Family cash balances, income, and expenditures trends through 2021 - JPMorgan Chase
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Family cash balances, Fiona Greig income, and expenditures Erica Deadman Pascal Noel trends through 2021 A distributional perspective MAY 2021 Abstract The COVID-19 pandemic resulted in unprecedented eco- with faster spend-down for families that are lower-income, nomic changes that impacted families’ financial positions. younger, or working in essential industries. Furthermore, the In this report, the JPMorgan Chase Institute uses adminis- mechanisms for initial pandemic balance increases differed trative banking data to assess checking account balances by family income. Despite greater job losses, low-income in conjunction with household income and spending. We families experienced balance increases driven by increases analyze activity from 1.7 million families who were active in income, due in part to government supports. High-income checking account users between December 2018 and families, in contrast, experienced balance increases despite January 2021 to understand changes in household finances decreases in account inflows because of large concurrent during the COVID-19 pandemic. We find that cash balances decreases in account outflows. Altogether, our results temporarily increased by roughly 70 percent after the offer new insights into families’ financial lives and cash arrival of stimulus payments in April 2020 and January 2021, balances during the COVID-19 pandemic, and contribute with lower-income and younger account holders experi- to the ongoing understanding of the economic impacts encing the largest balance increases on a percent basis. of the pandemic and associated government supports. Balances fell continuously after the stimulus payments, About the Institute The JPMorgan Chase Institute is harnessing the scale and scope of one of the world’s leading firms to explain the global economy as it truly exists. Drawing on JPMorgan Chase’s unique proprietary data, expertise, and market access, the Institute develops analyses and insights on the inner workings of the economy, frames critical problems, and convenes stakeholders and leading thinkers. The mission of the JPMorgan Chase Institute is to help decision makers—policymakers, businesses, and nonprofit leaders—appreciate the scale, granularity, diversity, and interconnectedness of the global economic system and use timely data and thoughtful analysis to make more informed decisions that advance prosperity for all.
Table of Contents 3 Executive Summary 16 Finding Five Income increased for low-income house- holds despite greater job losses, due 6 Finding One in part to government suppor ts. Cash balances increased temporarily by roughly 70 percent in April 2020 and January 2021, after the arrival of stimulus payments, but 19 Finding Six fell continuously after those payments. After initial spending decreases across the income distribution, low-income families rebounded to higher-than-usual spending 9 Finding Two while high-income families remain roughly at Households with lower star ting liquid balances—low- parity with prior trends. er-income and younger account holders—experienced the largest balance increases on a percent basis, but lost those initial balance gains the fastest. 25 Implications 12 Finding Three 26 Appendix We see faster spend-down in liquid assets among families with essential workers. 28 Data Explanation 14 Finding Four 35 References Initial pandemic balance increases for low-income families were driven by increases in account inflows, while high-income families offset decreases 36 Endnotes in inflows by larger decreases in outflows. 38 Acknowledgements and Suggested Citation
Executive Summary Finding 1: Cash balances increased temporarily by roughly 70 percent in April 2020 and January 2021, after the arrival of stimulus payments, but fell continuously after those payments. Balances increased across the distribution, with greater relative percent changes at the low end Year-over-year percent change in weekly balances National Emergency EIP distributed fro declared Mar 13, 2020 Treasury Jan 4, 2021 EIP distributed from Treasury Apr 15, 2020 100% 80% 60% 40% 20% 0% Jan 4 Feb 8 Mar 14 Apr 18 May 23 Jun 27 Aug 1 Sep 5 Oct 10 Nov 14 Dec 19 Jan 23 2021 End of week 25th percentile Median 75th percentile 90th percentile Source: JPMorgan Chase Institute View text version Finding 2: Households with lower starting liquid balances—lower-income and younger account holders—experienced the largest balance increases on a percent basis, but lost those initial balance gains the fastest. Low-income families saw the greatest year-over-year percent balance gains, but depleted those gains faster than high-income families Year-over-year percent change in median weekly balances, by income uartile National Emergency EIP distributed fro declared Mar 13, 2020 Treasury Jan 4, 2021 EIP distributed from Treasury Apr 15, 2020 100% 80% 60% 40% 20% 0% Jan 4 Feb 8 Mar 14 Apr 18 May 23 Jun 27 Aug 1 Sep 5 Oct 10 Nov 14 Dec 19 Jan 23 2021 End of week 1st income quartile 2nd income quartile 3rd income quartile 4th income quartile Note: We assign households into inco e quartiles based on their total labor inco e fro 2019. Households in inco e quartile 1 earned between $12,000 and $30,296 in labor inco e; quartile 2 households earned $30,296 to $44,955; quartile 3 households earned $44,955 to $68,896; and quartile 4 households earned ore than $68,896. Source: JPMorgan Chase Institute View text version Family cash balances, income, and expenditures trends through 2021 Executive Summary 3
Finding 3: We see faster spend-down in liquid assets among families with essential workers. Finding 4: Initial pandemic balance increases for low-income families were driven by increases in account inflows, while high-income families offset decreases in inflows by larger decreases in outflows. Balances grew initially due to increased account inflows and decreased outflows Year-over-year percent change in median account inflows, Year-over-year percent change in median account outflows, y income quartile (four-week periods) y income quartile (four-week periods) National National Emergency EIP distributed Emergency EIP distributed declared from Treasury declared from Treasury Mar 13, 2020 Jan 4, 2021 Mar 13, 2020 Jan 4, 2021 EIP distributed from EIP distributed from 40% Treasury Apr 15, 2020 30% Treasury Apr 15, 2020 30% 20% 20% 10% 10% 0% 0% -10% -10% Feb 22 Apr Jun Aug Oct Nov Jan 23 Feb 22 Apr Jun Aug Oct Nov Jan 23 2020 18 13 8 3 28 2021 2020 18 13 8 3 28 2021 End of four-week period End of four-week period 1st income quartile 2nd income quartile 3rd income quartile 4th income quartile No e: We assign households in o income quar iles based on heir o al labor income from 2019. Households in income quar ile 1 earned be ween $12,000 and $30,296 in labor income; quar ile 2 households earned $30,296 o $44,955; quar ile 3 households earned $44,955 o $68,896; and quar ile 4 households earned more han $68,896. Source: JPMorgan Chase Ins i u e View text version View text version Finding 5: Income increased for low-income households despite greater job losses, due in part to government supports. Finding 6: After initial spending decreases across the income distribution, low-income families rebounded to higher- than-usual spending while high-income families remain roughly at parity with prior trends. 4 Executive Summary Family cash balances, income, and expenditures trends through 2021
