Re: how to compare living standards


"Jim Blair" <jeb@xxxxxxxx> wrote

Consider 5 quintiles. <with TP notes>

Time 1 (1990) <per-capita = 150/5 = 30>

Top =50 units <166% of per-capita>
Second=40 < 133%>
Third =30 <100%>
4th =20 < 66%>
Bottom=10 < 33%>

Time 2 (2000) <per-capita = 223/5 = ~45

Top =100 units <222% of per-capita>
Second=50 <111%>
Third =40 < 88%>
4th =25 < 55%>
Bottom=8 < 32%>


Q Did the "rich get richer"?
A In some sense.

Q Did the "poor get poorer"?
A In some sense.

Q What does this say about "mobility"?
A Nothing


Right. You merely illustrate rising inequality. If this was real data from
a real country about which you have no pre-conceived notions, that's all
this "data" would tell you.

If you were serious about investigating whether this hypothetical country
has "income mobility" or a rigid 5-caste system, you would want to know what
fraction of the individuals in each quintile in time_1 shows up in each of
the quintiles in time_2.

Suppose that in this strange country there's a weird lottery every 10 years.
Each individual draws a random number from 1 to 5, and is assigned to the
corresponding quintile for the next decade. You, the investigator, know
nothing about any lottery. You only see the resulting income data.

Does that data tell you anything about "mobility" in that strange country?

-- TP








.

Quantcast