Make a weighted distribution

[mr Zakeri]

I used to distribute national taxes X to households in Households Expenditure and Income Survey for measuring progressivity. I distributed all tax components by weight (distribution 1) and then divided them to weight for making household level values (distribution 2).
when I compare these two distribution, the share in deciles are different, and also when use them for progressivity analysis by weight, the results are so different! what is wrong? Please! Regards.
my syntax is:

*Allocating national tax
**Set national tax
gen totaltax= 10*10^9
**Allocating tax total tax to households
format tax %20.0fc
summ tax, format
replace tax=0 if tax==.

gen wtax= hhwt*tax
egen totwtax=sum(wtax)
gen basetax=wtax/totwtax
gen alltax=totaltax*basetax
table decile, c(mean alltax)/*(weighted distirbution 1)*/

*National tax tax calculation
gen ntaxtax=alltax/hhwt
table decile, c(mean nalltax)/*(national distirbution 2)*/

[Phil Bromiley]

You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex. Also, we normally use our real names (first and last) on this listserve.

Without more information (e.g., what the different variables are), it is hard to know exactly what you've done. You have several different variables but how they relate is not obvious from the names. I would wonder about gen basetax=wtax/totwtax and gen alltax=totaltax*basetax. It looks like something should be additive in one of these equations. But, there seem to be simple multiplicative relations among many of these variables.