The Economist’s other example is proposition 13 from 1978 which restricted property taxes in California. The article suggests that this proposition has lead California to near ruin. Again, I am not so sure. One the one hand California definitely needs to get its finances in order. It probably makes sense to exclude public finances from citizen initiatives. On the other hand, how else can the citizenry restrict the size of the state? What would California be like without proposition 13?
And why exactly, other things being equal, would the citizenry wish to restrict the size of the state? That is, given a proposition that is ex hypothesi valuable, efficient, and productive in itself (if it were not, it could be voted down on its merits without the need to introduce the overriding principle) but that increases the size of the state, why should the populace reject it?
Political differences like this aside, Sam's proposition that
There are some questions on which the Australian public is just as qualified as parliamentarians in Canberra to make a decision
is either trivial or unworkable, depending on what agency gets to decide which these questions are. Sam presumably, from the framing, agrees that there are some questions on which the Australian public is not as qualified as parliamentarians in Canberra to make a decision; the issue, then, is how these are to be excluded. As Sam makes no gesture towards a solution of this issue, he is clearly adopting the radical simplification beloved of constitutional reformers, the pocket benevolent despot.
Again, as some comments point out, the citizen-initiated referendum idea rests on the assumption that decisions come in discrete packages, which they don't. I once attempted to find consensus in a deliberative assembly on a large and complicated initiative by putting it as a series of votes - do you want A or B; given A, do you want A1 or A2: given A2, do you want A2i or A2ii, and so on, the whole thing falling back into chaos when it became clear that people voted for A2 on the assumption (say) that the eventual vote would be A2iibIV6, not A2iibIV5, and if the vile miscreants of the other side rammed through A2iibIV5 would then roll back their votes to A1 or even B to avoid this outcome. It's Arrow's theorem, basically.