Our Views and Comments
The Signal and The Noise
As I was a bit puzzled by the timing of the price action on Italian assets following the forming of a new government, I have been doing a bit of analysis on Italian CDS and government bond price behavior. Even with the benefit of hindsight, the timing of market reactions is very difficult to rationalize. In fact, only just as the election results became known, it was clear to everyone that there would be a very high probability that at least one Euro-sceptical party would become part of the government. Now in those days in early March, yield on Italian government bonds went lower, while the 5 year CDS was pretty stable. For the whole month of April, 10 year bonds were trading more or less stable, while CDS spreads even tightened a bit.
Now if we take a step back, CDS prices reflect the probability the market assesses to a specific event – a default by the Italian Treasury (for a given recovery rate). If the market would behave rationally, one would expect it to act accordingly a kind of Bayesian probability re-set each time a new piece of information comes up which is relevant for this assessment (Bayesian probability calculus starts from a given probability, and is re-computed each time a new piece of information becomes available). For a very clear but thorough explanation, I would refer to Nate Silver’s gem of a book, “The Signal and theNoise”. Clearly, early March, no re-calibration of risks on the basis of new information was made (at least not in a negative way).
Game of Thrones - Italian Style
However, as from around May 10th, international newspapers start reporting about the likely formation of a populistic government. One also sees a surge in internet searches for words and themes around Italian politics. It is from thispoint on that we see Italian assets taking a first dive. In the weeks after this, bond spreads, CDS, Italian stocks, … would all react promptly to any new information during the “Game of Thrones” featuring Mattarella, Cottarelli,Savona,….
It seems that the markets chose to “forget” about the risks which were implied by the outcome of elections until the dangers became imminent and close by. There is a fantastic experiment on this.
When presented the below video, the viewers are asked to focus on the ball, for example by getting asked how many times it gets passed by the players in the whiteT-shirts.
At the end each viewer is asked to submit the number of passes she or he counted. When asked whether they noticed something unusual (the ape!) only a limited proportion of viewers have actually spotted it. This is exactly what happened in the market in March and April. The ball on which the investment community had its sight was the risk of a US-Chinese trade war. Or at some point it could have been tensions with North-Korea, or other hot topics of the moment.
Type I and Type II Errors
So it is fair to say that most market participants were sleeping as far as it concerns Italian political risk (except for a few savvy investors such as Alan Howard ofBrevan Howard, whom allegedly made very handsome profits by shorting Italy). It would be easy to say that the answer is that in any future potentially tricky situation, one should short all there is. We do not want to replace type II bytype I errors (a type 1 error would be where you predict a crisis and then nothing happens – a type 2 error is where you fail to predict this, and a crisis occurs anyhow). I believe the right answer is to really follow up each day how probabilities of events, and their potential impact evolve. Given what just happened, how likely is it that a populist coalition will prevail? And if this is the case, what assets would be impacted by their policy? And so on.
It is as if you challenged Mike Tyson, and then during weeks you would just forget or ignore what a dangerous fighter he is, until you are actually in the ring with him. When you are in the market, you constantly do all it takes to assess risks and likely scenarios. Just the way you would constantly wonder what your opponent's weaknesses could be and train to exploit those. In the example of Iron Mike that won’t help you neither, but the people who know me know that I like to use sports (especially boxing) as an illustration.
So the point I wanted to make is that even in widely scrutinized markets (such as European government bonds and EUR FX rates), most market participants only take into account the information which is “really in their face”, while failing to adapt their expectations on any news which is relatively sub-radar. As it seems now we could have elections in Spain pretty soon, on top of the planned polls in Mexico and Brasil, it will be interesting to see how those develop….
Italian Elasticities…. Lessons about loss aversion
Italian Elasticities…. Lessons about loss aversion
Extremely Loud & Incredibly Close
A few weeks ago, we reflected on the timing of moves in Italian assets, and related market instruments following the political turmoil in Italy (see article on LinkedIn or on https://www.mrm-advisory.be/our-views-and-comments). The conclusion was that markets really only tend to take notice of events when they become “extremely loud & incredibly close” (the catch-line is borrowed from Jonathan Safran Foer’s novel, which otherwise has no relationship with this subject). And market participants mostly choose to act as if things don’t happen as long as they don’t make headlines.
In today’s piece, we want to dwell on the price elasticity of events…. In other words, why do some news items trigger huge price movements, while others (which at first sight seem at least as fundamentally important) only cause moderate fluctuations? And even more importantly, do good and bad news lead to the same degree of market movement?
Up & Down
Below, we have split the period between May 28th and June 15th in 4 parts, each linked to a wave higher on lower in the yield on 10-year Italian government bonds. Yields had already moved substantially higher in the days before Monday 28th of May, but it was at that moment that markets really got in panic mood, after Italy’s president refused to appoint a Eurosceptic (one might say anti-Euro) finance minister. The “red” periods indicate a higher yield, and hence lower bond prices, while the opposite is true for the “green” periods.
Red Fields indicate a streak of higher yields/lower bond prices, Green the opposite.
Basically, on May 29th, at some point, the 10-year yield was more than 70 basis points higher than at the close of the day before – and the 2 year yield made its biggest daily move since 1992, bigger than any one-day move during the eurozone debt crisis. The sheer size of the swings, both up and down, has been largely attributed to regulation, which has eroded liquidity by forcing banks trading desks to reduce their risk taking capacity (banks used to have far larger trading books, which allowed them to “buffer” any wild price swings). But this is the subject for another article.
