Two recent surveys of thousands of Americans come to broadly different conclusions on how people used their economic impact payments — their stimulus checks — part of the CARES Act that became federal law in March. Taxpayers received up to $1,200 apiece if they made less than $99,000 — $198,000 for joint filers — last year or the year before, plus $500 per child up to $3,400 for a family of four.
One survey from academic economists finds stimulus recipients by and large saved or paid down debt. Another survey from the Bureau of Labor Statistics finds higher percentages of people who spent or planned to spend their stimulus checks. One consistency between the two: People with less liquidity — those without much cash in the bank or little access to credit — tended to part with their stimulus rather than save.
What’s liquidity?
Think of cash-or-credit-on-hand like a pitcher. If there’s no water in the pitcher, or if the water is frozen, nothing will come out. If the water in the pitcher is liquid, it’ll pour. As for the “frozen” part of the analogy, think of a household vehicle. A car might be worth a good deal of money, but it might also be hard to sell at full value if its owner needs cash now to pay rent. Monetary value is frozen between the physical good and the ability to sell the good. It’s why economists sometimes refer to cars and other high-value items that may be difficult to sell as “illiquid assets.”
Nearly a quarter of U.S. households have bank accounts that might be called “very illiquid,” with less than $400 cash on hand, according to a 2018 analysis from economists at the Federal Reserve. People who are older, wealthier and married have more savings, “but even many of these families do not have more than three months of liquidity,” write the authors, Neil Bhutta and Lisa Dettling. People with at least a bachelor’s degree have assets that are, on average, more liquid than households with some college or a high school diploma or less, according to a nationally representative survey of more than 12,000 U.S. adults that the Fed conducted in October 2019 and updated in May 2020.
Yet even among those who graduated college, racial and ethnic groups differ drastically in their liquidity. The Fed’s analysis found 6% of white respondents with at least a bachelor’s degree said they couldn’t fully pay their current month’s bills. That figure pops up to 11% when they were asked to consider an emergency $400 expense. Meanwhile, 15% of Black respondents with at least a bachelor’s degree couldn’t fully pay their bills before a sudden $400 outlay, and that figure jumps to 23% after. For Hispanic respondents with at least a bachelor’s degree, it’s 12% before and 21% after.
Take one: Tend to save
The first recent analysis of how U.S. taxpayers spent their stimulus checks comes from an August National Bureau of Economic Research working paper — “How Did U.S. Consumers Use Their Stimulus Payments?” — by economists Olivier Coibion at the University of Texas at Austin, Yuriy Gorodnichenko at the University of California, Berkeley and Michael Weber at the University of Chicago.
The authors analyzed roughly 12,000 online survey responses from the July wave of the Nielsen HomeScan Consumer Panel, which asked how the stimulus checks affected consumer spending About 46,000 individuals were invited to participate. The final sample represents a response rate of 26%. That’s a good rate compared with other private surveys, according to the authors: “For example, Qualtrics estimates that their average response rate is 5% to 10%.”
Before getting to the results, a quick side note: Often in studies that sample people, the demographics of the sample won’t perfectly match the larger population researchers want to study. Say researchers want to learn something about the entire U.S., but the percentage of women in their sample is greater than the percentage of women in the nation. The researchers would put their fingers on the scales, so to speak, and count each response from men more than once. Each response from a man might be counted as a response from one-and-a-half men, or two men — whatever multiplier needed so the sample data reflect the population being studied. This very simplified scenario is an example of a common statistical adjustment called sample weighting. The authors of the NBER paper perform this adjustment so that responses from populations underrepresented in the sample are weighted more than those that are overrepresented, making the final analysis more closely representative of the overall U.S. population.
The results? Only 15% said they would spend most of their checks, 33% said they would mostly save, and 52% said they would pay down debt. On average, participants said they would spend or plan to spend about 40% of their stimulus money. But break the data down further and the “average responses mask significant differences across households,” the authors write. Lower-income households, for example, were more likely to spend their checks. So were people in households without much liquidity, as were individuals who weren’t working and had given up looking for work.
