Rates of Gloom
Posted by Possum Comitatus on October 25, 2007
It’s equation time!
So how important is the cash rate on the Oppositions primary vote?
We’ve modelled this before over here in terms of the effect a rate change has on the primary votes of major parties and how long it takes for that effect to flow through, but today we’ll take a different approach entirely.
What we’ll do today is make a simple model that can explain the effect (if any) the cash rate level has on the ALP primary vote by taking into account the ALP leadership and any linear time trends that may be floating around in the mix.
So to start with we’ll be using the period from 1999 through to the present, and the basic model we’ll be looking at is:
Where:
ALP = the ALP primary vote at time t (where t is the month).
Cash Rate = the RBA cash rate at time t.
Time = a time variable that has a value of 1 in January 1999, 2 in February 1999 etc right through to its current value of 106 in October 2007.
DummyRudd = dummy variable which has a value of 1 for the periods of the Rudd leadership and zero for all other times.
DummyLatham = dummy variable which has a value of 1 for the periods of the Latham leadership and zero for all other times.
So what we are modelling here is how value of the cash rate explains the level of the Oppositions primary vote, but where we also account for the two leadership periods of Latham and Rudd having an effect, as well as allowing for a linear trend to manifest in the vote.
Once we run the regression model, our results come out as:
Which, for the non-geeks can be visualised as this:
[Note: I’ve ignored the autoregressive nature of the polling series so that the results can be better visualised for the non math-geeks here]
The “Fitted Primary Vote” in the graph is our modelled vote based on the regression equation.
So what to make of it?
For every 1 point increase in the cash rate, the ALP primary vote increases by 3.5 points. So if rates were to rise by 0.25 in November, the ALP primary would be expected to rise by 0.9 points, or around 1 point.
It also suggests that the Rudd leadership (or what has happened during the period of the Rudd leadership) has boosted the ALP primary vote by around 8 points, and that Latham boosted the ALP primary vote by about 2.5 points (which is fair enough considering that Latham took over from the dire polling position Crean had created).
All up, our variables explain about 75% of the movement in the ALP primary vote over the period 1999 through to the present.
If we now run the equation again, but this time using the ALP two party preferred vote, remove the time trend variable and use the period from December 2002 through to today (simply because my TPP figures before that are in a database that I’ve stuffed up and need to rebuild 😉 ) we get:
This gives us a visualised model of:
Again, our variables here explain about 75% of the movement in the ALP two-party preferred vote, the cash rate has nearly identical influence on the TPP here as it did to the primary vote in the first equation, Latham again lifted the TPP vote about the same as he lifted the primary, but this time the Rudd leadership has increased the TPP vote only around 3.5%. This is expected as remember the ALP TPP had been growing constantly since the last election, it was their primary vote that got the big boost under Rudd.
So all up, if rates rise by 0.25 points, the ALP can expect their primary vote to increase by about a point, and their two party preferred vote to do likewise.
Possums Pollytics 
pligg.com said
Rates of Gloom « Possums Pollytics
It’s equation time!
So how important is the cash rate on the Oppositions primary vote?
We’ve modelled this before over here in terms of the effect a rate change has on the primary votes of major parties and how long it takes for that effect to fl…
Mark said
Whoah — I’m a bit rusty, but have you fallen into the “dummy variable trap” …
http://en.wikipedia.org/wiki/Dummy_variable
i.e. you can’t have the Rudd dummy PLUS the Latham dummy PLUS the constant intercept term …
Mark said
Sorry, ignore that, I’ve just realised that you are using a period that takes in Beazley and Crean as well. Mea culpa — the Possum is correct!
Possum Comitatus said
Dont forget Beazley is the baseline.
wantok said
Both amazingly good fits, Possum, well done. (Though it’s somewhat depressing to think that the electorate could be that simple-minded and predictable… climate change recalcitrance, WorkChoices, a million civilian deaths in Iraq, and so on make so little difference?)
Possum Comitatus said
Wantok, I’m sure those things all make a difference.We’ve modelled the effect of Workchoices before, for instance, and found the effect quite profound.
It’s just that a lot of things separately can explain a given long term poll movement, so collectively they all probably play their part.
