First release

- As shown in the paper, we propose a simulated general method of moments (SGMM) for the SAR model (see function smmSAR and Section 2 of our vignette).
- We can now estimate the maximal bias of the instrumental variable estimator (see Section 1.1 and 1.2 of our vignette).

- We provide a smoother simulator of adjacency matrices in the SGMM approach.
- We add weights to the probit/logit network formation model.
- We allows the use of an initial probit/logit estimate of \(\rho\), where the observed part of the network is assumed non-stochastic in the MCMC. This is a quite different from using an initial probit/logit estimate as prior distribution of \(\rho\). In this latter case, \(\rho\) is updated using, among others, information from the observed part of the network. In the first case, \(\rho\) and the unobserved part of the network are updated using information in \(y\), where the initial estimate acts as prior distribution of \(\rho\). Information from the observed part of the network is not used to update \(\rho\). This information is included in the initial estimate.

Adjustments with Eigen 3.4