. Uninformed speculation It is entirely plausible that there are speculators who, despite having no extra information from exit polls, will still have the courage of their convictions and enter speculative positions as a result. Furthermore, there may also be traders who have no conviction about the final result but simply believe that a skittish market is likely to panic at some point during the day. If they were to rationally sell in expectation of this it might become a self-fulfilling prophecy.

Market interpretation: If you believe this option then there is no useful information in the price action and the risk-reward of trading in response to this move is unappealing.


Whilst it might initially appear to be a simple task, conducting an unbiased exit poll is surprisingly difficult. An appropriate methodology for conducting an exit poll for a UK general election has been honed over many years; however, that methodology is not possible for the referendum since it relies heavily upon returning to the same polling locations sampled in the previous election and using the change measured to estimate the swing in the vote.

Notwithstanding these concerns, it is clearly possible in theory to conduct an exit poll for a referendum. Furthermore, exit polls for referendums in other countries have proven to be surprisingly accurate. Of particular note are the exit polls conducted for French referendums on the Maastricht treaty (in 1992) and the European constitution (in 2005). These exit polls were eerily accurate – particularly when you consider how close the 1992 vote was (51% to 49%).

We believe that sharp moves in the price of Cable on the day of the referendum are quite likely. However, on balance, we believe that such price moves will be either very noisy signals of the referendum outcome or totally uninformative. The prudent approach on the day will be to resist the temptation to trade in response to any such price moves as the risk-reward trade-off is likely to be very unappealing

These notes are intended for information purposes only and are a small sample of the institutional content we post daily within our Trading Hub including full research notes, flow reports and trade desk commentary with trader views

Posted in Forex Analysis, forex institutional research, tagged with on