Brexit pricing precedents: an empirical study

As the Brexit referendum approaches, the market appears utterly incapable of taking its eyes off of the gaudy risk premiums on offer.  Indeed, having apparently blown out liquidity in the sterling options market (the one week forward one week implied volatility is currently around 33% mid), the market may have simply moved on to more liquid forms of volatility protection such as the VIX, as a few people have pointed out.

Wherever there are obscenely high implied vols, however, you can be sure that there are vol sellers lurking close by.  One of the first things that you learn as an option trader is that "high" levels of implied vol are not necessarily "expensive", just as low vols are not always cheap.   The speed with which you learn this lesson separates the quick from the dead.

In any event, the numbers are shiny: 2 week vols at 26% imply a straddle of 575 pips, and that 33% forward 1 week referenced above implies 520!  These numbers all seem very large, although perhaps slightly less so when one considers that the range of the past two weeks (when we've just been trading off of dodgy opinion polls) has been 610 pips!

What's a reasonable expectation of realized vol moving forwards relative to those levels of implieds?  Macro Man decided to comb the dusty library of his memory to examine a few other examples of (largely political) event risks to see what was priced...and what was realized.  Of course, it is impossible to know what sort of profitability could have been attained from being long (or short) options without knowing what sort of hedging would have been done, which is of course unprovable.

Nevertheless, we can still distill some insights into looking at market behaviour around key event risks and comparing it with option pricing.  Macro Man adopted the following methodology: assume that a 1 week straddle was purchased 3 business days before the event risk in question and held to maturity (2 business days after the event.)  Compare the spot range over the life of the option to the premium paid, as well as the closing price relative to the strike price.  (Note that he used the NY close rather than the 10 am NY expiry time for simplicity's sake; then again, he used the NY close 3 days before the event to set the strike price as well, so the maturity is truly 1 week.)

He looked at the following events, in chronological order:

* The Quebec independence referendum of 1995
* The French referendum on the EU constitution in 2005
* The Turkish Supreme Court decision on whether to outlaw the AKP party in 2008
* Kuroda's first big BOJ decision in 2013
* The Scottish referendum in 2014
* The UK general election of 2015

The results are set out in the summary table below:

There are a few general observations that we can make.  In each and every case, the spot range of the week straddling the event risk exceeded the straddle price, though that is hardly surprising.   It's not exactly going out on a limb to suggest that the range in cable will be greater than 520 pips between next Monday and the week after.  Were it to match the dampest of damp squibs (the Scottish referendum), that would represent a range of 675 pips or so on the week.   Matching the average range of just over 2.5 times the straddle price would represent a range of 13 big figures.

Perhaps the best case scenario for would-be option sellers is the first one on the list, the Quebec referendum.   Looking back on it, it's pretty remarkable how tame the run up was given the constitutional ramifications of the issue and the closeness of the vote (50.6% vs 49.4%).   Then again, USD/CAD was a super low-vol currency pair back then and the market was rather less developed than it is now.

There is a course a risk of a sampling error in this study.  Macro Man tried to pick events where there was either notable narrative and/or pricing heading into the event or a notable reaction afterwards.   So events that would look pretty tame in the rearview mirror such as Lula's election in 2003 didn't make the cut. 

By the same token, there are no examples in the study that have quite so much priced as the Brexit referendum.  It's perhaps notable, however, that the highest vol in the study- last year's UK general election- comfortably realized what had been priced into the implied vol market.

The ultimate issue, of course, is one of surprise and uncertainty.   The example above that perhaps best fits the current scenario is the AKP case, where the market was "hoping" for a certain outcome but far from certain that it would materialize.   There are important differences- the timing of the court decision was not on a preset schedule but known in generally rough terms- but the magnitude of the relief rally does perhaps provide a useful analogue for what might happen in the case of a "Remain" vote (looking at the rally relative to vol, not just in an absolute sense.)

While the EUR/USD sell-off after the French voted down the EU constitution comfortably exceeded what had been priced into markets, it's important to recognize that at 8% vol, nothing had been priced.

Ultimately you have to make your own decision on whether to buy or sell or not to play.   From Macro Man's perspective, given the closeness of the polls and the abject lack of liquidity he could easily see an 8 big figure move in either direction over the next couple of weeks, depending on the outcome of the vote.   Nothing that he observed in this study suggested to him that this is an outlandish view.

Obviously, if the polls shift dramatically and decisively towards remain over the next week, then the expected ex-post move will change....but then again, so will the ex-ante price.  Regardless of which way you're betting, hopefully the data above help you reach a more informed investment decision.

'