At the end of 2014, U.S. household retirement assets totaled awhopping $27 trillion (according to the latest editionof the Federal Reserve’s Flow of FundsReport.) Those assets, the great majority of which enjoy somedegree of tax advantage, are spread across various sectors of thesystem: IRAs, pension plans, 401(k) plans and so on.

The single largest slice is IRAs, at $7.3 trillion.Employment-related plans account for most of the remainder, withprivate, federal and state/local government defined benefit pensionplans respectively representing $3.1 trillion, $3.3 trillion and$4.9 trillion. Defined contribution plans total $6.3 trillion (thegreat majority of which is in private sector plans.

A few observations:

1. Public sector vs. private sector At least $9trillion of the $27 trillion total represents public sector workers(possibly more, to the extent they also hold IRAs and annuities atinsurance companies.) Yet (according to the CongressionalResearch Service) the public sector represents only around 16%of total employment. This implies that the average public sectorworker has more than double the retirement assets of the averageprivate sector worker. A large part of that is due to differencesin coverage: a much greater proportion of the private sectorworkforce has no retirement savings vehicle at all. It’s certainlythe case that the question of coverage is a hot topic amonglegislators at present.

Even allowing for the significant part of the private sectorworkforce with no retirement savings at all, the public sectorseems to fare better; this conclusion is consistent with theCongressional Budget Office’s own findings in 2012 that “onaverage, the [non-wage] benefits earned by federal civilianemployees cost 48 percent more than the benefits earned byprivate-sector employees with certain similar observablecharacteristics.”

2. Defined benefit vs. defined contribution Public sectorretirement assets are heavily concentrated in defined benefit (DB)plans, while the majority of private sector assets is in definedcontribution (DC) plans. In a DB plan, it is the employer who gainsor loses from fluctuations in investment values, while in a DCplan, that investment risk falls on the individual. Private sectoremployers have been moving away from DB plans toward DC for manyyears, not least because they have become increasingly unwillingand/or unable to bear that investment risk. Hence private sectorassets in 401(k) and other DC plans now total $5.4 trillion, morethan 40% above their value at the end of 2010.

3. Some parts of the system are more efficient than othersIt is obviously desirable that the retirement system works ascost-effectively as possible. Once it became clear that the primaryretirement savings vehicle for the private sector had become DC,not DB, expectations rose. As Josh Cohen argued in the latest issueof Russell Communiqué: “DC plans are institutionalprograms—start treating them that way.” He’s not alone in thatview: that’s why there’s been so much focus on fees, forexample.

DC plans are, however, easier to run cost-effectively than the oneslice of the pie above that has grown faster than they have overthe past 4 years: IRAs. IRA growth has been driven by rolloversfrom qualified plans (which is where most IRA assets start out).But IRAs serve individuals not institutions, so costs tend to behigher. The changing shape of the retirement system does not makeit easy to drive overall costs down.

Building pressure for change

In a recent blog, I described (in the context of the WhiteHouse’s budget proposals) how budgetary considerations, a growingfocus on coverage, questions of efficiency, and several otherpressure points are leading to a growing likelihood of change in the retirement system. A numberof proposals are circulating at the federal and the states level. Alot is in play: a system that has grown to $27 trillion cannotexpect to avoid legislative attention.