Error in stock_yogo_reccomender about cragg

[beniaminogreen]

I don't think it's possible to run the test with one endogenous variable and one instrument. The critical values for the test are derived from this paper, and are only provided for cases in which there are more instruments than endogenous variables (critical values are on pages 39-40).

If you have one instrument and one endogenous variable, I think you should be able to use the Cragg-Donald statistic, or the first-stage F statistic which should be identical in this case.

It's been a while since I have looked at the Stock + Yogo paper (and I only read the paper closely enough to write the code for it when I was developing the package), so I am happy to revisit this if you think it's a genuine issue with the package.