Best execution in UK listed cash equities – wholesale banks

This multi-firm review will be of interest to:

• Banks and brokers who provide securities dealing and execution.
• Investors in securities who rely on banks and brokers to execute on their behalf to deliver the best possible execution outcomes. This includes asset managers who also have obligations in relation to best execution.  

While this review focused on UK listed cash equities, some findings are relevant to other products.

Best execution is fundamental to investor protection, maintaining trust in UK markets and attracting investment and trading activity. Banks have a key role in delivering best execution for clients who rely on them to act in their best interests.

We last carried out a thematic review (TR14/13) of wholesale banks’ best execution practices in 2014. That review identified a risk that best execution was not being consistently delivered to all clients.

Since then, market practices in the equities market have continued to evolve. In particular, firms increasingly rely on execution algorithms and smart order routing technology. Market dynamics have also shifted, with liquidity moving towards bilateral trading, including systematic internalisers (SIs). There has also been a move to increased trading during closing auctions on exchanges.

In the context of regulation, the Markets in Financial Instruments Directive (MiFID II), which came into force in 2018, has extended the best execution rules and regulatory framework for SIs in equities.

This review looked at 8 wholesale banks active in UK listed cash equities dealing and execution.  

This sample provided broad coverage across the market and included large and mid-size banks.  

MiFID transaction reports submitted to us show the UK equities market (excluding over-the-counter) in Q3 2025 had notional volumes of around £1.3 trillion and 185m transactions. The banks in our sample accounted for around 38% of the notional volumes and 67m transactions in this market.  

Our assessment considered:  

1. Whether banks had suitably scoped their best execution obligations.
2. The strength of banks’ governance and oversight.
3. The robustness of banks’ best execution monitoring and management information (MI).
4. Whether banks are identifying and managing conflicts of interest appropriately when internalising client orders.  

We assessed these areas against our:

• Best execution rules set out in COBS 11.2A.
• Conflict-of-interest rules set out in SYSC 10.

Our starting point was to analyse information in banks’ transaction reports and their own data to understand arrangements and outcomes.

We followed this up with meetings to explore individual banks’ approaches. We also spoke with buy-side representatives and trade associations.

We were encouraged to find stronger practices compared to our 2014 thematic review.  

Of the areas we assessed, banks generally had strong practices in assessing the scope of best execution. We also found no evidence that internalisation was damaging client outcomes.

Banks’ monitoring of best execution was able to identify good and poor outcomes. We also saw evidence of banks taking action to address examples of poor outcomes. In contrast, the quality of management information (MI) to support senior management oversight was variable. We found some MI was comprehensive, but we also found examples of it being either too high level or overly complex.

While we observed some good practices in governance and oversight, this was also where we found the least progress made since the 2014 review. Specifically, we found the need for improvement in the challenge from the second line of defence. The banks who were the strongest in this area had empowered their compliance functions, supporting them with the right data and tools.

We have given individual feedback to the banks in our sample, highlighting where we have identified good practice and areas for improvement. We’ve highlighted some of those examples in this report for firms to consider. We will continue discussions with wholesale banks on their approaches to best execution.

During our review, we received feedback from firms on the rules and guidance covering the scope of best execution and the requirements to get express consent, outlined in COBS 11.2A.24R. We will consider this feedback in any future review of our rules.

Furthermore, in July 2025, we published CP25/20 seeking views on market effectiveness in light of recent developments in market structure. 

We announced changes to the systematic internaliser (SI) regime for bonds and derivatives, structured finance products and emissions allowances in PS25/17. We continue to consider feedback and conduct further analysis to bolster our view on equity markets. This analysis considers the structure of trading and the liquidity provided by SIs, utilising MiFID II transaction data and order book information. In parallel, we’re also exploring execution outcomes across different venues. 

If we consider that proposing rule changes is appropriate, we will consult on those changes in the first half of 2026.

In quote-driven markets – for instance, in most bond markets but also in parts of the equities markets, such as the exchange traded fund (ETF) or convertible bond markets – banks often deal on own account.

When a firm deals on own account in a quote-driven market, the European Commission’s four-fold cumulative test can be a useful analytical tool to help the firm assess and evidence whether it is acting on behalf of the client (‘best execution obligation’).

The test considers:

• Who initiated the transaction.
• Market practice and ability to ‘shop around’.
• Levels of price transparency within a market.
• Information provided by the firm and any agreement reached with a client.

All the firms in our sample used this test, evaluating different characteristics of the order and the market, although in different ways.

In line with the test, dealing on risk in quote-driven wholesale markets does not usually involve acting on behalf of a client. This is because pricing information in these markets is generally readily available and the market convention is for clients to shop around for quotes from different dealers. Therefore, clients in these markets rarely rely on the dealer for pricing or other aspects of the trade which could create a best execution obligation.

