Editor’s note: This commentary was written and published before the coronavirus pandemic began to disrupt life and work in America and around the world.

The Census Bureau recently announced it wants advice on ways to develop more accurate measurements of poverty—a welcome and much-needed change. 

Year after year, the Census Bureau reports that more than 30 million Americans live in poverty. Yet it is widely acknowledged that the way government measures poverty is deeply flawed.

The first question is: What does it mean to be poor in the U.S.?

According to the government’s own data, the average American family or single person, identified as poor by the Census Bureau, lives in an air-conditioned, centrally heated house or apartment that is in good repair and not overcrowded. They have a car or truck. (Indeed, 43% of poor families own two or more cars.)

The home has at least one widescreen TV connected to cable, satellite, or a streaming service, a computer or tablet with internet connection, and a smartphone. (Some 82% of poor families have one or more smartphones.)

By their own report, the average poor family had enough food to eat throughout the prior year. No family member went hungry for even a single day due to a lack of money for food. 

They have health insurance (either public or private) and were able to get all “necessary medical care and prescription medication” when needed. 

Reality of Poverty vs. Perceptions

None of this matches the public perception of poverty created by the media. Images of dilapidated trailer homes or drug-infested neighborhoods, full of boarded up buildings, are a staple of media poverty reports. 

These conditions are quite real, and we should be concerned about people who do live in them. But they are, fortunately, not the norm for poor Americans. Fewer than 1 in 10 poor people lives in a mobile home, and 9 out of 10 poor families report no vacant or abandoned buildings in their neighborhoods.

Still, subgroups among the poor do experience substantial financial stress and deprivation. About 7% of poor households report missing a rent or mortgage payment in the prior three months; 2% have had utilities cut off due to nonpayment. And 11% report having delayed or failed to get dental or medical care sometime during the year for lack of money. 

But the majority of those defined as poor by the government do not experience material hardship. Of course, their lives are not a stroll down Easy Street; their finances are often uncertain, and they strain to make ends meet. But the average living conditions among the government-defined poor are well removed from “poverty” as the term is ordinarily understood. 

By using inaccurate measures that pool together those who actually experience deprivation with those who do not, we deflect attention from those who truly need assistance.

Why the disconnect between how the poor actually live and the routine Census Bureau claims of widespread poverty in the U.S.?  

Take for example its consistent report that 1 in 6 children are poor. Government spends over $220 billion on cash, food, and housing aid for low-income families with children. This is two and a half times the amount needed to eliminate all child poverty in the nation. How can so many children remain poor?

The answer lies in the methods the Census Bureau deploys to measure poverty. It defines a family as poor if its “money income” lies below the poverty income thresholds ($25,926 for a family of four in 2019). But “money income” excludes nearly all the benefits provided by means-tested welfare programs, including food stamps, the Women, Infants, and Children food program, Medicaid, housing aid, the earned income tax credit, and the refundable child credit.

Of the $220 billion in means-tested spending on cash, food, and housing for families with children, the Census Bureau counts only 5% as “money income” for purposes of measuring poverty. Under this measurement, government could double welfare spending on poor children and measured poverty would barely budge.

Consider the situation of a single mother with two children working full time at federal minimum wage. After taxes, this mom has only $13,853 in annual earnings. But in nearly all cases she also would receive benefits from food stamps, the earned income tax credit, the refundable child credit, and child nutrition programs worth around $12,600.

Her combined income from earnings and benefits would be around $26,500, about 30% above the poverty threshold for a family of three.

But there is more. In the average state, her family also would receive Medicaid coverage worth another $10,000. If she also gets housing aid (as do roughly 1 in 4 low-income single parents), her combined economic resources from earnings and benefits would reach $47,400.

Despite this financial assistance, the Census Bureau would count only the original $13,853 as income, then dutifully inform the public that the mother was desperately poor. Poverty statistics of this sort are misleading and harmful to rational discourse.

Better Approaches to Measure Poverty

The Trump administration wants to improve poverty measurement by adding a new measure. Two approaches are promising.

The first would be based on self-reported spending or consumption by the poor. This approach circumvents the underreporting of welfare benefits that plague the normal Census Bureau income surveys. When a poor household spends resources received from the earned income tax credit, food stamps, or WIC program, the value of those benefits shows up in the household spending tally. 

The government has surveyed spending on poor households annually since the 1980s; year after year, the government reports spending around $2 for every dollar of “money income” that the Census Bureau says they have—no surprise, given their flawed measurement system. Using self-reported spending figures rather than the bureau’s “money income” formula would cut the U.S. poverty rate roughly in half. 

The second approach would link Census Bureau surveys to the administrative data from welfare programs. These data record the actual benefits received by individual households during the year. This commonsense linkage would give policymakers something they’ve never had before: accurate information on the extent and depth of the current welfare state. (As with current surveys, the privacy of individual households would be maintained.)

A recent study by prominent poverty researcher Bruce Meyer of the University of Chicago linked Census Bureau data to the administrative records for the Temporary Assistance for Needy Families, food stamp, and subsidized housing programs in New York state. It shows that families received far more assistance than is commonly understood—to the extent that using accurate information from these three programs alone cuts the poverty rate of single mothers in half. 

Additional data from the earned income tax credit; the Women, Infants, and Children food program; and other nutrition aid programs, if available, would show even greater reductions in measured poverty.

Developing Better Policies on Poverty 

Accurate information is essential to crafting effective anti-poverty programs. Faulty surveys that ignore nearly all of the current welfare state exaggerate the extent and severity of financial deprivation in the nation.

This is not helpful: It leads to a misdiagnosis of poverty, the misallocation of resources, and policies often irrelevant to the real problems facing society and the poor.

The accurate measurement of poverty should not serve as a pretext for simply slashing welfare spending. Instead, it should provide a better assessment of actual need.

Accurately measuring poverty also could facilitate the development of longer-term policies that address the real underlying problems facing low-income communities: the collapse of marriage, deindustrialization in smaller cities, and slow wage growth for lesser skilled workers.

This approach would truly benefit society and the poor.