Economics Job Market Rumors Topic: A Modern Gauss-Markov Theorem
https://www.econjobrumors.com/topic/a-modern-gauss-markov-theorem
Economics Job Market Rumors Topic: A Modern Gauss-Markov Theoremen-USMon, 30 Jan 2023 17:27:17 +0000http://bbpress.org/?v=1.0.2<![CDATA[Search]]>q
https://www.econjobrumors.com/search.php
Economist on "A Modern Gauss-Markov Theorem"
https://www.econjobrumors.com/topic/a-modern-gauss-markov-theorem/page/18#post-8622488
Wed, 04 Jan 2023 11:10:10 +0000Economist8622488@https://www.econjobrumors.com/<blockquote><blockquote>
<blockquote>Lihua Lei's proof involves the argument that the symmetric positive def nxn matrices span all of nxn Euclidean space. Why would this be the case? Linear combinations of symmetric matrices are symmetric after all?</p></blockquote>
<p>Have not read the paper yet, but what had been on Twitter was speculation and no proof.
</p></blockquote>
<p>Then read the paper instead of writing useless comments on EJMR.
</p></blockquote>
<p>Why would that comment be useless? Has nothing to do with the paper, but with the Twitter claims...
</p>Economist on "A Modern Gauss-Markov Theorem"
https://www.econjobrumors.com/topic/a-modern-gauss-markov-theorem/page/18#post-8622419
Wed, 04 Jan 2023 10:34:37 +0000Economist8622419@https://www.econjobrumors.com/<blockquote><blockquote>Lihua Lei's proof involves the argument that the symmetric positive def nxn matrices span all of nxn Euclidean space. Why would this be the case? Linear combinations of symmetric matrices are symmetric after all?</p></blockquote>
<p>Have not read the paper yet, but what had been on Twitter was speculation and no proof.
</p></blockquote>
<p>Then read the paper instead of writing useless comments on EJMR.
</p>Economist on "A Modern Gauss-Markov Theorem"
https://www.econjobrumors.com/topic/a-modern-gauss-markov-theorem/page/18#post-8622308
Wed, 04 Jan 2023 08:58:16 +0000Economist8622308@https://www.econjobrumors.com/<blockquote><p>Lihua Lei's proof involves the argument that the symmetric positive def nxn matrices span all of nxn Euclidean space. Why would this be the case? Linear combinations of symmetric matrices are symmetric after all?</p></blockquote>
<p>Have not read the paper yet, but what had been on Twitter was speculation and no proof.
</p>Economist on "A Modern Gauss-Markov Theorem"
https://www.econjobrumors.com/topic/a-modern-gauss-markov-theorem/page/18#post-8621955
Wed, 04 Jan 2023 04:59:43 +0000Economist8621955@https://www.econjobrumors.com/<p>Looks like Lihua Lei and Wooldridge finally produced a paper. It covers some old ground but unifies and extends some of the known results.</p>
<p><a href="https://twitter.com/CavaliereGiu/status/1610194841064947712" rel="nofollow">https://twitter.com/CavaliereGiu/status/1610194841064947712</a></p>
<p>The main point was covered on Twitter early in 2022: there are no nonlinear, unbiased estimators in the original setting that Hansen covered.
</p>Economist on "A Modern Gauss-Markov Theorem"
https://www.econjobrumors.com/topic/a-modern-gauss-markov-theorem/page/18#post-8375191
Wed, 12 Oct 2022 12:24:59 +0000Economist8375191@https://www.econjobrumors.com/<blockquote><p>
Nice exhibit of your intellectual level. Over and out.
</p></blockquote>
<p>You started calling confused folks here morons, and now you seem to be upset at them calling you back, even appealing to their intellectual level? Lol.</p>
<p>This thread is too much fun. I am going to get some popcorn.
