The Last Look…
Posted by Colin Lambert. Last updated: July 20, 2021
“Intent is a difficult concept to demonstrate…the benefits arising from pre-hedging can be very difficult to show”. Therein lies the challenge, and ongoing problem, with the concept of pre-hedging.
Those quotes are from the Global Foreign Exchange Committee’s “commentary” on the role of pre-hedging, which was released at the end of last week (it can read in full here) – the same paragraph also states, “It is also clear that pre-hedging may not always lead to a better outcome for the liquidity consumer.”
What is tantamount to a guidance paper on pre-hedging could potentially significantly shift the order management landscape in FX markets, not least because it maintains what I sense has been a trend towards having the buy side take more responsibility for the decision-making process.
For a start, the last-quoted sentence also muddies the waters a little with regard to the FX Global Code’s actual guidance which states in Principle 11, “Market Participants may Pre-Hedge for such purposes and in a manner that is not meant to disadvantage the Client…”, because, as noted by the GFXC, nobody really knows whether it will disadvantage the client or not. Deliberate attempts to shift the market ahead of an order are difficult enough to identify (unless they are truly dumb), therefore a cloud of uncertainty will always exist over the practice. Whether or not the FX industry needs such uncertainty over a practice seen in some markets as front running is probably for individuals to decide – I think we all know where I stand!
The problem is one of commercials – just as I have railed against (pre) hedging orders for the Fix not being part of the calculation, how do you TCA a transaction that is pre-hedged? Are individual trades to be picked out as part of a pre-hedging operation? That could prove very onerous in an operation where prices are being distributed in their thousands, but only some transactions actually form the pre-hedging (after all, an LP could decide to pre-hedge an order but do significantly more volume than required in the normal course of business – not all trades in a currency pair can be a pre-hedge).
I think we need to rid ourselves of a couple of misconceptions. As noted in the GFXC paper, there are doubts in the industry over whether pre-hedging actually leads to tighter spreads. Given how most pre-hedging appears to be in relation to orders where the direction is known, I think the spread is irrelevant. As a trader, in a 50-55 market for a large selling ticket, I could make 40-41 – great spread, but absolutely no benefit to the client.
Equally, can we please drop the naïve argument that selling or buying in the market has no effect on price? There are dozens of smaller players out there, purporting to be LPs or market makers, who exist only to jump in front of other orders. Try putting a price in one of the primary venues and see what happens – this is nothing new, an e-FX head at a bank demonstrated this to me a decade ago – new orders or trades have an impact and that inevitably makes it more likely a level will be hit, which is to the detriment of the client.
There are doubts over whether pre-hedging leads to tighter spreads…I think the spread is irrelevant. As a trader, in a 50-55 market for a large selling ticket, I could make 40-41 – great spread, but absolutely no benefit to the client.
Of course, it should be pointed out that some clients have driven the industry to this point – not the first time such a thing has happened – through their insistence of no slippage on their orders. That’s an unrealistic demand in most circumstances and therefore I feel the Code could offer guidance in this area, namely, leave your stop loss at will, but do not insist on it being executed “at market” without either accepting slippage. It’s either that or you accept pre-hedging.
A significant change from the GFXC’s paper could be in how some clients execute their larger trades, namely they may have to consider changing their best execution policy. A decent proportion of the buy side still operates on a “competitive quotes” basis – at the basic end of the scale this is four phone calls to four LPs, at the sophisticated end its some form or aggregated liquidity pool or an RFS/RFQ multi-dealer platform.
The GFXC paper discusses – and this is something I have previously raised in this column – what happens when a customer asks four LPs for a price to execute a large ticket. In what can only be construed as a warning to the buy side, the paper states, “Asking for a large number of quotes also increases the risk of information leakage and may not lead to an optimal outcome for the liquidity consumer.”
