Try to hold these two thoughts simultaneously: (i) Greens voters all come from the ALP, not the Liberals; (ii) Greens voters are so uncommitted to the left that they’re likely to preference against Labor because of the mining tax.

What, you can’t do it? You think that if they’re ex-Labor voters, then they’re mostly from the hard left and likely to be against the mining companies — or conversely, that if they’re that worried about the mining tax, then probably a fair few of them used to be Liberal voters?

Then you evidently don’t have the mental flexibility of former Labor senator and now pundit John Black, who in Saturday’s Australian argues both propositions, apparently oblivious to the contradiction between them. He might almost have been taking lessons from the White Queen, who boasted of the ability to believe “as many as six impossible things before breakfast”.

Black is quite correct in identifying the relatively high socio-economic status of Greens voters, which means that their economic interests cut against high-tax policies. Not everyone votes their economic interests, and Greens voters especially would be among the least likely to do so, but it’s not unreasonable to think that might have some effect on their preference flow.

But this only represents a gain to the Coalition if the Greens voters are all (or almost all) coming from Labor in the first place, so this is what Black argues: “The Greens vote is now moving former Labor voters across to the Liberals in the same way that the old Democratic Labor Party moved Catholics into the mainstream Liberal ranks 50 years ago.”

Nor does he think this is a recent phenomenon; he claims that in 2007, “the Greens took more primary votes from Labor candidates than they gave back in second preferences”.

Is there any evidence for this? Well, no. The only thing Black has is the fact that the drop in Labor’s primary vote since the last election is roughly equal to the rise in the Greens vote (according to the latest Newspoll, confirmed by this morning’s Nielsen poll).

But is it not equally plausible to think that, in addition to some movement from Labor to Greens, there are also groups of people moving from Labor to Coalition, and from Coalition to Greens?

I’m at a loss to explain how Black missed this simple point. It could be part of the broad Labor campaign to denigrate the Greens, painting them as both extreme left-wing splitters and pawns of the Liberals, or it could be just carelessness.

Many people on both sides seem to have a lot invested in the idea that Greens voters come from Labor, not from the Coalition. No doubt the majority do, but both their geographical distribution and their socio-economic profile (as far as we know – remember it’s based on survey data) suggest that a significant and probably increasing number are ex-Coalition voters. The fact that Nielsen finds an increased preference flow to the Coalition corroborates that.

Both polls show a large drop in Labor’s primary vote (8% in Newspoll, 10% in Nielsen); it simply defies belief to think that almost none of that is going to the Coalition. Much more likely that Labor is leaking to both Coalition and Greens, and that the Greens are also picking up some of the Coalition vote.

To play a role analogous to the DLP of the 1950s and ’60s, a party has to harvest voters from one side and deliver them to the other side as preferences. It’s not impossible that the Greens will do that, but it’s not intuitively plausible and there’s no actual evidence of it happening.