Help with Perron-Rodriguez unit root test

[hvalera]

Dear Tom,

I am trying to do the unit root test of Perron and Rodriguez (2003) on rice export prices. I used the @perronrodriguez.src procedure, but I did not get any results after implementing it.

I don't know how to go about resolving it. Any suggestions? There is my code and data. Thanks in advance for any help.

Kind regards,
Harold

[TomDoan]

Pierre Perron did not post the code for doing the Perron-Rodriguez paper so when we weren't able to hit the published results (with a well-known public data set, so it wasn't a data problem) we didn't really put much effort into the @PerronRodriguez procedure. On revisiting it, we found that the differences were minor enough to be almost certainly be due to slight (asymptotically negligible) differences in calculations. Note that anything that requires a double search over break points and lag lengths will have *some* situations where a very slight difference in an underlying calculation results in a discrete change, since both the break point and the lag length are discontinuous.

We have an updated @PerronRodriguez procedure, which needs an updated @ADFAutoSelect procedure (which has the MBIC criterion added) and which produces a report output. The easiest way to get those is to use the Update Procedures operation on the Help menu.

Worked example from the Perron-Rodriguez JOE 2003