Last Look is a hotly debated topic in the institutional FX world. Have the GFXC’s latest recommendations gone far enough to mitigate risks? Vivek Shankar investigates.
The Global Foreign Exchange Committee (GFXC) has always ensured that the institutional FX markets function smoothly to provide market participants fair and transparent opportunities. The FX markets work on a system of self-governance, and from time to time, market participants need clarifications regarding processes. To this end, the GFXC’s Global FX Code serves as a framework that functions as the single reference for participant behavior in the institutional markets .
The committee instituted a review of the Code and following feedback from stakeholders, published a guidance paper on the Last Look period. Last Look is a much-discussed and controversial topic in the FX world. The guidance paper seeks to address the issues market participants face and ensure a fair execution process. Phil Weisberg, EVP Strategic Planning and Partnerships at oneZero, explains that opaqueness surrounding order fill times could lead to suboptimal liquidity functions. The new recommendations from the GFXC together with analytics that are now available help to shed light on this matter for both LCs and LPs.
“Traditionally, it’s difficult for customers to visualize the fill times on their orders and differentiate how much of that was due to the latency of their ecommerce, aggregation or portal, and how much was due solely to the LP,” he explains.
The behavior of LPs during the Last Look period and the impact on prices LCs (Liquidity Consumers) receive is one of the major issues the GFXC has sought to address. The question all market participants are asking is: Does the new guidance go far enough to mitigate them?
Changes to guidelines
Last Look has traditionally been a problematic area in the FX world thanks to the LP’s ability to reject trade requests that might adversely affect its profitability. Principle 17 of the GFXC’s Code explicitly disallows using the Last Look period for information gathering purposes.
Despite this, Last Look guidelines are subjective and needed clarification. In its guidance paper, the GFXC makes three recommendations. The first recommendation calls for LPs to ensure a fair and effective Last Look process by minimizing the period of uncertainty while the LP checks their trade request.
“Traditionally, it’s difficult for customers to visualize the fill times on their orders and differentiate how much of that was due to the latency of their ecommerce, aggregation or portal, and how much was due solely to the LP,”
Phil Weisberg
The GFXC states that “LPs should apply the price and validity check without delay and promptly make their decision to accept or reject a trade” (underlining in the original). Any additional delay to what is required to complete a price check is considered contrary to this guidance. Also, LPs should not use any confidential information provided by the LC for reasons other than the purpose specified. Crucially, the GFXC also specifies that the word “Clients” refers to all counterparties, whether anonymous or named.
The second recommendation calls for enhanced ex-ante disclosures. As a part of these disclosures, LPs must disclose whether their price check process is applied symmetrically or asymmetrically. LPs should also disclose the length of their Last Look window to the LC. Any liquidity sourced by the LP through “cover and deal” arrangements must also be disclosed.
The final recommendation calls for LPs to regularly disclose information that allows LCs to evaluate the handling of order requests. LPs should record trade rejection information and disclose rejection reasons at a high level.
While the intent behind these recommendations is clear, how effective are they in eliminating concerns about the Last Look period?
Lack of specificity
While the first recommendation clearly states that additional holding times beyond what’s required to conduct a price check is contrary to Principle 17, its effectiveness boils down to its interpretation by LPs and their lawyers.
In a press release, GFXC Chair Guy Debelle noted that “Liquidity consumers should ask questions of their liquidity provider’s last look process, and evaluate whether to trade with liquidity providers that are using last look.”
To ensure greater transparency, the GFXC recommends increased internal controls and managerial oversight. However, these controls were present previously, and it’s debatable how much of a change this is. If the LP doesn’t believe that asymmetric holding times and latency buffering are unfair practices, increased oversight doesn’t solve any issues.
Critically, the first recommendation doesn’t address selective trade processing by LPs and offering different Last Look periods to different clients. Such behavior points to the avoidance of losing trades (adverse selection), which is against the GFXC’s Code.
To better understand adverse selection, it’s crucial to note that Last Look (as it’s currently practiced) is a combination of two practices. The first is the LP’s right to accept or reject a trade request, a legitimate process to ensure LPs manage credit risk.
The second practice, called Additional Hold Time or AHT, is debatable. It involves holding client requests for a certain period before applying a price check and accepting or rejecting the trade request. This practice explicitly contradicts Principle 17 since there’s no reason to hold an order unless it’s for information gathering purposes.
