Correction, they used to import sometime ago. Now almost all Motorocycle brands (over 75% local brands) are assembled/made locally, including about seven Bajaj models. Our market is much smaller than India, so 100% local value addition has to make sense for volumes....this is not a Bhartiya nationalistic thing, it has to make economic sense...India has used CKD assembly in the past too in spite of its large markets.
Fair enough. I get carried away with some other members. There is really nothing too surprising about the model Bangladesh is following....it has a good track record worldwide. I have always said that the best projection/vision is the actual realisation that happens.
Either compare both country with same criteria and year or not at all.World bank data are very outdated even missing in case of BD.Not only BD but also in case of a lot of countries.You only have to look at their poverty rate chart to realize how reliable they are.They use different year for different countries,even different base year then list them to make a page.Look at the condition of this page.
http://data.worldbank.org/indicator/SI.POV.DDAY
World bank counted 77.6% of BD' population below 3.1 dollar (ppp) or annual 1132 dollar ppp when they themselves said in 2014 BD's ppp per capita 3123 dollar.It is impossible that nearly four fifth of the population have income less then one third of the average.No other indicator support that assumption.Our GINI co-efficient is not that high,it is actually lower than India.Even Nepal whose gdp ppp per capita they showed 2374 dollar have 48.4% of population below 3.1 ppp dollar.I am convinced they are seriously missing something in case of BD.I don't know what it is.
Well they get different years for different countries because full surveys for every country are not available each year (they are done by the country's own statistics bureaus etc.. and the data is used and analysed by the World bank and it normally gets published many years later in the series after vetting and standardisation or in this case, coefficient analysis + back scaling.
Now lets look at if the Gini makes sense from having such an income distribution.
Lets assume the scaling that World Bank established holds for income levels of 1.9 and 3.1 PPP
77.6% below 3.1 dollar so 1132 dollar PPP
43.4% below 1.9 dollar so 693 PPP
Lets get some calcs out of the way.
PPP per capita for 2010 = 2400 (not 3123 which is from a much later year, we have to use data all from the same year).
Population of Bangladesh in 2010 = 151.6 million
Total PPP economy size = 151.6 * 2400 = 364 billion
Lets assume each bracket's average income is 75% of the final bracket (skew factor) and see what the end result is.
So below 1.9 dollar, the total income estimate is: 693*0.75*0.434*151.6 = 34.2 billion
From 1.9 to 3.1 dollar lies a further 34.2% of people so income estimate using same assumption = 1132*0.75*0.342*151.6= 44 billion
So remaining population will have 364 - 34.2 - 44 = 286 billion total income in PPP and represent the remaining 22.4% of the population.....giving an average income of 286 billion / (151.6*.224) = 8400 PPP.
This gives the following segments of income vs population block:
Lowest 43.4% earn 9 % of total income
Low- Middle 34.2% earn 12% of total income
Middle-High 22.4% earn 79% of total income.
Lets simulate as a population with only 10 people (1 at each 10% interval).
I approximated their incomes as follows to get in the ball park of the component and total averages calculated and tested for two skew factors (0.75 and 0.85):
Sim 1) 350, 460, 580, 700, 800, 900, 1000, 3000, 7000, 9000
Sim 2) 550, 600, 650, 700, 900, 1000, 1200, 3500, 7000, 8000
Giving a GINI coefficient of around 50%.
http://www.peterrosenmai.com/lorenz-curve-graphing-tool-and-gini-coefficient-calculator
Sim 2:
I would imagine that if we get even better resolution of the spread within/between each band (especially person number 8) and account for the fact that regular Gini uses Nominal income....we can get coefficients as low as 30%. Thats mostly because PPP multiplier tends to influence the higher and middle income portions of the curve due to the scaling.
If you read this paper:
http://www-wds.worldbank.org/servle...58349_20090909092401/Rendered/PDF/WPS5044.pdf
You will see that when using PPP, the Gini coefficient of the world is around 0.7 (70%) in more recent years. So PPP and nominal Gini may not be useful to compare across each other but only within themselves.
I mean here a more nominal based measure of Gini gives the world gini average at around 0.5 in recent years:
http://www.conferenceboard.ca/hcp/hot-topics/worldinequality.aspx
So the scaling factor looks to be about 0.5/0.7 = 0.71 for conversion between nominal and PPP Gini (assuming linearity which it may not be). So a PPP Gini for Bangladesh of around .5 using this scale factor would suggest a nominal Gini of around 35% or so....which compares closely to the published Gini estimates for Bangladesh that are out there.
So Bangladesh may actually get a better poverty rate measure if we use nominal instead of PPP.
It is to be noted that Bangladesh national poverty line had about 31.5% below it in 2010....so the difference between nominal and PPP may account for that.
But then the debate ultimately stems from which measure is better, nominal or PPP and we have discussed that plenty times enough (i.e are you interested in conversion of what you consume to US dollars and associated price level scenario i.e nominal or are you more interested in local consumption volumes of goods and services i.e PPP)
I can try to use a given Gini and backtrack an income spread using a bit of estimation (with national total income information) and then derive what the poverty rate is from that. I will try later if you are interested and I got some time. It will however have to be nominal based I think.
