Dialing Down The Risk

George Rosenberger, Managing Director and Head of ConnEx, ConvergEx Group’s outsourced FIX connectivity solution, looks at some of the major themes in risk management and the value of client profiling.
What is the current state of risk management controls?
There have been a number of recent market events and new regulations that have highlighted why risk management continues to be a very important and ever-evolving topic in our industry. Sell-side firms need to invest in up-to-date risk management tools now to make sure that they have the right risk controls in place to protect themselves against an unexpected market event today or in the future. They also need to ensure that they have the right technologies and protocols in place to make sure they themselves do not send erroneous trades to exchanges or market centres.
A comprehensive risk management strategy should take into account fat-finger checks, credit checks, and checks at the execution system level as well as at the execution venue level. This three-tiered approach helps provide a comprehensive risk management framework.
First, it’s critical that firms have specific fat finger checks and credit checks in place at the FIX gateway level or as high up in their infrastructure as possible. When firms perform client profiling, it is typically done at that level and therefore it is logical to start applying risk management checks there as well. When an order is received from one of their clients, firms should already understand, through their client profiling, the type of trading the client does as well as their overall trading patterns including their normal volume, the number of orders that they send and the velocity of those orders. From a credit perspective, firms should also ensure that they have an appropriate buying power set for their clients and have an idea of how quickly their client consumes that preset buying power on any given day.
Once client profiles are in place, firms should create intelligent risk management checks to identify abnormal trading patterns from clients. If an irregularity is detected, the system should alert users so that they can pause the order, verify the order with their client and then resume it or cancel it as appropriate.
The next logical place for risk checks is at the execution system level — whether it be at a firm’s smart order router, algorithmic trading engine or exchange gateway. That’s important because not only are you protecting your client, but you’re also protecting yourself. Risk checks at the execution system level ensure that your systems don’t have a runaway algorithm/model that starts routing erroneous orders.
When creating execution system profiles, there are ways systems can help customers identify how an algorithmic trading service, smart router or trading model should work effectively. These systems can also help provide guidance with their decision-making when it comes to best practices and prevention of erroneous trading.
Finally, firms need to be cognizant of exchange or execution venue-based risk checks. These supplemental risk checks can greatly complement their own internal risk checks if thoroughly understood. Most of the execution venues today have some level of risk checks in place, whether it is buying power, price collars, singlestock circuit breakers, etc. The reason why exchanges and ECNs have this logic in place is to prevent market disruptions or dislocation.
So a comprehensive risk management strategy factors in all three types of risk checks: the fat-finger checks at the gateway level, order checks at the execution system level as well as risk checks at the execution venue level.
How fast is that environment changing?
The risk management landscape and regulatory environment are constantly changing.
On November 3, 2010, the Securities and Exchange Commission adopted Rule 15c3-5 that required brokerdealers to maintain risk controls in connection with their market access. Because Rule 15c3-5 is a principal-based rule, brokers-dealers are responsible for deciding the appropriate level of risk monitoring that they should perform. All brokerdealers need the traditional fat-finger checks such as order size, notional value, price checks, etc., but there are also other more sophisticated types of risk monitoring that must be considered as well.
One example of this type of advanced risk monitoring is duplicative order checks. How do you properly ensure that you’re not taking in duplicate orders from a client? Do you just look at the order ID FIX tag or do you implement more advanced matching logic that takes into consideration side, symbol, quantity and price over a predefined time period from the same client? These are implementation decisions that broker-dealers must consider. It can be very challenging to find the appropriate balance between protecting the client, protecting the firm as well as adhering to market regulations without causing unnecessary business interruption.
What are the principle challenges for you when it comes to this kind of system? What keeps you awake at night?
The principle challenge for risk management is keeping up with ongoing client profiling.
Client profiling is critical to a successful risk management system. It takes a thorough and thoughtful discussion with your clients to fully understand what to expect from them from a trading perspective. Each client has a very different trading profile. For instance, the trading profile of a long-only institution that has a trader making an investment decision from their order management system is much different than a trading profile of a client that operates a black box or algorithmic trading service.
Client profiling requires constant monitoring and fine tuning to ensure that what was agreed upon during the client onboarding process is consistent with the client’s current trading patterns. For instance, if you had agreed upon a trading profile with your client three years ago and asked them how much they expected to trade with you that could be a very different story from what their trading profile is today. Setting up static values for clients will no longer work. Client profiling needs to be adaptive and evolve as the client’s trading pattern changes over time.
Environmental conditions, monetary policies and geopolitical changes all cause market volatility. When market conditions change, firms need to have a dial to adjust everyone’s risk profile up or down to different bands. Think of it as a slide ruler where, in real time, operators can make decisions like “today is a very active day in the market, let me increase the risk bands by 3% to accommodate for it.” On slower days, that same operator may decide to tighten everyone’s profiles by 5%. Being able to adjust risk management controls, based upon market events while considering client’s core trading profile, is critical to helping customers safely achieve their trading objectives.
What keeps us awake at night is the unknown. What haven’t we thought of yet?
We engineered the ConnEx solution based on historical trends, but also created a flexible solution to accommodate for dynamic changes to market conditions. You don’t know what’s going to happen with market structure and with the global economy in the future. We have to be prepared for anything. Recently there was a very large US market maker that had a systemic error that resulted in a disruption to the market. This type of event caused unanticipated ripple effects throughout the industry. Firms have to be prepared to act quickly when those situations arise in order to ensure that they aren’t contributing to any further disruptions.
Immediately following that event, there was an elevated interest in risk management from the buy-side. Institutions want to better understand the risk protection policies that broker-dealers have in place. The challenge for the industry as a whole is how to apply a common set of risk parameters for a particular client when that client trades in multiple products through various asset classes and across different geographic regions?
We have developed a model that helps solve that challenge. By creating a client ID at the FIX gateway level, firms can perform consolidated risk monitoring across all business lines, asset classes and order types for a particular client, regardless of the trading region.
Thinking about where the markets might be in six months, a year, two years, etc; what are the long term trends in risk management?
I think that the industry could evolve to having some type of exchangelevel utility that all brokers must use. To some extent we have pieces of it today. In the US, every single market has some type of risk system in place whether it’s the single stock circuit breakers, price collars, velocity checks, etc. However, since brokerdealers are regulated entities, they are still required under 15c3-5 to perform adequate risk monitoring for their orders prior to submitting them to the various execution venues. Broker-dealers will never be able to fully defer that responsibility to the exchanges, even if an industrywide utility is developed.
So while an industry-wide utility could have certain benefits, I think down the road, we are still going to be in a situation where broker-dealers will continue to maintain their own risk systems and protocol. Broker-dealers have their own unique relationships with their clients and each brokerdealer has their own level of risk tolerance and capital requirements.
The challenge for everyone is how do firms that have limited resources keep up with the pace of change required to stay compliant with the regulations and the changes in our business? That’s where outsourcing comes into play. By leveraging a comprehensive vendor-based solution, firms can be in compliance at a fraction of the cost while saving on unplanned capital expenditures and focusing on core business goals.