Let’s say you had $100 and circumstances arose where you were forced to invest that money into one of two propositions:
The first, a risky strategy where you have a 10% probability of getting a $200 return, but where you also have a 90% probability of getting back only half of your initial $100 investment.
The second, a less risky strategy, is also available where you have an 80% probability of getting back $80, and a 20% probability of getting back $70.
To make matters more complicated, let’s say that once every three years you get to buy entry into a special club, but this club membership costs $100. And while you can invest the returns from these wagers over any three year period, club membership (which you really want) only accepts money deriving from these wagers.
Which strategy would you take?
Do you take the risky proposition with the small hope that you can end up with $100, or do you take the option with less risk, probably end up with a higher amount (but still not $100) and then invest this return for three years, so that you’ll have a better chance of buying the membership in three years time?
If we replace dollars with two party preferred vote share, we see the dilemma that the Coalition currently faces.
The more political risk that Turnbull takes to become Prime Minister, the higher the probability becomes that he will succeed – but at the expense of also having a much higher probability that he will fail, and fail in a big way. It’s the probabilities of “just failing” or “staying the same” that get hollowed out as the best and worse cases start to dominate the outcome probabilities.
This might sound a bit of a strange argument, so I built a flash app to visually demonstrate. Let me stress, the numbers in the application aren’t real – they merely exist to demonstrate the relationship between risk and reward in politics. The slider on the left represents “Political Risk” on a scale of 1 to 100, where the higher the number the larger the risk. The left axis represents the probability of any two party preferred result with that amount of risk, and bottom axis is the two party preferred result itself.
The “expected value” is simply the sum of the probabilities of each outcome attached to each two party preferred vote value – effectively telling us that if an infinite number of elections were held with that given risk level, that would be the average result. But as we can see, the average is probably less important than its composition – but it’s still worth watching how it changes with a given level of risk. Again, these numbers are for demonstration purposes and aren’t real (but regardless of the actual probabilities, the same relationships will hold)
As political risk increases, crash or crash through becomes the dominant set of outcomes. We saw that with Latham, Whitlam and Hewson – it rarely ends well.
Similarly, as we’ve seen with Ozcar, it didn’t end particularly well for Turnbull either – but Turnbull has an appetite for risk. Look at the way he went after Rudd on the Hu matter, another risky proposition, straight after getting his clock cleaned with a forged email.
While Turnbull may naturally (and by his historical actions) have a preference for risk, the Liberal party itself is a very risk averse institution. It avoids risks, often at all costs and to the point of sometimes looking irrational from the outside in.
Something will have to give – a risk preferring leader heading up a risk averse institution is a recipe for inevitable conflict. Turnbull will want to roll the dice if he has a 10% probability of becoming PM at the next election, even if that means a 90% probability of losing by such a large amount that it effectively destroys the functionality of the Liberal Party for a decade (as we’ve seen can happen in Qld, NSW and Vic in the States) as swathes of marginal seats – and potential future Liberal leaders in the process – fall like dominoes to Labor.
The Liberal Party has to make a decision on the spread of the two party vote they want to fight for at the election. The safe option is to jump through the hoops when it comes to the rhetoric of victory, but accept defeat and concentrate on saving as many marginals as possible to give them a chance of victory in 2013. The alternative is to let Malcolm loose and probably end up getting smashed – throwing the next 1 or 2 elections as a consequence and ending up like the Liberal Party in NSW and Qld in the early parts of this decade.
The popular theory is that Turnbull is safe and will fight the next election – I reckon that all depends on how many more high risk positions he plays for and loses between now and Christmas, and whether the Liberal Party as an institution sees itself as more important than Turnbull the man.
29 Comments
Interesting analysis but I think it overlooks two issues:
Do Australian’s want a risk taking Prime Minister? I suspect not so (due to the inherent conservatisim in Australians) so constant and obvious high risk strategies may overall reduce the risk of likelihood unless the “success” is so spectactularly bad for Labor (ie kiddie porn links or blatent corruption), the chance of success is probably more like 0.5%.
Secondly, I suspect that the probability of a win that you identify in your spiffy little toy (and it is spiffy) really only apply where the risk is attached to actual policy as demonstrated by the GST elections and Workchoices (although I suspect the Libs didn’t see the latter as being risky when they rammed it through). The chance at the moment of a high risk smear campaign working are so remote that really it has to be a policy that Turnbull pins his hopes to.
I suspect the chance of success from the high risk strategy is really only a chance of taking some of the shine off the Government and not a chance of a win at all. The more times this high risk strategy is tried and fails, the greater the tarnish that attaches to Turnbull and no matter how grimy Rudd ends up looking as a result of the smears, he will still be shiner than Turnbull.
