Saturday, 10 February 2018

The changing narratives of a market dump.

It’s February and I have been using Twitter more than the blog as, in effect, most of my thoughts are pretty simple and don’t need expounding but it is probably worth pulling everything together for the record and to evaluate the 'what happens next'

As regular readers know I have a regular concern about the way markets start new years, which can effectively be summarised

1- Markets take off in January en masse in the direction set by all those ‘2018 trades of the year’. This sets consensus.
2- The week after Martin Luther King day, or Martin Luther Turn Day as I prefer to call it, together with the first expires, can often trigger a turnaround.
3- The start of February sees a peak crisis in something - EM, Bank Balance sheets, whatever.
4- This causes first round damage in the assets associated with the assumed crisis.
5- This causes losses which need countering by selling other assets that are in profit
6- This sees a cascade unwind in anything that is leveraged and heavily positioned.
7- Narratives chase price moves but are usually later proven to be incorrect.
8- These February washouts of the consensus trades of the year slowly settle down and reverse, leaving March as the time to really put on your trades of the year.

With this in mind, I approached the third week of January with huge caution, switching holdings to cash, but Martin Luther Turn Day came and annoyingly it didn’t produce the falls, which was FOMO painful. However, month-end was looming and it wasn’t hard to calculate that, with bonds having fallen and equities have risen so much, rebalancing of assets in funds was going to see some very large selling of equities and buying of bonds.

The large size of equity selling occurred in the days running up to the end of the month but there didn’t appear to be the bond buying. However, the narrative of ‘just month end’ still accommodated the equity move leaving those long excused from their positions.

But the 'end of month’ narrative had an expiry date - the end of the month - and though this had passed we saw no bounce in stocks and the amplitude of intraday swings in equity prices was picking up (normally a turn signal).

Friday morning had me scratching my head as to why traded volatility wasn’t rising

We didn’t have to wait long. The US data showed a growth in wages. This flickered like a force 4 tremor on the seismometers along the San Andreas fault of inflation concerns. Bonds sold off again and stocks fell heavily, with inflation concerns triggering ‘just stop losses’ as the inflation story had been a back-burner narrative for a while. However, we know that ‘just stop losses’ is on a par with quoting Fibonacci levels in the league of ignorance of real reason.

With both bonds and equities lower, attention turned to the 'risk parity' sector with it now being blamed for the stop loss action. Risk parity funds switch between bonds and equities as historically when one falls the other rises, hence keeping risk levels constant, but now we had bonds and equities falling sharply. So, it was assumed, it must have been them. Even volatility was polite enough to move with VIX a whole point and a half up from 13.5 to 15.

Now a quick brief here to anyone reading this who is scratching their heads over what this volatility product thing is. Experts, please jump this.

Volatility is a mathematical measure of how fast and far a price moves. It is a historic measure as you need to know how far and fast something has moved to work out how far and fast it has moved.
But volatility is also the key ingredient to pricing an insurance policy. If you know how far and fast something has moved in the past you make ASSUMPTIONS as to how far and fast it will move in the future. This assumed future volatility, though psychologically referenced through price anchoring to past volatility is basically an informed guess.
The maths used to calculate insurance policies or options as we like to call them in finance, the Black-Scholes model, can pin down every variable (Exercise price, forward prices, interest rates, discount rates, time, etc) but there is always an unknown variable (otherwise we are saying the future is certain) and this unknown variable all boils down to one number called 'implied volatility’. And this is what is traded in options markets. 
The word ‘implied' should be the clue to the danger here because as this is the only effective unknown the equation, should the price of the option change, even though basic demand (say a corporate wants to hedge a large overseas future payment) then the implied volatility changes too. This does not mean that actual price volatility of the underlying asset will change even though the implied volatility has. Nor does it mean that, as volatility is so closely linked to implied probability in the equations, that any actual probabilities have changed. 
Basically, it may IMPLY that probabilities have changed but it doesn't mean they actually have. It is worth comparing this to CDS pricing where the same calculations are made as it too is an insurance product, where many confuse CDS prices with the actual probability of default.
So with this in mind we look at the next derivative of implied volatility - which we have already decided is a derivative of the maths of guessing where prices will be in the future - the VIX. This is an index of lots of specific implied volatilities from lots of different assets in different time frames. If you think it is simple then just look here to correct that view. Yet this index is bet upon in its pure form through a futures market and as soon as something can be bet upon as a future it is assumed to be clean and pure. As this future can be bought and sold and held in a portfolio the dangerous next step is to consider it as an asset rather than the complex hypothetical maths derivate of the future unknown that it actually is.
As this comfort grows funds are structured to hold these futures in ETFs and ETNs. and if the movement in these ends seems pretty small they are geared up by leverage to multiple factors in specialist ETFs. Finally, a product is offered a product that goes up when implied the VIX goes down - the XIV (as in VIX reversed, not the French king Louis XIV, though they both suffered excruciatingly painful deaths). 
As stock prices had been rising strongly for a year demand for hedges against them falling had dropped. And so implied volatility fell too. (I won't get on to the asymmetric behaviour of volatility but it is worth bearing in mind that historic volatility can be as high on strong up moves as it can be on down, but the way people associate implied volatility is that it goes up on price falls). 
This created a trend which was observed as an trend in a true asset and was sold as such. As soon as past performance can be cited as a reason to buy then a dangerous feedback loop evolves (see Bitcoin). Money poured into the trend with very little idea of what was actually being bought, indeed funds themselves started to sell volatility to increase there own performance having missed out on the stock moves. 
We even saw people study these products with technical analysis like an asset, adding trend lines, oversold/overbought lines and Fibonacci levels (see above) to it. The distance they had stretched from the reality of what drove these things had reached parsec levels. But it had already established itself as a legitimate hedging tool and had become embedded in countless financial products and bank and fund positions.   

