Brookings Institution: Savings Incentives for Families (2005-05)

I offer a link to this study only because an aspect of its procedures (not its findings) staggered me.

The study, conducted with the aid of H&R Block, wanted to see whether taxpayers would be more open to depositing their tax refunds into IRAs if some entity (in this case, H&R Block) offered them "substantial matches" to do so. About 15,000 low- and middle-income Block clients were randomly offered a 20 percent match on such IRA contributions, a 50 percent match, or no match (this would be the control group).

Surprise, surprise: Higher match rates increase the likelihood of IRA participation and contributions. Who's getting paid to figure this out, anyway?

But here's what made my head spin: In the control group (no match offered), a mere 3 percent of clients signed up for the IRA contribution. (I'm surprised it was even that high.) When a 20 percent match was offered, the take-up rate increased to 10 percent. When a 50 percent (!) match was offered, the take-up rate increased to 17 percent. Matches were limited to $200 (for the 20 percenters) or $500 (for the 50 percenters) per spouse.

With the average refund amount in the $1,580 area, I find it really amazing that only 17 percent of the clients who were offered a 50-percent match accepted the offer. (Actually, married taxpayers took up the offer almost 25 percent of the time, but even that seems absurdly low.) It's just one more piece of evidence, I'd say, that people have their tax refunds spent well before they receive them. Otherwise, how do you NOT accept an offer like that? Or even the 20-percent match?

Clients could even "split" their refunds, choosing to divert only part of it to the IRA if they wished. Heck — what about the possibility of making the IRA deposit, taking the match, and then cashing it out shortly thereafter, paying whatever penalties and interest were mandated? With a 50-percent match, at least, you'd probably still come out well ahead. (After the standard 10% penalty, plus any regular income taxes owed.)

Yeah, there are other variables that could play into this. The clients could be uncomfortable with using H&R Block as their IRA holding company, I guess, but I doubt that low- or middle-income families would give much thought to this aspect. Block's IRA account fees and such might be an issue, but I think these fees were waived for study clients.

Side note: The author of the above study says elsewhere that the median account balance for 401ks and IRAs held by 55 to 59 year-olds is roughly $10,000. That is downright scary, even when you consider that some of those folks likely also have "guaranteed" pensions through their employers.