Introduction More than a year after the declaration for families during the COVID-19 Second, we rely on direct-deposited of national emergency on March pandemic, and how does this differ paychecks to measure labor income 13, 2020, the COVID-19 pandemic across families—by demographic and industry of employment. Income continues to constrain economic characteristics and industry of received via other channels will not activity in unprecedented ways. employment? And second, what be captured in our measurements Restrictions on group gatherings and are the drivers of these observed of these concepts. In addition, our non-essential outings make certain changes in balance? Specifically, primary income metric reflects businesses untenable, leading to how have consumer income, take-home income, different from layoffs and closures. In other indus- spending, and debt payments gross or pre-tax measures of income tries, remote working has become changed during the pandemic? reflected in most public data sources. the new normal, leading to dramatic To answer these questions, we use a Overall, we find that liquid balances shifts in daily needs and routines for data asset based on the daily balances, were elevated during the COVID-19 employees. Schools have likewise inflows, and outflows of Chase per- pandemic but fell throughout the sec- struggled to reopen, with some sonal checking accounts1, from January ond half of 2020. Balances rose again forced back into remote instruction 2019 through January 2021. We focus with the second round of Economic after increases in infections. on a sample of 1.7 million families who Impact Payments (EIP) in January As families work to adapt to these have been active checking account 2021. These patterns exist across the frequently-changing conditions, users since December 20182. distribution of balances, but with a economic health and stability are Administrative banking data provides a larger initial rise followed by a faster top of mind for policymakers. The unique, high-frequency lens into con- subsequent decline in liquid balances most vulnerable among us have been sumer finances, with transaction-level for families with lower starting bal- impacted the most, with lower-income views into income and expenditures. ances, mainly lower-income families, earners facing higher unemployment Our ability to categorize those or those with younger primary account rates than their higher-earning transactions enables us to observe holders. For most families, balances counterparts (Ganong et al. 2020). The family-level income, which we can use grew despite elevated outflows government responded with massive to show our results across the income because of even larger increases in interventions to suppor t households distribution. Furthermore, we are able inflows, but high-income families financially by funding stimulus to isolate labor income and UI pay- experienced the reverse: balances checks, expanding and supplement- ments from other account inflows and grew despite decreases in inflows ing Unemployment Insurance (UI), observe industry of employment for a because of concurrent cuts in outflows. and providing relief on debts. subset of our sample. So too, are we In this report, we shed light on able to distinguish between spending families’ financial positions during and debt payments among outflows. the pandemic, and how changes We emphasize two main caveats to Liquid balances in behavior and policy have played our data asset: First, we use check- were elevated during a role. Using patterns in liquid ing account balances as a window the COVID-19 pandemic balances as a lens into household into families’ liquid balances but finances, we aim to understand how but fell throughout the acknowledge that a full liquid balance second half of 2020. the pandemic has affected financial picture would include additional vulnerability. Specifically, we focus types of accounts, including those on two main questions. First, how at other financial institutions. have liquid asset balances changed Family cash balances, income, and expenditures trends through 2021 Indtroduction & Data Background 5
Finding One Cash balances increased temporarily by roughly 70 percent in April 2020 and January 2021, after the arrival of stimulus payments, but fell continuously after those payments. Figure 1a shows weekly3 mean and percent changes, highlighting the large do not capture the full picture of median checking account balances year-over-year increase in balances balance changes. The distribution for two series: February 2019 through in April 2020. This account balance has shifted higher overall, as shown January 2020 and February 2020 growth was driven initially by sharp in Figure 2a. The largest dollar gains through January 2021. Balance declines in consumer spending in occurred at the upper end of the levels reflect a family’s purchasing March and April during the shutdowns distribution, with the 90th percentile power, and changes here shed light (Cox et al. 2020). Balances then increasing by more than $4,000 on changes in ability to procure increased steeply after April 15, between Januar y 2020 and January additional goods and services. when families started receiving EIP. 2021. On a percent change basis Average balances rose throughout Gains in average balances continued (Figure 2b), the year-over-year changes the spring, stabilizing after April throughout the year, ending roughly were largest for the lower end of the 15, when families started receiving 40 percent elevated in late December. distribution, owing to the progressive Economic Impact Payments4. Average Year-over-year changes in median nature of the stimulus payments, balances remained steady throughout balances follow a different trajectory, with the 25th percentile elevated by the remainder of 2020, increasing showing a clear downward trend after 95 percent at its peak in April 2020. again in January 2021 to $6,900, more the arrival of the stimulus payments. The lower end of the distribution also than $2,000 higher than the previous Specifically, median balances peaked reverted the most by year's end: in January. In contrast, median balances at 69 percent higher than 2019 late December, the 25th percentile was show a similar increase through May balances the week of May 3 and then 32 percent higher than it had been 2020, but a clear downward trend decreased throughout the rest of in 2019, whereas the 75th percentile throughout the remainder of the year 2020, ending 30 percent higher than was still elevated by 39 percent. This before increasing again in January 2019 balances in December. While change in rank-ordering occurred 2021. Compared with 2019, median median balances were again elevated twice: at peak balances in April 2020, balances were up by about $900 at by 69 percent the week of the second the lower ends of the distribution had their highest