WORKING PAPERS AND ONGOING PROJECTS

- "Asymptotic and bootstrap inference for generalize ACD models (with T. Mikosch, A. Rahbek and F. Vilandt)

The paper presents a full asymptotic and bootstrap theory for the generalize Autoregressive Conditional Duration (ACD) models. Despite the fact that these models have been out for 25 years and that are widely used by practitioners, a complete asymptotic theory was not available. Properties of the bootstrap are completely unknown. We provide it here - and we have many surprising results, including the fact that the tail index of the durations is the key player of the theory.

- Network Vector Autoregression (with M. Barigozzi and G. Moramarco) [abstract and pdf file]

Here we introduce the FNAR model - a new way to analyze time series in complex economic networks, like the web of trade and financial ties between countries. Or estimation strategy involves the extraction of latent network factors from the many layers describing the network structure. We've proven (theory/simulation) that our method is reliable and works well, even with huge amounts of data. We've also applied it to study how different countries' GDP growth rates are interconnected

- Comment on "Autoregressive Conditional Duration: a new model for irregularly spaced transaction data" (with T. Mikosch, A. Rahbek and F. Vilandt) [abstract and pdf file]

We show that strict stationarity and ergodicity alone are not sufficient for consistency and asymptotic normality of the QMLE of the ACD(1,1), and provide the needed additional sufficient conditions to account for the random number of durations. In particular, we argue that the durations need to satisfy the stronger requirement that they have finite mean.

-Time-varying Poisson Autoregression (with G. Angelini, E. D'Innocenzo and L. De Angelis) [abstract and pdf file]

Here we propose a new time-varying econometric model, called Time-Varying Poisson AutoRegression with eXogenous covariates (TV-PARX), suited to model and forecast time series of counts. We show that the score-driven framework is particularly suitable to recover the evolution of time-varying parameters and provides the required flexibility to model and forecast time series of counts characterized by convoluted nonlinear dynamics and structural breaks. We study the asymptotic properties of the TV-PARX and of the related MLE. We test the usefulness of the model in practice on COVID-19 data and US corporate defaults.

- MinP Score Tests with an Inequality Constrained Parameter Space (with Z. Lu, A.Rahbek and Y. Yang) [abstract and pdf file]

- Bootstrap diagnostic tests (with L. Fanelli and I. Georgiev) [In preparation]