By Albert B. Crenshaw
Sunday, May 15, 2005

In theory, 401(k) plans and other tax-preferred savings programs can provide a good retirement fund -- if the worker joins up, contributes a substantial chunk of his or her pay, makes the right investment choices and doesn't drop out or tap the account for non-retirement expenses.

That's the theory. The reality so far isn't pretty. Today, the median account balance - median means the midpoint, that half are larger and half are smaller -- of 401(k) and individual retirement accounts combined, for households headed by someone 55 to 59 years old, "on the verge of retirement," is about $10,000, the Brookings Institution's Peter Orszag noted last week.

But a couple of recent findings and developments may offer avenues that will help.

First, Orszag, Brookings colleague William Gale and several other experts, with the aid of the Pew Charitable Trusts and H&R Block, have been studying whether low- and moderate-income taxpayers would be more likely to divert at least a portion of their tax refunds into an IRA if they were offered a matching contribution of some amount (IRAs, unlike 401(k)s, have no equivalent of the employer's matching contribution). So, using real money provided by H&R Block, they arranged to offer varying matches to a random sample of Block clients who were getting refunds and looked at how they responded.

Now, you might think that what amounts to free money would be an obvious attraction, and that more free money would be more attractive, but some studies of 401(k) plans have suggested that higher matches don't necessarily produce higher participation.

But the new study found that "rational man" (and his updated counterpart, "rational person") is not entirely a figment of economists' imaginations. While only about 3 percent of taxpayers were willing to divert refund money into an IRA if they received no match, 10 percent of taxpayers contributed if given a 20 percent match, and 17 percent did so if given a 50 percent match.

"You see very significant increases, very statistically significant and meaningful increases, in participation rates as you move up the match rate from 0 to 50 percent," Orszag said at a meeting last week at which he described the results.

As the rate of match rose, so did the amount refund recipients contributed to their IRAs. The 3 percent who got no match but were willing to contribute anyway put an average of $856 into their IRAs, while those who got a 50 percent match contributed an average of $1,300 - and got the match on top of that.

The findings also contradict the conventional wisdom that low-income people won't save, or won't save in retirement accounts, Orszag said, noting that even among taxpayers receiving the Earned Income Tax Credit, a program aimed at low-income workers, the participation rate rose at higher match rates.

But the impact of the rising match in this study appears to have been much greater than what is seen in the federal Saver's Credit, which varies, depending on the taxpayer's income. There, getting a higher credit does correlate with increased saving, but not as much as with the Block match.

The Saver's Credit, a special program to encourage lower-income people to save for retirement, allows workers with incomes below certain thresholds to take a credit that reduces their taxes by as much as $1,000 ($2,000 for a couple) if they contribute to a 401(k) or similar plan or to an IRA.

But the Saver's Credit is very complicated, and Orszag and Gale suggested that the effect of information and advice in their experiment - the match was explained to taxpayers individually by Block return preparers - seems to have played a major role in making the match more attractive. Also, Block made the entire process of setting up the IRA, getting the refund and contributing some of it much simpler than it would be in real life.

Indeed, tax refunds today cannot be easily split. The Internal Revenue Service at this point sends refunds as a single payment, so a taxpayer who wants to save part and keep part to spend has to do that division himself.

But -- and this is the other item that could help retirement saving in the long run -- the IRS said recently it plans to offer taxpayers the ability to split their refunds and have part sent one place and part to another, beginning in the 2007 filing season, which would be for next year's taxes. The agency said, though, that it is still ironing out details of how a split would operate.

"It's particularly important that the system of split refunds work with IRAs because those are the individual accounts we now have," said J. Mark Iwry, former benefits tax counsel at the Treasury Department and nonresident senior fellow at Brookings.

With more than 100 million households claiming refunds averaging more than $2,000 each, the ability to direct a portion to IRA savings could have an important long-term impact on retirement saving, Iwry said.

The combination of the Block experiment and split refunds suggests ways to boost retirement saving among lower-income workers.

Obviously, H&R Block isn't going to pony up for matches outside this experiment, but the Saver's Credit does provide a precedent for government to take a role. Perhaps restructuring and expanding that credit and matching it up with split refunds and IRAs could help produce what various tax-favored accounts so far have not: a reasonable chance for a secure retirement.

-----

The Justice Department asked a federal court in Baltimore on Friday to bar John Baptist Kotmair Jr. of Westminster, Md., and his organization, "Save-a-Patriot Fellowship," from selling what the government claims are tax-fraud schemes. The department also wants the court to order Kotmair and Save-a-Patriot to turn over their customers' names, mailing and e-mail addresses, and telephone and Social Security numbers.

The government charges that Kotmair, Save-a-Patriot and a Save-a-Patriot division called the National Workers Rights Committee tell their customers - whom they charge upwards of $99 a year - that they are not required to pay federal taxes or file federal tax returns. Such arguments repeatedly have been rejected by courts across the country, the government said.