Some like to save, others don't

Stephen F. Venti of Dartmouth and David A. Wise of Harvard concluded that the primary reason for differences in retirement assets was differences in propensities to save. It is not unusual to see low-income households with high savings rates holding more financial assets at retirement than high-income households who saved a smaller fraction of their income.

[One] suggestion is to provide matching grants to low-income individuals.

In Saving Incentives for Low- and Middle-Income Families: Evidence From a Field Experiment With H&R Block,

about 14,000 low- and middle-income families in the St. Louis area were offered a 20 percent match on their contributions to an I.R.A., a 50 percent match or no match at all. Individuals could make a direct contribution or allocate part of their tax refund to an Express I.R.A. account offered by H&R Block.

Only 3 percent of the individuals who had no match — the control group — contributed to an I.R.A. But 8 percent of those with a 20 percent match rate contributed, and 14 percent of those with a 50 percent match contributed. The amount contributed was four times as much as the control group for the 20 percent match rate and seven times as much for the 50 percent match rate.

So what? even with 50% matching grants, three quarters still didn't contribute. And then, "four months after the initial contribution, over 90 percent of the individuals still kept the money in their I.R.A.'s." So less than a year later, nearly 10% withdrew the money?

Update
They didn't mention Save More Tomorrow.