Central Counterparties : Risk management of central counterparties focusing on the analysis of stress tests

My thesis focuses on central counterparties (CCPs) and the stress tests they perform to achieve a resistant operation and therefore increase market resilience. CCPs’ role is to take over the counterparty risk during trading. A CCP needs to operate a multi-level guarantee system that can absorb losses of clearing members’ defaults to fulfill its role. One of the main focal points of my research is to analyze how the size of the guarantee system changes and how the cross-guarantee undertaking changes between clearing members and markets if the CCP merges the guarantee systems of different markets. This question is essential from a financial stability perspective since the guarantee system’s size and structure will affect the loss-absorbing capacity of a CCP. Monte Carlo simulation was used to simulate a 30 years time series for three different products, which gave the basis for the Value-at-Risk-based margin calculation and the stress test-based default fund calculation. Results show that merging the guarantee systems will always decrease the total value of the guarantees because the margin will decrease, which cannot be offset by the increase in the default fund size. In conclusion, it is not optimal from the financial stability perspective to merge the guarantee systems. However, merging is more suitable if the CCP wants to provide cheaper services or if the clearing members are willing to cross-guarantee each other. In order to increase efficiency and to analyze and align the incentives of the CCP and its clearing members, I decided to use the model presented above to shows how the CCP’s own contribution in the default waterfall is affected in different market structures. The model needed to be simplified to perform a sensitivity test on the skin-in-the-game. Overall, the model is improved, and all three layers of the default waterfall are part of the stress testing, so results are based on the whole default waterfall rather than just a part of it and therefore reached more sophisticated conclusions. The proposed model shows how the CCP’s own contribution in the default waterfall is affected in different structures. It is an important question since this own contribution is being financed from the capital of the CCP. So the larger this contribution is, the larger the stake of the CCPs’ capital is risked. Results for the four different operations are as follows: • The merged market scenario is risky for the CCP since it offers the lowest value for the overall default waterfall, and also, the CCP has to provide the second largest SITG value compared to the fully or partially separated - IM structures. This setup is favorable for some members engaging in risky trading and also affects clearing members’ liquidity the least. In the long run, the CCP would need a tremendous amount of capital to support the system, mainly if it aims to protect non-defaulting members. However, to avoid resolution, the CCP relies on the fund provided by the non-defaulting members. Due to the heightened level of loss-mutualization, this setup is disadvantageous for members active only on one of the cleared markets. • The separated design gives an advantage to the CCP, but it is not beneficial for minor participants from the market because the clearing activity can become too costly. • The default fund level separation setup was the least favorable since it required the highest SITG, while the members had a higher default fund contribution obligation. • The partially separated - IM one was proven to be the most suitable for all stakeholders. It brings the benefits of a higher margin requirement and smaller SITG for the CCP, but members can profit from hedging and risk-mutualization on a default fund level, ultimately, this being the best compromise between parties.