However, there are some limited circumstances where in these markets a firm may be acting on the client’s behalf. For instance, where a firm provides a risk price in an illiquid market where the client is unlikely to have access to the instrument pricing or cannot shop around for quotes, the client may rely on the firm to price the transaction. As a result, the firm dealing on own account is effectively executing on behalf of the client and best execution applies.

We received feedback that assessing the scope of best execution obligations in quote-driven wholesale markets, where firms deal on their own account, can be complex. We will consider this in any future review of our best execution rules.

Our review found banks had a better understanding of the scope of the best execution obligations compared to our 2014 thematic review. Then, we found several firms sought a blanket exclusion from best execution obligations for dealing on quotes. 

Good practice in this review saw most banks assessing their obligations in quote-driven markets. They recognised that even when dealing on risk on request for quotes (RFQ) with non-retail clients, there may be limited instances where they may in fact be executing on behalf of a client, and so best execution applies. To evaluate for these outcomes, firms periodically applied the four-fold test to their activities.

We also previously found some firms excluded any order that had an element of client specific instructions from the scope of the best execution obligation. We saw no evidence of that during this review. However, as we set out below, client specific instructions were the most frequent reason for trades falling outside of the banks’ threshold for best execution.

Some banks considered direct market access (DMA) or sponsored access (SA) to be specific instructions. DMA and SA occur where the client places orders with venues directly using the bank’s infrastructure or trading code without the bank’s intervention. They therefore treated these as orders where they had satisfied best execution. In contrast, banks did not consider a client’s selection of algorithms to be specific instructions unless they specifically customised the algorithms for the client. We did not identify problems with these approaches. 

Nonetheless, our review identified some room for improvement. 

In quote-driven markets, we found some banks inflexible in their best execution framework. 

For example, we saw some banks apply the four-fold test only at a single point in time or with limited frequency without considering that market dynamics, conventions and client behaviour can change.  

Firms’ frameworks should be flexible enough to ensure best execution is applied and evidenced.  

All banks reassessed the application of best execution to their business at least annually as part of reviewing the Order Execution Policy (OEP) and execution arrangements more broadly. One bank reviewed the scope of its obligations quarterly.

We observed that banks apply a range of approaches to the scope review. While most did a high-level assessment of their trading activities against the factors included in the four-fold test, others applied a more detailed approach. 

One bank supported the review with attestations from different desks on their activities. The results of the attestations were reported to the best execution committee overseeing the review with supporting data which, for example, captured the number of quotes won by the desk.

Another bank conducted sample testing of transactions traded on an RFQ basis in non-equities markets to assess factors of the four-fold test post-event. This included reviewing the level of competition in the market and the type of client and counterparty activity.

Review of the scope of best execution obligations is fundamental to good outcomes. Firms should examine how and when they undertake these reviews against their execution outcomes, using a flexible framework allowing easy adaption to market and client activity changes. 

Best execution monitoring by banks in our sample consisted of real-time and post-trade monitoring. Overall, we found that banks’ scope of, and approach to, real-time and post-trade monitoring was fit for purpose and proportionate.  

Real-time monitoring

Real-time monitoring generally considered:  

1. Prices and volumes available for instruments at various venues.
2. Platform availability and connectivity.
3. Client order preferences.  

Banks’ real-time monitoring would typically use alerts, thresholds or predefined rules that identified, and sometimes stopped, orders being handled if the monitoring suggested best execution or client requirements were not being met. For the systematic internaliser (SI), banks would also monitor the quotes it streamed, ensuring they were in line with the BBO/ European Best Bid and Offer (EBBO). Most also validated this on a T+1 basis.  

Post-trade monitoring

Post-trade monitoring involved banks measuring the execution outcomes delivered across different execution factors and tracking trends across different types of executions, such as trades using a certain algorithm or trades executed on a certain venue.

Our rules outline several execution factors for banks to consider when executing a trade. These include:

• price
• costs
• speed
• likelihood of execution and settlement
• size
• nature  
• any other relevant considerations

All but one bank monitored delivery of best execution across these factors using a combination of widely adopted transaction cost metrics and proprietary measures, benchmarking outcomes against pre-defined tolerances.

Many banks highlighted implementation shortfall, the difference between the price of the instrument at the time the broker receives the order and the final execution price, or a variation of this, as a particularly important performance metric. Implementation shortfall was also an important metric for the buy-side. However, its relevance varied depending on the order’s objectives, and firms typically combined it with other measures to form a holistic view of execution quality.

All but one bank had implemented pre-defined tolerance thresholds within their monitoring. This ensured a consistent approach to identifying trades where best execution may not have been delivered. The outlier bank relied on traders to determine on a case-by-case basis whether best execution had been delivered.  

Relying on case-by-case judgement to assess execution quality makes it more difficult to identify areas where outcomes are not consistently the best, particularly given the complexity of the data and the expertise required to interpret it.

