I'm going to do this for 2004 simply because there's two useful government reports available for that year: The National Drug Strategy Report on The costs of Tobacco, alcohol and illicit drug abuse and The ABS's report on Tobacco Smoking.

The Facts:

• Estimated net health costs of Tobacco: $318.4 million (National Drug Strategy)
• Estimated net productivity costs of Tobacco: $8,009.1 million (National Drug Strategy)
• Revenue from Tobacco Excise 2004: $5,237 million (Federal Budget)
• Tobacco Excise per cigarette (Feb 2005): $0.22621 (Tobacco in Australia) [This is currently significantly higher]
• Number of smokers: ~3.5 million (ABS)
• Percentage of adults who smoke: 23% (ABS)
• Percentage of adults who are ex-smokers: 30% (ABS)
• Percentage of adults who are not smokers: 47% (ABS)
• Decrease in life expectancy due to smoking: 10 years (BBC)
• Life Expectancy: 81.2 years (Wikipedia)
• Cost of all aged pensions (2003-2004): $19 billion(APA)
• Number of aged pensioners (2003-2004): 1,866,000 (APA)
• Total number of people 65 and over (2006): 2,644,374 (2006 Census) [this will be slightly higher than the 2004 numbers which I use for most of this article, but the difference would be minor and should not affect the results significantly]

The Math

• Net Health Costs per smoker per year: $90 ($318.4/3.5)
• Net productivity losses per smoker per year: $2,288 ($8,009.1/3.5)
• Tax revenue per smoker per year: $1,496 ($5,237/3.5) [not counting GST or company tax on tobacco companies]
• Portion people on pension: .71 (1,866,000/2,644,374)
• Pension cost per pensioner per year: $10,182 ($19,000,000,000/1,866,000)
• Cost of ten years pension: $101,820 ($10,182*10)
• Average pension cost of extra ten years extra life: $72,292 ($101,820*.71)
• Approximate Life expectancy of smokers, ex-smokers, and non-smokers: 75, 80, and 85 years ([.23*75]+[.30*80]+[.47*85]=81.2) [This makes the assumption that smoking for a while and then quiting only decreases your life expectancy by half that of smoking for the rest of your life. Also I should have used age at death in the following analysis, but life expectancy was easier to find, and 81.2 results in nice round figures. Since age at death would have been lower this will result in a slight bias towards higher tax rates in the following analysis]
• Average life expectancy for a 20 year old who smokes for the rest of their life: 55 years (75-20)
• Total health costs over 55 years: $4,950 ($90*55)
• Total productivity losses over 55 years: $125,840 ($2,288*55)
• Total Tax revenue over 55 years: $82,280 ($1,496*55)

Estimates of net costs

There's three ways I can see to calculate the net of the costs, and the corresponding "correct" Tobacco Excise:

Simplistic

Simply take the net health costs per a smoker of $90 a year and use that. This has the advantage of being easy to work out, and I expect most people would consider it reasonable (until they found out that the resulting tax rate is a lot smaller than the current rate). "Correct" tax rate: $0.0136 per cigarette (1,496/90=0.22621/0.0136).

Aggressively anti-smoking

Pretend the entirety of the productivity losses are a burden on the tax-payer in the form of welfare (which is obviously wrong, many of the productivity losses are borne by the individual in question, but some portion will be borne by the tax-payer). Since we are including the welfare costs in this, we need to include welfare savings as well (namely the aged pension). Net life-time cost to taxpayer: $58,498 ($4,950+$125,840-$72,292); "correct" tax rate: $0.1608 per cigarette (82,280/58,498=0.22621/0.1608).

Aggressively pro-smoking

Pretend that none of the productivity losses are a burden on the tax-payer in the form of welfare (which is obviously wrong, many of the productivity costs are borne by the individual in question, but some portion will be borne by the tax-payer), but still include the aged pension. Net life-time cost to taxpayer: -$67,342 ($4,950-$72,292); "correct" tax rate: -$0.1851 per cigarette (82,280/-67,342=0.22621/-0.1851).

Conclusion

Obviously both the aggressively anti- and pro-smoking calculations are wrong, but they do provide some outer-bounds for the tobacco tariff (or subsidy) that can be justified by the "expense to the tax-payer" excuse. I suspect that the anti-smoking result is closer to the truth, and would oppose a subsidy regardless, but more information would be needed to confirm this.

I don't want to discuss the merits of this decision (I'm hardly qualified to make a judgement on that even if I did know all the facts), but rather I wish to point out that such decisions must be made in any healthcare system. In a user pays system the decision ultimately rests with the user themselves, while in a government run system the government must appoint a body or an individual to make these decisions. (In an insurance driven system such as in the US the conditions on when to make such decisions are established as part of the insurance contract)

Such decisions must be made because "the best healthcare money can buy" would cost the entire economic output of a nation to pay for. No matter how much you do there is always something more that could be done to improve the odds of survival or the comfort of the patient some small amount. As a result there must by rationing, which is where this and similar treat-or-not decisions come in. Someone has to make the decision, and such decisions must take into account budget limitations, or they will simply over-spend now and lack resources for future situations.

In any publish healthcare system people will die as the result of a government employee's decision to limit their treatment (either by cutting it off entirely, or restricting it to only certain procedures). Anyone who claims otherwise is either ignorant or lying to you (and perhaps to themself). (Note that in an insurance based system like in the US such decisions are still made, the difference being that the conditions for the decision are set as part of a contract rather than either via law or simple ad-hoc)

The fact that so many supporters of public healthcare regularly claim that such things do not happen tells me that they are either liars, or that they don't even realise that such decisions are a necessary evil, which means that they are not even close to qualified to design or run any sort of healthcare system.

