Morgan Stanley’s recent US$500,000 fine from the Commodity Futures Trading Commission (CFTC) for overbilling clients trading fees isn’t a large fine considering the FCM’s size. What is alarming is that Morgan Stanley is the fourth FCM the CFTC has fined for overbilling clients. Morgan Stanley was fined in 2014, Barclays Capital in 2016 and JP Morgan Securities earlier this year. The common reason: lapses in technology and procedures in matching up the exact number of trades, their values and attributes between internal systems and third-party providers.
In the case of Morgan Stanley, those shortcoming results in clients paying US$3 million more than they should have for trade execution fees for some futures deals from 2009 to 2016. “The smaller size of Morgan Stanley’s fine likely reflects the firm’s self-reporting of its mistakes to the CFTC,” says Mary Kopczynski, chief executive of 8of9, a New York firm specializing in regulatory technology. “However, the fact that Morgan Stanley was fined even though it discovered and corrected the errors on its own shows that the CFTC wants FCMs to improve their operational controls when it comes to client fees.” Morgan Stanley, which the CFTC says has reimbursed all affected clients and corrected its systems, declined to comment for this article.
According to the CFTC, Morgan Stanley’s automated proprietary fee and reconciliation system implemented before 2010 worked well when it came to making certain Morgan Stanley itself wasn’t overcharged any trading fees by exchanges. However, it wasn’t until 2015 that Morgan Stanley’s system was programmed to catch instances of where it might overbill clients. Until then, the system only flagged instances where Morgan Stanley was undercharging customers.
In its settlement with Morgan Stanley, the CFTC never disclosed what percentage overcharge the US$3 million overcharge represents. Operations managers at five FCMs — which have not been fined by the CFTC — estimate that on average there is only a 0.2 percent discrepancy between the amount customers should be charged and what they are actually charged. The difference almost always represents an overbilling, they say. The 0.2 percent could come to a hefty amount depending on the size of an FCM’s book of business.
“Morgan Stanley didn’t have proper internal audits and controls in place to catch its mistakes quickly enough,” says one operations manager at an FCM. “Quickly enough would ideally mean within a month or two,” says one operations manager. However, six months to a year is a more practical timeframe because of the amount of time it takes to complete an audit, he explains.
The nature of calculating trading fees lends itself to billing errors and fees for trading futures contracts are typically more complicated than those for equities. “Exchanges differ vastly in their fee structures depending on the type of product traded and rebates offered,” explains Paul Geiger, president of Theorem Technologies, a Chicago firm offering software-as-a-service for post-trade reporting and reconciliation for FCMs and commodity trading advisors. “One futures product could easily have over ten fee schedules.” The newly launched Theorem is a spin-off from Thales, an independent introducing broker.
FCMs, which clear and execute futures orders for introducing brokers, must incorporate every fee schedule and any changes into their billing platform so that the reconciliation can be done effectively. The fee schedule must also include any discounts offered large customers.
“Static data managers or reconciliation managers are often responsible for setting up the rules in the billing and reconciliation platform. If they don’t keep up to date on the rules, errors are inevitable” says Geiger. Those middle office managers must contend with a data tsunami. They could be receiving several hundred notices a month from exchanges about changes to fee schedules in a multitude of formats — either emails, faxes, or paper-based communiques.
Making the job even harder for middle office executives is that FCMs don’t receive a separate bill from each exchange reflecting the trading activity of each customer. Instead, FCMs will receive bills representing the total or aggregate number of trades executed. FCMs must then allocate the right amount of the total bill to each of their customers based on their trading activity for each financial product.
In its settlement with Morgan Stanley, the CFTC suggests that Morgan Stanley’s mistakes in three different catgories of trades were due to human errors in inputting the correct data on the fee schedules used by different exchanges. Unfortunately, Morgan Stanley didn’t catch the errors until years after they were made.
The largest amount of overbilling — US$1.5 million — started in 2009 and was the result of Morgan Stanley’s billing platform not correctly taking into account the unique fee structures of the Chicago Futures Exchange, the Italian Futures Market and the Osaka Securities Exchange for “give up” transactions. “The fee structure on the CFE, IDM and OSE differs from the more traditional clearing broker pays all fees rule utilized by most exchanges in that the trading fee is paid by the executing broker,” says the CFTC. Morgan Stanley didn’t catch the mistake, affecting over 100 customers, until January 2015.
The second largest amount of overbilling — US$1.3 million — started in 2011 when Morgan Stanley never adjusted its billing system to take into account the Chicago Mercantile Exchange’s reduction in fees. The overbilling, which affected more than 70 customers who traded “e-minis” or fractions of standard futures contracts on the CME, was not suspected until June 2015. It was further investigated in April 2016 as the result of a customer complaint.
