Financial markets approach tomorrow’s UK EU referendum vote apparently pretty confident that “remain” will prevail. While I’m hopeful of a “remain” vote – I’m one of the 90%-plus off economists who firmly think that Brexit would generate significant UK/international macro costs and might not end up solving immigration concerns (if single market access is to be maintained) – financial markets look overly sanguine. Specifically, several factors suggest that the chances of “remain” winning are closer to 50% than the 75-80% currently priced in betting markets, which financial markets are taking their lead from (generating GBP spot rises and surprising falls in GBP implied volatilities) whereas opinion polls remain knife-edge tight with over 10% of voters undecided. The much-anticipated last-minute swing to the status quo risks not materialising and it’s surprising that markets haven’t raised their bar more as the result approaches. Moreover, a very narrow “remain” win could yet generate political waves. So I’m expecting relatively-bumpy financial moves overnight Thursday and into Friday morning with the likely ebb and flow of the news as results in different areas are released. But trading may well now remain light until a strong sense of the result emerges – it’s too much of a coin-flip to place significant positions – which will likely not be until early Friday (probably 4-5am). Given the recent asset price recovery, the probably bounce on a “remain” vote is unlikely to be huge. Conversely all hell could break lose on a little-priced “leave”vote: 10-20% GBP falls of are easy to envisage (co-ordinated G10 FX intervention could follow), expectations of an MPC rate cut would solidify, inflation breakevens would likely fall and (bank) equities would tank.
Financial markets continue to focus on betting markets, not directly on opinion polls
As I’ve previously detailed, traders have continued to take their cues from betting market Brexit odds. These Brexit odds have fallen to around 20-25% from a high of around 45% a week ago (see Chart). Moreover, Ladbrokes reported that 95% of recent EU referendum bets were backing “remain”. And the second chart below makes the big picture point that GBP movements, the key barometer of Brexit risks, has continued to be closely aligned with the ebb and flow of the betting market Brexit odds.
But the financial market impact over the recent fall in betting market Brexit odds has been widespread, as many of the safe haven effects I discussed here have sharply reversed. In particular:
(1) Sterling spot FX rates have rebounded, with GBPUSD reaching a 5-month high (1.47) and EURGBP back below 0.77. And while my view that GBPJPY could challenge 147 after BoJ inaction proved accurate, it has subsequently also rebounded to over 153 (despite USDJPY remaining below 105 as I expected). I discussed here how the high cost of sterling short positions could limit GBP downsides, and this may have continued to provide some GBP support.
(2) That said, near-term GBP implied volatilities have surprisingly fallen back (see Chart above) as investors have apparently become more relaxed. That said, implied vols crecent correlation with betting market Brexit odds is considerably weakly than for spot GBP (see Chart below): vols remain extremely elevated in a historic context. And it’s also notable that near-term GBP downsides implied by risk reversals have continued to set fresh record negatives (see Chart). So the story seems to be traders lacking sufficient confidence to short GBP (with likely some short-covering and illiquid conditions accounting for the bounce on lower Brexit odds) but continuing to take out significant insurance against post-referendum GBP falls. Indeed, it continues to be notable that even longer-tenor GBP risk reversals remain very negative: no significant post-referendum GBP bounceback appears priced by options.
(3) Market expectations of Bank Rate have risen (1-year OIS by 7bp, 2-year OIS by 12bp), continuing their strong correlation with betting Brexit odds. So OIS rates now imply a 30% chance of a Bank Rate cut by end-2015, down from around 40% a week ago. And the latest MPC minutes finally acknowledged that Brexit was affecting OIS rates, after strangely arguing that they hadn’t in the May Inflation Report (see here). But MPC continued to stick with the line that rates could either rise or fall on a Brexit, which the market patently doesn’t believe. That’s probably because the upward pressure on inflation from a Brexit-driven GBP fall would be more transient than the downward pressure associated with the output hit (even if potential output also fell a bit).
(4) International government bond yields have risen back from their record lows, as safe haven flows have apparently reversed: 10-year gilt yields are up a chunky 20bp over the past week, while the 9bp rise in 10-year bund yields has taken them back into positive territory and the 13bp rise in 10-year Treasuries temporarily halts the downtrend apparent over the past year. But most notable is the 20-38bp fall in 10-year Euro-Area peripheral-Bund spreads, with the largest move in Portuguese spreads. That, alongside the retracement of the spike in EA sub-investment bond spreads (see ) should reduce ECB concerns about the effectiveness of their QE/credit easing bazookas.
