Update on approval ratings

Last month we had a look at Donald Trump’s approval ratings, noting that he seemed to have got some – but not very much – of the boost that you would expect leaders to get in a time of crisis, and that other world leaders mostly do seem to be getting.

Since then, some of that fairly modest boost has dissipated. According to FiveThirtyEight’s rolling averages, Trump is polling at an approval rating of 44.0%, down from its peak of 45.8% in early April but still historically high; prior to the health crisis, he had not reached that level since March 2017. Disapproval shows the same pattern: 51.7%, up two points from its low point but still clearly below the long-term average.

As I noted last time, by the standards of previous presidents these are very small movements. Both Trump’s supporters and opponents seem very set in their ways, with only a small middle group susceptible to change.

You might naturally assume that that rigidity will carry through to voting intention, as indeed it might. Certainly presidential elections seem to be getting less and less volatile: none of the last six elections has recorded a two-party swing greater than 5%, a run unprecedented since the nineteenth century.

But there are some signals the other way. While Trump’s approval rating has improved, the polls of voting intention – both for him and the Republican Party in general – aren’t nearly as good. Democrat Joe Biden is stubbornly holding onto a lead of about five points, and in the “generic” congressional ballot the Democrat lead has widened to eight points, almost back to the 8.6% gap recorded at the 2018 election.

Also, as Nathaniel Rakich reported last week, a set of recent Senate polls “imply a Democratic wave of truly epic proportions.” It’s too early to say whether they represent a real trend, but it’s certainly not good news for Trump.

Interestingly, despite its very different experience of Covid-19, Australia has a somewhat similar story to tell about polling. Trumpist prime minister Scott Morrison has enjoyed a big boost in his approval rating: in last week’s Newspoll, 66% say they are satisfied with his performance and 30% dissatisfied. In the beauty contest or “better prime minister” polling he leads opposition leader Anthony Albanese by 56% to 29%.

But voting intentions have not shifted. Morrison’s Coalition leads (two-party-preferred) by just two points, 51% to 49% – almost unchanged from last year’s election.

There’s ample room for speculation as to why a significant number of people might approve of Morrison’s performance but still not want to vote for him. But for essential background on the question, have a read of Kevin Bonham’s discussion last month about the relationship between voting intention, approval ratings and beauty contests.

Bonham’s headline, “Why Better Prime Minister/Premier Scores Are Still Rubbish,” will tell you what he thinks. He concedes some small relevance to approval ratings, but finds that the “better prime minister” (or premier) scores add nothing to our knowledge: “Better Leader scores add no useful predictive information to that provided by a regression based on polled voting intention.”

If anything, the data indicates that high beauty contest scores correlate (weakly) with a less sustainable lead in voting intention: “In the last 12 federal elections, the higher Better PM leads have been correlated with governments underperforming compared to their final 2PP polling, and small leads and deficits have been correlated with governments overperforming.”

As Bonham points out, this is a point that Peter Brent has been making for a long time, although Brent usually expresses it in terms of approval ratings – here he is on the subject, for example, back in 2006.

If that relationship holds, then Morrison’s lofty approval ratings stand to do him little good, and Labor can take genuine encouragement from its standing as regards voting intention. But with two years to run until the next election, neither side should really be taking any more than the most casual interest in the polls.

For Trump, on the other hand, the gap between approval and voting intention is a matter of much more immediate concern.

5 thoughts on “Update on approval ratings

  1. A highly popular PM can, I think, override their own party more than otherwise. Rudd did so, but shied at the crucial hurdle of a double dissolution over climate change. Turnbull chickened out and paid the price. In Morrison’s case, the big question is whether he actually wants anything enough to do this. If he wants a place in history, he could override the climate deniers (who are also the Covid deniers)>


    1. Thanks John. Yes, I think that’s true, but is it only because the party is using approval ratings as a proxy for the leaders effect on voting intentions? In other words, if the MPs knew that the leader wasn’t winning them any extra votes, would they be influenced at all by the fact that the leader was personally popular? (I realise the question is probably academic, because in practice they couldn’t know that.)
      Since Morrison strikes me as someone who needs to summon all his intellectual energy to tie his own shoelaces, I find it hard to imagine him striving for a place in history. But perhaps I’m too harsh.
      Also, while I agree the Covid deniers are also climate deniers, I don’t think the converse holds: there are plenty of climate deniers in & around the Coalition who show no sign of embracing Covid denialism.


    1. Is it possible that the reason that the predictive power of high approval ratings for voting intentions is that approval ratings are a more lagging indicator – did I approve of what the PM did last week rather than a leading indicator – would I vote for these guys in the future.

      The form of the question may almost cause this dichotomy with one focus on recent and short term views and one focussed way into the future.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.