In my two earlier posts on the "third source" (here and here) I discussed the appropriate test for classifying government powers and whether they could be treated as arising from statute (as, on a generous reading, Lord Sumption suggested in R. (New London College Ltd.) v. Home Secretary,  UKSC 51.
In this post, I want to expand on my previous suggestions about the classification and location of statutory powers by offering a concrete example.
Recently, the Supreme Court of Canada was shaken by allegations that former (and deceased) Chief Justice Bora Laskin engaged in inappropriate discussions with Canadian and British officials about the patriation reference. Justice Willard Estey was also fingered, but the Laskin allegations made much more noise.
The Court shifted promptly into gear and launched an investigation (which proved inconclusive) into the allegations. The investigation was internal only, but I do not think anyone's hackles would have been raised had the Court contacted members of Laskin's family to ask (but only ask) for any private papers in their possession or control.
Did the Court have a power to conduct such an investigation? Common sense suggests that the Court (or any other body) should be able to search for and reveal any relevant information it holds.
Principle suggests the same result. If the Court has no power, some other entity would have had to investigate. Given the absence of jurisdiction on the part of the Canadian Judicial Council, a body whose remit extends only to judges currently sitting, an inquiry would have to have been led by the executive or legislature. Adherents to any sort of separation of government powers would feel most uneasy about any such exercise. Put bluntly, either the Court had the power to investigate, or no one did.
Yet one would search in vain the Supreme Court Act for any mention of a power of investigations. Certainly, no express power has been granted to the Court. If the test for implied powers is one of reasonably incidental, internal investigations might on a (very) generous construction of the Act be legitimate. But if the test is one of necessity, then conducting investigations cannot be countenanced. That leads back to the choice of test outlined in an earlier post.
However, if the power to investigate is an implied power, then it can be exercised in a coercive manner (although I accept courts are often reluctant to legitimate certain types of coercion, such as the compulsion of testimony, in the absence of clear statutory authority). Perhaps people -- including those serving in the executive and legislative branches at the time of the alleged indiscretions -- could be compelled to deliver up relevant documents to the Court. That would be a very far-reaching power to imply.
Far better, I submit, to consider the power to conduct an investigation as a power which any statutory body has by virtue of its existence, springing from the third source (or whatever label we want to use). This is certainly true of internal investigations. I think it is also be true of non-coercive requests for cooperation from outside parties.
As to the question of where to locate the power, I think it would be odd to locate this power in the nature of the Crown as a corporation sole. If a power exists, it is much more plausible that it is a power of the Court as a statutory body (though not a corporation sole), exercisable only in the furtherance of the Court's functions, than a general power of the Crown shared by all statutory bodies.
As I suggested in a previous post, treating the power as arising from the Act has the added benefit of confining its exercise to the furtherance of the objects of the empowering legislation. The Court cannot go off leading investigations into things that are manifestly not its business.
In summary, the Supreme Court of Canada example suggests in my view that a tripartite division of statutory powers is appropriate, with a strict test for the implication of powers.