Here's a post from my colleague, Peter Coy:
Until now, evidence on the efficacy of 401(k) matches has been mixed. Some studies have indicated that workers are much more likely to contribute to 401(k) plans if the company offers to match their contributions. Other studies have found only a small effect. But all the studies have been too small to be conclusive.
Finally, some serious evidence is in, and it’s a surprise. A new study based on 500 retirement plans covering 740,000 workers concludes that the matches are extremely ineffective in getting more people to contribute. In the typical 401(k) plan, only 10% of non-highly-compensated workers are induced to save more by match incentives, concludes the study by Olivia Mitchell of the University of Pennsylvania’s Wharton School, Stephen Utkus of the Vanguard Center for Retirement Research, and Tongxuan (Stella) Yang of Penn.
It might strike most economists as unfathomable, but it appears that most American workers just aren’t motivated by the free money of a company match. Write the authors: “We are left with a strikingly strong level of savings aversion, apparently quite resistant to substantial economic incentives.”
If matches don’t work, what does? The authors point to work by others (like her Wharton colleague Brigitte Madrian) on the success of automatic enrollment, where workers have to opt out of 401(k) plans instead of having to opt in. Another choice: Automatic employer contributions to workers’ 401(k) plans, regardless of whether the employees contribute or not.
Of course, all this assumes that companies’ main goal is to ensure that all employees accumulate retirement savings. The fact is, some companies may regard generous 401(k) matches as primarily just another form of compensation for the middle and upper managers who are most likely to take advantage of the matches. Or the companies may be aiming to recruit young employees, who generally aren’t much interested in saving for retirement and wouldn’t be impressed by automatic employer contributions in 401(k)s. For these types of employers, a company match may work just fine.
One thing’s for sure, though. If you’re running a retirement plan that’s at risk of running afoul of the non-discrimination rules because it’s top heavy with high-income participants, don’t kid yourself that upping the employer match will get lower-income workers to join in droves. It just won’t happen.