point last spring, but EIP disbursement in January 2021, greater year-over-year percent gains those balance cushions decreased those gains seemed to diminish even than the higher ends, but that reversed to roughly $400 by the end of more quickly: by the end of January by the end of 2020, when the 90th December. Put differently, by the end 2021, median balances were up only percentile had the greatest gains over of the year, median balances had 47 percent relative to the previous the prior year and the 25th percentile lost 55 percent of the initial balance year. One likely reason for this is had the least. With the arrival of the gains in April. Balance increases in that the second round of stimulus second EIP, positions were once again January 2021 again had medians payments was much more compressed reversed, though quickly reverting to nearly $900 elevated relative to the than the first, with 96 percent of the original ordering by month’s end. prior year, but that quickly declined payments delivered by January 8.5 to less than $600 two weeks later. The stark difference between the To account for seasonal fluctuations, evolution of mean and median Figure 1b shows year-over-year balances implies that averages alone 6 Finding One Family cash balances, income, and expenditures trends through 2021
Figure 1a: Balances increased through early 2020, with Figure 1b: Year-over-year percent change increased sharply averages remaining elevated throughout the year and medi- for both mean and median balances in April, followed by ans falling from May onward; January 2021 saw additional increasing mean and decreasing median trends throughout balance upticks the year Weekly balances, February 2019 through January 2021 Year-over-year percent change in weekly balances National National EIP distributed Emergency EIP distributed Emergency from Treasury declared from Treasury declared Jan 4, 2021 Mar 13, 2020 Jan 4, 2021 Mar 13, 2020 EIP distributed from EIP distributed from Treasury Apr 15, 2020 Treasury Apr 15, 2020 70% $7,000 $6,000 60% $5,000 50% $4,000 40% $3,000 30% $2,000 20% $1,000 10% $0 0% Feb Mar Apr 1 May Jun Aug Sep Oct Nov Dec Jan 23 Feb Mar Apr May Jun Aug Sep Oct Nov Dec Jan 23 8 14 8 23 27 1 5 10 14 19 2021 8 14 18 23 27 1 5 10 14 19 2021 End of week End of week Mean 2020-2021 Median 2020-2021 Mean Median Mean 2019-2020 Median 2019-2020 Source: JPMorgan Chase Institute View text version View text version Figure 2a: Balances increased across the distribution, with Figure 2b: Balances increased across the distribution, with greater dollar changes at the high end of the distribution greater relative percent changes at the low end Weekly balances Year-over-year percent change in weekly balances National Emergency EIP distributed National declared from Treasury Emergency EIP distributed Mar 13, 2020 Jan 4, 2021 declared from Treasury Mar 13, 2020 Jan 4, 2021 EIP distributed from Treasury Apr 15, 2020 EIP distributed from Treasury Apr 15, 2020 $1 ,000 100% $12,000 80% $10,000 $8,000 60% $6,000 40% $ ,000 20% $2,000 $0 0% Jan Feb Mar Apr May Jun Aug Sep Oct Nov Dec Jan 23 Jan Feb Mar Apr May Jun Aug Sep Oct Nov Dec Jan 23 4 8 14 18 23 27 1 5 10 14 19 2021 4 8 14 18 23 27 1 5 10 14 19 2021 End of week End of week 25th percentile Median 75th percentile 90th percentile Source: JPMorgan Chase Institute View text version View text version Family cash balances, income, and expenditures trends through 2021 Finding One 7
These significant distributional to understanding the impacts of the We find that, consistent with Figure dynamics are not visible in public pandemic and policies on the typical 1b, more families experienced balance datasets that report only aggregate family, as trends differ for businesses.7 growth over the course of the pandemic statistics. Reported statistics often rely To understand what these overall than during the same months in the on averages, which we have shown to increases mean for individual families, prior year. Even with spending declines follow very different trajectories than we further explore within-family and government supports, roughly medians. While the typical household balance changes. Even with median 30 percent of families experienced steadily depleted their cash buffer balances elevated, some families could a 10 percent or more decline in cash throughout the second half of 2020, experience balance decreases and balances, but this share was lower in that pattern gets obscured in aggregate face economic hardship as a result. We the pandemic than in the prior year. In public statistics due to the influence of calculate family-level balance changes other words, while some families do high-balance households in driving the between February 2020 (just before experience balance declines during the means. Moreover, aggregate statistics the pandemic) and January 2021 (the pandemic, there were fewer such fam- on total deposits do not isolate the latest data in our sample). Figure 3 ilies than prior to the pandemic, likely outcomes of households as distinct shows the distribution of those balance due to several government interven- from other sectors, such as commercial changes compared to the prior year. tions and changes in spending patterns. customers.6 This separation is critical Figure 3: Fewer families experienced balance declines during the pandemic (between February 2020 and January 2021) than in the prior year Distribution of within-family balance changes between February 2020 and January 2021 compared to the prior year 52.5 50 40 Share of households (%) 35.4 30 26.4 21.5 20 18.3 14.7 12.8 11.5 10 4.9 2.1 0 < 50 [ 50, 10) [ 10,10) [10, 50) >=50 Balance change (%) 2020 2021 Source: JPMorgan Chase Institute View text version 8 Finding One Family cash balances, income, and expenditures trends through 2021
Finding Two Households with lower starting liquid balances—lower-income and younger account holders—experienced the largest balance increases on a percent basis, but lost those initial balance gains the fastest. Given that the pandemic resulted in considerably narrower gap than families ended January with balances greater job losses among low-income when balances were at their peaks in 61 percent elevated, compared with workers (Cajner et al. 2020) in the spring. With the delivery of the 32 percent for their high-income particular, we further explore liquid second round of stimulus payments counterparts. In sum, low-income balance changes by income and age.8 in January 2021, balances were up families’ balances have been the Separating first by income, Figure 4a 96 percent for the lowest earners, most impacted by stimulus payments, shows balance increases for all income and 43 percent for the highest. experiencing the largest percentage quartiles.9 The highest income quartile Though the gap appears to narrow increases and fastest declines. (families with take-home labor income more quickly this time: low-income greater than $68,896 in 2019) posted the largest dollar gains, with balances roughly $1,400 higher in May 2020 Figure 4a: All income groups experienced similar balance trends during the than May 2019 (Figure 4c). The lowest pandemic, though balance levels vary significantly by income