Stefan Sobernig
2013-Jun-21 10:39 UTC
[R] Gini coefficient sensitive to "skewness" of Lorenz curve?
Dear list, I have already posted to sci.stat.consult, however, I was hoping that there might be some experience in the R community on the following question:> I am wondering whether anyone has experience with variants of (parametric) Gini coefficients (or alternatives) which allow for reflecting/penalising the skewness of a Lorenz curve? > > For my part, I just stumbled across the following: > http://www.accessecon.com/Pubs/EB/2010/Volume30/EB-10-V30-I2-P146.pdf > > While appealing to me, I did not find any alternatives, nor any reported use of this variant; nor is there any account of its liabilities (e.g., strategies to estimate the parameters). >> ... >> P.S.: My background is not in economics/econometrics or the like, rather i am currently evaluating the use of inequality measures for software measurement. See for some background: > http://dx.doi.org/10.1109/ICSM.2011.6080798I'd appreciate any hint in this direction! stefan
David Winsemius
2013-Jun-21 19:39 UTC
[R] Gini coefficient sensitive to "skewness" of Lorenz curve?
On Jun 21, 2013, at 3:39 AM, Stefan Sobernig wrote:> Dear list, > > I have already posted to sci.stat.consult, however, I was hoping that there might be some experience in the R community on the following question: > >> I am wondering whether anyone has experience with variants of (parametric) Gini coefficients (or alternatives) which allow for reflecting/penalising the skewness of a Lorenz curve? >> >> For my part, I just stumbled across the following: >> http://www.accessecon.com/Pubs/EB/2010/Volume30/EB-10-V30-I2-P146.pdf >> >> While appealing to me, I did not find any alternatives, nor any reported use of this variant; nor is there any account of its liabilities (e.g., strategies to estimate the parameters). >> > > ... > > >> P.S.: My background is not in economics/econometrics or the like, rather i am currently evaluating the use of inequality measures for software measurement. See for some background: >> http://dx.doi.org/10.1109/ICSM.2011.6080798 > > I'd appreciate any hint in this direction!This is not really the correct place to pose such a question. You might instead read the Posting Guide to see why I say that, and then post to CrossValidated.com -- David Winsemius Alameda, CA, USA