As the Covid-19 pandemic rocks markets, ESILV students majoring in Finance try to make sense of the stock market roller coaster. Two teams worked remotely on a value-at-risk project under the guidance of Matthieu Garcin, a lecturer and researcher in Quantitative Finance at ESILV Engineering School.
Understanding current market fragilities and developing new strategies for well-defined market regimes are crucial lessons during this crisis environment. As a part of the “PI²” (industrial innovation) project, seven 4th-year students from ESILV, majoring in Financial Engineering have worked on a few applications of time-varying densities to Value at Risk (VaR) models.
The teams, carefully guided by Matthieu Garcin, a quantitative finance professor at ESILV, have created a research demo video, that they have shared on LinkedIn. The members of the teams are Ayoub Ammy-Driss, Akin Arslan, Thomas Barrat, Sarah Bouabdallah, Sana Laaribi, Brieuc-Marie Le Brigand, Jules Klein. All of them pursue a master’s in quantitative and computational Finance at ESILV.
“A financial crisis is an exciting topic of study to test the relevance of time-varying statistics. Were the distributions we had defined able to respond appropriately to this sudden regime shift? What could they tell us about the chronology of the crisis and a possible return to normal?” (Matthieu Garcin, professor at ESILV)
The co-authored research paper offers a view on the Covid-19 impact on the financial markets, notably on the differences between the epidemic’s effects on different areas, in terms of time and amplitude. The students deepened their analysis based on a comparison between the 3 primary stock exchange indexes, the United States’ S&P, China’s SSE and France’s CAC40.
“The representation of time-varying densities of stock index returns shows how quickly the stability of certain market statistics disappears, making forecasts more than challenging.” (Matthieu Garcin, professor at ESILV)
One of the groups had notably delved into alpha-stable densities model, and the other group, into a model of non-parametric densities.
“The lower the alpha, the fatter the tails of the distribution (the highest possible alpha, 2, corresponds to Gaussian returns). Before COVID-19, the alpha is high and stable. When the epidemic emerges in a region, the asymmetry parameter, beta, momentarily moves away from zero. Then, alpha gradually decreases.” (Matthieu Garcin, professor at ESILV)
The statistical term ‘fat tails‘ refers to situations in which extreme outcomes have occurred more than expected. This research examines the performance of different Value-at-Risk (VaR) models over the long-term during these unprecedented disruptions.