The born-to-rule set (in their own minds, of course) feel so humiliated, so galled, so utterly indignant at being kicked out of government that they would do practically anything to regain their rightful place lording it over “the rest of us”.
I therefore feel they will stick with the high-risk Turnbull, if nothing else desperate for the tens of millions of dollars he’s expected to put into campaign coffers.
If the gamble pays off, they will be happy campers, and set about wreaking revenge on those who tossed them out of power (Labor voters, the Labor Party, unions and “minorities”).
If, as seems more probable, they crash, as opposed to crash through, then that will be just all the more reason to hate that bastard, Rudd, for, won’t it. They will slink off with immense rage and hurt feelings to plot their future revenge against Labor voters, the Labor Party, Mr Rudd, unions, and “minorities”.
In defeat, malice. In victory, revenge.
Nice analysis but what constitutes risky behaviour within this context? Surely, smear and innuendo, tax cuts and Union bashing have always been staple fare for the Libs. So within the Liberal mindset employing such tactics is the non risky option for returning to power quickly.
Like the GFC maybe the Libs have not correctly calculated the true risk of their strategy. However, given they have dominated Federation for 80 of 110 years they would claim they know what they are doing.
Interesting thoughts.
By risk-taking I assume you mean the opposite of the small target policy Rudd used. An approach of trying to change the electorates minds not just give them what they want.
I think Australians want a PM with a big and new vision and set of ideas. The problem is getting 51% of people to agree on what that vision is.
I’m a big fan of your work Possum, but I’m not so sure about this one. As an opinion piece it’s fine, really good actually, Turnbull is a bit of an all or nothing operator and it’s questionable whether that is what the Liberals need at the moment, given that the best bet is as you say to try and hold on to as many seats as possible to build a sustainable and credible alternative government for when the voters inevitable tire of Rudd.
But there is an implication that there is some numerical rigour behind your arguement that is not really the case. You’re not dealing with any real world numbers or statistics here. For those that can follow that maths you present there isn’t any reward, since there is no real data or additional information to be gleaned from it ( I know you make the point that it’s a made up model, but then why present it in the first place?). For those that can’t follow the maths, it just re-inforces the kind of Elizabeth Farrellyesk smugness-in-ignorance that the educated innumerate like to exude.
You make a very good argument and argue it persuasively, it just doesn’t need (in this case) the hocus-pocus re-inforcement of some invented risk model that bears no resemblance to the real world. Leave it for what it is, a well argued opinion.
Poss
Another variable is that the Libs are not a homogeneous lot. In two leadership ballots Turnbull narrowly lost one and narrowly won one.
Crazy brave is tough when you have unanimous support, it is impossible in Turnbull’s position.
They will dump him.
One effect of a sense of entitlement (and thus a belief that the polls are wrong) would presumably to delay a leadership change until reality has sunk in. In other words, well into next year. Otoh, if it happens sooner, maybe that says that maybe more rational minds are at work. Maybe..
The assumption is of course there is someone to replace Turnbull. Someone who the rational number crunches in the Party think is more electable.
And the bit of game theory that Possum describes isn’t just applicable to Turnbull. Any leader of the Liberal Party probably faces the same dilemma. The only thing I have to add is possibly that higher risk strategies are better used sooner and lower risk strategies are better used closer to the election.
In any case I’d love to hear from anyone who can give a decent analysis of what difference a new leader would make to the LP anyhow. My take on it is that so long as the extremists remain, they are going to have real difficulties coming up with a coherent policy platform and one that present day Australia wants.
Personally, I *want* the LP to go down, and stay down, in the nastiest most undignified way. So who would I prefer to have as LP leader to suit my own Machiavellian desires? Possibly Turnbull. The reason being that if any of the ideological nasties like Abott or Minchin get a running, and lose badly, that will probably hasten the departure of the dinosaurs and thus the renewal of the LP. However, if Turnbull stays (and he could quite possibly stay after the next election), these people will linger on hoping to be next.
Even assuming a ‘Machiavellian’ desire for the Libs to fail, there still needs to be an effective opposition. While on balance I think the Rudd Government is doing a reasonable job so far, there are some significant issues that they do need to be held to account on, now and in the next few years. Unfortunately this current opposition aren’t doing that very effectively, instead opposing every issue out of course and grandstanding badly on a select few issues (i.e. simplistic approach to debt and deficit and demanding KRudd don the Batsuit and fly in to rescue Stern Hu…).
