How to generate a total and share of expenditure?

[Mohamud Hussein]

Hi,

Can you please tell me how I can generate a total value for the variable listed below and then create a share for each variable(e.g. wexp_grains/total)? An example would be useful.

 
wexp_grains
wexp_frveg
wexp_pulses
wexp_meat
wexp_fish
wexp_milkda
wexp_drinkbev
wexp_oilsfat
wexp_others

Thanks,
Mohamud

[Ivica Rubil]

*** begin code ***
egen wexp_total = rowtotal(wexp_*), missing
foreach x in grains frveg pulses meat fish milkda drinkbev oilsfat others {
gen share_`x' = wexp_`x' / wexp_total
}
*** end code ***

[Mohamud Hussein]

Thank you Ivica.

I took the approach you suggested but got a result that I did not expect. As you would see wexp variable from which I calculated the total and share expenditures resulted in 'no observations' despite the fact the wexp has a mean and range values.

Can you help me please to understand what is going on, and suggest an alternative approach to overcome the problem?

 
sum wexp_grains wexp_frveg wexp_pulses wexp_meat wexp_fish wexp_milkda wexp_drinkbev wexp_oilsfat wexp_others

Variable Obs Mean Std. Dev. Min Max

wexp_grains 74,936 2.13377 .7073587 1.211606 4.009885

wexp_frveg 124,895 1.789554 .7799143 .8157954 3.692125

wexp_pulses 27,755 1.529597 .5298207 .12 2.064826

wexp_meat 39,080 3.928044 2.320698 .5922232 7.026239

wexp_fish 11,816 4.106419 1.545911 2.055259 5.845754

wexp_milkda 23,752 3.583346 2.261793 .893132 6.690858

wexp_drink~v 31,444 2.808646 .969089 1.707988 5.07098

wexp_oilsfat 23,547 2.341372 .6586063 .9970102 3.128694

wexp_others 116,800 1.973933 1.794487 .5076035 8.734041

.

. *generating share of weekly expenditure food for each food category

. egen wexp_total = rowtotal(wexp_*), missing

. foreach x in grains frveg pulses meat fish milkda drinkbev oilsfat others {

2. gen swexp_`x' = wexp_`x' / wexp_total

3. }

(399,089 missing values generated)

(349,130 missing values generated)

(446,270 missing values generated)

(434,945 missing values generated)

(462,209 missing values generated)

(450,273 missing values generated)

(442,581 missing values generated)

(450,478 missing values generated)

(357,225 missing values generated)

. sum swexp_grains swexp_frveg swexp_pulses swexp_meat swexp_fish swexp_milkda swexp_drinkbev swexp_oilsfat swexp_others

Variable Obs Mean Std. Dev. Min Max

swexp_grains 74,936 1 0 1 1

swexp_frveg 124,895 1 0 1 1

swexp_pulses 27,755 1 0 1 1

swexp_meat 39,080 1 0 1 1

swexp_fish 11,816 1 0 1 1

swexp_milkda 23,752 1 0 1 1

swexp_drin~v 31,444 1 0 1 1

swexp_oils~t 23,547 1 0 1 1

swexp_others 116,800 1 0 1 1

.

end of do-file

[Ivica Rubil]

What does missing means in your data? Does it mean zero expenditure or truly missing data?

[Mohamud Hussein]

zero expenditure

[Ivica Rubil]

Then, before you run the code I suggested, just replace all missings for all expenditure variables by zero.
Because, if you don't, then the share of an expenditure category cannot be computed for those observations for which the amount of this very expenditure category is missing; whereas if missings are replaced by zero, it can be computed (it will be zero, of course).

[any owusu]

Hello, I am working with household survey data. how do I calculate household expenditure from the individual household member expenditures

[William Lisowski]

Welcome to Statalist.

This question isn't really a followup to the original topic. You will perhaps see a larger audience if you post it as a new topic in this forum by clicking on

General

and in the page that opens, click on the button labelled "+ New Topic" near the top of the page.

Also, if you have not already done so, take a look at the Statalist FAQ linked to at the top of this page for posting guidelines and suggestions.