Does policy matter for the Liberals?

If this morning’s Newspoll (pdf) on which party would best handle various issues is right, the Coalition’s policy change to support signing of the Kyoto protocol has seen it drop 10% to 15% as the party that would best handle the environment. That’s their lowest score on the evironment ever. Their decision to drop WorkChoices has seen their rating for industrial relations drop 7% to back where it was when the original WorkChoices was in force.

Their decision to defend the Howard government’s record on the economy has seen them drop 9% as the party that would best handle the economy.

Their unknown policies on a range of other issues have seen similar drops in health (9%), education (8%), water planning (7%), welfare and social issues (7%) and national security (11%).

So whether the opposition agrees with the government, disagrees with the government, or has no policies the results are much the same.

Issue ownership theory assumes that it is pretty rare that voters have more than the vaguest idea what policies parties have, and instead they infer which party would best handle issues from prior perceptions of the party and recent general performance. The Coalition’s leadership problems and chaos on the apology and industrial relations, plus the ongoing Ruddmania on the other side, are driving these results. The Liberals are still ahead on the economy and national security, as issue-ownership theory would predict based on past strong results, but those issues are experiencing the same trend as all the others.

In recent work I have been doing on issue ownership I have found that the recent general performance theory holds up much better for the Coalition than for Labor. I’ve studied this by looking at survey-to-survey movements in the two parties’ ratings as best to handle and seeing if they move together.

My statistical skills are rather limited, but generally if we compare movements in the Coalition issues of inflation, interest rates, and tax the correlations between them are above .8 (where 0 would mean no statistical relationship, and 1 would mean a perfect statisticial relationship). The exception is immigration, which correlates at around .6, at least partly due to some large movements around Tampa and the Rau/Solon debacles. There are a relatively small number of surveys on the economy, defence and national security.

The same thing can be seen with Liberal movements on Labor issues. They have .8 correlations between their rating movements on welfare and education, industrial relations, the environment, and unemployment.

Correlations with leadership satisfaction are a bit lower but still high, on around .7.

For Labor, there are high correlations on the Liberal economic issues, but on their own issues the correlations are all over the place and no more than .7 and as low as .13. The most common are in the .4 to .5 range.

The numbers needn’t match the Coalition’s (in reverse) because Labor can lose and gain from minor left-leaning parties as well as the Coalition. But based on the general theory that most voters know too little about the issues to make specific judgments, except for the main economic issues that have clear and well-publicised indicators and the occasional very high-profile story, we would still expect most issues to move closely together most of the time.

I don’t really have a good theory as to why Labor’s ratings seem to show so much more issue autonomy. Perhaps the Liberals, despite a long period in government, just don’t leave a strong issue impression on voters. That leaves them very reliant on strong leadership, and in deep trouble when they do not have it.

12 Responses to “Does policy matter for the Liberals?

  • 1
    Dave Bath
    February 20th, 2008 20:34

    > Coalition’s policy change to support signing of the Kyoto protocol has seen it drop 10% to 15% as the party that would best handle the environment.

    This might actually make sense. Maybe this is a judgement that (a) their policy changes are motivated by political perceptions (b) this is too important to be driven by anything other than fact-based decision making (c) any party that does this is unsuited to manage climate change.

    Then again, I may be overestimating the sophistication of respondents to the poll! Just a tad.

    On the other issues, the model I gave is no help.

  • 2
    Spiros
    February 20th, 2008 21:03

    “Perhaps the Liberals, despite a long period in government, just don’t leave a strong issue impression on voters.”

    Perhaps it’s because when they were in government, they adopted a die-in-the-ditch, over-my-dead-body attitude to Kyoto, the apology and WorkChoices, and now they are in opposition, their message is

    “Meh, that was just rhetorc. We never really believed that stuff anyway”.

    No wonder the voters don’t take them seriously.

    Next, they will cave in on Iraq, and the set will be complete.

  • 3
    Pedro S
    February 20th, 2008 21:10

    Did Labor drop all around when they were having leadership problems?

    Also, did the Libs pick up when they started off in office?

