Assessment of ‘closing phase’ in China’s equity markets

seconds

With Zoe Zhang and Richard Knight, Execution Quant Group, CLSA

Zoe Zhang, CLSA

At CLSA’s Institutional Brokerage Division, trading in China is a major focus for us at the Trading Desk. Our team at quants is always on the lookout for ways to improve performance; One way we can accomplish this is by optimizing execution during the “closing phase”, a distinct period in Chinese markets that begins on average 40 minutes before the closing bell.

The algorithmic suites need to dynamically detect the start of the closing phase and adjust trading patterns to optimize execution results during this critical period. That’s exactly what we do at CLSA. Once we see a pattern for the closing phase, we can investigate further, and during this phase, many stock-specific trading factors used by CLSA’s adaptive trading algorithm change markedly.

final stage identification

In Chinese equity markets, thin auction volumes on both the Shanghai and Shenzhen exchanges force traders who are targeting the close to start executing their orders quickly. This creates an observable subtle compositional change and initiates the typical closing phase.

We have studied a large number of microstructure metrics, attempting to identify features that show noticeable and repeatable changes up to the closing auction.

One such metric is order book refill time, or the average time it takes for a quote size to expire. The average exchange profile for this feature is shown in the figure below.

Using the multivariate adaptive regression spline (MARS) modeling technique, we can identify the point at which this subtle structure change occurs – on average, about 40 minutes before the trade close.

There are several China-specific microstructure constraints that lead us to believe that this feature is a good proxy for behavioral changes closely related to trade. These include regulatory restrictions on the placement-to-cancellation ratio; Odd lot fills; and mixed market data.

When calculating the total volume traded within the closing phase, we find that it corresponds to the mainland markets with other developed markets accounting for about 20% of the full day’s volume (compared to <1% in the actual closing auction). brings.

It is important to note that while the effects of the closing phase can be expected to occur at approximately the same time each trading day, the exact timing varies depending on the number of participants targeting different participants as well as the liquidity of the stock. But will be different.

implications for trading strategies

Our analysis indicates that the microstructure features within the closing phase are markedly different from those observed earlier in the trading day.

Richard Knight, CLSA

A notable difference comes from the evaluation of the market impact coefficient which is significantly lower than in the pre-close phase.

Some of the implications for algorithmic trading strategies are as follows:

It is insufficient to rely on daily averages to assess microstructure characteristics. The algorithmic suite should be able to detect and adapt to the intraday trading phases.

The low market impact in the closing phase should be taken into account when calculating the optimal schedule targeting the close, VWAP or IS benchmarks.

Real-time divergence in key microstructure metrics can be an important indicator of near-target trading volume. As such, they can provide unique and differentiated features to improve the performance of volume and alpha prediction models (see example of volume prediction models below).

Within the closing phase, strategies can more passively price orders, maximizing spread capture with less chance of adverse selection.

summary

By careful observation of microstructured trading metrics, it is possible to identify changes in participant behavior that we believe are related to trades that target the close. These changes define a closing phase that brings volume targets closer to those of other developed markets. Several major microstructural features show a marked change during this period; We believe that the key to execution strategies is to react to these changes in real time to ensure optimal performance.

At CLSA we believe that China’s business pattern is evolving, and market participants need to adapt. We have embedded this discovery not only into our close algo to improve performance against benchmarks, but also into our entire adaptive suite to provide a broader trading edge.

Disclaimer: This publication is subject to the terms and conditions of use and disclaimers and includes www.clsa.com/disclaimer.html

Be the first to comment

Leave a Reply

Your email address will not be published.


*