<div dir="ltr">Thank you very much for your prompt answer and help, Alexios.<div><br></div><div>I am running now some models and found some results that are puzzling me. </div><div><br></div><div>From another thread I saw one recommendation that we should "(...) <span style="color:rgb(0,0,0);white-space:pre-wrap">pass values in the external </span><span style="color:rgb(0,0,0);white-space:pre-wrap">regressor which are close in scale to the variance equation</span>", what makes sense to me... </div><div><br></div><div>However, I noticed that when I define the external regressors with values close to the return series, the coefficients calculated for the external regressors are non-significant, while I have a strong hypothesis that those should be relevant in the process. Nevertheless, I noticed that when scaling to an annual basis the external regressor coefficients becomes significant...</div><div><br></div><div><br></div><div>Attached follow a sample code (sample_gjr_extregressor.R) and simplified database (data.txt) with daily returns (r) and lagged realized variance (lag_extreg) I am using as inputs for the models.</div><div><table border="0" cellpadding="0" cellspacing="0" width="192" style="border-collapse:collapse;width:144pt">
<colgroup><col width="64" span="3" style="width:48pt">
</colgroup><tbody><tr height="19" style="height:14.4pt">
<td height="19" width="64" style="height:14.4pt;width:48pt"><br></td><td width="64" style="width:48pt"><br></td><td width="64" style="width:48pt"><br></td></tr></tbody></table></div><div>There are three models as below:</div><div><br></div><div>Model1: Standard GJR</div><div>Model2: GJR with lagged daily realized variance</div><div>Model3: GJR with lagged realized variance annualized (scaled by 252)</div><div><br></div><div>The table below summarizes the results found.</div><div><br></div><div><table border="0" cellpadding="0" cellspacing="0" width="1005" style="border-collapse:collapse;width:756pt">
<colgroup><col width="64" span="7" style="width:48pt">
<col width="86" span="2" style="width:65pt">
<col width="77" style="width:58pt">
<col width="94" style="width:71pt">
<col width="86" style="width:65pt">
<col width="64" span="2" style="width:48pt">
</colgroup><tbody><tr height="19" style="height:14.4pt">
<td height="19" width="64" style="height:14.4pt;width:48pt"></td>
<td width="64" style="width:48pt">model</td>
<td width="64" style="width:48pt">omega</td>
<td width="64" style="width:48pt">alpha1</td>
<td width="64" style="width:48pt">beta1</td>
<td width="64" style="width:48pt">gamma1 </td>
<td width="64" style="width:48pt"> delta1</td>
<td width="86" style="width:65pt">p-val_omega</td>
<td width="86" style="width:65pt">p-val_alpha1</td>
<td width="77" style="width:58pt">p-val_beta1</td>
<td width="94" style="width:71pt">p-val_gamma1</td>
<td width="86" style="width:65pt">p-val_delta1</td>
<td width="64" style="width:48pt">LL</td>
<td width="64" style="width:48pt">For_01d</td>
</tr>
<tr height="19" style="height:14.4pt">
<td height="19" align="right" style="height:14.4pt"><br></td>
<td>Model1</td>
<td align="right">0.011449</td>
<td align="right">0.131705</td>
<td align="right">0.89924</td>
<td align="right">-0.08386</td>
<td> NA</td>
<td class="xl64" align="right">0.001217</td>
<td class="xl64" align="right">0.000000</td>
<td class="xl64" align="right">0.000000</td>
<td class="xl64" align="right">0.000012</td>
<td class="xl64" align="right">0.000000</td>
<td align="right">-2033.96</td>
<td align="right">0.800619</td>
</tr>
<tr height="19" style="height:14.4pt">
<td height="19" align="right" style="height:14.4pt"><br></td>
<td>Model2</td>
<td align="right">0.011449</td>
<td align="right">0.131705</td>
<td align="right">0.899244</td>
<td align="right">-0.08387</td>
<td class="xl66" align="right"><span style="background-color:rgb(255,255,0)">1.23E-08</span></td>