John V K said
Ya basically the step function, of consumer spend and income recalibrated as a flow on disposables. Interest rates impact discretionary income sigificantly, the big argument on economics is not the hip pocket arguments that the punter has to factor with competing desires. So in effect the emotics have been wrong, here is a once a year perk and the interest rate nag continues from time to time.
Like my missus says when I get the tax returns done and speak about household macro economics how much of a refund are we getting (which tranlsates in her mind how much can I sneak off that tight wad miser bastard normally with great success may I add).
If Rudd wins the election and has any nouse he will be doing everything in his power to stop interest rates rising thru infationary pressure. I believe Saul Easlake has given advice as no doubt the great macro economist will very shortly after the election
Great stuff Poss I stand in postive awe of your prowess and will petition Green peace immediately for a moratorium on possum scalping and possum rib barbecues at school fetes across the wide brown land.
Cheap talk dont pay the mortgage or get the kids teeth fixed.
John V K said
Excuse me the big argument on macro economics is not the hip pocket arguments.
Welcome to the Coaltion nightmare.
Jessica said
Possum – what do you make of the Galaxy poll on the senate suggesting the Libs are on 38% and Labor 33%?
Could it really be true – will the Libs hold the senate?
Possum Comitatus said
Jessica, I think Senate polls are not completely worthless, but so close as to make the distinction a matter of semantics.
Senate polls wont be even marginally valuable until we know the result of preference deals much later in the campaign.
Grant Williamson said
All these old-school p-values…how about doing some multi-model inference / information theory analysis on this? What’s the most _important_ variable, interest rate or leadership?
Doug said
National Senate polls are worse than useless.ignore them they are not worth the paper they are written on
To be useful senate polls need to be taken a state/territory at a time – probably close to the election.
Then you have to take account of preference deals and even so you can get the sixth seat wrong. Tasmania at the last election – everyone was saying that Christine Milne would fail to get the final seat because of the preference deals – but those pesky below the line voting Tasmanians proved them wrong.
Andos the Great said
Hey Possum,
What sort of time period does it take for rate rises to cause the quoted rise in primary vote?
Jessica said
Thanks for that Possum and Doug… I wonder whether given Getup’s political leanings they have released these figures now with in order to scare people away from voting for the Government in the senate??
jamesj said
Not a statistician, economist or psephologist. Just a mug, but Labor’s 2PP is at heights never seen before – and just four weeks out from the election. Surely any real-world model can’t continue to show similar increases from their current position. I’ll gladly take a one point rise, though.
John V K said
There is one point and I’m not psephologist. Why is there a so called ceiling in voter positions. Like Antony’s green election calculator has a ceiling of 60% and previous commentator James statement about highs.
My thinking that say Adolph Hitler if he had survived the bunker resignation syndrome, would have had popularity ratings in the first 10. His opposition would be 90%.
Or is it just the nation has never been in this country in our history?
Don Wigan said
I’m too long removed from my basic stats knowledge, and too lazy to work out what you’re saying… but I’m happy to take your word for it. Adding another point to Labor’s primary sounds just the ticket!
As the then 90+yo George Burns responded to a gorgeous blonde apologising for brushing against him, “Not at all! You’ve just made an old man very happy!”
And your graphs are pretty.
Keep it up Poss!
caf said
Possum,
What happens if you put CashRate_t-n into the formula instead, repeating the regression for various values of n? My feeling is you would find the correlation peaking at some positive value of n indicating the typical lag time between a money market cash rate rise and the effect on people’s votes. This should be a more accurate model (possibly not much more accurate, though).
Also, a linear relationship between Cash Rate and voting intention seems unsupportable, at least if the Cash Rate becomes very large.
Possum Comitatus said
Caf,
A one period lag on the cash rate lifts the coefficient from 3.59 to 3.66. Lags greater than one period quickly decrease the explanatory power and the size of the coefficient.That’s consistent with other stuff we’ve done around here measuring the time it takes for interest rate changes to show up in the polls.Normally around a month.
The cash rate these days really operates through a limited window under normal circumstances.If inflation got to the point where the cash rate was heading past 10%, fiscal action and various legislative responses would be the order of the day one would think. Under those types of extreme circumstances, these sorts of relationships to the primary vote would simply breakdown.
