Income and Expenditure Accounts Technical Series
Constructing Provincial Time Series: A Discussion of Data Sources and Methods
3. Linking methodology
Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.
Data sources are ranked from newest to oldest (for example, Statistics Canada, Dominion Bureau of Statistics, academic sources, etc.), and the data are then linked using the most appropriate method.
The starting point for each variable is the current vintage of data. The current vintage corresponds with the price index, income and population variables produced on an ongoing basis by Statistics Canada. In the dataset, they are denoted by a superscript (A) to indicate their current vintage status.
When national accounts programs are revised, recent data tend to be most affected, and historical estimates, least affected. This is especially the case for growth rates. The relatively larger adjustment to recent data occurs as statistical systems react to structural changes in the economies about which they report.
In most cases, historical estimates are linked based on their growth rates; that is, the modern estimates are projected back through time based on the growth rates of historical estimates for similar variables. In the dataset, these estimates are denoted by a superscript (B) to indicate that they are historical projections. The historical estimates are treated as if they are indexes with a base value equal to the earliest data point from the modern data vintage. This approach works well because, while revisions may change level estimates, the growth rates for the economic aggregates employed are similar across data vintages.
The advantage of using growth rates rather than regressions to back-cast the time series is that the magnitude of historical events is preserved. Regression estimates can expand or contract the variance of historical estimates relative to current estimates based on what is typically a limited sample of overlapping observations. Moreover, the overlapping observations may not be particularly representative of previous periods.
In some cases, data are limited, and instrumental variable techniques (nearest-neighbour matches, regression-based approaches, constructed indicators) are used. These estimates are denoted by a superscript (C) to indicate the instrumental variable approach.