I don't have time to go into this now, but for a pretty persuasive case that Social Security is in crisis, or at least has some major problems that can't be waved away, go see Victor's work at Dead Parrots. Some of the better posts include:

1) This important clarification of some of the productivity assumptions in the Social Security report. If you do all the math, even increasing economic-wide productivity to a very high 2.1 percent doesn't get you to 75-year solvency. (Which would imply 2.5 percent productivity growth in the business sector.)

I'd add two comments. Spikes in the real-wage differential can produce real problems—it can boost the level of promised benefits for one age cohort (because their average wages go up), who then find that the next generations can't pay off the benefits (because productivity and hence the tax base drops again.) Also, I don't think high productivity growth will always give you positive real-wage differentials (which is what really improves solvency). The last few years have been a case in point.

2) This critique of the Diamond Orszag plan. (This post gets unfair at the very end, when he says that the plan "will likely be a drag on real GDP growth." The CBO report on the plan projected real growth of 4.07 under Diamond-Orszag, compared to 4.11 under the do-nothing plan, and 4.13 under privatization.)

However, the main liberal talking points still stand, I think. Even if we did nothing, the "intermediate case" scenario (exhausted Trust Fund, automatic benefit cuts in 2052) really isn't all that bad -- benefits in the future will still be higher than they are today. And Social Security is a problem that's very far off. Meanwhile, much more pressing and immediate problems exist today. If Medicare was under control and the Federal budget wasn't running massive deficits, I'd worry more about Social Security, frankly.

Also, as for Kevin Drum's cool graph, I want to note that under CSSS Plan 2, which is what Bush seems to be proposing, most people do get smaller benefits, yes, though low-income people actually get higher benefits than they would under the do-nothing plan. (Since benefits are boosted for low-earners.) It probably doesn't change anything, but it's worth pointing out.

UPDATE: As Brad Setser points out in comments, (among other good points) I was using the Trustee's data and Kevin was using CBO data. I'm not sure why the two models give such different results (the Trustee's analysis seems easier to follow.) My general point is still okay -- low-earners get screwed less under the privatization plan than everyone else -- but Kevin was right, in the CBO's analysis even low-earners do worse than they would under the "leave it alone" plan. Although the distinction between low-earner and average senior seems artificial. Any retiree who a) has medical bills and b) still has to pay rent or existing debt, needs every penny he or she can find.