Enemy Combatant said
“So all up, if rates rise by 0.25 points, the ALP can expect their primary vote to increase by about a point, and their two party preferred vote to do likewise.”
PC, your ongoing kindness to the mathematically challenged is noted with apprecition. I’ll be guided by you then, on the accuracy of Akaike and the precision of Schwarz.
———————————————————-
The things a bloke notices when he’s recently re-read Orwell’s essay, “Politics and the English Language”…..
Below is fresh from ABC online news homepage:
“Union boss Joe McDonald has been found not guilty of trespassing on a construction site in the Perth suburb of Joondalup”
Union boss ?!? I beg your pardon!? Or as one notes with increasing frequency on more libertarian weblog sites these days, “WTF!?!.”
Remember the days when OUR National Broadcaster didn’t reflexly adopt government slogans, the days when workplace representatives were unbiasedly referred to on the ABC as union “leaders”; the days before El Rodente’s minions, Ronnie Brunton, Janet Albrechtson et al. were slotted into Auntie’s boardroom?
The use of “union boss” in that context is about as “fair and balanced” as the yelp of a neo-conservative fox. Who frames the language, controls the message.
To stuff your mind into their mould, the bastards wear you down slowly. High rotation, in-your-face, psychobabble. They are as relentless as a vengeful pack of desert dingos. First they dog whistle, then they fox yelp. Before you know it, you’re double-thinking and doubleplusgood duck-speaking with the best of ’em.
More widely, the style of propaganda eg. Uber-Patriotism and Old Time Religion that worked well enough to get Bush The Imbecile twice elected, doesn’t seem to cut the mustard Downunder. This is to our eternal credit as a nation. Notice how quickly Team Rodent dropped the hot scone that was “Aspirational Nationalism”? It was all the rage a month ago with the Crosby Textor primed, El Rodente, but havn’t heard him or any of his Rat Army subordinates squeak of it for weeks.
(thanks to Uncles: Duke, Frank and Eric for the memories)
El Nino said
wantok @ 5 – this does not predict individual actions, but collective action (i.e. the electorate). Individually, we are all wonderfully unpredictable creatures. We don’t all respond ‘simple-mindedly’ to change in the cash rate, but at any given cash rate change enough of us respond to show a political effect. And it not necessarily the same individuals responding each time.
El Nino said
What Poss has unearthed on here is a mathematical expression of Jung’s ‘collective unconscious’.
Beach Ball said
All I know is that all this maths is hurting my brain!
Perhaps Poss, it would just be easier to say the government is totally F&#*@$.
PoliticalHeadkicker said
Pos, if all this means the Howard’s going to get his ass kicked even harder if interest rates go up again, then I love it.
Neil Cammack said
Enemy Combatant @21 – this isn’t the first time ABC News has used the term “union boss”. I was tempted to fire off an email last time I heard it, and should have, but memories of the ABC’s responses to previous complaints put me off. Invariably they’re defensive and unapologetic, no matter how red-handed you catch them.
I don’t know whether this is the product of an “if you’re not with us you’re against us” bunker mentality, or outrage at the very notion of criticism is part of the corporate culture there. The “union boss” thing may be just the product of lazy, unthinking journalism.
El Nino said
PoliticalHeadkicker, don’t confuse correlation with causality (but you are probably on the money).
feral sparrowhawk said
While I find this uplifting in the news that Howard’s in even more trouble than he seemed, I was also depressed by the idea that interest rates are so dominant in explaining people’s voting behaviour, rather than issues like climate change, Iraq etc.
However, it struck me that the shift in the primary vote is amongst the people who are currently swinging voters. For them rates may be more important than these other issues. It doesn’t mean that the same is true for everyone else. It could be that there are a lot of voters relatively uninterested by interest rates, but that they have already locked on to one party or another over other things.
Come a Rudd government, some of these may shift for all sorts of reasons, and rates may no longer be the most important aspect.
Harmless Cud Chewer said
This reminds me of the standard model of particle physics. If you throw in enough independent variables you can model just about anything 🙂 But does it have any physical meaning?
I don’t doubt interest rates do have an effect on voting intention, but the model you’d need would probably need (at least) two separate parts.
The first part would be modeling the immediate personal financial shock on the part of those actually sensitive to it.
The second part would be more delayed and indirect.
