Correlation breakdowns, spread positions and central counterparty margin models

The default of a member of the Nasdaq Clearing commodities market in 2018 and the Covid-19 events in 2020 brought the importance of appropriately measuring breakdowns in market correlation to the attention of risk managers at central counterparties. The sizable price dislocations registered on these occasions suggested that traditional risk models may not be fully equipped to capture such breakdowns. Because correlations are directly impacted by the statistical properties of each variable, any model that lacks the capacity to deal with nonstationarity may inappropriately represent correlations or their alterations. Using an approach that combines a generalized autoregressive conditional heteroscedasticity model with dynamic conditional correlation (GARCH-DCC) to accommodate such properties, we aim to study correlation behavior during adverse market conditions and the potential impact on central counterparty margins. We propose a case study on energy commodities, with a specific focus on spread positions for the electricity market. The analysis suggests that correlation breakdowns are more frequent than traditionally expected. When different types of shocks (ie, those of September 2018 and March–May 2020) are considered, it becomes evident that while the magnitudes of the breakdowns may differ, their cycles present a number of similarities. We also recognize the potentially increased margin procyclicality that may be entailed by model corrections to deal with correlation breakdowns, highlighting the challenges of balancing margin responsiveness and stability during adverse market conditions.