The Greens and Xenophon are acting as a much more effective and intelligent opposition at present (no Fielding not you, you’re even worse).
The only member of the Libs I can think of who might be more electable than Turnbull would be Hockey, but he is ‘even worse’ as far as the Dinosaurs in the party go, so that’s not going to happen.
To get back into Government the Liberals need to do all that hard slog, policy reform, renewal etc etc and …time
For Malcolm to get into Government he needs something to happen, something extraordinary otherwise he won’t be leader (a member?) by the time that occurs. Taking risks is the only way he has any chance to turn it around.
As you say, Something will have to give
runawake
I’m beginning to think that he may dump them – no reason to think this other than cutting his losses may appeal to the business man in him. Turnbull must realise that the only reason he is still the Leader is because there is no alternative – I’m not sure that his ego will like the idea that he is really only the leader by default and not by choice.
What I was thinking about the other day was what the cliche “crash or crash through” actually means in practice – but also how it plays out practically as a matter of degrees, as one goes from “crash through” to the “crash” in terms of electoral consequences. That in turn got me thinking about risk – which in turn got me thinking about Ozcar and now this Hu business and how they’re all one and the same thing.
So instead of spending 1000 words to describe the mechanics of what I was thinking, a visualisation would hopefully be more effective.
Everyone here is right in their criticisms – “risk” is too undefined, the numbers are meaningless etc.
But I couldnt think of a better way to describe the consequences of mostly choosing the riskier option in strategy and tactics meetings in the Liberal Party, and how that flows through to probable election outcomes.Politics is always a matter of increments – so a ‘crash through’ style is, in practice, really little more than mostly choosing the riskier political option available at the time.
So what I was attempting to do was focus on that change from one position (always choosing the riskier option) to the other (always choosing the safest option) insofar as likely election results are concerned.
So at maximum risk – “crash or crash through” literally means one of two options – a win against the odds, or a terrible loss.
The opposite – the consequences of always choosing the minimum risk – effectively means both those extreme results are off the table and the actual result will be somewhere in between.
What I probably should have done is just had two explicit values on the bottom axis for the two party results – the best and worse case scenarios of 43 and 52, and then described the others as arbitrary values in between.
What was sticking in the back of my mind here was the question of whether always choosing the risky option (where risky is broadly defined, both in a policy and strategic/tactical political sense) hollows out the potential two party preferred results, leaving it more likely that the extremes will be obtained rather than the spectrum of values between them.
Bogdanovist at number 8 wrote:
The kind of block-headed conservative voter the Liberals are hoping to connect with (or retain, as the case is) understand little of, and care less for, nuance in political and / or economic debate. That’s democracy, for better or worse {sigh}.
Did you consider a Casino analogy, since for most people that’s their most tangible experience of quantified risk/reward? For example:
Imagine that you have 200 $5 chips when you sit down at the Roulette table. Now, you can either put 1 chip down on red for every spin of the wheel – under this scenario, you won’t win big, but the probability is that you’ll still be sitting there a few hours later, with your bankroll depleted a little. Alternatively, you could put 20 chips on lucky number 7 every spin – under this scenario, you’re most likely to bust out pretty quickly, but you *could* walk out in a few spins time with $700.
The problem is one of misaligned incentives. The Liberal party as a whole is aiming to still be at the table for the next few elections – but Malcolm knows he’s being kicked out of the Casino in half an hour anyway, so his incentive is to plonk his chips down now hoping to win big.
As it currently stands what the Libs do is almost irrelevent. The Labour lead is so huge and Rudd so popular that they would need both a series of moral scandals and a major crisis or two in the economy and health system.
The smart younger Lib operators are keeping a low profile until the electorate tires of Kev and his cabinet get complacent and screw up or he has a string of bad luck. Like Kev did, he stayed out of trouble until his party was in the ascendancy. You almost always only get one serious shot at election as leader.
The big issue I can see everything falling apart over is when CPRS is applied to the economy in 2011 (after the next election). The one off hit to prices, GDP, red-tape, messiness and screw-ups of a new system, etc…. will cost the incumbent a couple of %.
A 43% worst case vote is a little too low IMHO. It might make sense statistically but in reality, 45% is more realistic.
The problem for the Liberals is that at a time when they desperately need a two (or three) term strategy and a leader willing to sacrifice their personal ambitions for the long-term good of the party, their leader is focused on himself and willing to bet the house on voters being willing to kick the Rudd Government out after one term (which would require a fairly high level of sustained incompetence which neither Rudd or his ministers have ever demonstrated). To make this situation worse, if the Liberals kick Turnbull out they’re effectively writing off whoever replaces him as they’ll then go on to lose next year.
