Zusammenfassung:
When a company releases earnings results or makes announcements, a dominant sectoral wide lead-lag effect from the stock on the entire system may occur. To improve the estimation of a system experiencing dominant system-wide lead-lag effects from one or a few asset in the presence of short time series, we introduce a model for Large-scale Influencer Structures in Vector AutoRegressions (LISAR). To investigate its performance when little observations are available, we compare the LISAR model against competing models on synthetic data, showing that LISAR outperforms in forecasting accuracy and structural detection even for different strength of system persistence and when the model is misspecified. On high-frequency data for the constituents of the S&P100, separated by sectors, we find the LISAR model to significantly outperform or perform equally good for up to 91% of the time series under consideration in terms of forecasting accuracy. We show in this study, that in the presence of influencer structures within a sector, the LISAR model, compared to alternative models, provides higher accuracy, better forecasting results, and improves the understanding of market movements and sectoral structures.