income quartile (families making less than $30,296 in 2019) increased Median weekly balances, by income quartile balances by half as much, peaking National at a gain of $700 relative to 201910. Emergency EIP distributed declared from Treasury Due to their lower balance levels, Mar 13, 2020 Jan 4, 2021 the lowest-income families saw EIP distributed from Treasury Apr 15, 2020 the largest year-over-year gains in $5,000 percentage terms despite having the lowest dollar gains (Figure 4b). At $4,000 the peak in May, low-income families $3,000 had doubled their balances relative to 2019, compared with the highest- $2,000 earning families who were up by roughly 40 percent. However, these $1,000 gains were more quickly depleted for low-income families, losing 70 $0 Jan Feb Mar Apr May Jun Aug Sep Oct Nov Dec Jan 23 percent of their 2020 balance gains 4 8 14 18 23 27 1 5 10 14 19 2021 by the end of December. In contrast, the highest-income families lost 1st income quartile 3rd income quartile 48 percent of their 2020 balance 2nd income quartile 4th income quartile gains by the end of December. By Note: This chart shows the me ian checking account balance by income quartile. We assign househol s into the end of 2020, balances were up income quartiles base on their total labor income from 2019. Househol s in income quartile 1 earne year-over-year by 37 percent for between $12,000 an $30,296 in labor income; quartile 2 househol s earne $30,296 to $44,955; quartile 3 househol s earne $44,955 to $68,896; an quar tile 4 househol s earne more than $68,896. low-income families, compared to 21 percent for high-income families—a Source: JPMorgan Chase Institute View text version Family cash balances, income, and expenditures trends through 2021 Finding Two 9
Figure 4b: Low-income families saw the greatest year-over- Figure 4c: High-income families saw the largest dollar year percent balance gains, but depleted those gains faster gains during the pandemic than high-income families Year-over-year percent change in Year-over-year dollar change in median weekly balances, by income uartile median weekly balances, by income uartile National National Emergency EIP distributed Emergency EIP distributed declared from Treasury declared from Treasury Mar 13, 2020 Jan 4, 2021 Mar 13, 2020 Jan 4, 2021 EIP distributed from EIP distributed from Treasury Apr 15, 2020 Treasury Apr 15, 2020 100% $1,400 $1,200 80% $1,000 60% $800 40% $600 $400 20% $200 0% $0 Jan Feb Mar Apr May Jun Aug Sep Oct Nov Dec Jan 23 Jan Feb Mar Apr May Jun Aug Sep Oct Nov Dec Jan 23 4 8 14 18 23 27 1 5 10 14 19 2021 4 8 14 18 23 27 1 5 10 14 19 2021 End of week End of week 1st income quartile 2nd income quartile 3rd income quartile 4th income quartile Note: We assign households into inco e quartiles based on their total labor inco e fro 2019. Households in inco e quartile 1 earned between $12,000 and $30,296 in labor inco e; quartile 2 households earned $30,296 to $44,955; quartile 3 households earned $44,955 to $68,896; and quartile 4 households earned ore than $68,896. View text version Source: JPMorgan Chase Institute 10 Finding Two Family cash balances, income, and expenditures trends through 2021
We see a similar pattern for balance the most throughout the remainder of each increase in age group (Figure changes by age. Figure 5b shows 2020, and then see higher gains again 5a). Note that this is not merely an that younger age groups have higher in January 2021. These differences artifact of older account holders having year-over-year percent growth in liquid across age groups are driven by dif- higher income than their younger balances in the spring of 2020 than ferences in underlying balance levels. counterparts: the age gap in balances older age groups, deplete their gains Median starting balances increase with holds within each income quartile. Figure 5a: All age groups experienced similar balance Figure 5b: Younger families saw the greatest year-over-year trends during the pandemic, though balance levels vary percent balance gains, but depleted those gains faster than significantly by age older families Median weekly balances, by age group Year-over-year percent change in median weekly balances, by age group National National Emergency EIP distributed Emergency EIP distributed declared from Treasury declared from Treasury Mar 13, 2020 Jan 4, 2021 Mar 13, 2020 Jan 4, 2021 EIP distributed from EIP distributed from Treasury Apr 15, 2020 Treasury Apr 15, 2020 $4,000 100% $3,500 80% $3,000 $2,500 0% $2,000 40% $1,500 $1,000 20% $500 $0 0% Jan Feb Mar Apr May Jun Aug Sep Oct Nov Dec Jan 23 Jan Feb Mar Apr May Jun Aug Sep Oct Nov Dec Jan 23 4 8 14 18 23 27 1 5 10 14 19 2021 4 8 14 18 23 27 1 5 10 14 19 2021 End of week End of week 18-24 25-34 35-44 45-54 55-64 65+ Source: JPMorgan Chase Institute View text version View text version Family cash balances, income, and expenditures trends through 2021 Finding Two 11
Finding Three We see faster spend-down in liquid assets among families with essential workers. Another key aspect impacting a By observing earnings directly sized increases in their account family’s experience of the pandemic is, deposited into a family’s checking balances as a result of the first of course, employment. The pandemic account, we can classify the industry round of EIP. However, in both cases, has impacted working hours, working in which that family is employed.11 We workers in the essential industry—drug conditions, and layoffs, with changes focus here on two industries commonly stores, grocery stores, gas stations, distributed unevenly across different deemed non-essential in public health and discount stores—spent down types of employment. Restaurant ordinances—accommodation and food their cash balances faster than and retail employees saw widespread services, and department stores—in workers in non-essential industries— layoffs and unemployment, while comparison with two essential accommodation, food services, and workers in information-based office retail sectors—drug stores, and a department stores. Given that job jobs remained largely employed, able category comprised of supercenters, losses were more concentrated in to shift to telework (BLS 2020). And of grocery stores, and gas stations. non-essential industries, the slower course, workers in essential jobs—from We find these four sectors interesting depletion in cash buffers among healthcare professionals to grocery because they offer two pairs of non-essential workers is likely due to store cashiers—faced entirely different industries with employees who have the generous unemployment benefits changes during the pandemic. similar levels of account balances prior that workers received through July to the pandemic but who work in an as a result of the CARES Act and Lost essential (solid lines) versus non- Wages Assistance, which more than essential (dotted lines) industry. The replaced typical earnings and led to first pair, in blue, is accommodation considerable savings (Ganong et al. Workers in essential and food service workers versus drug 2020; Greig et al. 2021). On the other industries spent down store workers. The second pair, in hand, it is a stark reality that families their cash balances faster orange, is department store employees with essential workers exhibited a than workers in non- faster depletion of liquid assets. With versus other retail employees (grocery essential industries. the arrival of the second round of stores, gas stations, discount stores). Each pair begins with comparable stimulus payments in January 2021, account balances and sees similarly balances were again comparable for each essential and non-essential pair. 12 Finding Three Family cash balances, income, and expenditures trends through 2021
