Last month, an article in the Wall Street Journal argued that 401(k)s are dangerously underfunded:

Consider households headed by people aged 60 to 62, nearing retirement, with a 401(k)-type account at their jobs.

Such households had a median income of $87,700 in 2009, according to data from the Center for Retirement Research at Boston College, which derived this and other numbers by updating Fed survey data, at The Journal's request. The 85% needed for retirement would be $74,545 a year.

Experts estimate Social Security will provide as much as 40% of pre-retirement income, or $35,080 a year for that median family. That leaves $39,465 needed from other sources. Most 401(k) accounts don't come close to making up that gap.

The median 401(k) plan held $149,400, including plans from previous jobs, according to the Center for Retirement Research. To figure the annual income from that, analysts typically look at what the family would get from a fixed annuity.

That $149,400 would generate just $9,073 a year for a couple, according to New York Life Insurance Co., the leading provider of such annuities-- less than one-quarter of the $39,465 needed.

Unlike pensions, 401(k)s aren't guaranteed by the government.  And people can be even stupider about managing their 401(k)s than pension managers with beady eyed regulators peering over their shoulders: selling out of stocks when they crash, locking in the losses and then missing the rebound gains, or simply undercontributing.

This has led to a lot of nostalgia for the traditional pension, as Ron Browstein summarizes in our sister publication.  But the ever-provocative Andrew Samwick argues that this nostalgia is misplaced:

The other guests on the Marketplace segment, and the WSJ article, may in fact be correct that current near-retirees do not have enough accumulated to hit the standard of post-retirement income equal to 85% of pre-retirement income.  I think that's a ridiculously high target.  (The best research I have seen on this puts the amount of so-called undersaving at a much lower figure -- with perhaps 20% of people approaching retirement with inadequate resources.)  And when evaluating the title question for the post, it is the wrong standard.  The right comparison is to what people would have if they remained under a traditional pension plan, which itself may have been inadequate (even in the absence of cram down or reductions when the PBGC takes over).

My colleague Jon Skinner and I made that comparison in an article in the American Economic Review.  The result was that the projected distributions of retirement income were surprisingly similar under the old-style DB plans that were dominant in the 1980s and the 401(k) plans that supplanted them in the 1990s, assuming workers were covered by the same plan over a long career.  (The comparison was better for 401(k) plans when workers switched jobs -- vested deferred benefits under DB plans are often quite low.) 

The problem, Samwick concludes, is not that we used to have fantastic defined benefit plans, and now we don't--DB plans also take big hits when the market falls, and though they do have a government backstop which 401(k)s don't, they also make a mid-career job loss utterly catastrophic.  So there's no clear winner on the security side.

Rather, the problem is that we display little willingness to sacrifice the income we want now, in order to get the income we'll want later, when we're not working.  Much has been made of the Wisconsin unions' willingness to make higher contributions for their pensions and health care.  But that willingness seems to have only emerged after Republicans threatened to strip most of their ability to collectively bargain for those amenities--and would still involve paying less towards these benefits than the average worker in the private sector does.

Meanwhile, according to the Wall Street Journal, the average 401(k) contribution is 9% of worker income--including employer match.  This is unlikely to be enough, particularly when you assume that workers will have setbacks at some point in their careers that interrupt their contributions.

There's an almost delusional quality to our expectations of retirement.  Somehow, we think we're supposed to set aside 5% of our after-tax income, and have enough to live comfortably for twenty to thirty years without working.

Of course, thanks to the baby boom, that delusion was true, for a while.  The giant population bulge supported all the generations behind them by buying their houses and stocks, and paying the taxes, dividends, and interest payments that supported their elders in a comfortable retirement.  The problem is that we now think that this is something like a natural law, rather than a very temporary aberration.  And so thinks like the defined benefit pension become totemic--if we could only have them, we could have the comfortable retirement that went with them.

We're all going to be paying for this delusion for a long, long time.