In normal circumstances, the pricing of currency options should reflect, to a good extent, the probability of political upsets, particularly if those upsets have obvious macroeconomic implications. Here, Mark Richardson, and Steve Cain, fund managers in Henderson's Multi-Strategy Team, examine why, in the run-up to the French presidential election, this does not appear to be the case for EURGBP (the price of euros measured in sterling).
Given the two big political shocks that occurred in 2016, it is comforting to assume that the prospect of a continuation of the populist backlash is now being fully priced into liquid assets. In this article we illustrate why we do not believe this is the case.
Our work follows on from studies we conducted prior to both the UK referendum on EU membership and the US election1. In both those cases, as in this one, our analyses were centred around currency option market-implied pricing, and our conclusions were that the market appeared to be substantially under-estimating the implied probability of both Brexit and the election of Donald Trump.
The crux of this article is that we believe that an analogous mispricing exists in the EURGBP currency pair with respect to upcoming French election.
EURGBP: EU breakup proxy
In the recent Dutch elections, the Populist Party for Freedom did not manage to replicate the Brexit/Trump upsets of 2016. While Holland is not as systemically important to the broader European project as Germany or France, it is fair to say that the Dutch election was being treated by markets – at least to some extent – as a bellwether of the degree to which the populist wave had spread to European politics. Thus, perhaps unsurprisingly, the immediate impact of the Dutch result was a classic relief rally – bonds and stocks up, volatility down – indicating the apparent reduced risk of extreme occurrences happening in the near future.
But the threat of future electoral upset, and by extension the potential for a significant repricing of European disintegration risk, remains. The elephant in the room in this regard is the forthcoming French presidential election2.
According to our analysis, the risk of a shock outcome here appears substantially under-priced. To be concrete, we look in detail at the implied market pricing of EURGBP, which, given the imminent decoupling (via Brexit) of the UK from the European Union, represents a natural expression vehicle for divergences between the two economic blocs.
The price history of EURGBP is shown in Figure 1. We have marked the date of the UK referendum with a dashed vertical line since this corresponded to a notable overnight jump in the currency pair.
Figure 1: Price history of EURGBP
Source: Bloomberg, Henderson Global Investors, as at 22 March 2017.
It is not hard to assemble a plausible macroeconomic explanation for why EURGBP responded to the Brexit shock in this way. At a high level, it represents the immediate repricing of the prospect of a UK-Europe economic decoupling, the logic here being that the UK's ejection from the European trading bloc would lead to the imposition of trade tariffs and quotas, and thus making the cost of British goods and services more expensive from the perspective of a European trading entity. The manifestation of this in the currency has to do with the associated decreased demand for sterling. Further, one can also argue that there can be an additional feedback, or second-order effect, whereby the general weakening of the UK’s economic prospects leads to incremental decreasing demand for its currency. It is plausible to suppose that these factors, or more specifically, the market's anticipation of them, combined to produce the immediate Brexit currency shock.
Now let us engage in a thought experiment. Suppose Le Pen wins the presidency. All reasonable indications are that in such a situation she would quickly move to action her stated objective of withdrawing France from the single currency and the associated institutions of the EU. In such a scenario, it is plausible that a total break-up of the European project could begin. What would happen to EURGBP in such circumstances? It does not seem unreasonable to us that the mere prospect of France leaving the EU becoming reality – even if it did not subsequently come to pass – could result in a rapid reversion of EURGBP to near its pre-Brexit levels.
Quantifying the state of the world
Let us attempt to make this more precise by attaching some figures to the chain of reasoning. The following quantities are approximate in order to illustrate the logic of the argument, but one can easily imagine augmenting the exercise as required in order to make it more granular and precise.
We'll start by fixing 0.80 as a relatively conservative and attainable figure for EURGBP to reach following a Le Pen victory (note that this is some way above average pre-Brexit levels). The question is, in the event of a Le Pen victory, what would you estimate the probability as of EURGBP finishing below this, say, two weeks after the second round vote?
We would propose that, in the event of a Le Pen win, EURGBP falling to this level or below would be a probable outcome: perhaps 60% or 70% likely. But let us factor in conservatism and say that the probability is just 40%. Such a view would be expressed by the following conditional relation:
Now let us factor in the outright probability of Le Pen winning. There are several reasonable methods of estimation that one can employ for this purpose3. But in order to obtain a simple ballpark number, we use quoted odds taken from the website Betfair. Figure 2 displays the situation reported by the Betfair exchange at the time of writing4.
Figure 2: “Le Pen Wins” betting odds
Source: Betfair, as at 23 March 2017.