Call it the ‘do I trust this government?’ factor. The delay being mostly due to the effect of social networking. A *much* nastier model 🙂
By the way. Did you consider modeling the time dependent factor as a quadratic?
Lots of fun.. 🙂
-moo-
FB said
agreed HCC@28, introduce enough variables and you can model anything. There’s a challenge poss, point for point agreement!
PS: I always thought it was ‘maths’ not ‘math’ in oz.
Mercurius said
Heh. “DummyLatham”, hehe, “DummyRudd”, heh heh heh.
Who said stats is dull?
Kramer said
Pos, I wish I had an actuarial qualification of some description, because I find that pretty difficult to understand. However, it is pretty clear that there is going to be a direct correlation between Rudd’s results in the polls and a likely rate rise in November.
I don’t have a mortgage and it is going to hurt the Howard government, so it doesn’t affect me that much. They dug their own hole by politicizing the RBA’s role in 2004. They deserve everything they get.
On another matter, it looks like another one of Howard’s stalwarts has thrown in the towel:
http://www.theaustralian.news.com.au/story/0,25197,22642864-7583,00.html
I would not have believed that this article was legitimate 6 months ago. Now, it’s clear that mutiny is afoot within the grand MSM navy (The Australian) and their is wide spread dissension in le grande armee’s ranks (The Daily Telegraph, The Herald Sun).
The war is almost over and Napoleon (Howard) is preparing his troops for one last, desperate offensive. I can hear the ABBA song now…
Kramer said
Whoops, my mistake. All he is doing is bagging Labor and the ABC. Thanks for restoring my faith Greg!
Sorry if this was off-topic!
Dangerous said
Wow – that’s a real dummy spit of an article!
George said
Kramer, yeah, I read it and thought “what the heck are you on about” – just another Greg “I’m-living-on-another-planet” Sheridan “analysis” (and I use the term loosely!). But yeah, at least he’s capitulating.
I was also reading an article in the Age by Ross Gittins http://www.theage.com.au/articles/2007/10/23/1192941062380.html about the possibility of Rudd being another Howard.
The two articles together make for interesting thoughts on democracy in this country.
To me, the point is this: in some western democracies today there is no real choice when it comes to the major political parties. Look at the US, Canada, Australia, etc. The me-tooism hasn’t been invented by Rudd. It’s been embraced by oppositions throughout the globe regardless of their political bent (whether centre-leftist are in government or the centre-right).
A true democracy, in practice, is one that can meet the needs of almost every individual. We don’t have this when political parties pitch for the vote of a specific percentage of the population. True democracy is not when 50% of the primary vote creates a landslide, puts the control of the senate into one political ideology/view and gives that party the right to call on “a mandate”. A mandate to govern for the whole country when only half the population votes for you? Amazing! Even more disturbing, politics in Australia is based entirely on adversarial tactics. Instead, in a real democracy we should be striving for consensus and collaboration.
I think the Liberal party will be annihilated in this election (no surprise there, and Sheridan in his dummy spit basically admits this). The 2PP will deliver Labor a large majority and if not now, very soon control of the senate. Labor will be in control of the whole country, Federal, State, etc. If all views within the Labor party are given a voice and a truly democratic form of government ensues, then it could be a whole new direction for Australia.
If however the Labor party in government are anything like the Liberals have been for the past 11 years (and in particular the last 5), then what will eventuate is what has happened in many European democracies (like the German Bundesta). No one political party takes power; instead, many political parties share power, outstanding individuals are pooled from each party for the good of all citizens, and the running of the economy (so obsessed over by some in the West) is relegated to real economists, academics, policy makers, etc. outside of the influence of political agendas. Political parties like the Greens will become a true alternative force and it may be that we the people no longer accept this ridiculous system of preferential votes and marginal seats.
In the end, Rudd may be what-you-see-is-what-you-get. But those beyond Rudd (and Howard) know that what both sides offer is a fine tuning of a system that is barely ever altered substantially. It is this fine tuning that many people are waking up to – we need a new mechanism of government. Either Rudd/Labor will have the vision to embark on such a road, or their inaction (and ultimate me-toosim) will force the electorate to reconsider their “rusted on” mentality.
