Dr Ian Marsh
The past few years have seen a startling change in the way forex dealers trade. My study of UK-based traders conducted in March 1998 gave electronic brokers more than 50% of the market, up from just 6% five years earlier. Both the bilateral telephone network and traditional voice brokers are losing out. E-brokers are no doubt even further ahead now.
One major difference between trading systems is the amount of market-wide information revealed by the e-brokers. These give the best price on both sides of the market and some indication of other reasonably competitive quotes (in the form of the effective price that would be paid for a $10 million deal). In academic-speak, the e-brokers offer much higher quote transparency than bilateral telephone dealing.
A series of experiments performed at the University of Amsterdam suggests some surprising consequences of altering the amount of information revealed to the market. The investigators persuaded small groups of professional securities market makers to trade a hypothetical asset. In the first group of experiments, the “true price” of an asset was set by the investigators, but was not disclosed to the market makers. In addition to the professional market makers, computerised customers also initiated trades. Half of these customers were “informed” in that they bought (or sold) at the best possible price if the market price was below (or above) the true price of the security. The other customers were uninformed and randomly bought or sold.
The actions of the informed customers helped to push the market price towards the true price. The market makers learned the true price by trading with potentially informed customers and by observing the market price.
The microstructure of the market was then varied between experiments. In the “low quote transparency” setting, market makers could obtain quotes only by calling other dealers, corresponding to the FX bilateral telephone network. In the second scenario, all active quotes were displayed on the computerised terminals used to deal (“high quote transparency”), so revealing even more information than the current generation of electronic screens.
Comparing structures, the investigators found that, as expected, both uncertainty and the amount of time spent searching for the best price were reduced when there was high quote transparency. With less time spent searching, trading volume was higher. Market makers also actively quoted tight spreads to attract informed customers, and avoided being on the wrong side of the market by shading quotes once they had traded with customers.
The real surprise was that market makers used smaller price adjustments in the transparent market, which meant that the market price converged more slowly to the true price than when traders were forced to use a bilateral network.
Another set of experiments took the analysis one stage further. Using the same framework, one of the market makers was told the true price and so became an “insider.” Even with an informed insider, convergence to the true price was still slower in the transparent market. Importantly though the insider’s profits were higher in the transparent market as he had more time to acquire positions at good prices before the market converged on the true price.
Pre-trade transparency is one thing. What about revealing information about trades after they have taken place? FX traders only know details of trades they are a party to, but the investigators also experimented with revealing details of all trades (identities of buyer and seller, transaction size and price) to all participants. When this post-trade information was combined with high quote transparency, the market price converged to the true price faster than in any other scenario. Similarly, the insider’s profits were lowest.
The FX market is currently trading mainly via the e-broking networks, which give a high degree of quote transparency, but very little information about trades. This should result in lower spreads, high volume, and large profits for traders with good information at the expense of the uninformed. Any developments that reveal more information about trades will make the market more efficient, but will boost the profits of the “Pac Man” trader who plays the market like a computer game at the expense of the informed.
What remains to be explained is why the bilateral telephone network and the broking systems currently co-exist – more than 90% of respondents to my survey used both brokers and the telephone network.
The fact that dealers, who claimed to base their own trading on customer orders, conducted a much lower proportion of their trading via e-brokers, suggests that strategies centring on dealing systems may exist.
Research at City University Business School will try to find out whether dealers are actively employing trading strategies that use a mix of dealing systems, or whether the market is simply in a transition state after which all dealing will be via e-brokers.
Dr Ian Marsh
is a senior lecturer at London’s City University Business School.