Electronic FX Marketplaces: Getting as Much Mileage as Possible

Electronic marketplaces and traditional models

Electronic foreign exchange (FX) platforms are not a new phenomenon. Request-for-quote platforms, which largely replicate the age-old telephone dealing protocol via single bank or multibank platforms, have been around for several years. More recently, the multicontributor marketplace structure, which provides important dealing advantages to FX market clients, has emerged and been integrated by top treasury departments. Marketplaces such as these, devoted to equities or fixed income, have existed for several centuries and electronic ones for more than 15 years. This article aims at describing how treasury departments can extract full value from a connection to an FX marketplace. Perhaps a description of what constitutes an electronic marketplace can be useful. To visualize the difference in clear detail, one might look at the 2 attached diagrams.

Traditional Dealer Market Model

Figure 1 shows a traditional model: dealers can use phone or electronic dealing systems. This model is still in use by slightly more than half of FX clients around the world. Note that clients do not deal with clients. Only banks supply liquidity.

Electronic Marketplace Model

Figure 2 shows an electronic marketplace model: Clients connect directly to the market, which is composed of banks as well as buy-side clients, which access the marketplace anonymously via a clearing bank credit line.

Market Intelligence and Pre-Trade Price Information

Unlike dealer markets, an electronic marketplace, which is anonymous and provides clients and dealers with the same trading capabilities and market data, is neutral. Neutrality is an essential feature that enhances the quality and statistical robustness of the prices and amounts displayed on the marketplace. For this reason, electronic FX marketplaces offer more robust market intelligence than traditional single or multibank platforms.

In addition to providing more robust market data, some of the more advanced electronic marketplaces provide more market data, including all bids and offers in the market with amounts available at each price. This full ‘depth-of-market’ data can be used to assess pre-trade market depth and conditions, as well as post-trade execution performance and benchmarking. This additional information provides the same advantages to FX traders as it does to equities traders, where it is more commonly available. This effectively means that ‘level one’ and ‘level two’ market depths are now available to the FX world. Instead of only a best bid and best offer, a marketplace shows a complete snapshot of the market. Imagine the possibilities offered by knowing at what price a large amount can be traded by viewing the sum of the amounts offered at various prices

In the near future, information about volume at various price levels, akin to ‘level three’ equities data, will be available too. Some electronic FX marketplaces already offer these features, so far seen only in the equity and fixed-income world, such as VWAP data. Although these are more about post-trade than pre-trade information, they have immense value in their own right. They open the door to an order execution era where users have full control of when, at what price and how their orders are carried out.

Best Execution and how to prove it

A number of institutional investors in the “real money” world are subject to stringent best execution rules imposed by regulators. One common means of meeting best execution standards is to always seek competing quotes at trading time, and, of course, to take the best of them. For other institutions, where this is not a regulatory mandate, it is followed as a matter of best practice and common sense.

In the pre-electronic era, procedures imposed on a treasury team were to call a specified number of banks and take the best price. This was resource-intensive at best, and at worst still did not guarantee best execution. Nowadays, this is taken care of through the design of the electronic marketplace. Prices from many banks and from many clients are available at any point of time. Indeed, the best marketplaces make clients’ quotes available to other clients, albeit from behind the anonymity and credit screen of a clearing bank, which further enhances liquidity. It is therefore physically impossible to trade at a price which is not the best among those available.

Best execution can also be a topic for large orders at multiple rates. More and more in this case, best execution is ensured by having the possibility to use algorithmic execution, not unlike what is seen in equity markets. The best electronic marketplaces can be accessed via either API or FIX gateways, and eclectic toolboxes are available to help users mount their own execution routines. Indeed, leading algorithmic trading platforms have recently added access to FX marketplaces to support their clients FX order execution strategies.

Measuring Transaction Costs

Let us first define transaction costs. Unlike the received view that transaction costs are only about brokerage, transaction costs really can be defined as the difference between the theoretically attainable best rate (see below) and the effective execution rate of the client, including the brokerage, market slippage and other costs. This theoretical best rate varies as a function of the amount to be dealt. Indeed, traditional FX trading does not incur brokerage, but does incur all of the other sources of cost mentioned above.

Because traditional FX trades executed electronically or by telephone are private, statistics necessary to investigate this theoretical best rate are not available. However, transparent FX marketplaces provide statistics of their liquidity in various amounts over a period of time, something nigh-impossible to obtain from a traditional model.

The effective traditional rate (i.e. the rate at which a client will deal using traditional models) will add four factors to the theoretical one: a possible credit spread, the effect of dealing with a lesser number of participants, the impossibility to bid or offer for the client, and possible reading by the quoting bank.

The effective electronic rate (i.e. the rate at which a client will deal using an FX electronic marketplace) will add two factors to the theoretical one: the effect of dealing with a lesser number of participants (although by design the number of participants is higher than when dealing traditionally), and the brokerage (note clearing banks effectively solve the credit spread issue, in addition to providing anonymity). Brokerages are typically a small fraction of a pip.

Answering the question of the value of an electronic FX marketplace then can be described as measuring the difference between the “effective electronic rate” obtainable in this marketplace, and the “effective traditional rate” (see below).

These two rates will of course vary from client to client, and not everybody will derive the same benefits from using an electronic FX marketplace. For a given user, the best practical way to measure the above effective rates is to use statistics derived from available feeds as well as dealt rates in various cases.

In how many live cases would the conclusion of such a study be in favor of a traditional model? The answer to this question will govern the future of electronic FX marketplaces. So far, volume trends seem to be on the side of the latter.