Married Couples Are Leaving 401(k) Money on the Table

What You Need to Know

Couples with dual access to DC plans should take full advantage of the bigger employer match before saving in the other spouse's plan.
This allows couples to build more retirement wealth without having to save more.
There is a positive correlation between not optimizing retirement savings and less durable marriages.

A significant share of couples are not effectively coordinating their contributions to workplace retirement plans, robbing the average household of more than 10% of their total potential annual contributions, new research finds.

This is according to a new analysis published by the National Bureau of Economic Research, which shows that a variety of factors drive what they refer to as contribution inefficiency. Some are based on a broader lack of financial literacy among the working population, according to the analysis, but others are more nuanced and behavioral in nature.

The new paper, “Efficiency in Household Decision Making: Evidence from the Retirement Savings of U.S. Couples,” was developed by Taha Choukhmane, of the MIT Sloan School of Management; Lucas Goodman, of the Treasury Department’s Office of Tax Analysis; and Cormac O’Dea, with Yale University’s Department of Economics.

According to the study’s authors, roughly a quarter of married couples in which both spouses have retirement plans appear to allocate their individual contributions in a way that fails to optimally exploit the employer match incentives available at the household level.

In other words, by maxing out whichever spouse’s employer match is more generous, these couples could generate additional long-term savings without having to make additional consumption sacrifices in the short term.

Notably, the researchers suggest this lack of coordination cannot be explained by inertia, automatic enrollment trends or simple heuristics. Instead, they identify a number of compelling indicators that suggest weaker marital commitment correlates strongly with the incidence of inefficient allocations.

Ultimately, the researchers say their work underscores the complexity of the retirement planning process, showing that both financial and behavioral factors must be considered by advisor professionals while helping clients prepare for life after work.

The 401(k) Arbitrage Opportunity

According to the researchers, nearly two-thirds of U.S. civilian workers have access to an employer-sponsored DC retirement saving plan, and more than four-fifths of these plans offer matching contributions made by the employer.

Match schedules vary substantially across employers, the analysis notes, and this creates an “ideal laboratory” to study the efficiency of households’ financial decisions. As the authors explain, the incentives created by the employer match are large and transparent — i.e., the match offers a clearly defined return on investment — and the most efficient allocation for a couple can be clearly defined.

For example, a couple should always contribute first to the account with the highest marginal match rate. If one spouse has a dollar-for-dollar employer match up to a cap, and the other spouse has a 50 cents-on-the-dollar match on their retirement contributions, then the efficient allocation at the household level is to fully exploit the match offered to the first spouse before making any contribution to the second spouse’s account.

To study whether married couples do indeed allocate their individual retirement contributions in a way that efficiently exploits the match incentives available at the household level, the authors created and leveraged a new dataset of the characteristics of employer-provided retirement plans covering a majority of those in open DC plans in the US. They then linked this employer data to administrative records on the retirement saving choices of employees.

In running the numbers, which cover filings provided by more than 6,000 DC retirement plans in the U.S. covering more than 44 million eligible employees, the researchers found that fully 24% of couples in the sample failed to exploit a “within-period intra-household arbitrage condition.”

Stated more simply, these couples are missing out on an opportunity to adjust their contributions in a way that would increase their retirement wealth without changing their current consumption.

Generally, these inefficient couples could either generate more long-term wealth or increase their current consumption at no cost to retirement wealth by simply reallocating existing contributions from the account of the spouse with a lower marginal match to the account of the spouse with a higher marginal match.

According to the researchers, this result is “remarkably robust” to a variety of restrictions one could apply to the sample, and the magnitudes of inefficiency are similar when focusing only on, for example, couples in which neither spouse is 55 years of age or older or couples with more substantial earnings.

Digesting the Results

The authors emphasize that the roughly three-quarters of couples who do not fail the efficiency test are not necessarily coordinating their retirement saving contributions. They may just happen to independently choose individual contributions that are consistent with household-level efficiency.

Thus, to provide a benchmark, the researchers generated two placebo samples. The first involved rearranging all the individuals in the sample of couples into new placebo marriages, such that every individual has a new “spouse” who shared similar observable characteristics with their actual spouse. The second was forming placebo couples out of unmarried individuals that resembled married individuals in the sample.

In these two placebo samples — in which there was by construction no coordination between synthetic spouses — the researchers found that 35% to 38% of couples failed to exploit an available arbitrage opportunity.

“Relative to these benchmarks, our finding of 24% of true couples failing to exploit an arbitrage opportunity implies that while some couples actively coordinate, a substantial share of couples do not coordinate when making their retirement saving contributions,” the researchers explain.

Consistent with this interpretation, the researchers estimate that the incidence of non-coordination drops by 13 percentage points around the time of marriage, and it increases by 12 percentage points after a divorce.

The researchers interpret these results as suggesting that, while the marital contract meaningfully generates coordination over retirement saving for some couples, there is ample evidence that departures from efficiency are widespread.

What It Means for Advisors

According to the paper, in some cases, the costs of this inefficiency are small, but the mean and median levels of the foregone match for couples who fail to exploit the intra-household arbitrage opportunity are substantial, at $682 and $350 per year, respectively.

The authors found the mean and median forgone match are, respectively, 13% and 9% of the total employee retirement contributions made by the households analyzed. In addition, inefficiency is persistent, such that more than half of couples with an inefficient allocation at one point in time still allocate their savings inefficiently four years later.

The researchers noted that the non-coordination is insensitive to the stakes of coordination and that it persists even when there is more than 5% of joint earnings at stake. They suggest this is evidence that the results are not driven by “rational inattention.”

They also found that couples do not systematically improve efficiency when they make active savings decisions, and that couples who are auto-enrolled in their DC plans are no more likely to save inefficiently than those who are not. They say this is evidence against inertia driving the results.

According to the authors, a second class of explanations involves inefficiencies in household — rather than individual — decision-making and a lack of cooperation inside the household. Consistent with this mechanism, they found that plausible proxies of the strength of marital commitment improve the efficiency of household decisions.

“Conditional on a couple’s observable characteristics, we find that the likelihood of failing to coordinate falls with the length of marriage and with the presence of a child or a mortgage, and it is higher for couples who our data show will subsequently divorce,” the authors explain. “We also show that non-coordination is lower for couples who we observe to have owned a joint bank account in the year prior to marriage — a plausible proxy for cooperation.”

Ultimately, the researchers interpret these results as a failure of spouses to collectively realize a surplus available to them in a particular period, which is inconsistent with the widespread assumption of efficiency in decision-making, and suggestive of a greater role for non-cooperative models in the study of households’ economic decisions.

“Our paper is related to a large and growing literature on intra-household decision-making,” they conclude. “There is substantial evidence that multi-person households do not behave like a single person maximizing a unique utility function. … In particular, the distribution of resources within the household, and other proxies of household members’ relative bargaining power, alter household choices.”