So what happened?

Monday saw the dormant VIX crack as leveraged trades in toxic products cascaded down the tree of hedges into ‘have to sell stocks’ which just pushed up volatility making it all worse.Volatility skyrocketed, both implied and actual. I am not going to quote a number it hit because different time frames have different values and I could just pick, as the press does, the most impressive. They were already performing statistical crimes by reporting that volatility had gone up 100% when it had gone up from 14% to 28% (that's up 14%).

Even with this massive 10% move in stocks the narrative of explanation was still not concerned that this was anything to worry about in the longer term. It had moved from 'just month end', through 'just stops on wage data' to 'just an explosion in mad dog leveraged lunatic esoteric products'.

Which is somewhat ironic as on this basis it would be clear that the crash had been caused by everything that posts 2008 enforced regulations and been put in place to rid us of - ridiculous leverage in products that are toxic, bought by punters who have no clue and sold by spivs who point at past performance to sell them. It was textbook.

Tuesday saw the ‘volocaust’ settle down as volatility products were now the assumed culprit and this wasn’t a reason to sell global risk. But even though VIX, which was by now as keenly watched for direction as stocks themselves (another example of the tail wagging the dog). A rally in stocks was assumed to be great news that everything was settling down and the 'volocaust' was passing, but massive moves up in the underlying asset can be as representative of chaos in a volatility market as much as down, as delta hedges in the underlying get more desperate. A rapid up move did not mean volatility was falling even if implied volatility was.

Global assets were now wobbling. If this was indeed just an esoteric product wobble then why on Tuesday night / Wednesday morning did Chinese and Japanese stocks put in such a battering? Moves like that in China, without anything else to look at, would normally of themselves because for a mini global rout (2015) but these were being studiously ignored.

Wednesday saw rallies, which made my China theory look iffy and reinforced the 'it's just vol' story - until the close, which was dreadful. Stocks were being dumped again.

One of the consequences of measures of volatility moving is that it affects how much leverage you can have in your portfolio. The lower the volatility in an asset the lower the assumption of risk in holding it. Value at Risk, or VaR models, dominate bank, traditional fund and, most importantly, algorithmic funds. When the number you use as a volatility input increase you have to reduce your holdings even if you still consider your base argument for holding them valid. It depends on the time frame of allocations, these can be instant in high-frequency models, to monthly for old-fashioned real money to really slow with retail. Value and volatility shocks linger in the darkest crevasses of portfolio management for ages. It's like oil on beaches after tanker spills.

Thursday saw another leg downwards as things were now starting to look global. China was still steadily falling, FTSE was now down 10% off the highs, the DAX was in serious trouble (the darling of the ‘euro-uber-alles' trade) and it was all starting to look as though we had a big mismatch between action and explanation. When that occurs things get really messy. Not understanding why something unpleasant is happening is the gasoline on the fire of fear.

Friday was interesting. What appeared to be fear emerged properly with even the 'just a healthy correction’ crowd looking a bit like the Monty Python Black Knight, yet the markets bounced into the close. There was no new news. Price is news (PIN).


Stepping back from all of the micro of the week, we could fit all of the above into the classic layout of the original 8 steps of a February wash out. It starts as a US concern, has accelerated and is now hitting global risk appetite with apparently dissociated assets in far off places being sold.

‘Inflation' may now be sprawled in colour on the billboard outside the cinema but inside the show is a classic black and white blow out of consensus trades.

The big what 'happens next? ' I'll ponder over in the next post "What happens next? The great global risk repricing." (Now posted here) but I am far from sure that this is over.

But before I go I’m going to be a bit unkind. But it does need to be said. There a lot of people out there who pride themselves on analysing the minutia of finance, looking for clues as to the next nuance of price moves or the odd basis point arb between trades, or clever sector switches, or curve trades. So it is with an evil gloat I see their detailed embroidery get burnt up in the house fire caused by the gallon of gasoline left in the oven that they failed to spot. At the end of the financial day, you are judged on PnL not PhD

1 comment:

Al said...

Great post Polemic. I'm afraid I switched Twitter off because it is deeply unsatisfying so I've missed your pearls wisdom.

It made me chortle when CBOE came out fighting in defence of their vol products this week saying they did not cause the market correction and that the association between them and markets is the other way. Of course it is true strictly speaking. Until you factor in their use in VaR models in a chaotic feedback doom loop of deleveraging. Same as with all these corrections.