OrthoPanels can now handle cases that enter the panel late (refreshment), not just drop outs. Late entrants can either have NAs for time entries prior to their joining, or not be present in the data at early times (possible only when using the formula interface).
Add a dataset of the dynamics of labour demand in the United Kingdom, based on Arellano and Bond (1991)
missing data only in the first wave of X doesn’t
affect the result.
handle case and time variables that aren’t 1:N and
1:T
OrthoPanels can now estimate the long-run effects via function LongRunEffects. Longrun effects can be lotted using the caterpilar_longRun function.