Growth Rates

[Niall Beirne]

Hello!

I am trying to calculate growth rates for income (both yearly and every 5 years), by district. The data is every 5 years, 2004, 2009, 2014 and I am struggling to generate these values. Apologies if it is very obvious, it has been a while since I used Stata.

I have attached an image below of the data.

Thank you!

[George Ford]

This is a cheat, but I'd construct a time variable that is just 1, 2, 3, and then xtset on that. Then you can just use lags.

egen time = group(t1)  //check to make sure you get the right result
xtset time district
bys district: g growth = ln(x)-ln(l.x)

[Niall Beirne]

Thank you George.

When implementing your code, I have the issue of district being a string variable. To tackle this I used the code:

encode district, gen(district1)
drop district
rename district1 district

However once this is done and I re-run your code, I receive an error code saying the data is not sorted. Any ideas?

[Jared Greathouse]

Niall Beirne Make a unique ID for district too

egen id = group(district)

cap which labmask

if _rc {
    
    qui net inst gr0034.pkg
}

labmask id, values(district)
​​​​​​​

[George Ford]

xtset your data?

bys district time: ...

[Niall Beirne]

Thanks guys, all sorted!

[Nick Cox]

Jared Greathouse gave code which can be shortened to

egen id = group(district), label

labmask from the Stata Journal has its uses, but it's not needed here.

Further, the option delta(5) on your time variable can be applied with tsset or xtset.