What is the current state of play for NDF algo uptake?
I would say that the uptake has been limited so far and that the adoption has been slow. There’s not many providers who are providing a true algo product, rather it’s more like a smart order router to sweep liquidity, rather than internalising flow.
There are lots of people devoting substantial amounts of time and resources to NDF algos but the end product is not there yet. This is a combination of lack of client demand and lack of liquidity. However, I expect NDF algos to grow more popular in time. This is something that clients should be asking for and something that LPs should keep working on.
Is it the lack of liquidity hindering NDF algo adoption?
Yes, the lack of liquidity has harmed NDF algo adoption absolutely, but so has the availability of market data.
NDF trades need to be reported to swap data repositories and the data feeds right now are not real-time enough for algos to efficiently estimate market volume. Without accurate real-time market volume data, the predictive value of implementation shortfall algos and VWAP algos is reduced.
However, data availability should improve overtime, as more customers are demanding NDF algos, which I think is the big hurdle. Some customers talk about it, and it’s a conversation that keeps happening over and over again in the market and banks are definitely working on it, certainly investing time and resources – but there’s no clear market leading provider and no clear group of banks that are leading the charge.
It’s the thing that everyone’s working on and a lot of people claim that they have them. But, it’s much easier to do an algo in G10 or deliverable EM than it is in the NDF space.
In the NDF market, how are entities evolving to improve trade efficiency?
A lot of people are running electronic NDF businesses, whether it’s on their single dealer platform or using it to market make in the one month space, and it’s definitely being done at both banks and non-banks. However, there’s a disconnect because the market is trading one month and a lot of the end users want to trade IM dates. Therefore, there’s definitely basis risk and the market needs people to step up to solve that problem.
Many large market makers now run electronic NDF businesses. Their existing eFX spot businesses and market makers internalise NDF risk and add liquidity to the market. The additional liquidity makes trading more efficient, although some in the market might argue that the ensuing ECN liquidity fragmentation in fact makes trading less efficient.
Most of the focus has been around enabling on swap execution facility (SEF) for North American users. You see places like CBOE setting up SEF 1M pools but the problem that locals don’t want to trade on SEF so the result is there isn’t a ton of liquidity on these venues apart from North American hours where EBS on SEF is dominant (during Asian hours and London morning off SEF dominates). Progress here has been limited thus far.
What is the NDF market depth in LatAm markets/currencies and when are developments in market data and electronic pricing expected in LatAm markets?
Brazilian (and Asian) NDFs are much better than Andean FX, with a key challenge for clients being that they want to trade to IMM dates and interdealer liquidity trades on 1M (BMF for BRL). Not all NDF algo providers have built ability to trade to those outright dates, because there is not huge demand.
In Chile and Colombia, there’s not enough liquidity for a robust algo business at the moment. Hopefully, liquidity will improve enough over the next few years to allow for a robust algo market. While in Peru, there’s so much less that trades on a regular basis. I’m sure some people would like for there to be enough demand, but it just seems like that market is too small and therefore hard to use.
Not all the Andean pairs have good data inputs for the 1M point vs APAC where there are central limit order books (CLOBs) and Elecronic Communication Network (ECN) that have good liquidity. Therefore, one of the main challenges for clients is to justify taking an NDF algo when the trading outcomes are not as predictive. One has to balance the risk reward. In general, the less liquid you go in a market the more principally driven it is, so we see that play out in NDFs where banks and clients deal more often on streams and RFQs.