By Misty Heggeness

When wives earn more than their husbands, a puzzling thing can happen: Husbands say they earn more than they are and wives underreport their income.

New Census Bureau research shows that the incomes couples report on Census Bureau surveys do not always match their IRS filings. The Census Bureau is working to improve the quality of reported earnings by comparing an individual’s survey response with their reported response from another source.

“Wage and earnings data underlie a majority of federal statistics on income, inequality and poverty, and are critical for understanding the pulse of the nation and overall well-being of individuals in society,” said Bruce Meyer, an economist at the Census Bureau and McCormick Foundation Professor at the University of Chicago Harris Public Policy School.

Social Norms Shape Answers
According to this research, societal expectations about the roles played in married-couple relationships may be a factor in what people report for their earnings. Social norms can drive expectations and behavior, including how we report information about ourselves to others.

“We made a critical finding that adds to the understanding of gender norms and the quality of income statistics, in particular, wage gaps among different-sex married couples,” said Marta Murray-Close, an economist at the Census Bureau and co-author of the study.

Researchers used survey responses from the Current Population Survey’s Annual Social and Economic Supplement. They found that when wives earn more than their husbands, husbands report earnings that are 2.9 percentage points higher when they respond to surveys compared to what’s in their tax filings.

Who Really Earns More?
Take a husband in Fargo, N.D., who lives in a household where the wife earns more. If he reported annual earnings of $30,000, he would tell the Census Bureau he earned more — an average $30,870 during that same year. That’s $870 more than his earnings as reported by his employer(s).

The gap is 1.5 percentage points lower for wives who make more than their husbands.

Let us say, for example, that the same man’s wife earns $40,000. On average, her reported earnings would be equivalent to $39,400, around $600 less than she actually earned. The overall impact on family earnings for this household is about $300 more than their employer-reported earnings.

Different Answers From Husbands And Wives
Does it matter whether the husband or wife responds to the survey? The answer is yes.

We analyzed survey reported earnings based on whether the response came directly from the survey respondent or a proxy, such as their spouse answering the question for them.

When a wife earns more, both husbands and wives exaggerate the husband’s earnings and diminish the wife’s. But, husbands overstate their own earnings less than wives do, and wives devalue their own earnings less than husbands do.

In other words, survey reports of earnings are more heavily influenced by gender norms when earnings are reported by a person’s spouse.

How Society Influences Earnings Data
These differences can impact our understanding of national statistics on income, inequality and poverty. In this analytical sample, around 1 in 4 couples (22.9 percent) live in nontraditional marriages where the wife earns more than the husband does.

As nontraditional marriages become more common, self-reported earnings data could become less reliable if individuals continue to be influenced by social norms in their reporting.

The findings demonstrate the importance of understanding societal norms and their influence on data collection and survey responses. Using alternative sources of data helps confirm the accuracy of surveys.

The world of big data and administrative records is becoming more important. The research also provides evidence that building systems and structures that rely on alternative data sources and mechanisms — rather than just self-reported survey responses — should continue.

Misty Heggeness is Senior Advisor for Evaluations & Experiments in Research and Methodology at the Census Bureau.