<td class="xl64" align="right">0.002944</td>
<td class="xl64" align="right">0.000000</td>
<td class="xl64" align="right">0.000000</td>
<td class="xl64" align="right">0.000012</td>
<td class="xl65" align="right"><span style="background-color:rgb(255,255,0)">0.999999</span></td>
<td align="right">-2033.96</td>
<td align="right">0.800635</td>
</tr>
<tr height="19" style="height:14.4pt">
<td height="19" align="right" style="height:14.4pt"><br></td>
<td>Model3</td>
<td align="right">0.004839</td>
<td align="right">0.066382</td>
<td align="right">0.69109</td>
<td align="right">-0.08387</td>
<td class="xl67" align="right"><span style="background-color:rgb(0,255,0)">0.001161</span></td>
<td class="xl64" align="right">0.633308</td>
<td class="xl64" align="right">0.037429</td>
<td class="xl64" align="right">0.000000</td>
<td class="xl64" align="right">0.010968</td>
<td class="xl65" align="right"><span style="background-color:rgb(0,255,0)">0.000451</span></td>
<td align="right">-2012.57</td>
<td align="right">0.778901</td>
</tr></tbody></table></div><div class="gmail_extra"><br><div class="gmail_quote">Any idea or suggestion on what might be happening?</div><div class="gmail_quote"><br></div><div class="gmail_quote">Thanks again in advance for any help with that issue.</div><div class="gmail_quote"><br></div><div class="gmail_quote">Best Regards,</div><div class="gmail_quote">Assis.</div><div class="gmail_quote"><br></div><div class="gmail_quote"><br></div><div class="gmail_quote"><br></div><div class="gmail_quote">2015-08-20 11:01 GMT-03:00 alexios <span dir="ltr"><<a href="mailto:alexios@4dscape.com" target="_blank">alexios@4dscape.com</a>></span>:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi,<br>
<br>
The answer is yes and yes. Add variable(s) lagged.<br>
<br>
Best,<br>
<br>
Alexios<div><div class="h5"><br>
<br>
On 20/08/2015 14:57, Assis Duraes wrote:<br>
</div></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div class="h5">
Hi,<br>
<br>
first of all I would like to thanks for the rugarch package. it is really<br>
useful and a very nice package.<br>
<br>
I am investigating the effect of external variables on conditional variance<br>
models forecasts. more specifically, I am would like to check if the<br>
addition of implied volatility and realized variance as external regressors<br>
on a GJR (1,1) model somehow enhance the daily volatility forecasts of it.<br>
<br>
Looking for a tool to help modelling it I found the rugarch package, and<br>
started looking into it. In fact, at this point, I believe I have a very<br>
basic question. but did not find an answer on previous posts or in the<br>
package documentation.<br>
<br>
Should I inform the external regressors matrix in model spec already lagged<br>
or not? I imagine, yes, since I did not find in any place where specify the<br>
lags for those regressors, but would like to confirm. In case affirmative,<br>
If I want to use a same variable with different lags I need to inform it<br>
multiple times, obviously with different lags, in external.regressors<br>
matrix, correct?<br>
<br>
My apologies in advance if it is explained somewhere, but as I explained, i<br>
search without much success..<br>
<br>
Thanks in advance for any help with that,<br>
Assis.<br>
<br></div></div>
[[alternative HTML version deleted]]<br>
<br>
_______________________________________________<br>
<a href="mailto:R-SIG-Finance@r-project.org" target="_blank">R-SIG-Finance@r-project.org</a> mailing list<br>
<a href="https://stat.ethz.ch/mailman/listinfo/r-sig-finance" rel="noreferrer" target="_blank">https://stat.ethz.ch/mailman/listinfo/r-sig-finance</a><br>
-- Subscriber-posting only. If you want to post, subscribe first.<br>
-- Also note that this is not the r-help list where general R questions should go.<br>
<br>
<br>
</blockquote>
<br>
</blockquote></div><br></div></div>