LPs justify AHT using a standard set of statements. Smaller LPs argue that it’s tough for smaller banks to price as fast as large players. However, a study conducted by XTX Markets last year noted that regional banks such as Santander, Credit Agricole, and Westpac do not apply AHT.
Latency is also an issue cited by LPs, especially orders received from clients in different parts of the globe. However, global banks such as JP Morgan and HSBC deal with latency by applying a tolerance limit to prices and do not use AHT. The recommendations now require LPs to disclose maximum and minimum Last Look times, which aids LCs. Crucially, the GFXC’s Debelle recently confirmed that no AHTs should be used in spot FX transactions.“LPs should apply the price and validity check without delay. Anything else that prolongs the last look window is contrary to the intent of the Code” he said. “Liquidity providers adhering to these principles and providing transparency about their practices should help to give their clients greater clarity about the process.”
While a hugely positive change, it remains to be seen how market participants, specifically LPs, respond to these recommendations and put them into practice.
Asymmetric Hold Times
The debate between symmetric and asymmetric price checks is a long-standing one. The GFXC’s second recommendation calls for disclosures, especially if the LP uses asymmetric hold times. While requiring the disclosure is a great move, there is concern that by allowing LPs to merely disclose this practice, asymmetric hold times are being legitimized.
Asymmetric price treatment has few supporters outside of LPs. With this practice, LPs reject any order that results in adverse selection and only fill orders when the market has moved in the LPs favor. Essentially, asymmetricity exists solely for the LPs benefit, and clients lose whatever the LP gains.
“Liquidity consumers should ask questions of their liquidity provider’s last look process, and evaluate whether to trade with liquidity providers that are using last look.”
Dr Guy Debelle
Another issue is that many LPs use asymmetric logic as the default setting for all clients instead of offering them the choice to opt-in. They justify this practice by explaining that long AHT windows maximize client fills. However, the opposite is true. XTX Markets discovered that price movements post trade requests are not random and uniformly move against the client’s interests. This means LCs experience costly slippage even when receiving symmetric pricing logic. Thus, the longer the AHT is, the worse off the client is.
The good news is that the GFXC explicitly recognizes that the choice of pricing logic adds complexity to client workflows. Disclosures regarding the background of asymmetric logic will help LCs understand the circumstances in which it is used. These disclosures will help sophisticated LCs evaluate the effectiveness of their price execution.
oneZero’s Weisberg believes this will impact LCs’ workflows. “This information will be helpful if the LC and LP have a relationship and are able to find suitable tradeoffs which lead to better outcomes for both sides.”
Better trade analysis
The GFXC’s third recommendation acknowledges problems with the Last Look process that currently exist. It explicitly states that LPs should disclose trade execution information, including rejection reasons, that allow LCs to evaluate the quality of execution they’re receiving.
Helpfully, the GFXC also highlights examples of processes that might indicate an LP neglecting Principle 17. The Last Look guidance paper notes that “High rejection rates with an unusually long or unpredictable last look window,” or “A consistently strong market reaction that occurs when dealing with a specific LP,” and high trade rejection rates might indicate a less than efficient LP.
It’s tough to attribute trade rejection to a single reason, but a combination of these occurrences, and in reasonable frequency, suggests that LCs ought to consider alternative execution means. “The longer it takes for an LC to definitively know if they are filled, the more time there is for market volatility to move the market away from them, which could result in a less optimal fill in the event of a rejection” explains Weisberg.
He also points out the utility of better trade analysis for LCs. “The market impact an LC sees post execution may matter if the LC has only traded part of their risk with the LP, because subsequent prices would move against them if the LP they chose hedges in a way that creates abnormally high market impact.”
For a long time, LCs have struggled to make sense of their execution data. Enhanced disclosures will make it easier for an LC to determine the best execution methods via sophisticated trade analysis.
The debate between symmetric and asymmetric price checks is a long-standing one
Small, but right steps
The GFXC’s recommendations have met with less than enthusiastic reviews amongst stakeholders. However, they’re undoubtedly a step in the right direction. By acknowledging the deficiencies in the last look process and working to develop processes that address both LCs’ and LPs’ concerns, the GFXC has had to walk a fine line.
The GFXC’s Code is an ever-evolving document, and there’s no doubt that the future will bring more changes that market participants wish to see.
