The proposal is far from definitive, so coming up with any firm price estimates is rather difficult. It's the same as what happened about a year ago when Bush announced a long term policy of expanding spaceflight to return to the moon and to go to Mars. "Pundits" reached deep into their anal data store and pulled out estimates of a trillion dollars or more with no basis in fact. The devil is in the details, and none of the details are set at the moment because the topic hasn't even began the legislative process.

From the sketchy info that has been published on Bush's proposal, up to one third (4% of the current 12.2%) of current SS taxes would be eligible for investment into private accounts if the taxpayer so chooses. The remainder would go into the existing SS program to do two things. First, it would continue to fund current retirees, and second, it would establish the basis for the baseline SS benefit. For retirees who have private accounts, they'll be eligible for a lower baseline SS benefit but will have their private accounts to make up and perhaps exceed the reduction. No one in the Bush administration is proposing turning the entire program into private accounts, so your $20 trillion estimate seems rather questionable.

As for means testing, what would be the criteria for denying someone who has paid in tens of thousands of dollars into SS any payout? Would it be anyone who has a private 401K or other pension over a certain amount, or something else? Would it be based on their net worth at retirement, leading to possible problems in areas with high housing costs where a person could be "house rich but asset poor"?

I'm not denying that some form of means testing may be necessary, but I'd like to hear about the basis of the test. Even then, unless you're talking about denying benefits to a LOT of people, you won't come close to making up the shortfalls. Under current projections, starting in 2018, the amount of money being collected in SS taxes won't be enough to pay the promised benefits. The shortfall will have to either be made up with money diverted from the general fund to repay those Treasury bonds, or by means testing (or a combination of both). Since the average SS payment is somewhere on the order of $1000 a month, you'd have to eliminate over 80,000 people for every billion dollars of annual SS benefits you want to save. It gets worse as the funding shortfall grows with each passing year. For example, to make up a $10 billion shortfall, you'd have to means test out about 800,000 people. It wouldn't take long before you'd be well past the "eliminate benefits to the rich" phase and well into cutting out people who are by no means rich.

Furthermore, as you go about denying people the benefits they've been promised for decades, you risk undermining public support for the program. Like it or not, everyone has been paying into the system (except reportedly several million government employees) with the promise that they'll be on the receiving end when they retire. If millions of people see SS as just another tax with nothing in return, then it'll be easier to oppose the program and push for its elimination altogether.

My wife and I are far from rich, but we've based out retirement planning on the assumption that we won't get anything from SS despite what the politicians say. Odds are, we're likely to be eliminated by means testing or, perhaps more likely, have a check issued with one hand and taxed away with the other. By planning this way, we're working hard to have enough in our 401Ks, IRAs and other investments to retire without SS. If we're wrong and do actually receive SS benefits, then we'll have a greater assurance that we won't outlive our investments. As it is, I have little faith in the long term viability of the SS program. It's a Ponzi scheme that's heading for collapse unless some serious changes are made. Simply tinkering around the margins won't be enough to save the program, and the longer Congress waits to address the problems, the worse the corrections will have to be.