The normally staid capital markets scene is undergoing a revolution in Japan as US and EU-style ECNs (electronic communications networks) appear. These can be seen as off-exchange trading systems that replicate the best features of the exchange û in this case the Tokyo Stock Exchange, which is the primary dealing place in Japan and where price discovery takes place. Companies list on the TSE after careful screening.
The ECNs started life as platforms for day traders during the hours the exchanges were officially shut. Although ECNs have existed in the US ever since Nasdaq launched its automated trading system in 1971, the proprietary trading systems used by ECNs were only introduced in Japan in 1998. Japannext started in 2007 and Kabu in 2006. Japan is unique in Asia in having ECNs, since the necessary regulations have not been passed in other countries.
The companies are getting a great deal of attention as the investment banking industry is in the process of making one of its periodic lurches from principal business to an agency business model.
ôBanks will not want to keep much risk on their books, and the best way to make money with that mindset is to focus on the brokerage business,ö says Nick McDonald of Instinet, a pure agency broker. Instinet is the grand-daddy of electronic brokers and initially had a massive global market share, but nimbler and better equipped platforms have forced the giant on the back foot. However, the company will seek to benefit from a strong wish by the buy-side to invest through non-conflicted brokers. Recall that Lehman BrothersÆ proprietary bets ended up causing huge losses for their hedge fund clients. Hedge funds had, for example, posted collateral with Lehman, which was acting as their prime broker. This was almost impossible to recover when the US bank went belly up.
Although the two Japanese companies have positioned themselves as neutral trading platforms, they are owned by banks and investment banks. Kabu is 40% owned by the Bank of Tokyo Mitsubishi UFJ, while 10% is held by Mitsubishi UFJ Securities. Japannext is 50% owned by Goldman Sachs Japan, with smaller stakes held by Credit Suisse, Merrill Lynch and now-defunct Lehman Brothers. In contrast, Instinet is owned by another broker, Nomura, which bought the electronic trading platform in 2006. As Nomura evolves after the acquisition of LehmanÆs Asian business, assuming more of a principal bias, itÆs not clear what the effect will be on Instinet and whether its æpureÆ agency broker status may be compromised.
The market share of proprietary trading systems in Japan has grown fast after the necessary regulations were introduced in 1998. According to figures from Kabu.com, the market share of these systems (there are several apart from Kabu and Japannext, including Daiwa, Monex and Matsui, although these are still small) has grown from 0.003% of the total trading market share to 0.226% in October 2007 û a very rapid rate of increase.
Volume growth is very important to the business model of both Japannext and Kabu, because this is a volume game. As prices per trade fall, they need to be compensated for by volume. Technology is a key enabler of volume, because innovations such as programme trading and algo trading involve millions of trades, which can be done at the click of a button instead of going through a traditional sales trader. If one player adopts programmed trading, it then becomes difficult for other players not to follow suit, since the alternative is to be outgunned. Algos such as VWAP (volume weighted average price), for example, enable investors to buy shares throughout the day as close to the average price of the stock as possible. Skilful execution, especially the ability to chop up market moving trades into millions of tiny trades, can make an enormous difference to the performance of investors. And never is that more true than when markets are in decline and every trick to compensate for declining performance becomes crucial. Other important advantages of these electronic platforms are much faster order execution and narrower tick-sizes than on the TSE, both of which tend to encourage further trading.
The Japanese regulatory environment is different to the European and US environment in some crucial ways. In the US, brokers have been forced to adopt smart order routing (SOR), which means that the brokers must get the best possible price rather than trap the trades internally. Their electronic systems send out feelers and execute the trade in whichever liquidity pool they find it. In fact, this regulation has been very beneficial for the biggest brokers, because it forces the smaller brokers û who canÆt afford the technology - out of the game. In Japan, however, the government is not enforcing SOR (to the annoyance of the biggest brokers) because they want to protect the smaller brokers.
Instinet has the lionÆs share of the proprietary trading business, but it has fallen to 60% as of the end of October from 89% in 2007. Japannext and Kabu are currently neck and neck with an 18% market share of total trade each, but regulations may upset that balance as well. Curious as it may sound, proprietary systems have been given a quota by the regulator on the share of total trading volumes they may seize. Rather unfairly for Kabu, its ceiling is far lower than that of Japannext. ThatÆs because KabuÆs trading system sets the price by the auction method û which is precisely the same method used by the TSE. As a result, it is caught in an agonising dilemma: the more it grows the essential volume, the more rapidly it closes in on the quota. Japannext, which (along with all the other off-exchange platforms) uses a different pricing system, has a far higher ceiling.
Both off-exchange platforms have recently switched from night-time trading to trading alongside the exchange during its opening hours. They are clearly taking market share, and the TSE is responding to widespread criticism of its clunking technology platform by introducing a major upgrade, the æArrowheadÆ trading system. However this will only be rolled out in 2010, by which time it may be already obsolete.
A final issue to ponder is the relationship the brokers who own the ECNs have with the original purpose of the proprietary trading systems. The systems were created by investors who were suspicious of imparting trading data to their bulge-bracket brokers because they knew the latter would carry out proprietary trading on the back of knowing what the investor was trading. The brokers still run ædark poolsÆ, which are pools of liquidity in which no information is divulged to those who trade through it. The question which arises is whether there is an information asymmetry between the brokers who can see whatÆs happening in the dark pools and also see whatÆs happening in the off-exchange markets (known as ælight poolsÆ.) For investors looking for a completely unconflicted platform, the answer may not be quite here yet.
¬ Haymarket Media Limited. All rights reserved.