</p>Economist on "A Modern Gauss-Markov Theorem"
https://www.econjobrumors.com/topic/a-modern-gauss-markov-theorem/page/17#post-8375039
Wed, 12 Oct 2022 09:58:06 +0000Economist8375039@https://www.econjobrumors.com/<blockquote><blockquote>
<blockquote>
<blockquote><p>>>Why is it natural to require unbiasedness *only under* homoscedasticity? That actually seems weird and unnatural, and Hansen's version is more natural.<br />
It is unnatural if you then go on to prove dominance of the OLS-estimator only for distributions with homoskedastic errors as Hansen does. Had he establisehd dominance of the OLS-estimator under any variance-covariance structure, that would have been another story. However, as we know this is impossible.<br />
Please read the papers before posting, m0r;0n.
</p></blockquote>
<p>Why don’t you F OFF
</p></blockquote>
<p>Triggered much?
</p></blockquote>
<p>Bo zo
</p></blockquote>
<p>Nice exhibit of your intellectual level. Over and out.
</p>Economist on "A Modern Gauss-Markov Theorem"
https://www.econjobrumors.com/topic/a-modern-gauss-markov-theorem/page/17#post-8375037
Wed, 12 Oct 2022 09:56:46 +0000Economist8375037@https://www.econjobrumors.com/<blockquote><blockquote>
<blockquote>>>Why is it natural to require unbiasedness *only under* homoscedasticity? That actually seems weird and unnatural, and Hansen's version is more natural.<br />
It is unnatural if you then go on to prove dominance of the OLS-estimator only for distributions with homoskedastic errors as Hansen does. Had he establisehd dominance of the OLS-estimator under any variance-covariance structure, that would have been another story. However, as we know this is impossible.<br />
Please read the papers before posting, m0r;0n.
</p></blockquote>
<p>Why don’t you F OFF
</p></blockquote>
<p>Triggered much?
</p></blockquote>
<p>Bo zo
</p>Economist on "A Modern Gauss-Markov Theorem"
https://www.econjobrumors.com/topic/a-modern-gauss-markov-theorem/page/17#post-8375029
Wed, 12 Oct 2022 09:53:17 +0000Economist8375029@https://www.econjobrumors.com/<blockquote><blockquote>>>Why is it natural to require unbiasedness *only under* homoscedasticity? That actually seems weird and unnatural, and Hansen's version is more natural.<br />
It is unnatural if you then go on to prove dominance of the OLS-estimator only for distributions with homoskedastic errors as Hansen does. Had he establisehd dominance of the OLS-estimator under any variance-covariance structure, that would have been another story. However, as we know this is impossible.<br />
Please read the papers before posting, m0r;0n.
</p></blockquote>
<p>Why don’t you F OFF
</p></blockquote>
<p>Triggered much?
</p>Economist on "A Modern Gauss-Markov Theorem"
https://www.econjobrumors.com/topic/a-modern-gauss-markov-theorem/page/17#post-8375022
Wed, 12 Oct 2022 09:50:31 +0000Economist8375022@https://www.econjobrumors.com/<blockquote><p>>>Why is it natural to require unbiasedness *only under* homoscedasticity? That actually seems weird and unnatural, and Hansen's version is more natural.<br />
It is unnatural if you then go on to prove dominance of the OLS-estimator only for distributions with homoskedastic errors as Hansen does. Had he establisehd dominance of the OLS-estimator under any variance-covariance structure, that would have been another story. However, as we know this is impossible.<br />
Please read the papers before posting, m0r;0n.
</p></blockquote>
<p>Why don’t you F OFF
</p>Economist on "A Modern Gauss-Markov Theorem"
https://www.econjobrumors.com/topic/a-modern-gauss-markov-theorem/page/17#post-8375017
Wed, 12 Oct 2022 09:48:29 +0000Economist8375017@https://www.econjobrumors.com/<p>Because you assume homoskedasticity for ols being blue. Why would you then relax this GM assumption for unbiasedness?
</p>