The New York Times reports the results of an experiment to see how well matching funds work as a mechanism to increase saving. The results show that matching funds cause an increase in the number of people who save and increases the average amount saved by each individual. This implies that add-on accounts with two features, government or employer matches and opt-out rather than opt-in provisions are an attractive option to pursue in reforming Social Security in a way that increases national saving. There is direct evidence that matching funds increase saving, and though the opt-out provision is not examined directly, the study does show that ease of making the contributions is important:
H&R Blockbuster, NY Times: … Half of all American households have little, if anything, saved specifically for retirement … conventional wisdom holds that there's no way to get people to save more. Happily, the conventional wisdom is wrong. … lawmakers should pay close attention to the results of an experiment that was conducted this year at 60 offices of H&R Block in the St. Louis area. From March 5 to April 5, some 15,000 H&R Block clients, most of them low- or middle-income, were offered free help setting up I.R.A.'s. They were randomly assigned to three groups: people in one group got a 20 percent match for I.R.A. contributions of up to $1,000; another group got a 50 percent match on such contributions. Still others - in the control group - were offered no matching funds. H&R Block put up the money for the matching deposits, eventually spending $500,000. The test was designed and evaluated by researchers from the Retirement Security Project, whose lead sponsor is the Pew Charitable Trusts.
The experiment generated two broad findings: First, offering a match not only causes I.R.A. participation to rise, but also increases the amounts people contribute. A total of 1,500 taxpayers chose to contribute to I.R.A.'s; participation rates were 3 percent in the control group, 10 percent in the 20 percent match group, and 17 percent in the 50 percent match group. The average contributions for the people in the match groups (not counting the H&R Block matching funds) were at least 50 percent as high as in the control group, which got no match. Second, the information provided by the H&R Block tax preparers and the ease of contributing greatly influenced the participants' decisions to save; most of the participants simply diverted portions of their tax refunds into their I.R.A.'s. Full details of how the test isolated these and other factors that bore directly on the savers' decisions are available at www.retirementsecurityproject.org. Although it seems like common sense that offering matching funds would increase I.R.A. participation and contributions, that hypothesis had never been rigorously tested before now. And Congress has never incorporated such direct matches into the savings incentives it provides. Instead, Congress's chief tax writer - Representative Bill Thomas of California - and most of his fellow Republican lawmakers continue to emphasize new tax-deductible savings plans and higher contribution limits for the current tax-favored accounts. Such incentives have failed in the past to motivate most taxpayers to save much, and they won't work now. A big reason for that result is that tax deductions and lofty contribution limits provide the most value to affluent, high-tax-bracket filers, not to low- and middle-income taxpayers. … Lawmakers in Washington could establish a generous and easily understandable I.R.A. match for a fraction of what it would cost to extend the Bush tax cuts for the wealthy. The evidence in favor of doing so is compelling. Then, when the ideological din abates, a future Congress can enact the reforms that are actually needed to strengthen Social Security after midcentury: modest tax increases and tempered benefit cuts, phased in over decades.