A detailed methodology note for these modeled liquid asset estimates is available here.
The model was created using regression analysis of the SFS data and these two variables from the Census of Population:
a) Higher incomes should accumulate more savings and create greater liquid assets.
b) Immigrants are generally attracted to economically dynamic regions, which are expected to have higher wealth holdings and liquid assets.
Models including the mean age of major income earners or the percentage of major income earners with a bachelor’s degree or higher education generally did not correlate as well as the variables indicated above.
Synthetic vs Composite vs Outlier estimates
The Survey of Financial Security (SFS) does not have a sufficient sample size to provide accurate estimates of liquid assets at the CD level. In CDs where at least 10 observations exist, an "inaccurate" estimate was made. The final estimates are a weighted average of this SFS estimate and the model estimate, which is called the synthetic estimate. The weighted average is called the composite estimate. In cases where there is no SFS estimate, only the synthetic estimate is used. The data table indicates whether the estimate is synthetic or composite.
Some estimates are labelled as Outlier. These estimates are, in fact, composite estimates. The label indicates that the SFS estimate was excluded from the estimation of the model parameters because they were considered outliers. However, the SFS estimate is still included in the final weighted average of the SFS estimate and the synthetic model estimate.