Duke University
Software and Data


The BootGraph program written by Stephen J. Perez, Kevin D. Hoover, and Piyachart Phiromswad implements the bootstrap analysis of causal graphs discovered by the PC and SGS algorithms as described in

Selva Demiralp, Kevin D. Hoover, and Stephen J. Perez,
"A Bootstrap Method for Identifying and Evaluating a Structural VectorAutoregression"Oxford Bulletin of Economics and Statistics 70(4), 2008, 509-533. Coauthors:  Selva Demiralp and Stephen J. Perez.

The program is originally written in Gauss, but there is also a version that runs in Ox.  Unfortunately, for reasons that are not yet clear, the SGS algorithm does not work in Ox.


Programs are currently configured to run sample data sets from Kevin D. Hoover, Selva Demiralp, and Stephen J. Perez "Empirical Identification of the Vector Autoregression:  The Causes and Effects of U.S. M2," forthcoming in the Festshrift for David F. Hendry.  Sample date set: 

     Text format (click to open in new window or right-click and save link) 

     Gauss .fmt format (right-click and save link).

Please report any bugs to Stephen J. Perez.



  • "Selecting Instrumental Variables:  A Graph-theoretic Approach." Coauthor:  Piyachart Phiromswad.
          Data Set;
          Data Documentation.