Three economists have a big, new idea for other economists: Recognize that the two people in a couple may have two separate sets of reasons for doing what they do.

Pierre-Carl Michaud and two other economists make that case in a new study, “Understanding Joint Retirement,” which is available from the National Bureau of Economic Research.

Even when the two people in a couple don’t have to, they often retire together, or within a year of one another. People may retire because of factors such as pay, employment, leisure preferences or health problems.

Policymakers generally look at the effects of retirement policies only on one member in a couple, not both, and that “unitary model” approach can keep retirement policies from working as intended, Michaud and his colleagues write.

For instance: Couples may not retire together simply because they enjoy spending more time together. Pensions, pay, a partner’s affinity for staying in the workplace, and many other factors may enter into the decision.

Another complication is that spouses may have an imperfect understanding of each others’ preferences.

The economists tested  ways of determining spouses’ behavior in making retirement choices, using answers to questionnaires rather than “revealed preference data,” or data based on actual household choices.

The economists found that using questionnaire answers along with the revealed preference data led to a better evaluation of household retirement decisions and the reasons behind them than using revealed preference data alone.

— Check out Social Security Timing: What Couples Should Consider — The Advisor and the Quant on ThinkAdvisor.