Whatever happens, this time around, the electorate is rejecting Howard and his ideology, or at the least that he has gone too far. Either way, rightly so. Rudd knows this – he is not winning because of any grand new vision and direction, but because he is promising to go forward in a softly-softly approach in some people’s eyes, and in others’ eyes he is seen as someone that may eventually offer real change. This doesn’t mean he is being embraced – in fact he is being sent a warning, especially by Labor supporters: “we are rejecting Howard as are many who have not voted for Labor in a long time, so don’t think we want more of the same”. Will he listen? Will Rudd the “visionary” and “social democrat” stand up, or will it be Howard Mark II? Time will tell.
And this is where Gittins’ article brings up a scary scenario – for those who are looking for true social justice, the prospect of a me-too Labor party in power will mean a long time coming before true democracy is achieved.
In the end I reject the idea that Rudd could possibly lie within the same dark cesspool that Howard has inhabited for the past 11 years, but like the dark side of the force, power can make you do strange things. I just hope the social democrat in Rudd will ultimately prevail. Bring on Nov-24 already!
Tim said
I am optimistic about the coming Rudd Government because he seems to understand two things. The first is how to get elected. The second is that we have an information economy – hence is emphasis on broadband, education, skills and the costs of Federal/State turf wars.
So, I think we can look forward to having a very clever and imaginative control freak as prime minister compared to an even bigger control freak who had non-existent creativity in combination with limited (though not under average) intellectual skills.
CL de Footscray said
Thanks Possum, most interesting. The relationship is clear, and reasonably explicable, even to the statistically challenged (like moi). The change voting intention is due to those who swing, presumably and it seems entirely plausible that their votes are determined in very large part (75% or so) by the hip pocket. The rest is all the other stuff. This just means Howard’s REALLY f*@ked.
George said
Tim, I hope so!
Enemy Combatant said
Neil Cammack, agreed; complaining directly to the ABC is a complete waste of time. Maybe with the change of “cultural emphasis” that a fresh govt. will bring, Auntie’s editorial policy as well as her “complaint booths” will be more empathically personned.
Roger said
I don’t personally want to get too deeply into the analysis – but appreciate the fact that others do. All I want is for the rodent and his claque (Downer, Abbott, Ruddock etc.) to be seriously humiliated. They have spent nearly 12 years pushing a radical and self centred agenda that has damaged a large number of Australians – many of whom are unable to retaliate. From my perspective this deserves all the humiliation that we can dish up. Roll on the revolution.
swio said
Imagine if Rudd had held off on his tax policy until after today? He could have promised half or less the tax cuts of Howard and then take the moral high ground saying that any more than that would put too much pressure on inflation and interest rates. After a week and a half of hearing about tax cuts it would have thrown Howard’s economic management credibility under a bus and as a bonus finally created a point of difference between Howard. Oh well, maybe they’ll do better next time.
estate5 said
Hi all…
Good to see the boots out for Sheridan. What a gutless turncoat, but at least more principled than Alb-reghchhgchghht-sen. (make the huck-a-loogie noise when you read that for super fun at home).
My take on Sheridan: a premiere whiner, shoring up his position to be able to say “I saw this coming”. My advice to him: Go start a right wing journal yourself, you nonce, rather than complaining about the Griffith Review all the time.
My take on culture wars: the very term is proof of the ideologue nature of the neo-con Liberals. The current electorate is rejecting overt financial ideology politics, as sure as a previous one did to Chifley and bank nationalization in 1949, or to Whitlam and his crash-through-or-crash Khemlani nonsense in 1975.
The fundamental failure in Coalition ranks? Where to start… Primarily, a lack of intellectual firepower, caused by ideological inbreeding over the last 25 years. Darwinian overspecialization leading to inability to think outside a box. As example, consider the brilliance of Joe Hockey’s mind. You get the point. I rant on all these points at length elsewhere, but that’s the meat of it.
My take on poss’ analysis: as ever, to the point. Go team!
