Em 12 de março de 2011 Peter <span dir="ltr">&lt;<a href="mailto:peter.summers@ttu.edu">peter.summers@ttu.edu</a>&gt; escreveu:</span><div><span dir="ltr"></span><br><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">

FWIW, in a univariate TAR model the data gets re-ordered by (say) ascending values of the threshold variable, then separate regressions get run on the 2 sub-samples. Could this be done with an &quot;msortby&quot; command followed by smpl commands?<br>

</blockquote><div><br></div><div>Dear Peter, based upon Example 4.3 from the RATS Programming Mannual by Walter Enders, I wrote a script that can be used to estimate a simple TAR model with 2 regimes. Maybe you&#39;d like to take a look in it.</div>

</div><div><br></div>Best,<br>-- <br><b><font face="tahoma, sans-serif">Henrique C. de Andrade</font></b><font face="tahoma, sans-serif"><br>Doutorando em Economia Aplicada<br>Universidade Federal do Rio Grande do Sul<br>

<a href="http://www.ufrgs.br/ppge" target="_blank">www.ufrgs.br/ppge</a></font><br>
</div>