PROVIDE Project Technical Paper 2005:1
this assumption various households now earn income from labour although the person-
level labour income variables for that household are all zero.
February 2005
In do-file inclabscaling.do a scaling factor is created that either leaves the person-
level inclabp_new variable in tact or scales it up or down so that the sum of the
inclabp_new variables in each household equal the household-level inclab variable. For
4,946 households inclab remained zero before and after the adjustments. Variable
inclab remained positive and unchanged for a further 7,428 households. For all these
households the inclabp_new variables remained unchanged. For 11,399 households the
inclabp_new variables were scaled upwards by an average factor of 1.43 due to changes
made to inclab. A further 2,410 households initially reported zero income from labour
but now had positive income figures. For these households the head of the household
was assumed to have earned that income. The original inclabp_new variable was saved
as inclabp_old and inclabp_new was scaled up to its new levels.
Figure 1 compares the new and old versions of person-level labour income. As
expected the average income is now slightly higher for all of the occupation groups,
except for farmers and unspecified workers. Many households that previously reported
zero income from labour were now added to these two groups due to the adjustments
made to inclab in fixing.do. Specifically the addition of income from home production
to inclab explains the increase in the number of agricultural workers. The new workers
added to this group obviously had a lower average wage than the rest of the agricultural
workers, which explains why the average wage drops. Most of the other ‘new’ additions
were allocated to the unspecified category, because these workers did not previously
report income and never specified an occupation category. The average wage of
unspecified workers drops for the same reason as the drop in agricultural wages.
4.2.5. Forming factor groups (newfact.do and newfact_old.do)
Do-file newfact.do creates a province-level occupation code variable called newfact.
This variable is similar to mergefact but disaggregates workers further by race and
province. It contains 88 different types of labour. The original occupation groups
mapped from the LFS 2000:2 are (1) legislators, senior officials and managers; (2)
professionals; (3) technical and associate professionals; (4) clerks; (5) service workers
and shop and market sales workers; (6) skilled agricultural and fishery workers; (7)
craft and related trades workers; (8) plant and machine operators and assemblers; (9)
elementary occupations; (10) domestic workers; and (11) not adequately or elsewhere
defined, unspecified. In some provinces certain of these province-race-labour sub-
categories are not well represented, in which case aggregate groups are formed by
merging an occupation group with another of similar skills level. Thus, high skilled are
61
© PROVIDE Project