As we can see, the "Le Pen Wins" price has been capped at 6 for some time (ignoring the initial low-volume "price-finding" period), which corresponds to the implied probability being floored at 20%. Note that this also represents a conservative take on matters: in reality the implied probability has floated in the 25%-35% range over the past few months. Nevertheless, adopting this figure as our estimate gives:
Assuming further (but again, not unreasonably) that the probability of EURGBP finishing below 0.80 in the event of Le Pen losing is negligible, we would then have the following simple estimate for EURGBP closing below our 0.80 target two weeks after the second round:
This is our bottom-up estimate. Keep it in mind, as we will come back to it later.
How were currency options markets pricing Brexit and Trump?
At this point it will help to recap how the options market was pricing Brexit and the US election at key points during last year. In the analysis that follows we will look at implied pricing roughly six weeks prior to the respective event, with the options in each case expiring two weeks after that. Therefore, we will in effect be examining two-month expiry options.
Figure 3 shows the situation for GBPUSD (the price of sterling measured in US dollars) leading up to Brexit. On the left is a mixture model decomposition of the GBPUSD implied distribution, roughly six weeks before the UK EU referendum, and on the right is a time series of GBPUSD taken from January 2016 to the present.
Figure 3: (left) GBPUSD implied distribution decomposition on 13 May 2016 for options expiring 8 July 2016; (right) GBPUSD time series
Source: Bloomberg, Henderson Global Investors, time series as at 22 March 2017.
The bifurcated nature of the implied distribution has the interpretation that markets were in some sense “bracing for impact” vis-à-vis Brexit. Though it is true that GBPUSD options were not pricing Brexit as a likely outcome per se (the associated implied probability in this decomposition turned out to be around 30%), there was still an embedded “implied Brexit GBPUSD level” of around 1.35, which one can see in the chart as the point about which the red “Brexit” distribution is centred. The apparent predictive ability of this decomposition was validated to some extent following the vote, whereupon GBPUSD dropped almost immediately from around 1.45 to 1.33.
Much the same can be said if we now look at USDMXN (the price of US dollars measured in Mexican pesos) implied pricing six weeks prior to the 2016 US election. Once more the implied distribution admitted a clear decomposition into two sub-distributions. Again, each of these sub-distributions (“Clinton” and “Trump” in Figure 4 below) were centred at locations which, at least in retrospect, gave credibility to the state-contingent forecasting power of the approach (i.e. as we can see in the time series chart, the immediate effect of the election result was to elicit an overnight move of USDMXN from around 18.5 to 20.5).
Figure 4: (left) USDMXN implied distribution decomposition on 19 September 2016 for options expiring 24 November 2016; (right) USDMXN time series
Source: Bloomberg, Henderson Global Investors, USDMXN time series as at 22 March 2017.
How are they now pricing the French election?
Now let us contrast this with current market-implied pricing of EURGBP options expiring two weeks after the second round of the French election (Figure 5). As in the previous cases, the implied distribution admits decomposition into two component states. However, the difference this time is that our two states are not separated to anything like the extent GBPUSD and USDMXN were in the cases of Brexit and the US election.
Figure 5: Current EURGBP implied distribution decomposition on 22 March 2017 for options expiring 21 May 2017
Source: Bloomberg, Henderson Global Investors, as at 22 March 2017.
This, in essence, is the reason we do not think that a Le Pen victory is properly being priced in EURGBP. The red "Le Pen" distribution in the chart above should, in our view, sit further to the left, centred more around the 0.80 level as per the arguments above, and consequently allocating more distributional weight to the left tail.
It is possible to make this more precise still. Earlier in the article we estimated the "Le Pen wins” contingent probability of EURGBP finishing below 0.80 two weeks after the second round vote at 8%. How does this compare with market pricing? The implied distribution puts this figure at a much lower level: approximately 3.6%5.
For the avoidance of doubt, we are not necessarily anticipating a shock outcome. Our point is rather that we find the option-implied moves of EURGBP for the weeks following the French election to be implausible relative to any sensible bottom-up scenario analysis.
If 2016 taught us anything, surely it would be that threats of political upset deserve to be taken seriously. This does not appear to be the case in some segments of the currency market.
1 Mark Richardson presented analysis at the CBOE Europe conference, September 2016.
2 The first round is scheduled for 23 April 2017 and the second round for 7 May 2017.
3 In fact, we have devised our own methodology which is detailed in the Brexit article mentioned earlier. It is also hinted at the analysis below in which we decompose option-implied distributions into sub-states.
4 Betfair adopts the convention of quoting odds in the form (n+1):1, rather than the more typically observed n:1 format. This means that a quoted price of 5 corresponds to an implied probability of 1/(5-1) = 25%.
5 Technically, one needs to take care comparing implied (that is, risk-neutral) probabilities to estimates of real world probabilities. The proper way to do this is to specify an entire belief distribution, together with the hedger's risk-preferences, and then act in such a way as to optimise the difference between what is believed and what is implied. The comparison we make here is thus meaningful up to these caveats.