The aim of this research project is to develop new nonlinear time series econometric methods with applications in the fields of financial markets, carbon and energy markets, and climate change. In particular, the focus of this proposal lies on the modelling and hypothesis testing of volatility spillovers across markets, and developing new econometric methods for high-frequency data. Furthermore, this project intends to extend existing econometric methodology and applying it to carbon and energy markets and to introduce new models capable to explain nonlinear features of climate data. New techniques in time series econometrics are needed to investigate the recorded information about the market and its characteristics in intraday financial datasets. Prior to modelling trade durations (or time elapsed between two consecutive transactions), any intraday periodicity should be removed from the data. We contribute to the literature by proposing a new modelling framework for adjusting diurnal variation in durations.

When data are sufficiently long, one inevitably expects periods of market turbulence and changes in volatility due to economic, political or social events. Recent studies have shown that the traditional stationary GARCH process for modelling and forecasting volatility of financial returns is not appropriate for fitting such data. An alternative is to explicitly introduce nonstationarity in volatility to take into account deterministic shifts in the long-run variance. The vast majority of these models are based solely on time series information, but very few models include economic information. One task of this proposal makes progress along this research line by extending the nonstationary volatility model of Amado and Teräsvirta (2013) using exogenous information. At the same time, we shall investigate whether long-run variances of a group (or a subgroup) of returns move together over time. This shall be done by testing for co-movements in the long-run volatility, and then by modelling common deterministic shifts in volatility. Investigating the volatility of carbon prices is also important because lower levels of volatility in carbon markets would hedge against price spikes (and crashes) and encourage policymakers to more ambitious climate change mitigation policies. Yet, empirical studies on carbon prices volatility remain scarce. In this project, we provide a comprehensive study in the price volatilities of major international carbon markets allowing for changes in the long-run volatility. Furthermore, this project also makes research progress in the growing field of climate econometrics by tackling climate problems using econometric methods.

We shall mostly use parametric techniques and likelihood-based procedures that are state of the art in the field of time series econometrics. The techniques shall be investigated by Monte Carlo experiments and illustrated using financial data, commodity prices, carbon prices and climate data.