I, Amit Kumar Pandey, hereby declare that contents of this report are outcome of my original work and learnings performed at CME Inc. in Default Management team. It is being submitted to IEOR department of The Fu Foundation School of Engineering and Applied Science for my 1.5 credit course named as Fieldwork. This report has not been submitted to any other institute for any other degree or whatsoever. I also want to declare that none of the data presented in this report is a copy of data from CME. Although, I have presented some hypothetical values to convey a concept. Working with CME this fall was a great opportunity for learning. I got a chance to learn the functioning of a clearing house and also to hone my skills in risk management tools and Excel (VBA). It was also an opportunity where I observed and worked on application of theories learnt during my MFE program at Columbia. So, I am thankful to my career officer of MFE, Tamar Hofer for the co-ordination so that I could receive this opportunity.
At CME, I want to extend my thanks to my supervisor, Asim Sana, Director in Default Management team who guided, challenged and provided resources to me. In spite of his busy schedule, he took time to diligently explain theoretical concept behind the job that he wanted me to do. At the same time he challenged me to finish technical parts of the job without much supervision. This method was helpful for me in developing a tendency of taking accountability of the job. I also worked under Nikola Djordjevic, Director in Default Management team. His quick analysis ability and technical skill was impressive for me. I learnt a lot from him. He was very open to listen to new ideas and accommodating. This gave me enough space to come up highly flexible solution to problems that are applicable in all scenarios. He also guided me whenever I was going wrong so keeping me on track all the time. It was privilege to work with him.
A very important part of my internship was Serena Dong, a 2012 batch Alumni of our MFE program and Associate in Default Management team. She was friendly and very helping. I worked on my most important project with her, I learnt many little thing about CME and its products that was used in my project. I want to thank her for all the things that I learnt from her. I also want to extend my gratitude to the Executive director of the team Sergey Arefiev. His style of concisely presenting a complex concept in few sentences was notice worthy. During several lunches with him, I learnt a lot about industry and CME business. I am also thankful to another teammate Roman Kogan, Associate in Default Management team, who suggested a very nice project related to margin calculation. This project not only helped me sharpening my VBA skill but also gave an opportunity to automate a long cumbersome process.
My fall internship at CME started on September 10, 2018 and ended on December 7, 2018. I worked with Default Risk Management team. This team is responsible for clearing the position of a defaulted firm in the event of default. This team also conducts semiannual default drill. These drills are basically mock of a real default, where we pretend that a real default has happened and we do all the needful that we would do in the event of real default. On daily basis we develop and test tools that are going to use in the event of default and keep track of risk position of member firms to stay on top of events that might lead to default. So, we keep track of margins and collateral and other metrics that can tell us about health of member firm. This team had total 6 members and 2 interns. I would categories my work at CME in following three broad categories – Developing tools to be used in default Model for Cross-margining arrangement between CME and DTCC Trade Engine Margin Reporter Participating in ‘October Mock Default Drill’. Testing of In-house platform for pricing and risk calculation named ‘Olympus’. Updating and improving ‘Default Management Manual’.
Now explaining the first bullet, I have developed three tools for CME, full description of them is provided in following pages. Here, I am giving a brief description. Cross- margining is an arrangement that takes place between two clearing houses. In this arrangement both the clearing houses together handle a portfolio and calculate its margin on combined basis. Cross margining is done for offsetting positions taken at different clearing houses by the same customer. Therefore, when we calculate the overall margin, it turns out to be less than the sum of the margin of each individual portfolio. My job was to create a tool that will be used to decide how much money will be transferred from one clearing house to the other clearing house in case if this cross margined portfolio defaults. This tool has been accepted by both the firms and has now become a standard for default management.