BTP- Buono del Tesoro Poliennale or Build To Plunge? (BTP is the commonly used name for Italian government bonds)
The interesting part of the table above is the elasticity of prices to the news events driving them. After the first two waves, on June 4th, one could argue that the outlook for Italy, and its relationship with the Eurozone was not in a worse state than before May 28th. A government was formed, without any fanatic anti-Euro finance minister, with a similar policy as was on the cards one week earlier. On top of that, the president’s action was a proof that institutional guardrail mechanisms actually did their job.
However, when looking at the price action, it is striking that it took 2 days for the yield to move 103 basis points higher, and then to came back by only 88 points over a period of 4 days. For the next 2 waves, we could tell a similar story – but the story is less clear as the 2nd wave of decreasing yields had less to do with the situation in Italy, and was rather part of a general move lower in global interest rates. So for some reason, the bad news triggered a violent and quick reaction, while a return to a similar state as before this event, led to a much slower and incomplete correction.
Thinking, Fast and Slow
In our first piece on Italy, we looked at the concept of probability expectations, in which I made reference to a great book – The Signal and the Noise, by Nate Silver. This time I would refer to another great book: Thinking Fast and Slow by Daniel Kahneman. He describes a number of cognitive features which prevent actors from making rational decisions. In this case, the most prominent one is loss aversion. Market participants attach more importance to negative expectations than to positive ones – hence the move lower in asset prices (higher yields) has more momentum than the opposite move. One could say that fear is stronger than greed.
Now, obviously there are many more aspects which explain market behavior, and the market for government bonds is driven by an incredibly complex set of dynamics (for example, the different buying/selling behaviors of different investor types such as Hedge Funds, Asset Managers, Central Banks, …). But next time we see a crisis sparking a sell-off in an asset class, we will be curious to see whether and how the corrective process will work!
(Trade) War Games
(Trade) War Games
In all comments about the action/reaction sequences in the trade fight between the US and China, we hear about the tit-for-tat tactics which both sides are using. For people who are not familiar with game theory, tit-for-tat is a strategy which would be recommended when playing a series of prisoner-dilemma-style games. It implies that one always plays nice (i.e. adopts a co-operative attitude), and only retaliates when the opponent first aggresses. While in the classical one-time play of the prisoner's dilemma the aggressive or non-cooperative approach is the dominant strategy, when playing a series of games a tit-for-tat strategy yields better results. Note: This was empirically tested by Robert Axelrod around 1980. He organized two tournaments where scientists could enter strategies, which would then go to play series of games against each other. The tit-for-tat strategy proposed by Anatol Rapaport came out victorious !
Serial versus one-off games.
A face-off between countries about trade tariffs can be modelled quite well using a prisoner 's dilemma scheme (it would lead us too far to discuss the pay off schedule in detail, but if anyone is interested in how it works, please send me a (private) message and we will discuss). And since it is a game which is on-going, it is one which is played in a series.
Now, indeed China adopts a disciplined tit-for-tat strategy, which is what one would expect to be the optimal choice. A tough play would only be adequate if this were a one-time game, which is obviously not the case. Then why would the US opt to go for a non-cooperative approach, leading in principle to a sub-optimal outcome for both parties?
Why would the US stay away from the optimal strategy?
We see three potential reasons to deviate from the ideal strategy :
The US have the impression that the current state is not the “neutral” point. They feel China is already in a non-cooperative position, for example by failing to respect intellectual property rights. This would imply that retaliation is the dominant strategy. In fact many people, even those who don’t approve of a trade war, believe this is indeed an issue.
The view of what precisely the best outcome is may be blurred. Most observers seem to take it for granted that the ultimate goal in this type of games is to maximize the long term economic benefits for one's country. President Trump is a politician. And politicians want to win elections, the rest is noise. In a boxing match, playing tit-for-tat would imply that both boxers circle each other without hitting hard - as this would mean nobody gets hurt or knocked out (here I go again with the boxing analogies). However, the public hates this, and want to see the fighters go hard. And both fighters have a competitive nature which makes them want to win. As The Donald wants to win and wants to please his voters, he hits hard!
The third possibility is that the President is behaving irrationally and erratically. Even if this may be the case, up to now, all his so called crazy moves in hindsight proved to be quite well thought through. What is seen by most Europeans (and liberal Americans) as outrageous behavior makes perfect sense to the better part of Trump voters. It could be that introducing a flavor of unpredictability and hardmanship increases one's credibility in later stages where tit-for-tat is played. In such a case it would make an adversary aware that if retaliations are necessary, they will come fast and hard.
Or is there another scenario? In our view, what the US are doing now, is using another Game Theory concept, called Brinkmanship. This is the art of pushing dangerous events to the brink of an active conflict (but not any further). The Cuban Missile Crisis is the best-known example of Brinkmanship in international politics.
It helps in this that the President of the US is seen as a bit of a loose cannon. In a game of brinkmanship, it helps if your opponents do actually believe that one is capable of behaving strangely. Winning through Tit-for-Tat requires players to be calm, friendly and reasonable. Winning through brinkmanship requires a player to be bold, arrogant and reckless (or at least be perceived that way). Guess which game suits the President of the US best?
In a trade war, one can also bend the rules of the game such that playing hard-ball becomes less risky. In the case of the Cuban Missile crisis, there was no room for mistakes. One slight false maneuver could have meant the end of the world. This is a trade war. Nothing can happen that can’t be un-done. In the past, tariffs were decided upon to remain in place for years. Donald Trump has shown that he is capable and willing to impose new tariffs in the blink of an eye, but can also reverse any decision very quickly if needed. Therefore, it is less of a gamble to act tough.
When playing a game of brinkmanship, it is impossible to tell what the outcome may be. It could lead to escalation and very bad outcomes. But as long as people are consciously playing a certain strategy and know what they want to achieve, it is still possible to reach an orderly truce at some point!