Thursday’s column provided a steady stream of comments and feedback with one question over-riding all others – what can be done to avert more flash events, especially in the Australasian window before the mainland Asia open?
I actually think the question should be, ‘what, if anything, should be done?’ because I remain unconvinced that what happened last week requires a radical rethink of how the FX market operates. This may come as a surprise to long-standing readers who may recall me advocating for the use of central bank volatility bands post-sterling flash crash, but the two events are different. In the latter the market moved something like 30 standard deviations, this month it was five.
To digress I ought also to mention how I am unconvinced about the use of standard deviations to measure events – everything has a different trigger and there may well be a case – such as on Brexit night – for a 20 big figure move. Market observers with experience have an instinct for what is a “bad” move and what isn’t and while everyone agrees what happened in October 2016 was bad, few feel the same about the recent move (often a trader’s P&L will dictate their view however!)
To go back to volatility bands, I do actually think that a central bank using a momentum-based liquidity providing strategy would still help markets and give people confidence that something really silly isn’t going to happen – but the bands need to be wider than just four big figures (or if you prefer, five standard deviations). To me it is as much about the size of the move than the speed, although in the modern market, as discussed in this week’s podcast, everything happens quickly. If the market moves, for example, five big figures (it can be seven or more or less, it’s really for those responsible for market stability to decide) in quick time the strategy can post a bid or offer to the primary venue. That will do it. Algos need data and if they see a reasonable bid or offer there they will react by very gently placing a bid or offer at or close to the same level (they do it with every other price why not this?) and momentum slows…and the move stalls (and rebounds probably).
The time this takes would allow everyone involved to properly assess the situation – was there news worthy of such a move? If so, then buy or sell into the reversal. Was there nothing of the sort? Take the opportunity to get some cheap positions on board.
But to go back to the crux of this column, I am not sure that anything needs to be done if the moves are of the scale of this month’s move. I fully accept it’s a tricky issue because questions would undoubtedly arise over the willingness of market makers to continue to price in illiquid windows. If they think the market could disappear 400 points they probably won’t.
Of course one can point to the days of old when a dealer made a price not knowing what would happen, and occasionally it went horribly wrong as a result, but there was not the sense of entitlement then that there is now. People think they have a right to make money when making markets and it has not always been that way and it certainly has never been the case that market makers make money on every trade (or even every day).
Saying this only serves to reinforce one’s dinosaur credentials, of course, but it should not be overlooked as an issue if people are to debate a more robust market structure.
It should also never be forgotten that so many people who are quick to bemoan a flash event, or the lack of pricing, or even the quality of pricing, often never make a price themselves. They think managing risk is easy because they find someone else to do it and they under-appreciate the skill and work of a market maker – in other words, there is a sense of entitlement on the liquidity consumer side as well!
On Thursday I noted that the algos behaved pretty similarly to humans back when markets were largely voice and to reinforce that view I find it interesting how plenty of traders made some really good money merely by having resting bids and offers in the market during the most recent flash event. Many of the orders were for customers and that was perhaps the basis for their own decision to also buy or sell, but in many ways human traders established a quasi volatility band with the help of their order book.
Now some may question, in this post-chatroom era, whether dealers should be using their order book for information, however it cannot be argued that in circumstances like this it didn’t help alleviate what could have been a much worse situation. As long as the customer orders are filled first I don’t see a problem with it, what it probably needs is for bank internal compliance teams to sign off on the practice – good luck with that!
Either way, I have spoken to traders who are now in the habit of deliberately placing bids and offers several hundred points away from the market to take advantage of a fat finger trade or a flash event. Their rationale is that they will know if the move is justified or not before the orders get hit and they deem the risk/reward good enough – and I for one would not argue with that. Back in the day a fat finger trade would often be rolled back as both sides knew there was an error – today there are likely to be so many trades before the error is noticed nothing can be done. To paraphrase an early Pink Floyd song, Careful With That Finger Eugene.
This willingness to be more active in the market may actually signal the start of a rebalancing in how the FX market works. If human traders can demonstrate the value of being able to assume risk then perhaps more risk will be allocated to them, especially at the banks. If that happens then liquidity probably improves – it would be good news for the automated market makers as well thanks to more – and probably better – data.
Whether or not those traders will be allocated risk, or more importantly provided with a reasonable window on the institution’s order flow (within strict rules over how it is to be used and regular checks that customers are in no way put second or even inconvenienced), is open to question. I have already noted my scepticism that any compliance team – ignorant as most are over how markets work and the value of risk holders therein – would okay such a framework, but there is perhaps a way.
The Global Foreign Exchange Committee is obviously wrapped up in the Global Code and broadening its reach. Perhaps though it needs to understand its wider brief as a guardian of the markets? Central banks make up a core of the committee and they are often responsible for market stability, including currencies, therefore why not use this forum to discuss what can, or should, be done about these mini-flash moves?
They may decide nothing, and that’s fine, but I would warn there will, one day, be wider consequences, because as markets become more inter-linked, a flash event in FX could easily spill over into fixed income and equity markets. When that happens the politicians sit up and notice and things get a little tricky when they inevitably ask the question, “why were you not prepared?”