Dear colleagues,

using a dataset on bilateral trade (i.e. country g exports to h leading to the node (g,h)), I need to use the delete-one-node jackknife variance estimator of Cameron and Miller (2014, formula 2.17 on page 9).

How can I set up a bootstrap command that does the trick, i.e. in each round drops one node, i.e. drops country G, and any pair with that country, i.e. (g,h) and (h,g)?

Many thanks in advance,
Lennart







Cameron, A. C., & Miller, D. L. (2014). Robust inference for dyadic data. Working Paper.
http://cameron.econ.ucdavis.edu/rese...ith_tables.pdf