FIX Trading Community

 

 Remember me

Register  |   Lost password?

Risk Management

Discussions > Risk Management > New Risk Management Discussion Forum

New Risk Management Discussion Forum

Complete message thread from old site

Program Office Forum Posts (Archive)
1191 days ago,(2011/01/19)

All,

This ‘Risk Management’ forum has been created as a centralized way for individuals to share feedback with the group on the recent published guidelines: http://www.fixprotocol.org/documents/5537/FPLEquityRiskControls_final.pdf as well as share any additional comments or questions regarding the implications of electronic trading on risk management.

Just as an FYI for this group, the following was posted on the General Q/A forum: http://www.fixprotocol.org/discuss/read/ebd5b50d.

Regards,
FPL Program Office

John Prewett
1191 days ago,(2011/01/19)

John Harris has raised objection to the "pause" concept as proposed.

There are two positions currently being held:
1. No order should ever be paused if it exceeds a risk check, it should just be rejected.
2. All orders should be paused if they exceed a risk check, so that a good decision can then be made.

John Harris, on behalf of the electronic trading community amongst others appears to support #1. The authors of the original document support #2 on behalf of medium/high touch customers.

Can't we all just get along?
:-)

A new FIX tag submitted with the order could be used to indicate the preference as to what behavior (hold or reject) is desired if the risk checks were violated. Then both sides would surely be satisfied.

Clearly there would need to be agreement over what the default behavior would be if this preference tag were omitted, but that would be a substantially less heated debate.

So I suggest the document is updated to show that any order can either be paused or immediately rejected, depending on the order's preference (as indicated by a FIX tag).

Thanks

JohnP

FIX Trading Community
1191 days ago,(2011/01/19)

[ original email was from John Harris - john.harris@bondmart.com ]
Program Office,

Thank you for your kind reply to my statement of opposition to the "pause" feature you advocate in your new risk-management guidelines.

As published, the stated objective of FPL's Equity Risk Controls paper is "to provide information around risk management and encourage firms to incorporate best practices in support of their electronic trading platforms." The document says that the recommended guidelines apply to "electronic orders delivered directly to an algorithmic trading product, or to a direct to market (DMA) trading destination" and suggests that adoption of the guidelines will "systemically minimize the inherent risk of executing electronic algorithmic and DMA orders."

In reading the paper, I did not perceive the argument for limited applicability of the guidelines that you set forth in your reply ("This class of order flow can be considered to be 'medium touch' where the client is expecting a relatively high level of service from their electronic sales/trading coverage"). Neither am I persuaded that "Algorithmic orders typically have long trading horizons."

Even if "pausing" invalid or suspect electronic orders is, as you suggest, "a relatively common practice in the Cash Equities space," best practice - identification of which is the purpose of the paper - may well be (and in my experience is) another matter. And, respectfully, if tested, I doubt that your assertion that manual intervention reduces client workflow would prove to be the case in actual practice.

As I noted in my original post, the pause feature is burdened with risks not addressed in the paper. Where is the audit capability either for repeated errors by the originator or the middleman? How is front-running risk to be managed? How exactly are human beings supposed to judge a priori the market impact of orders and how will the effectiveness of such judgments be evaluated? To the extent this impact is "adverse" (as suggested in the paper), from whose perspective shall such adversity be judged? Are not adverse circumstances for some beneficial for others?

Not raised in my prior but as long as we're on the subject: are the flow diagrams even correct? Do DMA orders really go through SORs before entering an exchange?

If pausing (and then manually processing) electronic orders is truly a "best practice," then I suppose I will always be guilty of inferior practice. And I know I would never hire a risk manager who pitches such a scheme to me.

I mean no disrespect to the members of the Risk Management Committee and appreciate how much effort has gone into this document already, but sincerely hope they will take another crack at these recommendations.

Best,
John

> All,
>
> This ‘Risk Management’ forum has been created as a centralized way for individuals to share feedback with the group on the recent published guidelines: http://www.fixprotocol.org/documents/5537/FPLEquityRiskControls_final.pdf as well as share any additional comments or questions regarding the implications of electronic trading on risk management.
>
> Just as an FYI for this group, the following was posted on the General Q/A forum: http://www.fixprotocol.org/discuss/read/ebd5b50d.
>
> Regards,
> FPL Program Office

FIX Trading Community
1191 days ago,(2011/01/19)

[ original email was from John Harris - john.harris@bondmart.com ]
Mr. Prewett's suggestion is a good one and even holds the promise of an enhanced revenue stream for those brokerage firms that support the new field - they can charge more for handling those orders that request manual intervention if risk parameters are violated.

:-;

> John Harris has raised objection to the "pause" concept as proposed.
>
> There are two positions currently being held:
> 1. No order should ever be paused if it exceeds a risk check, it should just be rejected.
> 2. All orders should be paused if they exceed a risk check, so that a good decision can then be made.
>
> John Harris, on behalf of the electronic trading community amongst others appears to support #1. The authors of the original document support #2 on behalf of medium/high touch customers.
>
> Can't we all just get along?
> :-)
>
> A new FIX tag submitted with the order could be used to indicate the preference as to what behavior (hold or reject) is desired if the risk checks were violated. Then both sides would surely be satisfied.
>
> Clearly there would need to be agreement over what the default behavior would be if this preference tag were omitted, but that would be a substantially less heated debate.
>
> So I suggest the document is updated to show that any order can either be paused or immediately rejected, depending on the order's preference (as indicated by a FIX tag).
>
> Thanks
>
> JohnP