Rate-weight analysis: A critique and a reformulation

Conclusion

We have suggested a reformulation of the technique of DNO for regional-national growth rate differential by separate rate (growth) and weight effects as it is shown that their technique appears to be biased and inconsistent. From an analytical point of view the main points requiring modification are primarily two: (a) the use of geometric mean of initial and final year weights by sectors violates aggregation consistency, and (b) normalization rules are not obeyed by DNO for regional-national comparisons since uniform weights are not applied across both the region and the nation. An empirical test is performed by a reformulation (which eliminates the above two errors) with the employment data by 11 sectors for two regions, Atlantic Provinces and Ontario, vis-a-vis Canada (the nation) for two time units, 1961 and 1971. Separate rate and weight effects are then identified for Atlantic-Canada and Ontario-Canada.