Kim Man Li, Head of Japan Institutional Electronic Trading, Bank of America Merrill Lynch, looks at ongoing microstructure and buy-side trader behaviour changes in Japan.
In recent months, we have seen several changes to Japanese tick sizes that have been gradually introduced to the market. Whilst these individual changes have been implemented incrementally, the collective milestones should not be overlooked as they have fundamentally altered the behaviour and usage of Japan’s favourite benchmark algorithm, Volume Weighted Average Price (VWAP).
Japan has a long history of small and measured changes to its environment. Each is seemingly compact, potentially trivial, but collectively all these measured changes are part of a very long-term vision or plan to help maintain Japan as a reputable investment centre to the global investment client base.
It has been nearly nine years since the fated Livedoor incident that caused a deluge of order flow causing the Tokyo Stock Exchange (TSE) to shut its doors early and subsequently re-evaluate its approach to the operational needs of its investors. Since this incident, the TSE has embarked on a multi-year, calculated and steady approach to rebuild its systems and reputation. In some respects, adopting a more measured approach allowed the TSE to be flexible with its incremental short-term plans. This strategy, coupled with relatively light-handed regulatory oversight, has allowed the TSE to evolve whilst adapting to additional financial shocks along the way, modifying its plans to accommodate the emergence of technically demanding businesses such as HFT trading, co-location participants, competition from technically superior PTS and alternative venues, increased data demands from algorithm trading and market data processing. We must not forget that this was also done with the backdrop of several releases of trading systems and a merger with the Osaka Stock Exchange to create the JPX along the way.
Considering these factors and some other more recent regulatory induced products, including the new JPX400 Index with strong in-built corporate governance metrics for its constituents, Japan’s future trading environment looks favourable.
Looking back specifically at the most recent changes, we correctly predicted that the latest tick size changes would cause some fundamental changes to the trading landscape.
Overview of TSE Tick Size Reduction:
Two changes occurred in Japan last year: Phase One in January 2014, which considered stocks where the price was over JPY3000 and the subsequent Phase Two in July 2014, which focused on stocks that were below JPY5000. In both cases the aim was to reduce the prevailing tick sizes to substantially smaller sizes in absolute terms. The combined result of these two phases amounted to changing the tick sizes for names which collectively made up to 50 percent of the TSE market turnover.
Traditionally, the desire to have the smallest tick size possible allows for less frictional cost when buying and selling shares. Where stocks had previously had large minimum ticks, the bid/offer queues were subsequently long as buyers and sellers refrained from crossing the bid/offer spread to pay an expensive price. With minimised ticks, the cost to taking liquidity is reduced and ultimately these cost savings are passed onto the ultimate investor, be it direct retail, or retail via institutional vehicles such as mutual funds or other investment strategies. It would also be fair to say that this was a competitive reaction to alternative PTS venues which already had capability to trade these names at the smaller tick sizes and had been offering these savings for some time.
The ability for the TSE to do this, however, was only possible due to the technical advances it had made over time. Whilst decreasing the minimum tick size leads to savings for the end client, it comes at cost for the exchange which will now see the available liquidity distributed more finely across the order book. Finer minimum tick sizes ultimately means an increase in total number of (smaller) orders, which in turn leads to more market data requirements and potentially more work from the exchange participants to cope with the demands of the increased data requirements. If you consider that in 2006, the TSE had to shut early when it neared then capacity of 4.5 million transactions, to the current day where we saw peaks of 30+ million orders back in Q1 2014, you can get a sense of the how much the systems have grown.
In terms of how the market microstructure has affected participants we see that for stocks with a tick size reduction, the following is true:
• Reduced spread (from 10.4 to 4.3bps)
• Reduced available liquidity at the bid/offer (from 590k to 33.5k)
• Reduced trade size and queue time which sho ws how long you must wait before being able to make your transaction
• Increased spread from 1.3 (at old large tick size) to 2.2 (at new small tick size)
• Increased trade frequency (in correlation to smaller trade sizes)
VWAP Algo Performance of Affected Names:
Algorithm providers have also needed to enhance their abilities in-line with the exchange’s microstructure changes. Some providers will have had experience trading on the PTS venues. Regardless of trading experience, given the larger trading volumes on the TSE, all brokers will have needed to ensure that they can process the increased market data given the more granular tick sizes. Additionally, it is important that algorithms can also read market data that are more than just a few levels below/above the current bid/offer. Given the ticks can be up to 1/10th of the original tick size, in order to see the same amount of liquidity you would potentially need to read 10 times the depth of book. Algorithm systems that may not be able to react to the increased data demands and read this level of depth in real-time will ultimately pay for this in loss of execution quality.
Narrowing of tick sizes have ultimately led to both improvements in interval-VWAP (IVWAP) performance, as well as performance versus arrival price. For IVWAP it’s simply the ability to trade at more granular levels. For arrival price, given that liquidity is distributed across smaller ticks, the ability to complete execution at less than the previous full-sized tick also leads to an improvement to this arrival price benchmark. If we look at the distribution of trade volume pre and post tick size change and adjusted to old tick size, we can see that more volume executes in a narrower band post change, i.e. orders are able to complete more quickly at more favourable prices.
To conclude, the TSE and indeed all exchanges continue to evolve their technology and structure to compete with their peers and ensure they deliver the trading functionality that investors demand. Brokers must continually invest to ensure that they keep in step with the regulatory and technological changes. Whilst each incremental change may seem small in isolation, the amount of technical investment needed by participant brokers can be very high. It is imperative that brokers continue to invest in-line with the market developments to ensure that they advance in step with the market and continue to provide optimal execution for their clients.