https://sourceforge.net/p/jmulti/bugs/2007-01-18T22:57:41ZRecent changes to bugs2007-01-18T22:57:41Z2007-01-18T22:57:41ZDavid W Carterhttps://sourceforge.net/u/cartdc/https://sourceforge.netb82cd7564038975a167e2c5404bc3e3ccf430b9b

<div class="markdown_content"><p>The forecast deterministics matrix in the Specify Forecast feature of the VECM Analysis is not arranged in the same order as the corresponding parameter matrix. For example, in a model with a constant, trend and monthly dummies you have the following orders:</p> <p>forecast deterministics matrix<br /> constant, monthly dummies, and trend</p> <p>parameter matrix<br /> monthly dummies, constant, trend</p> <p>This causes incorrect forecasts. I have only verified this problem in the VECM Analysis for the example listed above.</p> <p>Thanks.</p></div>

2005-07-29T11:34:53Z2005-07-29T11:34:53ZAnonymoushttps://sourceforge.net/u/userid-None/https://sourceforge.net0f38f306910fd75bcc3b5e545d3dcd982737a0aa

<div class="markdown_content"><p>problem with VECM chow test</p> <p>fix is coming soon</p></div>

2005-07-13T09:23:25Z2005-07-13T09:23:25ZMarkus Kraetzighttps://sourceforge.net/u/kraetzig/https://sourceforge.netc877ead7fde239118ca932c39072c57a6b010647

<div class="markdown_content"><p>sometimes the computation fails with a message <br /> "not all singular values could be computed"</p> <p>will be gone in next release </p></div>

2005-07-06T10:00:06Z2005-07-06T10:00:06ZMarkus Kraetzighttps://sourceforge.net/u/kraetzig/https://sourceforge.net8510d1d668ff52c1aa7ae14d71f3c9cba52ac21d

<div class="markdown_content"><p>subset search with the System SER for VECM models <br /> does not work. </p></div>

2005-06-02T11:28:54Z2005-06-02T11:28:54ZMarkus Kraetzighttps://sourceforge.net/u/kraetzig/https://sourceforge.net9f042667349d9014dd294cb920fc794e5e315c35

<div class="markdown_content"><p>in section "Eviews Compatibility Guide"</p> <p>case 2 and case 3 are mixed up, but the description<br /> makes it clear</p></div>

2005-05-27T18:14:55Z2005-05-27T18:14:55ZMarkus Kraetzighttps://sourceforge.net/u/kraetzig/https://sourceforge.net9da7f4711be5304513aba5464487f6b67982b15d

<div class="markdown_content"><p>dialog does not always result in expected settings<br /> changes, <br /> defaults are not properly set back for all settings</p></div>

2005-05-16T17:36:44Z2005-05-16T17:36:44ZMarkus Kraetzighttps://sourceforge.net/u/kraetzig/https://sourceforge.net012ac517d85e479e39c57dcd0eccc909d0847ed6

<div class="markdown_content"><p>JMulTi seems to freeze on some configurations at <br /> startup without giving any error message. Only the<br /> splashscreen is visible. This looks like a deadlock. So<br /> far not reprocucible on <br /> developer machine, but possible sources are being<br /> explored. </p></div>

2005-05-09T15:16:54Z2005-05-09T15:16:54ZMarkus Kraetzighttps://sourceforge.net/u/kraetzig/https://sourceforge.net2bd9e54e60d0ac3b06c0a0652ebe32247c7787f3

<div class="markdown_content"><p>Trend breaks are not treated correctly in both<br /> cointegration tests:</p> <p>S&amp;L gives an error</p> <p>Joh. Trace statistic is ok, but response surface is not<br /> computed although it would be possible</p></div>