Why would CMA restrict algorithmic trading? What impact on liquidity?

Logo of Saudi Capital Market Authority (CMA)

The Capital Market Authority (CMA) is moving to introduce new restrictions on algorithmic trading in the Saudi market as part of a wider plan to develop the rules governing such activity, enhance its safety and efficiency, and prevent adverse market effects or trading disruptions.

On Sept. 24, the CMA put its draft Regulatory Provisions on Restrictions on Algorithmic Trading up for public consultation for 30 days, ending Oct. 24. The provisions are scheduled to take effect in their final form on Nov. 1.

What is the CMA proposing?

A key element of the draft is a cap on the ratio of algorithmic trading orders to executed trades in a security at 20 orders for every executed trade during a single trading day.

The limit would apply to securities listed on the Main Market (TASI), excluding those classified as very highly liquid securities, with the ratio calculated at the individual trader level.

For example, if 1,000 buy orders are computed for a stock and 40 trades are executed, the ratio would be 25 orders per trade, exceeding the proposed limit, while 50 executed trades would bring the ratio down to 20 orders per trade.

The proposal does not impose a cap on the number of trades. Rather, it limits the relationship between the number of orders sent to the market and those that ultimately result in actual executions.

How large is algorithmic trading in the Saudi market?

Some market institutions have the technological infrastructure to access the market directly and execute orders automatically through services such as direct market access (DMA) and FIX, as well as execution algorithms including VWAP and TWAP.

Fahad Alhuwaimani, a financial markets and economics specialist and board member of Masar Alnumou Finance Co.

Fahad Alhuwaimani, a financial markets and economics specialist and board member of Masar Alnumou Finance Co., said algorithmic trading is no longer a limited activity in the Saudi market as it was several years ago, but has become an important part of the trading and liquidity infrastructure, adding that the new rules could have a tangible impact on order behavior in the market.

Tadawul’s previous data showed that high-frequency trading accounted for about 25% of average daily trading value, rising to 46% on some days, while algorithmic traders accounted for as much as 40% of daily liquidity, Alhuwaimani told Argaam.

The data dates back about two years, and algorithmic trading's contribution may have increased since then, he said.

Saad Althaqfan, a board member of the Saudi Economic Association (SEA)

Saad Althaqfan, a board member of the Saudi Economic Association (SEA), estimated that algorithmic trading currently accounts for as much as 25% of total trading in the Saudi market, noting that it remains below levels seen in global markets.

Algorithmic trading accounts for about 60% of trading in global markets under normal conditions and may rise to 80% during periods of elevated volatility, Althaqfan told Argaam.

There is still room for algorithmic trading to grow in the Saudi market, driven by technological advances, growing interest among local and international investors, and the development and modernization of the regulatory framework governing such activity, he said.

Growth in algorithmic trading could deepen the market by adding more orders and narrowing bid-ask spreads, thereby supporting liquidity, he added.

Why could a high number of orders be a problem?

It is important to distinguish between the number of orders and actual market liquidity, Alhuwaimani said, noting that a key characteristic of algorithmic trading, particularly high-frequency trading, is the submission, modification and cancellation of large numbers of orders, many of which do not result in actual trades.

Reducing the number of such orders would not necessarily lead to a proportional decline in liquidity, but could make the order book more reflective of actual buying and selling intentions, he said.

How does the 20-orders-per-trade threshold serve the local market?

Alhuwaimani said a threshold of 20 orders per executed trade represents a reasonable starting point, particularly given the exclusion of very highly liquid securities, adding that algorithmic strategies and market making in highly liquid stocks inherently require frequent order updates in response to changes in prices, supply and demand.

At the same time, Alhuwaimani said the limit could constrain some strategies in less-liquid stocks, particularly algorithms that continuously adjust quotes without securing rapid executions, stressing the importance of monitoring the results after implementation to assess the actual impact on trading strategies and market liquidity.

Althaqfan also said the proposed limit was appropriate, noting that 20 orders per trade was a suitable level, particularly with very highly liquid securities excluded, and that applying the limit to less-liquid securities could help mitigate risks arising from high order volumes during periods of elevated volatility.

Are there similar restrictions in global markets?

Rules under MiFID II in the European Union include mechanisms for monitoring the ratio of unexecuted orders to trades, with the ratio calculated for each member or participant and each financial instrument based on the number and size of orders.

In India, the Securities and Exchange Board of India (SEBI) applies an order-to-trade ratio framework to algorithmic trading, including measures aimed at curbing excessive increases in the ratio.

Such activity is regulated to varying degrees in a number of global markets, Alhuwaimani said, noting that the European Union, for example, uses the order-to-trade ratio as part of its framework governing algorithmic trading while taking into account the liquidity of financial instruments and the nature of market participants.

Global regulatory frameworks also include requirements for testing algorithms before deployment, monitoring them and setting order limits, as well as the ability to halt them in the event of a malfunction, he added.

What would change for brokerage firms?

The draft bill goes beyond the 20-orders-per-trade threshold and includes supervisory and operational requirements for market institutions offering algorithmic trading, including establishing monitoring systems and procedures, testing algorithms before deployment and monitoring their operation, as well as the ability to intervene and halt an algorithmic trading system when necessary.

Brokerage firms would also be required to maintain records and provide regulators with the requested data. The regulations would likely result in additional compliance costs for institutions using algorithmic trading systems, while some strategies may need to be modified to comply with the new requirements, Alhuwaimani said.

Is the CMA targeting algorithmic trading itself?

Alhuwaimani said the aim of the rules is not to restrict algorithmic trading itself, which has become an integral part of modern markets and contributes to greater liquidity, narrower bid-ask spreads and faster price discovery.

Rather, the objective is to curb inefficient practices, such as excessive flows of orders, modifications and cancellations that consume market-system capacity without necessarily reflecting actual liquidity, he said.

The CMA has also not linked the draft to a specific manipulation incident or violation. Instead, it has presented the initiative as part of efforts to develop the rules governing algorithmic trading and enhance the safety and efficiency of trading.

What could change after implementation?

The proposal could improve market efficiency and depth, with a focus on less-liquid securities potentially helping limit the effects of high order volumes in those securities, Althaqfan said.

Alhuwaimani said the rules could have a positive impact on the market if they succeed in reducing inefficient orders without adversely affecting market making or actual liquidity.