Here is a chart produced in the context of a project focused on income and inequality indicators referring to the EU countries - for details see Glossary : EU statistics on income and living conditions (EU-SILC).

The indicators presented here are the income quintiles (median income of groups representing 20% of the population, ranked by per-capita income). In other words, in order to determine the value of D10, we calculate the median income of the poorest 20% of the population; we can also say that D10 is the threshold separating the poorest 10% from the rest.

These are nominal, non-deflated, yearly values, expressed in the national currency.

The chart has two parts: the small multiples (left), and Lady-in-Red (LiR).

#### Small Multiples

The small multiples show variation in absolute terms of each quintile, compared to the value in the base year, i.e., the first one in the series :

Here we can already see that, for most countries, various parts of the income distribution move to a different rhythm. Usually, the left side of the distribution (D10 and D30) grow less, or decrease more, than the median income for the respective country over the same period; this means that the standard of living of the poor is going down.

#### Lady-in-Red

LiR shows the evolution of the quintiles relative to the population median. By using the median as the reference point, we can see whether the income distribution has contracted or dilated, giving a more tangible feeling of the evolution of inequality.

I also wanted to eliminate the effect of distribution shifts (the tide raising all boats, due e.g. to inflation), which do not impact inequality; in order to do that, I subtracted the movement of the median (the central value of the distribution) relative to the value in the base year.

The final formula for the data points depicted by LiR is :

The value of the data points for the central value of the distribution becomes a constant, which allowed me to align the medians for all countries along a common axis. This, in combination with using the same scale for all countries, is facilitating the comparison of country distributions.

There is also a country zoom-in version of LiR, with the same format (included 2 examples).

#### A few words about the choice of representing time on the vertical axis for LiR.

One of the requirements of the chart was to present all 28 EU countries, thus the need to arrange them vertically, which in turn restricted the vertical real estate available for each country.

Placing time on the traditional horizontal axis did not create any issues with the small multiples, where the scales are country-specific and all series start from zero.

In contrast, the distance between the series of the LiR is comparatively larger than the series' ranges, so cramming everything in the same vertical space would obscure the individual evolutions, and in most cases the series would look parallel. An additional advantage of placing time on the vertical axis for LiR was already mentioned : distributions can be aligned by their central value, facilitating inter-country comparisons.

Data source : EU SILC

Year in original data is the survey year, not the income reference year

*See also another way of looking at income inequalities.*Data source : EU SILC

Year in original data is the survey year, not the income reference year

## No comments:

Post a Comment