The article analyzes complex models to understand how different factors affect the correlations between multiple time series data. The researchers use a combination of factor models and stochastic volatility models to estimate these correlations. They develop a method to efficiently estimate the parameters of the models and compare different versions to determine the number of factors needed. The researchers apply their methods to real-world datasets of exchange rates and stock returns, providing a practical way to analyze high-dimensional models of stochastic volatility.