(5) International inflation breakevens have also rebounded a little from recent lows on lower perceived Brexit risks: this again attests to the significant international impact of Brexit risks (hit to growth and inflation pressures). The moves are biggest for the UK, with 5y5y breakevens up 10-20bp depending on the instrument (see Chart above). But EA 5y5y breakevens are also up 5bp, although they remain worryingly below 1.4%.
(6) UK and Eurozone equity prices have also been rebounded by 4-5% alongside the perceived lower Brexit probabilities, with the boost to bank stocks being even larger at a relatively-starting close to 9% rise in Eurostoxx bank equities over the past week. That said, EA bank equites remain around 20% below their start-2016 levels, with even the S&P and FTSE All Share only now just back their start-2016 levels.
The danger, of course, is that all of these more positive market developments would likely again reverse (sharply) on a surprise Brexit vote. Indeed, while a remain vote will likely see further bounces, given the 20-25% Brexit probability apparently priced, a decisive further extension of recent positive developments may well depend on a reasonably-decisive “remain” win (akin to the Scottish referendum). A very narrow “remain” win could generate political turbulence in the UK and perhaps in other European countries. Spain’s General Election re-run on Saturday is poorly timed and electorates in other EA members could yet conceivably follow the UK lead in pressurising for their own referenda (given low EU approval ratings).
Betting market odds aren’t perfect: plenty of room for a surprise Brexit vote!
So how risky is market’s apparent faith in betting market odds as a guide to the referendum outcome? On the one hand it’s completely understandable for markets to be pay substantially more attention to the betting/prediction markets than the opinion polls, given their better performance predicting the Scottish referendum and the 2015 General election. But while they’re the best thing available, betting markets are nevertheless imperfect and may not pick up all the nuances and uncertainties associated with the vote. As the usual investment disclaimer reads “past performance is no necessary guarantee of future performance“. So I’m a bit surprised that the market doesn’t seem to have raised the bar on betting odds required to reassure as the referendum nears. Specifically:
(i) While concerns about lack of liquidity in such betting markets are ameliorated by the recent spike in trade volumes (see Chart), it’s hard to know how representative they are of actual voters. My suspicion is that participants in such markets could be a little biased towards higher-income voters. If so, the bettign market odds risk not encompassing lower-income voters’ concerns about the impact of immigration on their incomes etc.
(ii) the shift in betting market odds closely coincided with “remain” regaining a small lead in the opinion polls (see Chart). In other words, the fall in betting market Brexit odds risks being an amplified version of the smaller opinion poll turnaround.
(iii) betting markets, and traders, seem to be anticipating that, as per the Scottish referendum, there will be a last-minute swing (of undecided voters) to “remain”. But unlike the Scottish referendum, there have been no last-minute bribes to change voters views and more popular newspapers are backing Brexit. While overall no side scored a knock-out blow in the TV debates, it’s nevertheless worrying that the biggest cheer in last night’s BBC debate followed Boris Johnson’s “independence day” rallying cry (on 11 June I jokingly tweeted that he could do this).
While opinion polls have their definite downsides, the pollsters claim to have improved their methods (time will tell!) and the big picture is that they continue to suggest that the result is too close to call (see Chart above). Indeed, whattheukthinks poll of polls put remain on 51% versus leave on 49%. So only a 20-25% betting market Brexit probability looks overly-sanguine, despite opinion poll’s foibles.
Moreover, the impact of several important factors are very hard to call: there is ample room for a surprise “leave” win (even if many of their arguments are fundamentally incorrect):
(a) relative turnout rates – specifically whether the “remain” campaign has sufficiently energised their supporters to actually vote, to balance off the likely high turnout of “leave” supporters. A key factor is the high turnout rates or older voters (disproportionate “leave” supporters, despite Cameron’s appeal yesterday) versus the typically lower turnout of younger voter (disproportionate “remain” supporters). So the remain camp will be hoping that the latter, and voters in general, haven’t bought the “leave” campaign’s “project fear” accusations. The worry is that some earlier polls suggested, strangely, that the “leave”campaign was slightly more trusted than the “remain” one (although both were pretty un-trusted!)
(b) the attitudes of traditional Labour Party voters: I discussed here how the Labour Party’s low profile in the debate was concerning, especially since lower-income voters will likely be most concerned about the impact of . While the Labour party has raised it’s game more recently the issue is whether it that’s too little too late and/or how many labour voter want to “teach Cameron a lesson”, even though the evidence suggests that they have more to lose from Brexit.
(c) which way the around 10% of undecided voters swing: I discussed above that the usual last-minute swing to the status quo can’t necessarily be relied upon. Of course, given the very narrow “remain” lead in the polls it could only take a small majority of undecided’s voting “leave” to tip the balance.