Conclusions

View the most recent version.

It is often argued that low-income rankings are not robust. Using the stochastic dominance approach, this paper provides a robust way to compare regional low income without arbitrarily selecting a low-income line. We reach the following conclusions.

First, by applying dominance tests for the FGT class of poverty indices—up to third order condition—between regions, we show that, in most cases, dominance relations can be determined and regional low income can be ordered for a wide range of low-income lines.

Second, it reveals that low-income rankings based on commonly used Canadian low-income cutoffs (LICOs) are not robust. An illustration for Newfoundland and Labrador (N.L.) and Ontario shows that the opposite outcomes can be concluded when different low-income lines are chosen. Also, the LICOs only compare the headcount at one low-income line, while ignoring the depth and intensity of low income. The methodology used in this paper offers a more informative and revealing understanding of the distribution of low income. For instance, it is found that N.L.'s low income dominates that of Alberta at the second order, despite that the former has a significantly higher LICO headcount rate than the latter.

Third, in 2000 British Columbia (B.C.) ranked the highest in low income, as B.C. was first-order dominated by all other provinces. Quebec and Manitoba were in the second and third places in low-income rankings, respectively, with Saskatchewan/Alberta fourth, Ontario sixth, and N.L./Nova Scotia seventh. New Brunswick and Prince Edward Island were provinces with the least low income in Canada, as their low-incomes dominate all other provinces at first order.

Furthermore, this paper also demonstrates that dominance results are sensitive to assumptions made to defining equivalent income and the concept of income. Generally, dominance results are robust to the choice of equivalence scales, while rank reversal occurs when alternative cost-of-living deflators are used. Switching from an absolute to a relative low-income concept has virtually no effect on low-income rankings for B.C. and the Atlantic provinces but not in the case for other provinces. The findings urge closer scrutiny on underlying assumptions. Finally, the answer to the question "Where is low income greatest in Canada?" goes to B.C., at least for 2000. The result is robust for all scales, regardless of the choices of poverty lines, cost-of-living factors, equivalence scales and an absolute or a relative low-income concept.