Figure 6a: Employees working in essential industries Figure 6b: Employees working in essential industries depleted their balance gains faster than those in non- depleted their balance gains faster than those in non- essential industries essential industries Median weekly balances, by industry o employment Year-over-year dollar change in median weekly balances, by industry o employment National National Emergency EIP distributed Emergency EIP distributed declared from Treasury declared from Treasury Mar 13, 2020 Jan 4, 2021 Mar 13, 2020 Jan 4, 2021 EIP distributed from EIP distributed from Treasury Apr 15, 2020 Treasury Apr 15, 2020 $2,000 $1,000 $800 $ ,500 $600 $ ,000 $400 $200 $500 $0 $0 -$200 Jan Feb Mar Apr May Jun Aug Sep Oct Nov Dec Jan 23 Jan Feb Mar Apr May Jun Aug Sep Oct Nov Dec Jan 23 4 8 14 18 23 27 1 5 10 14 19 2021 4 8 14 18 23 27 1 5 10 14 19 2021 End o week End o week Accommodation and food services Drug stores Department stores Grocery stores, gas stations, and discount stores Source: JPMorgan Chase Institute View text version View text version Family cash balances, income, and expenditures trends through 2021 Finding Three 13
Finding Four Initial pandemic balance increases for low-income families were driven by increases in account inflows, while high-income families offset decreases in inflows by larger decreases in outflows. In Finding One, we document a large close to pre-pandemic levels by the lowest-earning families, outflows increase in household checking end of the year, before peaking again were elevated above their pre- account balances at the star t of the in January 2021. In contrast, account pandemic baseline through the end pandemic, followed by a depletion inflows for families in the highest of 2020. In contrast, outflows for the of gains during the remainder of income group decrease relative to highest-earning families remained 2020, with an increase again in 2019 in April and May. We observe depressed throughout the spring January 2021; in Finding Two, we a pronounced spike in total inflows and summer, peaking in July, and show that balance increases were in late June for all income groups, ending the year slightly elevated greatest for high-income families, due to irregularities in the alignment relative to 2019, though less so than and low-income families depleted of paycheck timing for families with at the star t of 2020. In January their balance gains the fastest. In different pay cadences.12 Inflows 2021, with the receipt of the second order to explain these trends, we stabilize $265 above 2019 levels round of EIP, all income groups turn our attention to the mechanisms by the end of the year, roughly 50 exhibited large spikes in outflows. that determine balance levels and percent lower than pre-pandemic In summary, while liquid balances separately assess changes in account status, when families started the increased for families across the inflows (labor income, government year with inflows $540 greater than income spectrum during the early supports, transfers in from other 2019 levels. As with the lower-income period of the pandemic, the ways in accounts) and changes in account families, January 2021 brings an which inflows and outflows contributed outflows (spending, debt payments, inflow spike to the highest earners. to those balances differed across the transfers out of the account). Account outflows show a marked income spectrum. For most families, We show trends in median inflows decrease in April 2020 for each balances grew despite elevated (Figure 7a) and outflows (Figure income group. Given that a national outflows because of even larger 7b) in four-week intervals through emergency was declared on March 13, increases in inflows. But high-income January 2021, by income quartile. and most states went into lockdown families experienced the reverse: We see marked differences in these by the end of the first week of April, balances grew despite temporary trends by income level, with the these trends may represent decreases decreases in inflows because of highest income quartile displaying in spending based on the constraints concurrent cuts in outflows. After the distinct trends compared with imposed by public health measures. second round of stimulus payments the remainder of our sample. Outflows rebounded in the following in January 2021, across the income All three of the lower income quartiles periods, peaking in July13 for every spectrum, families increased outflows. exhibit an inflows peak in April 2020, income group before decreasing The remaining findings will further coinciding with the first round of to settle at a lower steady state decompose inflows and outflows into stimulus payments. Inflow levels for the remainder of the year. key sub-categories to understand remain elevated for the remainder of These trends vary somewhat by which factors drive these distinctions. the summer before gradually falling income group, however: among the 14 Finding Four Family cash balances, income, and expenditures trends through 2021
Figures 7a & 7b: Balances grew initially due to increased account inflows and decreased outflows Year-over-year percent change in median account inflows, Year-over-year percent change in median account outflows, y income quartile (four-week periods) y income quartile (four-week periods) National National Emergency EIP distributed Emergency EIP distributed declared from Treasury declared from Treasury Mar 13, 2020 Jan 4, 2021 Mar 13, 2020 Jan 4, 2021 EIP distributed from EIP distributed from 40% Treasury Apr 15, 2020 30% Treasury Apr 15, 2020 30% 20% 20% 10% 10% 0% 0% -10% -10% Feb 22 Apr Jun Aug Oct Nov Jan 23 Feb 22 Apr Jun Aug Oct Nov Jan 23 2020 18 13 8 3 28 2021 2020 18 13 8 3 28 2021 End of four-week period End of four-week period 1st income quartile 2nd income quartile 3rd income quartile 4th income quartile No e: We assign households in o income quar iles based on heir o al labor income from 2019. Households in income quar ile 1 earned be ween $12,000 and $30,296 in labor income; quar ile 2 households earned $30,296 o $44,955; quar ile 3 households earned $44,955 o $68,896; and quar ile 4 households earned more han $68,896. Source: JPMorgan Chase Ins i u e View text version View text version Family cash balances, income, and expenditures trends through 2021 Finding Four 15
