The objective of this project is to improve existing state-of-the-art financial econometric methods by quantifying an important determinant of asset prices, the arrival of economic information such as press-releases and newswire and newspaper items, and by using thus improving estimates of financial risk. First, this RP is based on our earlier joint paper (Engle, Hansen and Lunde 2012), which provides empirical evidence on how the idiosyncratic volatility of a company’s asset price is related to the public news flow. Secondly, methodologically this project will start from simple extensions of the Dynamic Conditional Correlation models, which are a special case of a very general model class review in Patton (2012). These models are specified by GARCH equations for conditional variances, marginal distributions for each series, and a conditional copula function. The copula function is allowed to be time-varying by letting its parameters be functions of past data.
This research project aims to extend such Copula GARCH models by augmenting these models with news arrival indicators relevant to a particular pair of financial assets. This type model augmented with news arrival indicators helps to understand the economic forces driving correlations between financial assets and promises improved forecasts of volatilities and correlations. A central theme in this project is how to handle and translate the enormous amount of news items available to the analyst.
The main result is that our new financial econometrics methods for risk management are valuable to financial institutions and investment managers, who daily use quantitative risk management models for portfolio management and pricing of financial derivatives. Improved forward-looking estimates of how financial prices change will have a direct impact on how financial institutions can prevent losses on portfolios of financial assets traded in the financial markets. Moreover, methodological developments are attractive scientifically, and such a publication will be submitted to top finance journals in finance, econometrics, and statistics.