In response to EnergyPedant, the CPRS is a little more complex than that. Its geared to avoid a sudden increment to prices – rather to spread it out. And it comes with a lot of political-pain-relief in the form of subsidies and direct compensation to low to middle income earners. Personally I don’t like the scheme, but only because it distracts attention from direct investment in alternative energy and its very messy compared to a straight out carbon tax, but that’s for another forum. The point here is the only comparable event was the one off hit to prices incurred by the GST and so far as I can tell that event in and of itself made little real difference to voting intentions.
Hey Poss, any chance you could find a political analogy to the mathematics of “Deal or No Deal” ?
Actually, this is quite an interesting way of thinking about the internal politics of political parties and I think Caf is on to something with the misaligned incentives idea.
I’m going to leave out the numbers because they would just be made up. And keep in mind my analysis is also only partial.
The liberal party dilemma is a classic principal-agent problem. The party (and its key power brokers) are the principals and Malcolm is the agent.
The principal is not just wanting to maximise the risk-weighted probability of winning the next election, but also places some weight (decreasing as it discounts the future) on winning subsequent elections.
The agent also places some weight on future elections (a very close loss for example wouldn’t necessarily mean the party gets rid of him) but discounts the future more heavily than the principal. In addition to weighting the future differently, the principal and agents both have different risk preferences. The principal has to consider additional factors as well. First there is an imperfect information problem – how does it tell whether bad polls are due to the agent or due to other factors? Second, the party has the option of replacing the leader with someone else. And so will replace Malcolm only if an alternative leader, in its assessment, would offer a higher risk-weighted probability of winning the next and subsequent elections. Of course, this is a difficult calculation because it cannot know in advance how an alternative leader will peform (and so will apply some discount to the probability of their winning elections). In the case of the current liberal party it isn’t clear there is an alternative leader that would do much to lift their current standing in the polls – Rudd would wipe the floor with Hockey, Bishop or Abbott….I think this demonstrates the real selfishness of Costello – he is the only leader they have that can minimise the losses and set the party up for a decent tilt at the following election – but he won’t do it because he is too afraid to go down as an electoral loser.
So, really it demonstrates one of the reasons why politics and political economy are so complicated There are information problems everywhere and it isn’t possible for principals to draw up credible contracts for their agents so that the incentives are aligned.
Perhaps the best bet for parties is to be very wary of leaders that have different preferences than the parties they lead. It is a recipe for turmoil.
I think there is then a big difference between Whitlam on the one hand, and Hewson, Turnbull and Latham on the other. The latter three were not really creatures of the culture of their respective parties. In a sense their elevation to the leadership reflected a roll of the dice for parties that were desperate to win, but didn’t properly consider the consequences (inability to control their choices). They were moments of risk loving in otherwise conservative institutions.
Whitlam only appears to be a crash through or crash leader in hindsight on the bad (final electoral) outcomes associated with his leadership. He had the misfortune of governing through an economic shock more profound than what we are currently experiencing. But he wasn’t taking the party in a direction it didn’t want to be led. It had been out of power for a long time and actually wanted to undertake significant reforms. I think it is for that reason that Whitlam remains popular in the party today, despite the decimation in the 1975 election.
Finally possum, if you were interested, you could write down a formal mathematical political economy model that took into some of these factors and then used polling information to simulate the choices and decisions faced by parties and their leaders. You could probably get it published in a decent political economy journal!
The logic here is very sound. The risk of keeping a damaged leader is that he is tempted to use extreme measure in order to regain his standing. There is no way Turnbull can be satisfied merely holding the fort until after the near-inevitable election loss. He’s a proud bloke and will feel demeaned by that expectation. Expect him to take some big risks but not the sort of risks that paint him as risky in the eyes of the electorate, at least if the bets succeed.
His attachment to the party is not deep (a la Tony Abbott) and he will not worry too much about doing lasting damage to the party, except insofar as that damage might affect his own reputation.
A very interesting thread – as always in politics there is always so much incomplete information on which party apparatchiks can act. Another variable is “who is voting for the leader?” – most Senators are fairly safe in their positions, but how many backbenchers in marginal seats are willing to sacrifice themselves for the greater good?
In 1983 Labor took Hawke over Hayden – I still feel Hayden probably would have won, but I think they were so desperate to make it a “slam dunk” they took a calculated risk which in the end probably gave them a better buffer for the next few elections. Before Keating ousted Hawke it seemed Labor would probably lose the next election – were the “swinging votes” in the party room swayed to give Keating a narrow win by party altruism or self-preservation?