  • 4
    Andrew Norton
    February 20th, 2008 21:39

    Pedro – Unfortunately there was no Newspoll issues survey until September 1996; compared to January 1996 Labor were down on most issues but well up on leader satisfaction (the masses like Beazley more than Keating). The Liberals were up, particularly on their own issues. I haven’t gone through Labor’s period of leadership problems, though I can see what looks like the Keating leadership challenge to Hawke causing general downward trends.

    One overall problem is that my data is mostly from the Howard years; no data at all before that for some issues. After a few more years of surveys we can get a better idea of whether Labor does really demonstrate the pattern described.

  • 5
    lorenzo aka erudito
    February 21st, 2008 05:55

    I suspect the Liberals do have less of an ongoing corporate identity on issues than Labor. After all, the Libs have varied in living memory between the “keep things running” Party and the “great big changes” Party, and between “quite different” and “me too”. So they probably need a Leader who signals clearly where they are at a given time.

    Labor is more collectivist in its policy processes, so less variable, and can generally be expected, in the issues it dominates, to be in favour of government delivery of services and, more generally, the progressivist cultural pieties. Hence it’s greater issue autonomy.

  • 6
    Byron
    February 21st, 2008 07:37

    If “issue ownership” can’t be related to electoral performance – what does it matter?

  • 7
    CazZa
    February 21st, 2008 08:56

    3 months in, do issues polls matter? Libs don’t appear to have policies at present but keep on sticking balloons in the air and seeing how the wind blows.

  • 8
    STT
    February 21st, 2008 10:13

    Could it be that the previous Liberal team had a lot more credibility than the reshuffled shadow cabinet?

    What I am getting at is that when people are asked ‘Who could better manage the economy’?, they might well have had a higher opinion of Howard/Costello was than Nelson/Turnbull.

    If there is a perception in the electorate that the best talent in the Liberal party has gone to the back bench (or lost their seats), then it would be expected that their ratings would drop across the board. So it is not so much a matter of parties ‘owning’ issues as people within parties gaining credibility on certain issues.

    Out of government, I suspect that it will be hard for the Liberals to regain credibility – it will be more a matter of the government losing theirs and the Libs gaining ground by default. If there is a recession, expect the Liberals economic rating to improve overnight, with or without policy.

  • 9
    Club Troppo
    February 21st, 2008 10:49

    […] Andrew Norton wonders if policy matters for the Liberals. […]

  • 10
    Winton Bates
    February 21st, 2008 11:21

    Andrew, I wonder whether the difference in correlations might be a government versus opposition phenomenon, rather than a Coalition versus Labor phenomenon. The main opposition party has to try to look sufficiently responsible to be electable as a government. On particular issues it thus tends to be vulnerable at various times to minor parties that can appeal to voter support without having to worry about things like how policy proposals will be funded.

  • 11
    Andrew Norton
    February 21st, 2008 12:43

    Winton – Within the limits of the data (far more years of Labor in opposition than the Liberals), Liberal issue movements correlate even more strongly in opposition than government. That makes sense. People know less about oppositions than governments.

    Yet for Labor the correlations are even weaker in opposition than government. This may be because economic indicators are most likely to move together, so there is an early 1990s recession effect in the data for their government years.

    Byron – The US research suggests that that combination of policy mood and issue ownership does matter, but the can draw on far more polling data. Here I think ownership does matter, but more clearly at state than federal level. At federal level, the complicating factor is the economy. It’s both a threshold issue – parties have to be trusted on the economy to win – and a performance issue, so that objective data influences opinion as well as party stereotypes. Though the Liberals are generally seen as better, Labor was ahead when Newspoll started asking the question in 1989. When the economy is neutralised as an issue, Labor’s ownership of other issues becomes a real asset for them.

  • 12
    2 tanners
    February 21st, 2008 13:27

    Without going into the differentials per se (my stats are pretty sloppy too) I would be thinking that the Libs’ slide is an opposition effect – oppositions are less credible as managers, because they aren’t the manager.

    If you like, it’s almost an extension of the ‘Oppositions don’t win Government, governments lose it’ theory. As such, it may well be a one-off, but a long lasting one.