If you look at legislation of other countries, you'll see specific provisions that include things, for example, like the method of selection of the chief statistician, that there should be a selection committee, what the criteria for selection are, as opposed to the situation in Canada where the chief statistician is appointed indefinitely, but at pleasure. The chief statisticians in other countries are appointed for fixed terms, on good behaviour, which gives them greater protection.
The distribution of the powers under the act tend to be based on how and what, so the “what authorities” appropriately belong to the government, the minister to the political level deciding what statistics are required. But on the how—the statistical methods, the analysis, the dissemination—in order for the data to be credible, there shouldn't be political intervention in those matters, and therefore ideally those powers should be—