The control number V0102 of PNAD does not correspond to the code of municipalities of IBGE. How to unite the banks?

Question

Hello!

I’m working with food safety supplements from pnad04, pnad09 and pnad13 IBGE and two other IPEA banks, one on municipal GDP ipea.pib and other on investments in social programs by municipality ipea.ps. With this, I’m trying to form a large database in which I can integrate the variables that interest me from the three banks using merge(pnad13, ipea.pib, ipea.ps, by="V0102"), where V0102 is the variable "control number" that represents the code of the municipalities in the PNAD. That is, I work with the municipalities as reference. The problem is that by transforming the data, dictionary and input downloaded from IBGE into the PNAD database, V0102 has 8 numbers when it should only have 7! This problem has already been reported in comment before, but the given suggestion does not solve the problem: I cannot reproduce the government data of PNAD 2013 from microdata. The official code of the Brazilian municipalities has 7 numbers, as stated in the IPEA banks and this link: http://www.msystemmirassol.com.br/imagens/Codigos_IBGE.pdf.

pnad13[1,3]
[1] 11000015

# Este é o código de Alta Floresta do Oeste (RO) na PNAD (contém 8 números)
# Mas nos outros dois bancos o código de Alta Floresta do Oeste (RO) tem 7 números:

11000015 %in% ipea.pib$V0102
[1] FALSE
11000015 %in% ipea.ps$V0102
[1] FALSE

# Se retirarmos um 0 "extra" chegamos a seu valor verdadeiro:

1100015 %in% ipea.pib$V0102
[1] TRUE
1100015 %in% ipea.ps$V0102
[1] TRUE

# Outro exemplo:

35008270 %in% pnad13$V0102
[1] TRUE
length(which(pnad13$V0102==35008270))
[1] 18

# Este é o código de Buritizal (SP) e na PNAD está com 8 números
# Porém este código está errado e o código correto de Buritizal (SP) segue abaixo:

3508207 %in% ipea.pib$V0102
[1] TRUE
35008270 %in% ipea.pib$V0102
[1] FALSE
3508207 %in% ipea.ps$V0102
[1] TRUE
35008270 %in% ipea.ps$V0102
[1] FALSE

Supposedly it would be because UF and V0102 share characters from the database. Still, deleting UF of the dictionary the problem is not solved. Has anyone ever worked with the PNAD using the municipalities as reference (control variable)? Any suggestions? Thank you very much.

Answer 1

This variable v0102 is not the municipal code in the PNAD. Municipalities are not disclosed because the research is not representative at this level. Therefore, it is not possible to join the PNAD with another basis having as key the municipality.