Revisions to 2006 to 2011 income data
Examining longer-term trends with SLID and CISExamining longer-term trends with SLID and CIS

View the most recent version.

The previous sections described revisions to the 2006-2011 SLID data to make it more comparable to data from CIS. This section provides advice to users who wish to look at longer trends in income statistics using pre-2006 data^{Note 1}. Because the revision mainly affected low-income results, this section concentrates on trends in low income. Graphs comparing selected revised and unrevised low income trends are presented in the annex.

To summarize trends in low income in the revised and unrevised data, Table 5 compares average low-income rates (using the LIM-AT and LICO-AT) from 1996-2000 with those from 2007-2011^{Note 2}.

In Canada overall, the average low-income rate based on the LIM-AT increased by 0.2 p.p. before the revision and 0.8 p.p. after the revision, while the low-income rate based on the LICO-AT fell by 4.8 p.p. before the revision and by 4.1 p.p. after the revision.

Because the revisions to low-income rates tended to be upward, changes in low-income rates over time are in more of an upward direction in the revised data. That is to say, increases over time are usually larger in the revised data and declines are smaller. This is as expected given the pattern of revisions in low-income rates for 2006-2011 discussed earlier in this note.

For example, in Ontario, the change in the average low-income rate based on the LIM-AT between the two periods increased from 1.8 p.p. before the revision to 2.8 p.p. after the revision. In Calgary, the corresponding change went from -2.8 p.p. to -1.8 p.p. as a result of the revision. However, in most cases, conclusions drawn with the revised data would be similar to those with the unrevised data. Cases where there is a large increase (or decrease) in low income tend to be the same in the revised and unrevised low income series.

Conclusions regarding trend changes in low income based on the LICO-AT do not tend to change as a result of the revision. In most cases, low-income rates based on the LICO-AT show declines in both the unrevised and revised series.

The fact that in most cases conclusions drawn using the revised and unrevised statistics are unchanged leads to the recommendation that income estimates before 2006 remain suitable for analyzing long term, cyclical trends, and can be used with the recalibrated results from 2006 to 2011 and the CIS results from 2012 and beyond.

• Previous
• Next