Last look, lack of time-stamp and internalisation of flows - Which practice is most open to abuse?

Last month's publication by The Fair and Effective  Markets Review FEMR, set out 21 recommendations to help restore trust in the wholesale Fixed Income, Currency and Commodity\ (FICC) markets.

Recommendation 4b stated:

...."As part of that work, improve the controls and transparency around FX market practices, including ‘last look’, ‘time stamping’ and ‘internalisation’"

the report suggested that the Bank of International Settlements (BIS), and national central banks including the Bank of England lead that effort, fully covered in section 4.3.3 of the report.

I thought it would be worth briefly looking at these three practices and then gathering some very unscientific feedback from readers in the form of quick poll, and so below I have set up a poll, asking the same question to both buy-side and sell-side readers, with probably too simplistic a question, which is: "Which of the following practices are most open to abuse, resulting in sub-optimal client execution?".

Last look: I recently commented on last look, suggesting probably over simplistically, that last look was some way a 'free option' for liquidity providers (LPs), allowing them to accept a trade if they liked it and reject it if they didn't. Of course it's far more subtle than that, and I was immediately (and rightly) picked up by readers - which is actually great, as it shows there are people out there at the sharp end of in this case market making, who actually read my posts - many thanks, always grateful for that. The feedback was excellent, and the point made was that in a sense last look reflects the trade-off cost of the round trip latency 'risk' where:

"the liquidity consumer can accept a certain % of rejections and or some hold time in return for the Liquidity provider (LP) maintaining the tightest price possible. Or the LP can widen the price to build in the slippage."

Certainly then some valid reasons for the practice of last look, although perhaps there are cases, even on single bank platforms, where LPs need to be more explicit to the price taker in terms of the parameters around last look, slippage, and rejection ratios. In extremis, certainty of execution is always preferable to chasing prices due to high rejection rates.

Internalisation: I have written a few times on the benefits to the bank and to the customer of bank's ability to internalise flows and recently quoted by EuroMoney in The great internalisation debate, and again in extremis, internalisation does (in theory), enable banks to offer clients more robust, possibly tighter pricing which can't be a bad thing.

Time stamping: Here I fail to see any valid reason why client orders/trade requests should not have a complete and auditable time stamp history. Ideally one that shows the life cycle of the order as it progress from the time and prevailing rate at which the order/trade request was first received by the bank (whether that was by phone, fax, email, or etrading system), through to when it was entered into the system and executed (in many banks it's the other way round, executed then entered into the system), and the fill rate notified to the client.

Lack of transparency in this area opens clients to risk of abuse, with orders being filled and then 'rates of the day' type fill rate being applied that is towards the worse end of the day's range (a practice noted in the FEMR report).

So, in summary, I can see genuine reasons why banks use, and how clients can benefit from last look and internalisation of flows. However, I struggle to see any reason for the lack of time-stamping client orders and trade requests, as it's a practice that offers most opportunity for abuse, and acts against the interests of the customer.

I am surveying readers on another blog, and asked which of the three practices was most open to abuse. Not surprisingly, I am getting very different answers from those working for sell-side firms, than those working for buy-side firms.

Practice most open to abuse, % of replies from Buy and Sell side firms

1) Lack of time stamping trades/orders
Sell-side: 65%
Buy-side:    0% 

2) Last Look pricing
Sell-side:  35%
Buy-side:  100%

3) Internalisation of flows
Sell-side: 0%
Buy-side:  0%