Why does the simpler way to computing the Gini coefficient work?

50 Views Asked by At

The Gini coefficient is often used in economics to calculate inequality.

According to this website, the formula for calculating Gini coefficient is

$G = \frac{\sum_{i=1}^{n} \sum_{j=1}^{n} |x_i - x_j|}{2n^2\bar{x}}$

However, if all values are arranged in ascending order, computation is quicker:

$G = \frac{2}{n^2\bar{x}} \sum_{i=1}^{n}i(x_i-\bar{x})$

Why does the last formula work?

1

There are 1 best solutions below

0
On BEST ANSWER

The first formula has no $i=j$ contributions and is a sum of $i\leftrightarrow j$-symmetric terms, so can be rewritten as $\frac{1}{n^2\bar{x}}\sum_{i>j}(x_i-x_j)$ if we adopt the convention $i>j\to x_i>x_j$. It suffices to prove $\sum_{i>j}(x_i-x_j)=2\sum_i (ix_i -i\bar{x})$. The term $x_i$ appears on the left-hand side $i-1-(n-i)=2i-1-n$ times, and on the right-hand side $2(i-\sum_i \frac{i}{n})=2i-\frac{2T_n}{n}$ times where $T_n:=\sum_{k=1}^n k=\frac{n(n+1)}{2}$ is the $n$th triangle number. You can verify these coefficients agree.