Statistical modelling of multivariate and complex extremes

Financial time series consist of speculative prices of assets such as stocks, foreign
currencies or commodities. Risk management is intended to guard against risk of loss due to a fall
in prices or financial assets held. Extreme losses are multivariate in nature and can cause havoc
for the people. There is a clear need to have good statistical methods to estimate the likelihood of
occurrence and impact of such events. The central theme of this thesis is modelling dependence of
multivariate extremes in time series based on a range of recent methodological developments. This
thesis will develop novel statistical models that encompass a broad class of extremal dependence
structures and can be used to model shocks in financial systems and large claims in insurance.
Applications are also envisaged in the environment