Record Details

Peaks Over Threshold Vs. Lognormal Estimates Of The Czech Household Incomes

Research Journal of Economics, Business and ICT

View Archive Info
 
 
Field Value
 
Title Peaks Over Threshold Vs. Lognormal Estimates Of The Czech Household Incomes
 
Creator Cabla, Adam; University of Economics in Prague
 
Subject —
Peaks over Threshold, Parameter Estimates, Quantile Estimates, Income Distribution, Czech Households
C16, D31, O15
 
Description Income distributions are usually long-tailed and the right tail is often important part of income inequality metrics, but it is also problematic part of income distribution to be modeled. The POT method is theoretically well established method for modeling tails of unknown underlying distribution and thus candidate to become complement of the standard estimates. The article deals with the problem of parameter estimates using deHaan and CME methods and comparing the resulting quantile estimates with the one-distributional fitting. All of these estimates are done for the net money incomes of the Czech households. The results shows, that due to the data problems deHaan method usually gives.more robust estimates than CME method. The log-normal distribution usually fits the data well up to the quantile x0,995 but for the rest of the distribution, the GPD is better fitting distribution. The peaks over threshold method is then useful only for the genuine extremes and even there its estimates depends on the quality of data.
 
Publisher English Time Schools & Overseas Education
 
Contributor —
 
Date 2012-03-10
 
Type info:eu-repo/semantics/article
info:eu-repo/semantics/publishedVersion
Peer-reviewed Article
 
Format application/pdf
 
Identifier http://ojs.journals.cz/index.php/RJEBI/article/view/240
 
Source Research Journal of Economics, Business and ICT; Vol 4 (2012)
2047-7848
2045-3345
 
Language eng
 
Relation http://ojs.journals.cz/index.php/RJEBI/article/view/240/244
 
Rights Copyright (c) 2012 Adam Cabla
https://creativecommons.org/licenses/by/3.0/