Publications - Gustavo Fruet Dias en-us PURE Extension (Web Department) 30 <![CDATA[Price Discovery in a Continuous-Time Setting]]> Dias, G. F., Fernandes, M., Scherrer, C. We formulate a continuous-Time price discovery model and investigate how the standard price discovery measures vary with respect to the sampling interval. We find that the component share (CS) measure is invariant to the sampling interval, and hence, discrete-sampled prices suffice to identify the continuous-Time CS. In contrast, information share (IS) estimates are not comparable across different sampling intervals because the contemporaneous correlation between markets increases in magnitude as the sampling interval grows. We show how to back out the continuous-Time IS from discrete-sampled prices under certain assumptions on the contemporaneous correlation. We assess our continuous-Time model by comparing the estimates of the (continuous-Time) CS and IS at different sampling intervals for 30 stocks in the United States. We find that both price discovery measures are typically stable across the different sampling intervals, suggesting that our continuous-Time price discovery model fits the data very well.

Research Fri, 01 Jan 2021 14:50:46 +0100 b480adc1-fff5-4c50-9868-8641279e9448
<![CDATA[Estimation and Forecasting in Vector Autoregressive Moving Average Models for Rich Datasets]]> Dias, G. F., Kapetanios, G. We address the issue of modelling and forecasting macroeconomic variables using rich datasets by adopting the class of Vector Autoregressive Moving Average (VARMA) models. We overcome the estimation issue that arises with this class of models by implementing an iterative ordinary least squares (IOLS) estimator. We establish the consistency and asymptotic distribution of the estimator for weak and strong VARMA(p,q) models. Monte Carlo results show that IOLS is consistent and feasible for large systems, outperforming the MLE and other linear regression based efficient estimators under alternative scenarios. Our empirical application shows that VARMA models are feasible alternatives when forecasting with many predictors. We show that VARMA models outperform the AR(1), ARMA(1,1), Bayesian VAR, and factor models, considering different model dimensions.

Research Mon, 01 Jan 2018 14:50:46 +0100 38c5434a-743d-440c-8e34-25c191159dc7
<![CDATA[The time-varying GARCH-in-mean model]]> Dias, G. F. Research Sun, 01 Jan 2017 14:50:46 +0100 abc49657-f6a8-4bf2-bb8d-027e38d38fd8 <![CDATA[Forecasting long memory series subject to structural change: A two-stage approach]]> Papailias, F., Dias, G. F. Research Thu, 01 Jan 2015 14:50:46 +0100 c383c813-4f67-41df-a827-0a5bfdd28f24