Incidentally, swio above: just you watch for a “Costello Black Hole” to be found lurking in the Treasury accounts. I’m sure someone’s able to whip up the right numbers to justify a Hawke-like rejection of all promises, given a “failure” of the former Treasurer (as was done to Howard). This would be a failure made more believable by recent rate rises… This “failure” might, for example, be hidden in ceasing to accurately and separately report transactions in and out of the consolidated fund for the past few years of Howard rule *smirk* For amusement, check out…
http://www.theage.com.au/articles/2003/07/30/1059480404533.html
adam said
ps… estate5 just above is actually me, poss, logged in under blog name. whoops.
mark-sydney said
Hi Possum
Not sure if you’re able to do anything about this but I’ve noticed that your RSS Feed isn’t currently updating…..on my PC it’s still showing ‘The Dog Ate My Narrowing’ as your number 1 post
Just letting you know
cheers
KC said
Adam
The interesting point in the Age article is the $58 billion in Govt debt still outstanding under the Howard government.
When you take into account the $9 billion budget deficit that Howard left labor in 83 which equals about $30 billion in todays terms then the libs have actually retired about $2 billion in net debt over 11 years.
Not much when you consider the amount they have cut in health, education housing and social services. Where has all the money gone and what have they been doing over the past 11 years.
Bruce said
Hi Possum,
Warms the cockles of my heart… to see some regression analysis that is.
However, I have a few issues with the first model.
Firstly, I’m amazed you’ve fitted 5 parameters to 106 points and got p<=0.0001 on all of the estimates.
Secondly, the first model predicts a long term decline in Labor Primary vote of -0.075% per month. Over 105 months this is nearly -7.9%. This seems implausible.
In the 105 months, the cash rate has gone from 4.75% to 6.5% (if I’m reading the graphs correctly). This incresae in rates of 1.75% when multiplied by the rate coefficient of 3.6 gives +6.3%, not quite countering Labor’s long term decline. So the first model relies overwhelmingly on the Rudd-factor to overcome this difference (-1.6%) and give the current high figure.
The Rudd effect in the 2nd model seems more plausible.
Possum Comitatus said
The time variable is negative because I needed to include something that could mimic the honeymoon effect of both the leadership dummy variables and the interest rate rises themselves. Now I could have used an autoregressive approach with that, but I deliberately ignored the autoregressive nature of the polling data to try and make it easier to follow for the non-stats people.
So the time trend variable, with it’s negative coefficient substitutes well for the purposes intended.
This isnt a model of the opposition vote in entirety (I’m not sure such a thing could actually be created), it’s a simple model to highlight the relationship between interest rates and the ALP vote.
A proper model of the ALP vote would use PM dissatisfaction as that accounts for interest rate rises well.
The Rudd factor in the primary vote model is actually representative of reality.The ALP primary was going nowhere until Rudd gained the leadership.The increase in the ALP TPP since the 2004 election was happening as a result of Coalition voters moving across to the minor parties and giving the ALP preferences.If you click on the Swinging Voter Movement link in the sidebar on the right, you see how that played out visually.
Rudd boosted the TPP for the ALP, but by nowhere as much as he boosted the ALP primary.
Possum Comitatus said
Bruce, let me also add, on the p-values, I normally wouldn’t add 5 variables to that few observations, but there’s a few things in the polling data that have such rock solid relationships to the vote series of the parties that it makes it wash out OK. Leadership changes and interest rates for the ALP, workchoices, interest rates, the GST and the One Nation effect for the Coalition all have large, profound effects. Their significance overpowers the loss of parsimony when dealing with a not large, but not insubstantial number of observations.
Bruce said
I see you point about model complexity. To me it needs an autoregressive (perhaps integrated) moving average model with a decaying step change to deal with the honey moon effect of leadership changes. This would be a minor statistical project and would even have many statisticians wondering what was going on.
El Nino said
Possum,
Is the t statistic in the primary vote regression telling us that the cash rate was more important in driving up the ALP primary vote under Latham and Rudd?
If so, would that explain why economy scare tactics aren’t working as well this time (i.e this time there is more than just the cash rate driving high primary vote).
El Nino said
[Is the t statistic in the primary vote regression telling us that the cash rate was more important in driving up the ALP primary vote under Latham and Rudd? ]
That should read “driving up the ALP primary vote under Latham than Rudd”
Oops.
this said
this webpage
pcmaggio said
Just came across this website showing various ways to make some extra cash, deepwebsites.org. Anyone tried any of these? These look interesting.
butterwings1 said
Just found this website offering various ways to make some extra cash, deepwebsites.org. Anyone tried any of these? They look interesting.