2006-06-29

Poor people's living standard

How do the poor people fare in different countries? I am sure there are such stats. I just couldn't locate them. Without the official data, I need to make some quick approximations.

The basic idea is quite simple. I would use the GDP/capita number (
List of countries by GDP (PPP) per capita) for each country, plus the income disparity numbers ( World GINI and income inequality data from wiki, where there are ratios of the top 10% vs bottom 10% of the population, and also top 20% vs bottom 20% -- see also GINI index) and estimate the GDP/capita (a proxy for income/capita) of the bottom 10% (and that of bottom 20%) percentile of the population in each country.

Technical notes:
  • The same GINI index may give different values for the income ratio of 0-10% group vs 90-100% group, because the distribution curves are different (i.e. the curves between 11% and 89% are different)
  • PPP is used, because for the poor people, they buy the basic goods with lowest quality (or food with no clear quality distinction), and there is only small difference in quality
  • The income disparity ratios in wiki were collected in different years. I had to assume that they did not change and applied them to the 2005 average GDP/capita numbers
The mothod I used is an approximation and not very precise. So takes these numbers as indicative only. I assumed the average of the top 10% and bottom 10% equal the average of the overall population. So if the top/bottom ratio is 10, then the bottom 10 percentile average = national average x 2 /(10+1). In reality the curve if far from symmetric and my numbers are almost definitely wrong. Perhaps the average of the logarithms of these numbers would equal the national mean. Anyway, this is for illustration purpose only. Systematic error will carry through all countries. So the relative values should still sort of make sense.
  • e.g. If the average of top and bottom are at 30% more than the national mean (instead of the mean), we just have to shift the numbers of all countries up by roughly the same factor. So the relative values are still (partially) true, thought one would expect that the amount of 'shifting' will be different for different GINI distribution.
The table below shows the results, quite a few facts I found are counter-intuitive
  • Due to higher income disparity, standard of living for poor people in China are not much better than those in Pakistan! (despite almost 3 times difference in national average)
  • Poor people in Russia are most well off in my (arbitrary) list of developing countries (well, Russia is not really a developing country. Its 2005 GDP is also helped by the commodity boom)
  • Poor people in Japan are much better off (more than 2x!) than their counterparts in US, althought US average is 35% higher than Japan's.
  • South Africa, Botswana, Chile and Russia have similar GDP/cap. But poor people in Russia make almost 9x those in Botswana and 3.5x of S Africa, 5x Chile
  • Indian dalit lives are (3x) better than those in Rio de Janeiro scum
  • Almost all ex-Soviet Bloc countries (except a few poor central Asian Stans) enjoy low GINI index and decent lives. Socialism is not that bad? (But will deteriorate in another dacade? see China, right below)
  • China, on the other hand, is among the most capitalistic countries, with a GINI index similar to that of US. So, please, don't let me hear the word "communist China" any more.

Country GINI 20% ratio 10% ratio
GDP/cap (PPP) Poor 20% Poor 10%
Developing






Mongolia 30.1 9.1 17.8
2,175 431 231
Kyrgyzstan 34.8 5.5 8.6
2,088 642 435
Vietnam 37.0 6.0 9.4
3,025 864 582
Pakistan 33.0 4.8 7.6
2,628 906 611
China 44.7 10.7 18.4
7,204 1,231 743
India 32.5 4.9 7.3
3,344 1,134 806
Indonesia 34.3 5.2 7.8
4,458 1,438 1,013
Romania 30.3 5.2 8.1
8,785 2,834 1,931
Russia 31.0 4.8 7.1
11,041 3,807 2,726
Nigeria 50.6 12.8 24.9
1,188 172 92
Tanzania 38.2 6.7 10.8
723 188 123
Kenya 42.5 8.2 13.6
1,445 314 198
Uganda 43.0 8.4 14.9
1,617 344 203
Ethiopia 30.0 3.4 6.6
823 374 217
Bostwana 63.0 31.5 77.6
11,410 702 290
S Africa 57.8 17.9 33.1
12,160 1,287 713
Venezuela 49.1 17.9 62.9
6,186 655 194
Brazil 59.3 26.4 68.0
8,584 627 249
chile 57.1 18.7 40.6
11,937 1,212 574
OECD






UK 36.0 7.2 13.8
30,470 7,432 4,118
US 46.6 8.4 15.9
41,399 8,808 4,899
Germany 28.3 4.3 6.9
30,579 11,539 7,742
Sweden 25.0 4.0 6.2
29,898 11,959 8,305
Japan 24.9 3.4 4.5
30,615 13,916 11,133

Finally, this estimate is promted by Curzon's obseravtion that rural people in Vietnam may have a higher standard fo living than rural people in Yunnan and Sichuan in China. Let's make another approximation here, by adjusting the China number using its provincial GDP vs national GDP ratio (2001 data) (this is almost certainly wrong, because the disparity is double counted in the national GINI ratios. but hopefully the overlapping is small). Yunnan and Sichuan GDP/cap is 64-68% of national average. So the poorest 10% earns about 2/3 of the national average, i.e. about $500 (actual number should be slightly higher, given the "overlap factor" and picking up of inland economies from 2001 to present). While the poor in Vietnam makes $582. That is 16% above Yunnan peasants.

So Curzon's observation is supported by facts. The Vietnamese peasant he saw perhaps make considerably more than Vietnam's average. Because Sapa is a border town with tourism and trade incomes. But even if not, the Vietnam bottom 10% still makes some 10-20% more than bottom 10% in Yunnan and Sichuan. Chinese government's recent efforts in alleviating the pains of the peasants are belated, but they are neccesary, and imminent.
Categories:

2 comments:

J. Otto Pohl said...

Very interesting, but how do you factor in non-income contributions to economic well being? Poor people in the UK for instance have full health coverage while those in the US often have none. This is a big difference.

Sun Bin said...

J,

i did not, as my numbers are only 1st order approximation.
it should be factored it, eg, by estimating the healthcare cost (eg using insurance premium), subtract it from GDP, convert into per capita value, and add back to the income of each person. This would raise the income # of the poorest, because the healthcare potion of the GDP is evenly divided (previously it is divided based on weight of each person's income).

alternatively, you can use the after-tax income and add back the various government subsidies.