“Other groups that were more likely to report spending most of their checks were those living in larger households, men, Hispanics and those with lower education,” the authors write. Black Americans, older people, and people with mortgages were more likely to use their checks to pay debt. Paying down debt may not immediately improve liquidity, but if it brings a borrower one month closer to closing a loan, that’s one more month of liquidity in the future.
Those who didn’t want to spend their stimulus payment had to choose whether to save or pay down debt. In that group, higher-income respondents were more likely to save, the authors found.
To be sure, apart from online spending, there have been fewer easy opportunities to spend money in recent months. Restaurants and bars have closed or are operating at reduced capacity. More people are working from home — and spending less on gas, car repairs, new cars and public transportation. Lack of consumer spending is a big reason the nation’s gross domestic product for April, May and June fell like a rock, to the tune of nearly 10% compared with the first three months of the year.
That GDP drop likely would have been worse without the stimulus checks. Personal savings jumped sharply in April and remained historically high through July, though that doesn’t specifically speak to whether people saved or parted with their stimulus.
Take two: Tend to spend
Back to the narrower question of how people used their stimulus, a survey analysis published Aug. 26 from economists at the Bureau of Labor Statistics appears, at first glance, to paint a different picture from the Coibion, Gorodnichenko and Weber findings.
The BLS analysis is based on the Household Pulse Survey, a joint project of the Census Bureau, BLS and other federal agencies. The online survey conducted June 11 to 16 garnered responses from roughly 73,000 people about whether they received stimulus checks and how they spent them. The BLS economists — Thesia Garner, Adam Safir and Jake Schild — also weight their sample so it’s nationally representative for number of adults, education, sex, age, ethnicity and race. They extrapolate the survey results to estimate how adults in the U.S. writ large used their checks.
Another quick side note: “stimulus check” is a colloquialism for stimulus payments. Most eligible taxpayers received their stimulus money directly into their bank accounts. Physical checks went to most everyone else who didn’t receive tax refunds in 2018 or 2019 via direct deposit and who didn’t provide the Internal Revenue Service their bank information. About three-quarters of taxpayers got a refund in 2018 and 2019. People without bank accounts who received physical checks would have had to turn to alternative financial services, like check cashers, to get their money — and likely be subject to service fees, reducing the final take-home amount of their stimulus.
The topline findings from the BLS survey are that an estimated 84% of adults said they received or expected to receive a stimulus check. Among all respondents, 59% planned to mostly spend their stimulus on personal expenses, such as food and utilities, 13% for debt and 12% said they would save.
Non-Hispanic Black Americans at 83% were most likely to use their stimulus to pay for expenses. Non-Hispanic white people, at 65%, were least likely to spend the money on expenses. Hispanic and Asian people were in the middle at 78% and 74%, respectively.
“These results are in line with ethnic minorities experiencing higher rates of unemployment, and therefore, needing the stimulus check to replace regular sources of income,” the authors write. How “expenses” is defined is in the eye of the beholder. In its questionnaire, the Census Bureau suggests “expenses” might include spending on things like food, clothing and shelter — but buying groceries is a necessity while ordering takeout is a nice-to-have. In other words, the survey doesn’t discern whether people spent their stimulus money on things they needed or things they wanted.
The BLS results appear to depart from the Coibion, Gorodnichenko and Weber working paper. But there are two things worth noting. First, in both analyses, lower-income, Black and Hispanic respondents were overall more likely to part with their stimulus cash, either by spending it or paying down debt.
Second, the BLS analysis is based, again, on the Household Pulse Survey, which has had a consistently low response rate since it began in late April. While the mid-June survey gathered a good number of responses — 73,000 or so — more than 1.1 million people had been invited by email, text or both to participate. That’s a response rate of just a few percentage points, compared with other federal surveys that have response rates approaching or exceeding 50%. The tradeoff is that the Household Pulse Survey has a quick turnaround and is cheaper to conduct than in-person or mail surveys.
