9 min read 2 Dec 20
Summary: The term ‘episode‘ doesn’t lend itself to marketing: it doesn’t translate well into European languages and if you type it into a search engine it will bring up more about television shows than investing.
The concept itself is also a challenge, behavioural finance doesn’t give the comfort of sounding scientific, and deciding what is, and is not, an episodic opportunity can seem highly subjective.
A good example of this is the Turkish lira. Despite a sharp rally in recent weeks (after changes to policy rhetoric, personnel and interest rates), the currency remains weaker versus the US Dollar than it was in the middle of 2018.
And yet, our view is that the opportunity was more episodic in 2018. This is indeed a partly subjective assessment, but by looking at some of the key elements behind this view we can shed some light on both the situation in Turkey itself, and the process of episodic investing.
One area that can be looked at quantitatively is the speed of moves. It is not evident in the long term charts above, but the extent of short term moves in 2018 was greater in 2018 than more recently.
By contrast the more recent pattern of price moves has been one of steady deterioration (excluding a period of currency intervention). This might suggest that market participants are taking more time to consider fundamental developments from a longer term perspective, rather than simply panicking at near term losses:
This is not to suggest that simply because moves are rapid, that they must be ‘behaviourally driven.’ A rapid move can be entirely justified if there is an equally rapid shift in fundamentals.
However, when you see 14% declines in a single day, followed by 8% the next, as we did in Turkey in August 2018, you have a higher than usual chance that investors are ignoring any kind of long term view. Whether it be human fear that takes hold, or forced selling driven by risk models, the actions of myopic market participants can frequently enhance the returns on offer for those with even a modestly longer time horizon.
As I noted in 2018 the Lira weakness drew huge amounts of attention. Not only were people talking about the moves themselves, but events in Turkey were being used to explain market moves elsewhere.
Looking at our machine learning-based news analysis below we trace can the pattern of commentary 2018. The dots show Turkey related news stories and the colours reflect articles that the algorithm sees as related (we have tried to roughly characterise these by giving them topic names):
As the currency weakened (shown in the top right) the number of stories and the overlap of different topics increases dramatically, as shown by the lower scatter plot.
At the same time, sentiment on the upper scatter plot becomes almost uniformly negative (as shown by the prevalence of red dots). This resonates with our own subjective consideration of the tone of commentary at the time.
In 2018 some saw further Turkish deterioration as inevitable, risk models and investment committees were ‘black listing’ the asset even after (because?) it had already sold off, and clients were contacting investment managers to check whether they had exposure. Although the tone today is still highly negative, it is far more balanced than it was.
By contrast, the level of attention on Turkish weakness today is far lower and less all encompassing: there are fewer stories, and they are less clustered.
A simple Bloomberg search of stories with ‘Turkey’ as a key word supports this observation:
As the speed of price moves, these forces also cannot be considered in isolation. The pandemic, or even the US election will have dominated recent news coverage, and had they been taking place in 2018 Turkey would potentially have got less attention as a result.
However, if we believe that news coverage can impact investor sentiment just as much as it reflects it, then no matter what the reason, lower coverage today may suggest that investor decision making is less likely to have emotional drivers than in 2018.
Lastly, we should consider valuation. Insofar as you can assess value for a currency, the level of ‘carry’ (primarily driven by interest rate differentials and liquidity) is probably our best guide.
In 2018, policy rates were higher and carry spiked amidst the depths of the episode (and again in early 2019):
Since then policy rates have declined, reflecting both Erdogan’s own views on the value of rising rates to prevent inflation, but also the broader pandemic trend of emerging market policy makers maintaining easy policy into currency weakness.
This has meant lower levels of carry, alongside high degrees of volatility associated with currency intervention in the summer. Insofar as carry is a reflection of the compensation that investors are demanding for currency risk, this combination of lower and more uncertain carry, suggests that the currency is less attractive from a behavioural standpoint.
None of the above mean that the Turkish lira is unattractive today. We have just seen a ten-day period in which you could have made over 7% if you had timed the recent rally (and got out at the right time).
But such gains would have been predicated on forecasting particular events, namely policy changes, personnel, and rhetoric. Episodic investing on the other hand rests upon identifying trades that offer the best chance to perform well in a range of scenarios, because they were already mispriced relative to fundamentals as they stand, not what we may think (or hope) they are going to be in the future.
In the case of Turkey, you were able to generate returns in the aftermath of 2018 even though the fundamental picture remained extremely poor. Today Turkey does offer some value and could potentially do very well, but the deeply episodic characteristics that were in place two years ago are far less evident.
The value of investments will fluctuate, which will cause prices to fall as well as rise and you may not get back the original amount you invested. Past performance is not a guide to future performance.