As Weisberg points out, “Educating clients about the market microstructure is always positive and will likely lead to better outcomes for both LCs and LPs. Ultimately win-win situations can be created if the LC and LP collaborate on the formation of a liquidity function, enabling the smoothest possible risk transfer which is consistent with the LC’s execution preferences.”
Does new Last Look guidance go far enough to allay FX Market concerns?
Last Look is a hotly debated topic in the institutional FX world. Have the GFXC’s latest recommendations gone far enough to mitigate risks? Vivek Shankar investigates.
The Global Foreign Exchange Committee (GFXC) has always ensured that the institutional FX markets function smoothly to provide market participants fair and transparent opportunities. The FX markets work on a system of self-governance, and from time to time, market participants need clarifications regarding processes. To this end, the GFXC’s Global FX Code serves as a framework that functions as the single reference for participant behavior in the institutional markets .
The committee instituted a review of the Code and following feedback from stakeholders, published a guidance paper on the Last Look period. Last Look is a much-discussed and controversial topic in the FX world. The guidance paper seeks to address the issues market participants face and ensure a fair execution process.
Phil Weisberg, EVP Strategic Planning and Partnerships at oneZero, explains that opaqueness surrounding order fill times could lead to suboptimal liquidity functions. The new recommendations from the GFXC together with analytics that are now available help to shed light on this matter for both LCs and LPs.
“Traditionally, it’s difficult for customers to visualize the fill times on their orders and differentiate how much of that was due to the latency of their ecommerce, aggregation or portal, and how much was due solely to the LP,” he explains.
The behavior of LPs during the Last Look period and the impact on prices LCs (Liquidity Consumers) receive is one of the major issues the GFXC has sought to address. The question all market participants are asking is: Does the new guidance go far enough to mitigate them?
Changes to guidelines
Last Look has traditionally been a problematic area in the FX world thanks to the LP’s ability to reject trade requests that might adversely affect its profitability. Principle 17 of the GFXC’s Code explicitly disallows using the Last Look period for information gathering purposes.
Despite this, Last Look guidelines are subjective and needed clarification. In its guidance paper, the GFXC makes three recommendations. The first recommendation calls for LPs to ensure a fair and effective Last Look process by minimizing the period of uncertainty while the LP checks their trade request.
The GFXC states that “LPs should apply the price and validity check without delay and promptly make their decision to accept or reject a trade” (underlining in the original). Any additional delay to what is required to complete a price check is considered contrary to this guidance. Also, LPs should not use any confidential information provided by the LC for reasons other than the purpose specified. Crucially, the GFXC also specifies that the word “Clients” refers to all counterparties, whether anonymous or named.
The second recommendation calls for enhanced ex-ante disclosures. As a part of these disclosures, LPs must disclose whether their price check process is applied symmetrically or asymmetrically. LPs should also disclose the length of their Last Look window to the LC. Any liquidity sourced by the LP through “cover and deal” arrangements must also be disclosed.
The final recommendation calls for LPs to regularly disclose information that allows LCs to evaluate the handling of order requests. LPs should record trade rejection information and disclose rejection reasons at a high level.
While the intent behind these recommendations is clear, how effective are they in eliminating concerns about the Last Look period?
Lack of specificity
While the first recommendation clearly states that additional holding times beyond what’s required to conduct a price check is contrary to Principle 17, its effectiveness boils down to its interpretation by LPs and their lawyers.
In a press release, GFXC Chair Guy Debelle noted that “Liquidity consumers should ask questions of their liquidity provider’s last look process, and evaluate whether to trade with liquidity providers that are using last look.”
To ensure greater transparency, the GFXC recommends increased internal controls and managerial oversight. However, these controls were present previously, and it’s debatable how much of a change this is. If the LP doesn’t believe that asymmetric holding times and latency buffering are unfair practices, increased oversight doesn’t solve any issues.
Critically, the first recommendation doesn’t address selective trade processing by LPs and offering different Last Look periods to different clients. Such behavior points to the avoidance of losing trades (adverse selection), which is against the GFXC’s Code.
To better understand adverse selection, it’s crucial to note that Last Look (as it’s currently practiced) is a combination of two practices. The first is the LP’s right to accept or reject a trade request, a legitimate process to ensure LPs manage credit risk.
The second practice, called Additional Hold Time or AHT, is debatable. It involves holding client requests for a certain period before applying a price check and accepting or rejecting the trade request. This practice explicitly contradicts Principle 17 since there’s no reason to hold an order unless it’s for information gathering purposes.
