My two thoughts on the post in question were:
1) If frog has, as s/he appears to claim, gone through the NZ Super Fund Equity Holdings line by line and determined which parts of it are ethical and which are not, that is very impressive. It is a long list.
2) We should be careful of assuming that investing in WalMart (or by the same argument, developing countries) is bad because their workers aren't well-off. As Paul Krugman once said, bad jobs are better than no jobs.
However, I wanted to make this a post on the general concept of ethical investment.
I think that national investment funds should not invest in the following situations.
1) Where the market and/or Government has incorrectly priced externalities. E.g. highly polluting industries that are not subject to Coasean bargaining or socially efficient Pigovian taxation. This will encourage firms to restrict externalities themselves - see this interesting paper for further explanation. Of course, this is not a new idea.
2) Pursuant to the first, in situations involving repugnant transactions. We should not invest in dealers of child sex-slaves, for example. This is not at all a clear criterion, however.
In the 'pro' column, there doesn't appear to be a morally relevant difference between doing something ourselves and paying someone else to do it. A gangster that orders a drive-by shooting is culpable, even if he doesn't carry it out himself. I think child sex-slavery should be minimised, so facilitating the raising of capital for its continuance would be inconsistent.
In the 'con' column, People's moral compasses don't always completely overlap. You'll notice that with child prostitutes I dishonestly set myself up a very clear case. For example, frog from frogblog thinks that we should not invest in weapons companies. Many libertarians think that guns provide important protection against a tyrannical Government. Yet both their money would be in the pool. How should the moral criteria be set for a fund which constitutes money from the entire (tax-paying) populace?
To borrow one of my favourite Rawlsian concepts, it seems reasonable to have a moral framework determined by a loosely defined overlapping consensus of morality. Even if the majority consider an investment morally appropriate, we should not force the minority to invest in it, as it is forcing them to be morally inconsistent(*). We should only invest in things generally agreed on to be morally appropriate. It seems reasonable to set a (quite low) lower bound on this - racists might object to investment in firms owned by non-Europeans, for example. We can probably bring ourselves to ignore their cries.
Some issues are less clear-cut than this, for example investing in sweatshops. This is where we can borrow a second idea from Rawls, the well-known concept of reflective equilibrium. Essentially, on issues where there are clashes in interpretations of the situation but a broad agreement on many principles, we can solve the issue by argumentation, to ascertain which party is more consistent in its principles. We might agree that we want to maximise both the utility of the sweat-shop workers (or of their country in general) and our own through profit-making, with probably a higher weighting on the former. It wouldn't be unreasonable to rank them lexicographically, in fact. It seems then that we could refine any further intuitions we might hold (that we shouldn't give money to exploitative factory owners, etc) with recourse to evidence as to the agreed basic principles.
One final point: Does ethical investment cause your portfolio to make less profit? It seems obvious that it would, as you are selecting investments for criteria that are independent (or even negatively correlated with) profit. There appears to be mixed evidence on this score. But that is a topic for another time.
* This is more controversial than it seems, if you think through the domestic political implications. The point needs more expansion. The blog post however does not.