Saudi Arabia Adopts Algorithmic Order Cap; Proposed IPO Bid Changes Await Approval

Saudi Arabia's Capital Market Authority has adopted a limit on algorithmic order traffic in Main Market securities. A separate plan would make institutional IPO bids and underwriting commitments harder to withdraw. The first has moved beyond consultation. The second has not been shown to do so in the CMA material reviewed for this update.

Both address displayed demand, but at different stages of a security's life. In continuous trading, an algorithm can add or remove orders before a trade occurs. In an IPO, institutions can submit bids that help set the offer price before they pay. The CMA wants more discipline in both processes. The public evidence for the size of either problem remains thin.

The CMA's 29 October notice on algorithmic trading says its board adopted requirements beginning on November 1, 2026. Institutions have until December 31 to adjust their arrangements. That is an implementation period for an adopted measure, not a sign that the measure is still a draft. The separate IPO consultation closed on October 22 and described November 2 as an intended effective date if final provisions were approved. The consultation alone does not establish that approval.

The November 1 limit lacks a published baseline

From November 1, the final CMA notice sets a 20:1 ceiling on the ratio of algorithmic orders to executions during the open session, measured for each registered trader and each Main Market security on a trading day. The notice names the high, medium, low and very-low liquidity groups. It does not name the very-high group. It also calls for supervisory systems, testing, monitoring, alerts, records and the ability to stop an algorithm when necessary.

That is more specific than a general duty to control trading systems, but it is not a complete technical rulebook in the public notice. The earlier September consultation proposed a 20-order ceiling and linked it to the registered trader rather than each underlying client. Its draft text counted entered orders, twice the number of amended orders and cancelled orders, then subtracted one; it divided that total by executed trades plus one. Those are the mechanics of the September proposal. The October notice does not say whether that exact numerator and denominator were adopted.

The distinction matters to a broker running several clients' strategies through one registered trader. A ratio calculated across that trader can make one client's order changes consume capacity that another client expected to use. The public notice does not resolve how a partial fill counts, whether amendments and cancellations retain their proposed weights, or how an institution should manage several unrelated algorithms under one registration. Those are implementation questions, not evidence that the adopted 20:1 ceiling is optional.

Saudi Tadawul Group's 2024 annual report said algorithmic traders contributed up to 40% of daily cash-market liquidity. That makes the potential market effect material, but it cannot tell us how many trader-security days breach 20:1. The CMA's notices do not publish the distribution of existing ratios by liquidity group, the share of orders cancelled or amended, or the share of executions attributable to the traders the limit would constrain.

A cap can reduce messages that create a misleading impression of depth. It can also discourage a legitimate trader from updating a quote when prices move. Without the pre-rule distribution and a clear counting method, the public cannot tell whether 20:1 targets a small tail of excessive traffic or changes ordinary liquidity provision more broadly.

The IPO book is still a separate proposal

The existing bookbuilding instructions already require financial advisers and institutions receiving bids to verify a participant's solvency and ability to pay before registering its order. They permit participating institutions to amend or cancel bids during bookbuilding. The issuer must sign an underwriting agreement before subscription begins.

The September IPO consultation proposed tighter commitments. It would limit acceptable evidence of bid capacity to cash or cash equivalents, make participation orders binding for payment by the subscription deadline, and require the underwriting agreement to take effect before bookbuilding. The underwriter's obligation to buy all offered shares would begin when bookbuilding starts. If that ownership would breach listing requirements, the issuer would not list and the underwriter would still buy the offer. These are proposed changes, not current obligations established by the consultation.

Moving an underwriter's commitment forward changes when risk is carried, even if a heavily subscribed offer ends with little or no stock in the underwriter's hands. Arabian Mills' September 2024 offer announcement reported institutional coverage of about 132 times. That figure shows gross demand for one historical offer. It does not show how much of the book was backed by cash, how many bids were cut before allocation, or what an underwriter ultimately held. One successful book cannot measure the market-wide risk the proposal is meant to address.

The draft also asks the issuer to publish at least a year of forward-looking performance indicators and the financial adviser to examine those forecasts. That could give investors more to test against the price, but it increases the importance of stating assumptions and distinguishing a forecast from a promise. Again, the consultation describes a possible rule, not an enacted disclosure duty.

Show the cost and the failure

The adopted trading ratio and the proposed IPO commitments share a policy logic: make displayed demand more credible by attaching a cost to orders that do not turn into trades or funded subscriptions. Credibility is a plausible benefit. Its size has not been demonstrated in the public material.

For algorithmic trading, the CMA could publish order-to-execution distributions by liquidity group, including the contribution of amendments and cancellations, and explain the final calculation. For IPOs, the useful measures are bid reductions, cancellations, failed payments, underwriter take-up and realised losses. These would let readers distinguish a targeted repair from a rule that makes all market participants carry a higher cost.

Saudi Arabia has made the trading decision. The IPO decision requires its own adoption evidence. The next test is whether either measure improves the quality of displayed demand without making useful liquidity needlessly scarce.