Detecting Granular Time Series in Large Panels

Abstract

Large economic and financial panels often contain time series that influence the entire cross-section. We name such series granular. In this paper we introduce a panel data model that allows to formalize the notion of granular time series. We then propose a methodology, which is inspired by the network literature in statistics and econometrics, to detect the set of granulars when such set is unknown. The influence of the i-th series in the panel is measured by the norm of the i-th column of the inverse covariance matrix. We show that a detection procedure based on the column norms allows to consistently select granular series when the cross-section and time series dimensions are large. Importantly, the methodology allows to consistently detect granulars also when the series in the panel are influenced by common factors. A simulation study shows that the proposed procedures perform satisfactorily in finite samples. Our empirical studies demonstrate, among other findings, the granular influence of the automobile sector in US industrial production.

Granular Instrumental Variables

“…Procedures containing elements of GIVs A few papers have explored the idea of using idiosyncratic shocks as instruments to estimate spillover e↵ects, such as Leary and Roberts (2014b) 59 See Brownlees and Mesters (2020) for a potential way to extend this approach when G is unknown.…”

Estimation and inference in spatial models with dominant units

“…3 It is worth noting that in the current paper we assume W is known and focus on estimating the spatial parameters. In cases where information on direct connections of the network is unavailable, there exists a related literature that uses large panel data sets (with both n and T large) to detect which unit has the largest (when equals or is close to unity) from the pattern of correlation in the data without needing to know W. See, for example, Parker and Sul (2016), Brownlees andMesters (2018), andKapetanios et al (2019). In a related literature, also consider estimating using large panel data sets when W is not known.…”

Estimation and Inference in Spatial Models with Dominant Units

“…3 It is worth noting that in the current paper we assume W is known and focus on estimating the spatial parameters. In cases where information on direct connections of the network is unavailable, there exists a related literature that uses large panel data sets (with both n and T large) to detect which unit has the largest (when equals or is close to unity) from the pattern of correlation in the data without needing to know W. See, for example, Parker and Sul (2016), Brownlees and Mesters (2018), and Kapetanios et al (2019). In a related literature, also consider estimating using large panel data sets when W is not known.…”