The second tool called ‘Trade Engine’ was a fully automated VBA tool to create a trade register of desired risk profile. Trade Register can also be seen as a portfolio. We calculated risk profile using DV01 (Dollar Value per Basis Point). Purpose of this tool is to take a random Trade Register as an input and create an output register that has desired risk profile. I devised a method using optimization to change the notional of each trade in such a way that resulting portfolio has the risk that we want. The third tool, so called ‘Margin Reporter’ is used to calculate margin required per firm and per currency. We also started to create weekly margin report for default management team. This margin reporting is helpful in keeping track of riskiness of different currencies and different firms. So that default Management team is well aware of the areas where most of the risk lies. This tool is also been used to decided ADMC
team members. ADMC i.e. ‘Active Default Management Committee’ is a team comprised of traders from different member firms for different currencies. We keep AMDC members only for those currencies and firms where most of the risk is. It means that ADMC members come from high Margin area. So, we rank margin currency -wise and then firm-wise to find out whom to include in ADMC team. Next and very important part of my internship was to participate in October Mock Default Drill. Default drill is a big event at CME that is been organized at every six months. In this event traders are invited from all the clearing member firms. With the help of these traders the default management team hedges the position of the defaulted firm and then we sell those positions by using bidding process. The money that is lost in the default is covered from the default management waterfall.
Another major part of my internship was to help my senior Asim Sana in testing our in-house valuation and risk calculation platform ‘Olympus’. I regularly did run calculation of many currencies and portfolios and downloaded their risk reports to ensure performance of ‘Olympus’ was up to mark. I suggested several improvements and pointed out few faults to make this platform more efficient. Another job that I performed at CME was to improve their manuals to be used to manage default. I was assigned to test the lucidity of manual by following them and doing all the processes listed in them. I improved manuals on numerous occasions to make it clearer for a user. Cross-margining was an idea developed by the OCC (Option Clearing Corporation) in 1989.1 Purpose of this methodology is reduce the systematic risk. The way we do it is that we try to identify the offsetting positions taken at different clearing houses by the same customer. After we have identified these positions we can allow intermarket hedges and that will lead to a reduced margin.
We know the concept of sub-additivity in risk. This concept say that sum of risk of two portfolios is higher than the risk of a portfolio that is a combination of these two portfolios. Reason for this is that there might be any offsetting positions between these two portfolios. In a more Mathematical manner we can define these offsetting positions as positions with negative correlation. One of the challenges of such a portfolio is the distribution of losses after this portfolio has defaulted. CME and DTCC are into one such arrangement, however, they didn’t yet had any highly detailed model that can be used in the event of default. Earlier this October, both the firms started a high-level meeting to make some detailed rules and model for this purpose. My task was to create a model for the money distribution. I took into account following two type of margin that were available at CME and DTCC for the default management: Initial margin: This is the collateral amount that a clearing member should post with the clearing house before it can initiate a position with the clearing house. The amount of initial margin depends on the product and its volatility.
Variation margin: This is the amount of sum a clearing member should post with the clearing house to bring the margin level at a required level after the member’s positions have been marked to market on a daily basis. Therefore, variation margin is also called the mark to market margin. For the purpose of confidentially, I can’t disclose much detail of the model. However, I will provide some overview of it. The main idea behind the distribution of loss between the two firms was that Firm A will not bear a loss on the positions of Firm B beyond the level of margin money that is has from initial and variation margin. So, that means if firm A has some margin left after bearing its own losses then it will share this money with firm B but it will not share any additional money from its own pocket. Also, if firm A has more money left after handling its own losses and this money is more than the losses of firm B then firm A will give only the amount equal to firm B’s losses.
I created a Vba based application in which if we feed current position details of both firm A and firm B then it will give an output that will tell how much money needs to be transferred from firm A to firm B or from firm B to firm A in different phases of the default. Benefit of the Cross-Margining model: This tool is first of its kind that can be used in the event of default. It has made the distribution of money between very fair and brought everyone on the same ground. An important use of this tool is that we can test all the possible future scenarios in it and can discuss if the future distribution of money in different scenarios is fair or not. Also, I kept the design of the tool very simple so that its concept remained self-explanatory to the user.