It’s impossible to quantify the effect of nonresponse. Maybe results from the BLS survey analysis wouldn’t have changed much if the response rate had been a quarter or more. Maybe they would have changed substantially. By definition, it’s unknowable because there is no data on the substantial missing segment of people who didn’t respond to the survey. This is called nonresponse bias. As noted in the technical documentation, and as Safir confirmed to Journalist’s Resource, it’s a data-user-be-aware situation.
Put another way, the BLS researchers don’t know what they don’t know about the hundreds of thousands of people who didn’t respond to the survey. That said, at least one recent piece of academic research published in Evaluation Review looked at phone and in-person evaluation studies, which take more time to conduct and are usually more detailed than online surveys, and found the “costly pursuit of a high response rate may offer little or no reduction of nonresponse bias.”
Liquid savers
Nonresponse bias aside, several other recent surveys align with the notion that people without much liquidity tended to spend their checks. First up, results from the two most recent waves of a consumer finance survey from the Federal Reserve Bank of Philadelphia, with one wave conducted May 1 to 12 and the other from June 5 to 16. Each wave had about 3,400 unique adult respondents.
Of the 79% of respondents from the June wave who received stimulus checks, about 60% got their money through direct deposit, 15% by physical check and another roughly 5% by prepaid debit card. The report doesn’t break down by race or ethnicity who received physical checks, but, according to the May 2020 Federal Reserve report, 14% of Black adults, 10% of Hispanic adults and 3% of white adults don’t have bank accounts. The biggest change from May to June in the Philadelphia Fed surveys had to do with whether people planned to use their stimulus money for savings or for general purchases.
For May, 41% of respondents said they would save and about 51% said they would spend; for June, 43% said they would save and 50% said they would spend. Not a huge shift. But because people with lower incomes received their stimulus checks first, it’s a trend that supports the idea that the June respondents were “more likely to see the [stimulus] payment as a supplement to existing funds or an opportunity to save, rather than a necessity for day-to-day expenses,” writes study author Tom Ankana, a senior research fellow at the Philadelphia Fed.
An additional study worth knowing is a May NBER working paper that focuses on people with lower incomes who received stimulus checks. The paper is from a group of business and finance professors: Scott Baker at Northwestern University, R.A. Farrokhnia and Michaela Pagel at Columbia University, Steffen Meyer at the University of Southern Denmark and Constantine Yannelis at the University of Chicago. The authors analyze spending data spanning August 2016 to April 2020 from roughly 5,700 people who use SaverLife, a non-profit that aims to help people save money. Users link existing financial accounts to SaverLife online, which incentivizes saving through cash rewards and other prizes.
In the 10 days following a stimulus payment, spending rose $0.25 to $0.35 per dollar per household. SaverLife users tend to skew lower-income, with an average self-reported income of $25,000 a year and a median of $141 in their SaverLife-linked financial account. Lower-income households within the sample and those that lost the most income due to the coronavirus recession were more likely to quickly increase spending after getting their stimulus checks.
But the biggest factor the authors found that predicted more spending?
Liquidity.
“Liquidity plays the most important role, with no observed spending response for households with high levels of bank account balances,” they conclude.
Every individual’s decision on how they used their stimulus checks will have been unique to their situation. Given unemployment rates that were — and remain — historically high, a lot of Americans needed cash when their stimulus checks came through. Still, there will have been wealthier people who spent their entire stimulus on consumer goods, like a new television or furniture, and people with lower incomes who saved their stimulus for a rainy day. But research so far suggests, on aggregate, people who needed cash spent it and people who didn’t need cash saved.
Given the findings detailed here, including the breakdown of financial liquidity by education, race and ethnicity, it’s a short leap to suppose three things: Stimulus checks helped a lot of households stay afloat, the bank accounts of white households came out ahead of Black and Hispanic households, and the bank accounts of the well-educated grew compared with those of the less educated.
In short, the government cash likely benefitted everyone — but the already liquid got more liquid.
Expert Commentary