Finding Five Income increased for low-income households despite greater job losses, due in part to government supports. We define a measure of total income as Labor income—defined as income families, though even families in the all non-transfer inflows into a family’s received via paychecks direct-deposited highest income quartile saw substantial checking accounts. This consists of labor into a family’s checking accounts—follows increases. This is evidence that lower-in- income from direct deposit paychecks, a different pattern14. At the start of the come households faced a greater risk of UI income, tax refunds, and all other year, median labor income was elevated job loss. This may also be evidence that non-transfer inflows. (See Figure A1 in 3 to 4 percent relative to 2019 levels lower-income households dispropor- the Appendix for detail on each income for families in the highest three income tionately benefited from the CARES Act category by income quartile.) Figure 8a quartiles, but somewhat depressed for expansion of unemployment eligibility shows that median total income follows the lowest earners (2 percent down). to self-employed and “gig workers” via similar trends as median total inflows in In late March, labor income began to the Pandemic Unemployment Assistance Figure 7a: highly elevated in April 2020 decrease relative to 2019 levels for all (PUA) program. Large numbers of relative to 2019, depleting most of those income groups, reaching a low of 10 to previously ineligible families in the low- initial gains by the following month, and 15 percent below 2019 income in late income group could contribute to the continuing downward before reaching May. By late July, labor income reached larger increases in UI recipiency rates a steady state throughout the fall and a steady state for the remainder of we observe for low-income families. winter of 2020. The lowest-income the year, with the lowest three income families experienced the greatest quartiles slightly below 2019 levels (1 to increase in total income in 2020, relative 2 percent down), and the highest-income to their earnings in 2019; at the upper families somewhat more depressed Direct deposit end of the income distribution, median relative to 2019 (2 to 4 percent down). UI recipiency in our income was roughly at parity with 2019 Along with the observed decrease in sample increased sharply levels from May 2020 through year’s labor income relative to 2019, the rate following the national end. In January 2021, families across of direct deposit UI recipiency15 in our emergency declaration the income distribution experienced a sample increased sharply following in March. marked increase in total income with the the national emergency declaration in arrival of the second round of stimulus. March (Figure 9a). This increase was most dramatic for the lowest-earning 16 Finding Five Family cash balances, income, and expenditures trends through 2021
Figure 8a: Low-income families experienced the greatest Figure 8b: Low-income families experienced the greatest year-over-year percent increase in total income during the year-over-year percent decrease in labor income during the pandemic early months of the pandemic Year-over-year percent change in median total income Year-over-year percent change in median labor income, by income quartile (four-week periods) by income quartile (four-week periods) National National Emergency EIP distributed Emergency EIP distributed declared from Treasury declared from Treasury Mar 13, 2020 Jan 4, 2021 Mar 13, 2020 Jan 4, 2021 EIP distributed from EIP distributed from Treasury Apr 15, 2020 Treasury Apr 15, 2020 40% 10% 30% 5% 20% −0% 10% −5% 0% −10% -10% −15% Feb 22 Apr Jun Aug Oct Nov Jan 23 Feb 22 Apr Jun Aug Oct Nov Jan 23 2020 18 13 8 3 28 2021 2020 18 13 8 3 28 2021 End of four-week period End of four-week period 1st income quartile 2nd income quartile 3rd income quartile 4th income quartile Note: We assign households into inco e quartiles based on their total labor inco e fro 2019. Households in inco e quartile 1 earned between $12,000 and $30,296 in labor inco e; quartile 2 households earned $30,296 to $44,955; quartile 3 households earned $44,955 to $68,896; and quartile 4 households earned ore than $68,896. Source: JPMorgan Chase Institute View text version View text version Family cash balances, income, and expenditures trends through 2021 Finding Five 17
Conditional on receiving UI payments, provided a supplement of $600 per Act of 2021 reauthorized the $300 median payment amounts are roughly week from April through July for every- weekly supplement beginning after equal across income groups, with only one receiving unemployment benefits. December 26, 2020. We observe the lowest income quartile distinct, While FPUC originally expired on July changes in UI amounts consistent at a slightly lower conditional median 31, 2020, states paid out an addi- with these policy changes: conditional level (Figure 9b). The CARES Act tional $300 per week in Lost Wages median UI payments increased in expanded unemployment benefits via Assistance in September and October, May 2020, reverted to pre-pandemic the Federal Pandemic Unemployment and the Coronavirus Response and levels in August, and increased Compensation (FPUC) legislation, which Relief Supplemental Appropriations slightly again in January 2021.16 Figure 9a: UI recipient rate increased dramatically between Figure 9b: UI amount for recipient families was elevated March and May 2020, and has been slowly falling since May through July 2020, and again in Jaunary 2021 UI recipiency rate, by income quartile (four-wee periods) Conditional median UI amount, by income quartile (four-wee periods) 15% $3,500 $3,000 $2,500 10% $2,000 $1,500 5% $1,000 $500 0% $0 Feb 22 Apr Jun Aug Oct Nov Jan 23 Feb 22 Apr Jun Aug Oct Nov Jan 23 2020 18 13 8 3 28 2021 2020 18 13 8 3 28 2021 End of four-wee period End of four-wee period 1st income quartile 2nd income quartile 3rd income quartile 4th income quartile Note: We assign households into inco e quartiles based on their total labor inco e fro 2019. Households in inco e quartile 1 earned between $12,000 and $30,296 in labor inco e; quartile 2 households earned $30,296 to $44,955; quartile 3 households earned $44,955 to $68,896; and quartile 4 households earned ore than $68,896. Source: JPMorgan Chase Institute View text version View text version 18 Finding Five Family cash balances, income, and expenditures trends through 2021
Finding Six After initial spending decreases across the income distribution, low-income families rebounded to higher-than-usual spending while high-income families remain roughly at parity with prior trends. In Figure 10, we track household families in the lowest income quartile Spend increased for all income quartiles spending in four-week intervals. We surpassed pre-pandemic spending, in January 2021, after the arrival of the define spending as checking account peaking at levels more than 10 percent second stimulus payment in the first outflows to cash, paper checks, utility greater than the prior year in June and week of January. Families in the lowest or telecom payments, school fees, July, decreasing slightly though still income quartile increased their spending debit card purchases, and other spend elevated by year’s end at 7 percent to 20 percent higher than the previous excluding tax payments. This measure over 2019 spend. The highest-earning year. In contrast, this increase boosted differs from card-based spending families gradually achieved parity with the spending of the