Employment Growth through Labor Flow Networks
Data from panels A and B were fitted using maximum likelihood estimation. Due to the unusual magnitude of the scaling parameter estimated for panel B, we do not think it is a power law. However, other skewed distributions do not produce better fits under the Kolmogorov-Smirnov criteria. We used kernel regression to identify critical regions in panel C. Estimations in panel D were made with OLS. Panel E shows the universe of firms in Finland. Only 1% of the edges are drawn. The size of the node represents the degree. The color identifies firms with the same k-core index. The image was produced with the visualization tool LaNet-vi and it shows the organization of the LFN into a core-periphery structure. Groups of firms are less tightly connected as we move from the center to the outside rings.