Skip to main content

Network Effects

Network Effects and Quantile Magnitude​

Initially, network effects were not considered, although these could be incorporated by expressing the magnitude of a quantile in terms of others. This is why Gabaix argues that a network structure is unnecessary, as it appears in cross correlations (co-movements). For our analysis, the network structure results in a given micro shock distribution. We examine the consequences of this distribution on aggregate levels.

  • Part of the network can be attributed to fixed effects (FE).
  • Another part can be assigned to the covariance matrix of σϵ\sigma_\epsilon.

Importance of Cross Covariance​

Ignoring cross covariance is akin to "throwing the baby out with the bathwater." Hypothesized network effects, as discussed by Acemoglu and others, would be captured in the cross covariance terms. This approach differs from that of Gabaix and Acemoglu, where the focus is merely on the trace.

By understanding these nuances, we can better appreciate the complex interplay between network structures and their effects on economic models. This reflection encourages a more rigorous examination of results and methodologies, fostering a more robust scientific discourse.