highest-income measures publicly available from 2019 levels, but ended the year with families to roughly 4 percent above the other administrative data sources (e.g. spending depressed relative to both prior year. (For a full decomposition Opportunity Insights) in that it includes 2019 and their pre-pandemic activity. of the categories of spend by income cash, checks, and electronic payments quartile, see Appendix Figure A2.) but excludes credit card transactions, which we treat separately. In Box 1, we show a card-based measure of spending Figure 10: After initial spend decreases, spend increased after both sets of in two-week intervals, which includes stimulus payments, especially for lower-income families debit and credit card transactions, as a benchmark to those sources. Year-over-year percent change in median spend, by income quartile four-week periods) In February 2020, families were National spending slightly more than they did Emergency EIP distributed declared from Treasury over the same period in 2019: median Mar 13, 2020 Jan 4, 2021 spending was up roughly 2 percent EIP distributed from Treasury Apr 15, 2020 for the lowest-earning families and 4 20% percent for families in the remaining 10% income quartiles. Across the income distribution, families decreased 0% their spending dramatically in April, with the highest-income families −10% dropping the most relative to 2019. Spending was down 10 percent in −20% Feb 22 Apr 18 Jun 13 Aug 8 Oct 3 Nov 28 Jan 23 April 2020 for families in the bottom 2020 2021 End of four-week period income quartile, and 18 percent for the highest-earning families. 1st income quartile 2nd income quartile 3rd income quartile 4th income quartile Spending for all income groups began Note: We assign households into inco e quartiles based on their total labor inco e fro 2019. Households in to rebound in May, after the arrival of inco e quartile 1 earned between $12,000 and $30,296 in labor inco e; quartile 2 households earned $30,296 to $44,955; quartile 3 households earned $44,955 to $68,896; and quartile 4 households earned ore than the first round of stimulus payments. $68,896. Over the course of the summer, Source: JPMorgan Chase Institute View text version Family cash balances, income, and expenditures trends through 2021 Finding Six 19
Box 1: Benchmarking our spend measure In an effort to explain changes While our checking-account- we plot a series of total card in families’ liquid balances, based metrics include purchases spend (i.e. debit card and credit the spend metrics we report in made via debit card, they do card transactions) for our this finding represent money not capture the spending that sample, in two-week intervals. flowing directly out of checking families conduct via credit Results are qualitatively accounts. This can make for card—only the payments consistent with trends captured difficult comparisons with made to those cards from in other administrative external benchmarks that rely the checking account. To datasets based on debit and on a combination of debit and compare our trends with external credit card spending.17 credit card transactions. benchmarks, Year-over-year percent change in credit card and debit card spend, by income quartile (two-week periods) National Emergency EIP distributed from EIP distributed fro declared Mar 13, 2020 Treasury Apr 15, 2020 Treasury Jan 4, 2021 20% 10% −0% −10% −20% −30% −40% Feb 22 Apr 18 Jun 27 Sep 5 Nov 14 Jan 23 2020 2021 End of two-week period 1st income quartile 2nd income quartile 3rd income quartile 4th income quartile Note: We assign households into inco e quartiles based on their total labor inco e fro 2019. Households in inco e quartile 1 earned between $12,000 and $30,296 in labor inco e; quartile 2 households earned $30,296 to $44,955; quartile 3 households earned $44,955 to $68,896; and quartile 4 households earned ore than $68,896. Source: JPMorgan Chase Institute View text version Given public health measures in total income, and, in particular, decreases for each income group that implicitly restricted many the generous government supports over the course of the year: roughly 1 types of discretionary spending, through stimulus and unemployment percentage point for the bottom two it is perhaps not surprising that insurance, which disproportionately income quartiles, and 2 percentage spending decreased in the wake of benefited low-income families. points for the top two quartiles by the national emergency declaration. Next, we turn to trends in credit card the end of November. The decreases And yet, many families’ spending, payments, which reflect trends in reversed in December 2020 and particularly that of low-income both spending and credit card debt January 2021, adding roughly 2 families, exceeded pre-pandemic repayment. In Figure 11a, payment percentage points for each income spending levels throughout much incidence rates were relatively quartile. Trends in median payment of 2020 and par ticularly in January steady throughout the year, though amount for families making credit card 2021. This is consistent with the trends we observe very minor and gradual payments18 in a given month (Figure 20 Finding Six Family cash balances, income, and expenditures trends through 2021
11b) diverge by income: families in the when the low-income families hit their card-based spending measure in Box 1 lowest income quartile started the year post-stimulus peak. Payment amounts shows that combined debit and credit 7 percent higher than 2019 payments, rebounded somewhat to end the year card spend was down year-over-year increasing to peak at 24 percent by 10 percent elevated, and in January for high-income families, but elevated May before decreasing again to end they were again elevated by 16 percent for low-income families. For each the year 13 percent up relative to after the second round of stimulus. income quartile, year-over-year credit December 2019. In January 2021, Interpreting these trends, we note that card payments in Figure 11b are ele- payments from the lowest-earning changes in payment amount can indi- vated more than year-over-year card families were 26 percent elevated, cate a family changing their spending spend in Box 1, suggesting that families after the arrival of the second stimulus behavior (how much they transact on are paying down their debt more payment, echoing the peak in May their credit card), or changing their rapidlly. Indeed, evidence of improved after the first round of stimulus. paydown behavior (what proportion of credit health—including declines in Families in the highest income quartile, the balance they pay each month). We credit card utilization and past-due in contrast, started the year 14 percent cannot precisely distinguish between debt—has been documented more higher than in 2019, but reversed to these in our data. However, our broadly (e.g., Urban Institute 2021). slightly below 2019 values in May, Figure 11a: Credit card payment rate was relatively steady Figure 11b: Conditional credit card payment amount was during the pandemic depressed during the pandemic for the highest earners and elevated for the rest; all income groups experienced further elevation by the end of 2020 Incidence