Of course none of this gives a pass to the people who deliberately use a very emotionally loaded phrase ("death panels") rather than a simpler, less emotionally charged term, but when it comes to people who use emotionally manipulative phrases versus people who are blatantly lying to you; the ones that are actually telling the the truth come off as a little more trustworthy.

*. "Death panels" is a rather emotionally manipulative name for such things (Adult Guardian is just as bad in the opposite directions though), a better term would be something like "Treatment Decision Panel".

The Social Compact

The most important part of The Social Contract is Chapter VI of the first book, since this chapter is where Rousseau explains the conditions of the "Social Compact" itself, at least as he sees it. All the previous chapters build up to this chapter and all the following chapters build upon it. If the arguments presented in this chapter are in error the book as a whole collapses and becomes purely an interesting intellectual curiosity (there are some sections that still maintain some value in isolation). Unfortunately I would argue that Rousseau's understanding, as expressed in this chapter, is in error, and such severe error as to be un-salvageable.

First let us examine Rosseau's own words (emphasis mine):

The clauses of this contract are so determined by the nature of the act that the slightest modification would make them vain and ineffective; so that, although they have perhaps never been formally set forth, they are everywhere the same and everywhere tacitly admitted and recognised, until, on the violation of the social compact, each regains his original rights and resumes his natural liberty, while losing the conventional liberty in favour of which he renounced it.

These clauses, properly understood, may be reduced to one - the total alienation of each associate, together with all his rights, to the whole community; for, in the first place, as each gives himself absolutely, the conditions are the same for all; and, this being so, no one has any interest in making them burdensome to others.

And as to why this must be so:

... if the individuals retained certain rights, as there would be no common superior to decide between them and the public, each, being on one point his own judge, would ask to be so on all; the state of nature would thus continue, and the association would necessarily become inoperative or tyrannical.

What man would give every part of himself to society; not just his body and his efforts, but even his opinions, since Rosseau claims that men judging between themselves and society would cause the contract to become inoperative? It seems that only an absurdly trusting man would join such an association willingly and that most men would only join such such an association by force. As Rousseau himself argues earlier (in Chapter III) force does not make right, with the result that such a contract, created by force, would be invalid.

I doubt that men sufficiently trusting and naive to join such an association exist even in small numbers, let alone the numbers necessary to form a society, but for the sake of argument let us assume that they do. Would not at some point these people judge the actions of society (it does not seem to matter whether they judge them as right or wrong in Rousseau's opinion), or at least their descendants? At which point society either ceases to exist or reverts to tyranny.

It seems extraordinarily unlikely that such a society would come into existence or last for any length of time if it did. Rousseau's writings on how to best structure society are consequently irrelevant to any real society, and are only of interest as a thought experiment.

The Marks of a Good Government

As another example in a less important way Rousseau also errs in his discussion of the best way to measure the whether a government is good or ill:

But if it is asked by what sign we may know that a given people is well or ill governed, that is another matter, and the question, being one of fact, admits of an answer.

...

For my part, I am continually astonished that a mark so simple is not recognised, or that men are of so bad faith as not to admit it. What is the end of political association? The preservation and prosperity of its members. And what is the surest mark of their preservation and prosperity? Their numbers and population. Seek then nowhere else this mark that is in dispute. The rest being equal, the government under which, without external aids, without naturalisation or colonies, the citizens increase and multiply most, is beyond question the best. The government under which a people wanes and diminishes is the worst. Calculators, it is left for you to count, to measure, to compare.

This is obviously absurd. For example using the birth and death rates listed on Wikipedia the best government in the world is Mali and the worst is Ukraine, China is somewhat worse than Australia or the US (which are about the same), but significantly better than Canada, and Mexico is significantly better than the US, even though huge numbers of Mexicans risk their lives every year to try to get from Mexico to the US. While it probably made a bit more sense in the time's prior to birth-control, it still doesn't work. A government could simple require each woman who wasn't currently pregnant or nursing a child to visit the local "love shack" once a week. This could easily double the actual birth rate of a nation, yet I'm sure most people would agree that it would actually make the government worse.

Slavery and Taxes

As I mentioned earlier, though the central premise of this work is nonsense, there are some individual chapters that are still of some small value in isolation. Chapters III and IV of the first book are of some value as arguments against slavery and "might makes right", though Rousseau does make several assumptions that I disagree with (in fairness I suspect he may have covered these more in his Discourse on Inequality which was written earlier than The Social Contract, but which I haven't read). Another section that stands out is the first part of Chapter XV of the third book, where he states that he "... hold[s] enforced labour to be less opposed to liberty than taxes." I don't know if I agree with him on this, but I do know that if we where forced to work one day out of two for the Government our taxation burden would be much more obvious, and I suspect more likely to be opposed.

Conclusion

Rousseau's Social Contract is really only of use as an intellectual curiosity. Due to his fundamental misunderstanding of human society any attempt to structure a government based on his recommendations would be a mistake, quite probably a tragic mistake. If you're reading it to expand your intellectual horizons, as I was, you may gain something (particularly from the first book), but if you want to improve your knowledge of good government you will be sadly disappointed.