Putting aside the wonderfully understated phrasing – I think most of us know there is no “may” about it – what do the vast swathe of customers who use competitive quotes as the backbone of their best ex policy do now? The paper offers five alternatives;
- Use an algo
- Have the order executed via the agency model
- Ask for a two-way price to conceal the direction
- Specifically ask that no pre-hedging takes place
- Clearly indicate the RFQ is exclusive
Here are five observations:
- A lot of buy side firms don’t have the capability or permission to execute via algos – they are not allowed to take market risk
- Agency style execution again means the client holds the market risk
- (As noted by the GFXC) a two-way RFQ probably will involve a wider spread, more to the point, a lot of business is likely to be predicted by a regular LP of the customer
- LPs will be defensive quoting for a ticket they know is in competition – there is a trust issue over whether none will use the information
- Aside from the historical issues with “only you”, this approach throws the best execution policy out of the window
The paper states that if an LP “anticipates in good faith (original italics) that the liquidity consumer will accept the quote”, then pre-hedging (subject to the broader guidelines in place) is acceptable. The problem is, how does an LP know whether they are going to win the trade or not? They don’t know the other prices being viewed by the client and, more to the point, prices are changing by the millisecond – the provider of the best bid or offer will change while the client interrogates the market and makes their decision (if the client just hits best price immediately there would be no pre-hedging of course).
So if four LPs are asked, and all conceivably have a chance of winning the trade – the only way an LP would know they are not going to win it is if thy deliberately quoted not to do so – then all four are going to pre-hedge. If they all do so for, say 30% of the trade, the order is already over-hedged in the market.
Clients using a competitive quote mechanism under their Best EX policy are going to have to either change how they operate, or accept the market will move ahead of any large order they execute.
Of course, say the advocates of pre-hedging, this means that when the trade is actually transacted and the winner goes to hedge out the remainder of the risk in the market, there will be no market impact because the other LPs will be covering their own pre-hedging.
Really? Are we (that word again) that naïve? First of all, the client has to specifically tell the other three LPs they haven’t won the trade, not something a lot of them do in a timely fashion, and secondly, a risk engine with a short position, is not going to aggressively place bids in the market when the trend is down – the logic doesn’t work that way.
Clients using the competitive quote mechanism under their Best Ex policy are going to have to either change how they operate, or accept the market will move ahead of any large order they execute. Neither is that great an outcome.
The GFXC has done a lot of hard work on what s an incredibly complex subject, and to return to the statements at the top of this piece, I would humbly suggest that “difficult” is not the word, when it comes to proving intent and the benefits of pre-hedging, “impossible” is. During a press conference with senior GFXC officials last week I asked why one of the P11 Illustrative Examples was being removed and the answer was insightful. Effectively, having read the example through the prism of the latest guidance, nobody actually really understood it! That highlights just how difficult the subject matter is.
For me, I think we have to go back to the commercials and, sorry to do this to everyone, to an “old school” way of doing things. The GFXC encourages open discussion between LP and customer over pre-hedging, as it should. In a lot of cases, the pre-hedging element of a trade is not fully taken into account when establishing the rate. Partly because the client, if it is a stop loss at 60 is happy with a fill at 60, even though the actual rate might have been 60.5, and partly because, as noted earlier, it can be hard to pick out transactions that are specifically part of a pre-hedging strategy.
Notwithstanding that difficulty, however, in my mind, if the LP says to the client they think the trade should be pre-hedged, then it becomes a matter of trust. If the customer believes the LP is genuine in the need to pre-hedge (and the vast majority are), then they should accept the strategy but the final rate should be calculated with defined pre-hedging trades included. It’s akin to “working an order” as it used to be done.
The transaction can be TCA-ed (independently) to ensure the LP did not try to “ramp” the market and the customer has to accept that they will hold some (but significantly reduced compared to an agency order) market risk.
The fact is, pre-hedging is just too hard to quantify and, importantly, justify. I understand this borders on “advice” which many banks shrink away from, but they are most often using solid, robust analytics for the basis of their decision to pre-hedge, so share those with the client and let the client have the benefit (or, very rarely, cost) of the associated activity.