LPs justify AHT using a standard set of statements. Smaller LPs argue that it’s tough for smaller banks to price as fast as large players. However, a study conducted by XTX Markets last year noted that regional banks such as Santander, Credit Agricole, and Westpac do not apply AHT.
Latency is also an issue cited by LPs, especially orders received from clients in different parts of the globe. However, global banks such as JP Morgan and HSBC deal with latency by applying a tolerance limit to prices and do not use AHT.
The recommendations now require LPs to disclose maximum and minimum Last Look times, which aids LCs. Crucially, the GFXC’s Debelle recently confirmed that no AHTs should be used in spot FX transactions.“LPs should apply the price and validity check without delay. Anything else that prolongs the last look window is contrary to the intent of the Code” he said. “Liquidity providers adhering to these principles and providing transparency about their practices should help to give their clients greater clarity about the process.”
While a hugely positive change, it remains to be seen how market participants, specifically LPs, respond to these recommendations and put them into practice.
Asymmetric Hold Times
The debate between symmetric and asymmetric price checks is a long-standing one. The GFXC’s second recommendation calls for disclosures, especially if the LP uses asymmetric hold times. While requiring the disclosure is a great move, there is concern that by allowing LPs to merely disclose this practice, asymmetric hold times are being legitimized.
Asymmetric price treatment has few supporters outside of LPs. With this practice, LPs reject any order that results in adverse selection and only fill orders when the market has moved in the LPs favor. Essentially, asymmetricity exists solely for the LPs benefit, and clients lose whatever the LP gains.
Another issue is that many LPs use asymmetric logic as the default setting for all clients instead of offering them the choice to opt-in. They justify this practice by explaining that long AHT windows maximize client fills. However, the opposite is true. XTX Markets discovered that price movements post trade requests are not random and uniformly move against the client’s interests. This means LCs experience costly slippage even when receiving symmetric pricing logic. Thus, the longer the AHT is, the worse off the client is.
The good news is that the GFXC explicitly recognizes that the choice of pricing logic adds complexity to client workflows. Disclosures regarding the background of asymmetric logic will help LCs understand the circumstances in which it is used. These disclosures will help sophisticated LCs evaluate the effectiveness of their price execution.
oneZero’s Weisberg believes this will impact LCs’ workflows. “This information will be helpful if the LC and LP have a relationship and are able to find suitable tradeoffs which lead to better outcomes for both sides.”
Better trade analysis
The GFXC’s third recommendation acknowledges problems with the Last Look process that currently exist. It explicitly states that LPs should disclose trade execution information, including rejection reasons, that allow LCs to evaluate the quality of execution they’re receiving.
Helpfully, the GFXC also highlights examples of processes that might indicate an LP neglecting Principle 17. The Last Look guidance paper notes that “High rejection rates with an unusually long or unpredictable last look window,” or “A consistently strong market reaction that occurs when dealing with a specific LP,” and high trade rejection rates might indicate a less than efficient LP.
It’s tough to attribute trade rejection to a single reason, but a combination of these occurrences, and in reasonable frequency, suggests that LCs ought to consider alternative execution means. “The longer it takes for an LC to definitively know if they are filled, the more time there is for market volatility to move the market away from them, which could result in a less optimal fill in the event of a rejection” explains Weisberg.
He also points out the utility of better trade analysis for LCs. “The market impact an LC sees post execution may matter if the LC has only traded part of their risk with the LP, because subsequent prices would move against them if the LP they chose hedges in a way that creates abnormally high market impact.”
For a long time, LCs have struggled to make sense of their execution data. Enhanced disclosures will make it easier for an LC to determine the best execution methods via sophisticated trade analysis.
Small, but right steps
The GFXC’s recommendations have met with less than enthusiastic reviews amongst stakeholders. However, they’re undoubtedly a step in the right direction. By acknowledging the deficiencies in the last look process and working to develop processes that address both LCs’ and LPs’ concerns, the GFXC has had to walk a fine line.
The GFXC’s Code is an ever-evolving document, and there’s no doubt that the future will bring more changes that market participants wish to see.
As Weisberg points out, “Educating clients about the market microstructure is always positive and will likely lead to better outcomes for both LCs and LPs. Ultimately win-win situations can be created if the LC and LP collaborate on the formation of a liquidity function, enabling the smoothest possible risk transfer which is consistent with the LC’s execution preferences.”