Keating “rolled the dice” a bit in the 1993 campaign, but then he was up against another high roller in Hewson!
I often think of an analogy from the film “Flight of the Phoenix” [the original one is better] – Jimmy Stewart [later Dennis Quaid] uses his second last Coffman engine starter to “clean out” the cylinders – a very risky move as then they only had one left. Was Latham a “clean-out” starter – a narrow leadership spill winner (with lots of enemies in their party) burned in a general election to set up a better chance for the next election, and is Turnbull repeating history?
I’m not sure about the Latham example above. It is impossible to know, but my feeling is that the 2004 election was very winnable by Labor, and rather than Latham being the annointed the fall guy for an inevitable loss, I think it was a gamble that didn’t need to be taken. I think the long period in opposition after the Hawke/Keating years made the Labor machine a bit too impatient and they took the gamble on a leader that *might* have been able to give Howard a real walloping (and for a good few months this looked like being the case) instead of a small target approach that had a good chance of seeing them just flop over the line.
I’m not sure Turnbull understands at all that he takes risks, what we are witnessing is someone with uncontrolled arrogance and chronic bad judgment. The commentators think he is crazy brave when in fact he might just be ignorant.
Turnbull may have been able to do well in a time of the global boom but I am beginning to feel that in another economic era he would be a perpetual bankrupt.
During the tension of an election campaign it is possible that this bad judgment will be come magnified and will become one hail Mary pass after another and the arrogance will out itself in spades the more he falls behind.
I agree that Turnbull represents a huge risk to the Liberal Party. There is no way Turnbull can win no matter what he does or what dice he throws, Rudd would have to lose by his own actions. But Turnbull in desperation with desperate act could cost the Liberal Party big time.
The Liberal Party should accept that it can’t win and that is should play it safe with a leader less likely to blow up in their faces.
Bogdanovist…
Labor “rolled the dice” with L*th*m, many thought he would win
The “Campaign” is wot done him in, he failed at this crucial point, simple as that.
Rudd kept it together during (The Campaign) and well…
Enough said
I’m “relaxed and comfortable” about the next election
Following on ‘the campaign’s the thing’ view above, Turnbull , Bishop, and Abbott are all likely to screw up big time under the pressure of an election campaign. Shrek is the most likely aspirant to be able to keep it together during an election.
Anyway you look at it though… the Liberal Party is currently unelectable, regardless of what they do. The only remaining issue is how big a losing margin will they experience at the next election. They are also a write off for the one after next, and at that one, they would need a huge turnaround merely to get a whiff at the one after that.
Really, the next Lib PM is not appearing on our screens at the moment.
I think Turnbull is a risk-taker, but I also think that the Liberal Party (always feel like putting quotes around the ‘Liberal’) is in paroxysms of desperation at the moment because it has been replaced by a government that (despite a few little differences in emphasis) is basically on the same patch of political turf. If the Labor Government had turned out to be left of centre the Liberals could just have waited for middle Australia to come home to them (or at least 51%+ TPP). As it is it’s pretty clear that almost everything that Labor are doing is what a releected Howard Government would have done, which is why nothing the Liberals are doing is having any resonance. They fear being displaced from their own niche.
calypto – it’s like V8 Supercars – the only differences are the badges, a bit of bodywork, and the rear wing. Underneath the blueprint is identical.
Bogdanovist…
Yes, I didn’t really mean Labor insiders did the Latham thing deliberately, but there was this “Latham Problem” in the party until he got made leader, and it would only go away when he contested an election as a leader, win or lose.
People too young to remember would not see that the Hawke situation was not dissimilar – he was a bit of a destabilising force, and if he had grabbed the leadership and lost (seems ridiculous now, but not to those who had seen 1980) he would have been totally finished, as was Latham.
What can you do to get back quickly – this is a question I have been thinking about for a few years. Paradoxically a big loss or even two may be part of the answer. Labor 1975-1977, Vic Labor 1992-1996. The first loss is bitter but the second loss forces you to really take stock – on personnel, policy, and leadership. Conversely, the “unusual” narrow loss for Federal Labor in 1998 was probably the worst result for them in the long run, as was Peacock’s effort for the Coalition in 1984.
I wonder if there is a belief among some Liberals supporting the Turnbull leadership that the risk is not really that great and that no matter how incompetent, confused or rash they are there must be a Narrowing based on the great Orstrayan love of all underdogs, and they won’t really get trowelled worse than, say, 54-46.