Defined-Contribution Pensions Are Cost-Effective

Source: Josh B. McGee, Manhattan Institute, Civic Report, No. 100, August 2015

From the summary:
In recent decades, U.S. private-sector employers have increasingly offered retirement benefits through defined-contribution retirement (DC) plans. The share of workers who are offered a retirement plan through their employer and who participate only in a DC plan has increased—from 16 percent in 1979 to 69 percent in 2011. Yet the vast majority of American public-sector workers (75 percent) still earn retirement benefits under a defined-benefit retirement (DB) plan.

The relative merits of DC plans and DB plans have long been debated. Many public-sector employers have recently considered placing new employees in a DC plan; but only two states, Michigan and Alaska, as well as a handful of cities, currently use a DC plan as the primary retirement savings vehicle for new employees. When state and local governments have considered adopting a DC plan for new employees, they have encountered significant opposition from organized labor, managers of current public-retirement systems, and the cottage industry of consultants that supports public DB plans.

Critics of DC plans argue that DB plans are more cost-effective because the latter deliver higher investment returns and convert retirement savings into annuities. This paper investigates whether such assertions hold up to empirical scrutiny.
Key findings include:
DB plans are not structurally more cost-effective than DC plans. Claims of the superior efficiency of DB plans—underpinned by false assumptions and a neglect of pension debt as a significant cost driver—are not supported by empirical evidence.
DC plans achieve similar investment returns. Between 1995 and 2012, average estimated ten-year performance differences between DB and DC plans—at the mean, median, 25th, and 75th percentiles—were less than half a percentage point and were generally not statistically significant. Bottom-performing DB plans outperformed bottom-performing DC plans; top-performing DC plans outperformed top-performing DB plans. Since 2000, performance differences have further narrowed.
DC plans can—and do—offer annuities. The limited availability of annuities among private-sector DC plans is largely the result of misguided federal regulation discouraging their provision. Nevertheless, a number of private-sector firms provide annuities under their DC plans. And most public-sector employers—which do not face regulation hostile to annuities—provide annuities at favorable prices under their DC plans.
Pension debt is a significant cost driver for DB plans. DC plan critics generally ignore the cost of carrying pension debt—one of DB plans’ largest cost drivers—in their DC-DB plan comparisons. For example, carrying a pension debt equal to 10 percent of liabilities would increase annual cost as a percentage of payroll by around 70 percent; carrying a debt equal to 20 percent of liabilities would increase annual cost by around 140 percent.
DC plans are a good option for providing retirement security. Most current DC plans include a number of plan features—including well-designed, diversified, professionally managed investment products—that automatically place participants on a secure retirement path. DC plans can also solve many of the political-economy and benefit-design problems associated with DB plans.
Related:
Manhattan Institute Research Highly Flawed, Irrelevant
Source: National Institute on Retirement Security (NIRS), Press Release, August 17, 2015

A new study released last week is so fundamentally flawed that it is irrelevant to the retirement security debate…

NIRS stands by its research, Still a Better Bang for the Buck: Update on the Economic Efficiencies of Pension Plans, which is inaccurately criticized in the McGee study. The numbers in the NIRS study add up, and it remains a credible and accurate retirement study. It was conducted by a respected pension actuary with both public and private sector experience; it is based on empirical research on investment behaviors of individuals; and was carefully reviewed by a committee of experts.

The McGee study:
Claims that DB plans are not structurally more cost-effective than DC plans. But, NIRS data and empirical evidence shows otherwise. DB plans can deliver a target retirement benefit at half the cost of a DC account.
Says DC plans get similar investment returns as DB plans. But, the analysis fails to use public pension data, and it conveniently ignores asset allocation shifts in private sector pensions due to “frozen” pensions.
Indicates the DC plans can offer annuities. But, few offer annuities and even fewer retirees choose annuities because they are costly. The author conveniently ignores these costs or seeks to buy them from the DB plan.
Says pension debt is a significant cost driver for DB plans. But, this not relevant to the economic efficiencies of DB plan – just like funding shortfalls and asset leakage are not relevant to measuring the efficiency of DC accounts to deliver adequate income replacement.
Indicates DC plans are a good retirement security option. DC plans can be well designed, but the one public DC plan that might come close to the cost efficiencies of public pensions relies on the state DB plan to provide lifetime income. Luckily, this state reopened the DB plan to new employees and most of them actively make elections to join the DB pension.