2023-12-13 10:50:40 +00:00
|
|
|
# Measures
|
|
|
|
|
|
|
|
|
|
TODO:: separation of shares, inequality indices and ratios
|
|
|
|
|
|
|
|
|
|
**absolute** shows magnitude of difference between subgroups of pop -> focus on absolute welfare
|
|
|
|
|
|
|
|
|
|
**relative** shows proportional difference between subgroups of pop -> focus on welfare in relation to others
|
|
|
|
|
|
2023-12-13 14:04:19 +00:00
|
|
|
## Direction
|
2023-12-13 10:50:40 +00:00
|
|
|
|
|
|
|
|
### Vertical
|
|
|
|
|
|
|
|
|
|
Gini coeff, Theil, entropy measure, atkinson indexes
|
|
|
|
|
|
|
|
|
|
### Horizontal
|
|
|
|
|
|
|
|
|
|
mean differences between groups; ratios between better off/worse off
|
|
|
|
|
|
|
|
|
|
GGini; GTheil; GCOV
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
## All indicators
|
|
|
|
|
### Absolute
|
|
|
|
|
|
|
|
|
|
* quantiles (division by 5) / deciles (division by 10) / centiles (division by 100)
|
|
|
|
|
|
|
|
|
|
* Absolute Gini index
|
|
|
|
|
* ?
|
|
|
|
|
|
|
|
|
|
* Standard deviation
|
|
|
|
|
* deviations from norm in *absolute* terms
|
|
|
|
|
|
|
|
|
|
### Relative
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
* ratios of quantiles to each other (division of various quantiles)
|
|
|
|
|
|
|
|
|
|
* Gini coefficient:
|
|
|
|
|
* based Lorenz curve - population percentage versus accumulated fractions of wealth
|
|
|
|
|
* is area between lorenz curve and perfect equality
|
|
|
|
|
* needs: individual/household income/wealth; pop size
|
|
|
|
|
|
|
|
|
|
* Theil index:
|
|
|
|
|
* how much income distribution (person x receives y from total income) is away from perfect uniform distribution
|
|
|
|
|
* log -based
|
|
|
|
|
* can be disaggregated into subgroups
|
|
|
|
|
* Theil L > weighting factors are population groups; Theil T > weighting factors are fraction of appropriated income
|
|
|
|
|
|
|
|
|
|
* Atkinson measures:
|
|
|
|
|
* measures cumulative deviation of each income in relation to average income value in distribution
|
|
|
|
|
|
|
|
|
|
* Palma ratio:
|
|
|
|
|
* dividing share of income/wealth held by top 10% by bottom 40%, thus high = bad
|
|
|
|
|
|
|
|
|
|
* s80s20, s40s20 ratios:
|
|
|
|
|
* ?
|
|
|
|
|
|
|
|
|
|
* Concentration index:
|
|
|
|
|
* divide whole pop by 100 to find percentiles
|
|
|
|
|
* then you know how many ppl are in each group and add their total income
|
|
|
|
|
|
|
|
|
|
* Foster-Greer-Thorbecke-class distribution (@Sotomayor2021)
|
|
|
|
|
|