Some banks reviewed the tolerance threshold for outliers at least annually, with one bank reviewing quarterly. Where there was a formal review, good practice evidenced both business and compliance challenge. At one bank, the thresholds hadn’t been revised since their introduction several years ago, raising concerns about their effectiveness, given changing market conditions and the firm’s arrangements over time.

Banks adopted varied approaches when reviewing outlier trades to establish root causes.

Two banks checked all their outlier trades, one bank checked all trades beyond a certain limit of their tolerance threshold and the others took a sample-based approach. Trends from this analysis were typically shared with the banks’ best execution committee or working groups.

We found that banks who regularly analysed outlier trades demonstrated more actions to continuously evolve their arrangements than those which did not.

The biggest driver of outlier trades for most banks were client specific instructions where best execution is limited to the aspects of the trade not covered by the instruction. 

Banks did not undertake any specific monitoring or analysis of execution quality for different client types to ensure their execution arrangements were effective for all clients. This is because they did not expect to see different outcomes. Since their algorithms were generally available to all clients, any performance enhancements the banks made benefited the entire client base.  

We assessed whether banks’ execution arrangements delivered different outcomes for clients based on their importance to the firm. We did this by reviewing banks’ execution performance metrics between the top 80% and bottom 20% of revenue generating clients. 

We were satisfied there were no significant differences in the overall outcomes between the 2 client groups.

All banks reported having both formal and informal client feedback channels in place. 

Formal channels included data-driven performance reviews of the execution service banks had provided at specific intervals. These bespoke sessions were typically reserved for the banks’ larger clients, where the smaller clients would request their Transaction Cost Analysis (TCA) report monthly, quarterly or on request.

Banks told us performance reviews were a valuable source of feedback. This was particularly the case when they included information from clients that allowed the banks to rank themselves against their peers and made suggestions for improvement and new products. Most of the banks in the sample did not capture the formal performance review feedback in their MI, nor present it to their committees. This was because the form of the review was tailored to the client, making comparison difficult. 

All banks acknowledged it was rare to receive client complaints. Some said they automatically included any complaints in the MI reported to oversight committees. All had formalised procedures for handling client complaints.

The quality of MI produced by banks varied. 

Some banks produced comprehensive MI covering both arrangements and outcomes, while at others it was high-level. 

Banks with strong MI presented to senior managers were best able to evidence challenge and compliance with their best execution governance and frameworks. 

Where MI was limited, we observed weaker challenge and scrutiny by senior management. In one case, a bank whose MI only covered outlier analysis, reported that its committees had not generated any substantial discussion or actions to improve the framework.  

Limited MI reduces transparency and makes it harder for senior management to identify risks, validate outcomes and drive improvements. Good practice involved producing meaningful, outcomes-focused MI that enabled effective challenge and informed decision-making.

Banks with the most comprehensive MI available to senior managers included information covering:

• An overview of the cash equities market and regulatory landscape during the relevant period.
• Working group escalations and updates.  
• Outlier trend analysis, including themes and rationales.  
• Volume trends in the firm’s use of different execution venues.
• Algorithm volumes, performance and testing results.
• SOR performance, including hit rates (the amount of liquidity the SOR is successfully capturing), fill rates, latency, testing results and rationale for any changes.
• Technology incidents.  
• Client complaints and feedback.  
• New product approvals.  
• Number of client consents for off venue trading and those restricting the publication of limit orders (COBS 11.2A.24R and COBS 11.4.1.R)
• Consideration of any OEP changes.  
• Use of third-party brokers.  
• Open and closed actions with action owners and time to complete.

Banks with less comprehensive MI primarily focused on presenting outlier analysis. High-level MI at committee stage was typically associated with banks that had sub-committees or working groups which undertook detailed analysis. While firms should avoid unnecessary duplication between the different committee levels, we saw more evidence of challenge at banks where senior management and compliance had more detailed information.

An area where some banks had challenges was in monitoring and reporting MI for their high-touch business.  

For instance:

• One bank had not implemented continuous monitoring of best execution for high-touch trades and so no metrics or information on this was included in the best execution committee packs.  
• Another bank did not have processes capable of producing MI for client order handling instructions that came through in a non-systematic way, such as through phone calls, email or chat.  

These banks have addressed, or are addressing, these issues.  

Firms should ensure they have procedures that allow them to monitor all in-scope trades and related client order handling instructions.

To ensure compliance with best execution obligations in SI executions, all banks in the sample applied a range of controls.

These included:  

• Implementing SOR logic and execution algorithms designed to objectively assess and deliver optimum outcomes across a range of venues.
• Regularly reviewing execution quality for internalised orders, including benchmarking against executions at external venues.
• Real-time monitoring of execution quality across both SI and external venues.
• Using information barriers to limit the visibility of internalised flow to reduce the risk of front running client orders.
• Providing staff training and guidance on order execution practices.