You have no choice but to pay taxes (and thank goodness our tax rates are as low as they are). You have no choice but to obey the laws of the land. There is no such thing as 100% free choice.
We have ample evidence in other countries that a long term sustainable pension scheme is one that pays for itself without being a burden on future generations of tax payers.
Let me give you something to think about. I did this quickly for fun. Let's say that you have $100,000. And you decide to self insure. You need a bit of math and excel skills to follow this argument.
According to the SingStat website
http://www.singstat.gov.sg/publications/...and_deaths
The proportion of males still alive at age 95 given you are alive at age 65 (i.e. probability of survival conditional on age 65) is 0.06. This is not a low number. However, I will assume you can plan your self-insurance until then.
I have seen posters assume they can make 5% returns if they self-insure. This cannot be a riskless portfolio. I model this as a mean return of 5% with a standard deviation of X%. e.g. if X is 1%, it means that 2/3 of my returns are between 4 to 6% (assuming normal distribution).
On a excel sheet, I basically use the norminv(rand(), mean, stddev) function to model the random returns in any one year. I then can simulate N number of scenarios (basically a monte carlo) and find out how many scenarios out of N do I run out of money by age 95.
On a $100K sum, I will assume an annual withdrawal rate of $8000 (which is similar to current estimated CPF life payouts on 100K of capital), I also assume that you invest the 100K at age 55 without any withdrawal until age 65 - again similar to CPF life.
My N is 16380 scenarios (basically about the width of the excel sheet :-)).
(A) For a mean return of 5% and a stddev of 3% (ie 2/3 of returns fall between 2% to 8%), I get 258 scenarios in which you have a 0 or lower principal at age 95 (1.57% of total scenarios).
(B) For a mean return of 5% and a stddev of 1%, I get no failures. with a stddev of 2%, I get 0.07% failure rate.
© For a mean return of 5% and a stddev of 4%, I get 1029 failures or 6.28% of total scenarios.
Now, statistically speaking you have a 6% probability of surviving until age 95 at the time you start withdrawing money. This means, you have a small but non-zero chance of running out of money for any reasonable risky condition.
What's the intuitive reason for this? Well, 8K is 8% of 100K. So when you draw a $8K pension, you may be drawing down capital. Even in bad years, you still draw down $8k, eating into your capital. That's why the results vary even though your mean return is still 5% per annum.
The assumption of normality in your return profile is, i would say, probably optimistic. It should show high kurtosis - i.e. you need to assume black swans. For example, even if you assume your average return is 5%, there may be entire stretch of years in which your return is very negative (I do not need to reach further back than 2008 for an example), but you still need to have your income to live on.
In conclusion, even if you assume a higher rate of return than an annuity's investment float, you will need to keep a buffer for the lean years. This may result in a withdrawal rate scarcely better (if at all) then with an annuity and with no certainty that you won't outlive your capital.