rate o credit card payments, by Year-over-year percent change in conditional median credit card income quartile ( our-week periods) payment amount, by income quartile ( our-week periods) National National Emergency EIP distributed Emergency EIP distributed declared from Treasury declared from Treasury Mar 13, 2020 Jan 4, 2021 Mar 13, 2020 Jan 4, 2021 EIP distributed from EIP distributed from Treasury Apr 15, 2020 Treasury Apr 15, 2020 100% 30% 80% 20% 60% 10% 40% 0% 20% 0% −10% Feb 22 Apr 18 Jun 13 Aug 8 Oct 3 Nov 28 Jan 23 Feb 22 Apr 18 Jun 13 Aug 8 Oct 3 Nov 28 Jan 23 2020 2021 2020 2021 End o our-week period End o our-week period 1st income quartile 2nd income quartile 3rd income quartile 4th income quartile Note: We assign households into inco e quartiles based on their total labor inco e fro 2019. Households in inco e quartile 1 earned between $12,000 and $30,296 in labor inco e; quartile 2 households earned $30,296 to $44,955; quartile 3 households earned $44,955 to $68,896; and quartile 4 households earned ore than $68,896. Source: JPMorgan Chase Institute View text version View text version Family cash balances, income, and expenditures trends through 2021 Finding Six 21
Next, we turn to other debt payments, enrolled in mortgage forbearance, during the pandemic likely varied which includes payments to and roughly one third of homeowners by their debt compositions. mortgages, auto loans, student loans, in forbearance continued to make In our sample, the incidence of debt and other installment loans (excluding payments (Farrell et al. 2020). In payments was steady at the start of credit card payments). During the contrast, forbearance on federal the year, with 62 percent of families pandemic, forbearance was offered for student loans was proactively in the highest income quartile making most types of household debt. Loan implemented, automatically stopping payments compared with 26 percent servicers implemented forbearance electronic student loan payments of families in the lowest income protocols in different ways for two for borrowers unless they opted quartile (Figure 12a). Payment rates key types of debt—mortgages and out. In short, while a variety of began falling in late March, reaching student loans—resulting in different forbearance options were available a new steady state by May, at roughly outcomes. Mortgage forbearance was during the pandemic, the ease of 53 percent for the highest-earning made readily available to homeowners enrollment and resulting uptake rates families, and 20 percent for the lowest- experiencing self-reported COVID- varied by debt type, so the extent to earning families for most of 2020. related hardship. Despite the lack of which families used forbearance to Notably, debt payments increased documentation requirements, fewer reduce their debt payment burdens in January 2021 after the arrival of than 10 percent of homeowners the second stimulus payments. Figure 12a: Debt payment rate decreased in the spring of Figure 12b: Conditional debt payment amount was depressed 2020 for all income quartiles during the pandemic for the highest earners and elevated for the rest; all income groups experienced further elevation by the end of 2020 Incidence rate o debt payments, Year-over-year percent change in conditional median debt by income quartile ( our-week periods) payment amount, by income quartile ( our-week periods) National National Emergency EIP distributed Emergency EIP distributed declared from Treasury declared from Treasury Mar 13, 2020 Jan 4, 2021 Mar 13, 2020 Jan 4, 2021 EIP distributed from EIP distributed from Treasury Apr 15, 2020 Treasury Apr 15, 2020 80% 30% 60% 20% 40% 10% 20% 0% 0% −10% Feb 22 Apr 18 Jun 13 Aug 8 Oct 3 Nov 28 Jan 23 Feb 22 Apr 18 Jun 13 Aug 8 Oct 3 Nov 28 Jan 23 2020 2021 2020 2021 End o our-week period End o our-week period 1st income quartile 2nd income quartile 3rd income quartile 4th income quartile No e: We assign households in o income quar iles based on heir o al labor income from 2019. Households in income quar ile 1 earned be ween $12,000 and $30,296 in labor income; quar ile 2 households earned $30,296 o $44,955; quar ile 3 households earned $44,955 o $68,896; and quar ile 4 households earned more han $68,896. Source: JPMorgan Chase Ins i u e View text version View text version 22 Finding Six Family cash balances, income, and expenditures trends through 2021
Conditional on making a debt payment, Higher payment amounts among Finally, Figure 13 decomposes average median payment amounts were steady lower-income families could represent outflow changes into changes in at the start of the year, at parity with families using their excess cash to pay outflow components, including spend, 2019 amounts (Figure 12b). Conditional down debt. Alternatively, it could reflect transfers, and debt payments for payment amounts began rising in late a selection effect rather than actual pay- each income quartile. We see a clear March for the bottom three income ment changes, if families with lower debt increase in spend for families in the quartiles, in coordination with payment payments were disproportionately likely highest and lowest income quartiles, incidence decreases. In contrast, to pause payments during the pandemic. mirroring median results from Figure families in the highest income quartile Indeed, balancing the sample to include 10. In addition, we see that outflow decreased their debt payments in April only families who made payments in transfers, while marginally elevated to 7 percent below 2019 levels. Across December 2020 results in relatively relative to the prior year throughout the income spectrum, payment amounts steady conditional medians throughout the pandemic, see an additional peaked in late May after the arrival of the year, confirming that selection increase in January 2021. These trans- the first stimulus payments in April, effects play a dominant role in the actions represent transfers out of fami- suggesting that families used some increase in conditional payment amount lies’ checking accounts into another of their stimulus funds to make larger for the bottom three income quartiles. account—possibly into some form of debt payments. Throughout the rest of In other words, the increasing trend in savings or investment accounts, or 2020, payments for the lowest three conditional payment amounts reflects checking accounts at another bank. income quartiles remained elevated the fact that, among lower-income Overall, the second stimulus disburse- by at least 10 percent relative to 2019, families, those with higher debt pay- ments in January 2021 appear to have rising further in November to roughly ments were more likely to continue to been used fairly soon upon receipt— 20 percent elevated through January. make their payments while families with either on spending (Figure 10), credit High-income families also finished the lower debt payments were more likely to card paydown (Figure 11b), or transfers year with elevated debt payments, by stop making payments altogether. out to fund savings or other accounts. roughly 5 percent in January 2021. Family cash balances, income, and expenditures trends through 2021 Finding Six 23
You can also read