The third and smallest amount of overbilling — US$204,227 — took place between October 2012 and September 2014 when Morgan Stanley incorrectly added a fee of US$2.00 to some block trade transactions executed on the ICE Futures US. According to the CFTC, Morgan Stanley’s proprietary fee system only reflected the actual fee charged by the ICE and never incorporated the additional US$2 fee charged by Morgan Stanley so its reconciliation application could not have picked up the mistake.The error, affecting nine customers, wasn’t caught until September 2014.
Given that it could be easy to make a billing error and be hard to find it, what’s an FCM to do? Even if the right rules and adjustments are input, it more likely to avoid errors or catch them early if an FCM has one billing platform, one source of data, and a well-trained centralized static data department. Some FCMs rely on separate billing systems for each product and don’t have enough static data managers to keep up to date with changes to exchange fees.
The reconciliation application could be accepting conflicting data on executed orders from real-time front-end trade data order management systems and batch back-end systems, says Geiger. Ideally, the front-end system would be the best source as it has the most up-to-date data.
“The reconciliation application also needs the right information on customer agreements from the billing system,” explains Sonia Goklani, chief executive of Cleartrack, a South Brunswick, N.J. firm specializing in customized trading and clearing technology for the derivatives market. “If the billing system is fed the wrong data or no data on customer agreements, the reconciliation process will also fail.” Such could be the case if the FCM were to rely on multiple contract inventory systems or stored paper-based documents. A better solution: a centalized contract inventory system that would extract the key fee terms of customer agreements to transmit to the billing system.
Even correct data might not be enough to ensure an accurate reconciliation. Reconciliation applications must also be coded correctly to do two matches. The first would be to match up the exchange’s figure of the total number of trades executed with the aggregate of trades executed for each client. If the parts don’t add up to the same total figure given by the exchange, it’s time to take an indepth look at each trade for each customer.
Presuming an exchange’s figure for the total number of trades executed by the FCM match up to the FCM’s records for the total number of trades executed for all of its clients, then comes the second level of reconciliation. The exchange’s total bill must match up with the aggregate of the invoice for each client. If they don’t, it’s again time to take a dive into each client account.
However, even the best reconciliation application needs continual testing to verify whether it is working correctly. Yet traditional sample testing might not be good enough. “The sample size might not be large enough or, worse, diverse enough to catch mistakes,” cautions Geiger. Kopczynski suggests that FCMs do periodic internal audits consisting of a sample of trades executed on each exchange for each financial product over a several month period. FCMs also need to document their policies for who is responsible for reviewing the results of the internal audits, who is responsible for adjusting the reconciliation platform to correct any deficiencies and who will notify clients of any mistakes.
All five FCMs contacted by FinOps cited Societe Generale as the best example of an FCM that has made billing reconciliations a priority. It has given a central static data department ownership of the process. That department reviews results from periodic internal audits, helps the IT department make adjustments to reconciliation applications, verifies that the changes were made and informs FCM customer interfacing executives of any overcharges that must be immediately refunded.
Given that ensuring correct client billing takes so much time and effort, some third-party firms are capitalizing on the inefficiency by offering outsoured bill reconciliation services. Theorem’s software as- a-service can reconcile the data in an FCM’s front and back-end systems with the data provided by exchanges. The platform can then calculate the correct trading fees. Small, mid-tier and large FCMs as well as their clients have shown an interest, says Geiger, former chief information officer at introducing broker Thales. Theorem is a spin-off from Thales.
However, outsourcing isn’t a panacea. It doesn’t eliminate the FCM’s legal liability for any client billing errors, cautions Kopzynski. FCMs must still periodically test the accuracy of third-party results. Third-parties also won’t do all of the legwork involved with reconciliation. Theorem, acknowledges Geiger, doesn’t keep track of all the exchange fee schedules or changes to those schedules. Nor will it do any data entry. FCM clients must input the fee schedules of each exchange and make any changes to those fees by themselves.
However, Theorem’s Web interface allows static data managers or reconciliation managers to input the exchange fee schedules on their own, make any changes, and re-run the reconciliation process in close to real-time, says Geiger. They don’t need to ask separate software managers to do the work for them and re-run the reconciliation process in a slower batch mode.
Will empowering business line managers to handle more work ensure that FCMs catch more billing errors faster? Not necessarily, say reconciliation managers. Any errors in inputs will still lead to fee overcharging regardless of whose billing or reconciliation platform is used. As is the case with all post-trade processes, a combination of good technology and good process is the best antidote to operational risk. FCMs shouldn’t find out about a billing error from either the CFTC or a customer.
“Too often, FCMs spend far more of their budgets to improve their front-office trading and back-office clearing services, but forget about the middle office responsible for billings,” says Goklani. “The CFTC’s fines show why that’s a bad decision.”
So will an exodus of customers. “Given that the FCM business is so competitive, firms can’t afford to make billing errors,” says Goklani. “They increase the potential a client will take its business elsewhere.”