Heterogeneity in the Marginal Propensity to Consume

Summary

A Federal Reserve Bank of Chicago working paper, WP 2020-15, Heterogeneity in the Marginal Propensity to Consume: Evidence from Covid-19 Stimulus Payments, by Ezra Karger and Aastha Rajan, revised February 21, 2021. Using anonymized transaction-level bank account data from Facteus, the authors identify 22,461 recipients of Covid-19 Economic Impact Payments. They report that in the two weeks after a $1,200 stimulus payment in April 2020, consumers increased spending by $546, implying a marginal propensity to consume of 46%, and used an additional 10% of the payment to pay off debt. For the second round of payments in January, 2021, they report spending of 39% within two weeks. Reweighting to the U.S. population, they estimate the CARES Act's $296 billion of stimulus payments increased consumer spending by $130 billion within two weeks.

Summary drafted by a model from the document's text below and checked by script against that text before publication. It is a navigation aid, not a reading of what the document proves. Where AI is used

Full text

Federal Reserve Bank of Chicago

Heterogeneity in the Marginal Propensity
to Consume: Evidence from Covid-19
Stimulus Payments
Ezra Karger and Aastha Rajan

REVISED
February 21, 2021
WP 2020-15
https://doi.org/10.21033/wp-2020-15
*

Working papers are not edited, and all opinions and errors are the
responsibility of the author(s). The views expressed do not necessarily
reflect the views of the Federal Reserve Bank of Chicago or the Federal
Reserve System.

Heterogeneity in the Marginal Propensity to Consume:
Evidence from Covid-19 Stimulus Payments
Ezra Karger*

Aastha Rajan†

Federal Reserve Bank
of Chicago

February 21, 2021

We identify 22,461 recipients of Covid-19 Economic Impact Payments in anonymized transaction-level
bank account data from Facteus. We use an event study framework to show that in the two weeks following a $1,200 stimulus payment in April 2020, consumers increased spending by $546, implying a marginal
propensity to consume of 46%. Consumers used an additional 10% of the stimulus payment to pay off
debt. Consumer spending fell to normal levels after two weeks. Stimulus recipients who live paycheckto-paycheck spent 60% of the stimulus payment within two weeks, while recipients who save much of
their monthly income spent only 24% of the stimulus payment within two weeks. Spending patterns are
quite similar for the second round of stimulus payments in January, 2021, with consumers spending 39%
of their stimulus payments within two weeks and using an additional 14% of their payment to pay off debt.
Reweighting our data to match the U.S. population, ignoring equilibrium effects, and assuming a constant
MPC for each person, we estimate that the CARES Act’s $296 billion of stimulus payments increased consumer spending by $130 billion (44% of total outlays) within two weeks of stimulus receipt. A stimulus bill
targeted at individuals with the highest MPCs could have increased consumer spending and debt payments
by the same amount at a cost of only $246 billion.
JEL Codes: D04, D12, E21
Keywords: Covid-19, stimulus payments, high-frequency data, marginal propensity to consume

* karger@uchicago.edu; 230 South LaSalle Street, Chicago, IL 60604.
† arajan@frbchi.org; 230 South LaSalle Street, Chicago, IL 60604.

First version posted May 28th, 2020. We thank Facteus for providing us with access to their proprietary data for
use in this paper. And we thank Daniel Aaronson, Gadi Barlevy, and Peter Ganong for helpful feedback. The views
expressed in this paper do not necessarily reflect those of the Federal Reserve Bank of Chicago or the Federal Reserve
System.

Introduction
This paper measures the effect of Covid-19 Economic Impact Payments on consumer spending. The
CARES Act, signed into law on March 27th, 2020, provided for payments of up to $1,200 per adult and
$500 per child for most Americans in the United States earning less than $99,000 (or $198,000 for joint
tax filers). Americans with individual income less than $75,000 (and household income less than $150,000)
received the full payment, and payments phased out at higher income levels.1 The IRS directly deposited the
first payments into bank accounts on April 10th, 20202 and by May 11th, 2020, people had received more
than 130 million stimulus payments (worth $200 billion).3 The Joint Committee on Taxation projected that
these stimulus payments cost a total of $293 billion4 when the disbursement process was completed by the
end of the summer, or roughly $881 per U.S. resident.5 In January, 2021, the IRS began a smaller second
round of payments following similar eligibility rules.6
We use a new anonymized transaction-level dataset from Facteus describing spending behavior from
tens of thousands of primary bank accounts to precisely measure the immediate effect of Covid-19 Economic Impact Payments on consumer spending and debt payments.7 We begin by identifying 22,461 active
accounts in the Facteus data that received a stimulus payment from the IRS between April 10th and April
15th. In an event-study framework, we show that $1,200 stimulus payments increased average consumer
1 For more information about the CARES Act, see https://home.treasury.gov/policy-issues/cares.
2 See

additional details here:
https://www.wsj.com/articles/u-s-treasury-starts-sending-individual-stimuluspayments-11586566954. Conflicting reports from the IRS and news organizations claim that the first stimulus payments were deposited into accounts on April 11th, but that is not consistent with our data:
https://www.cbsnews.com/news/stimulus-checks-irs-deposits-first-wave-of-stimulus-checks-2020-04-12/
3 See https://www.washingtonpost.com/business/2020/05/11/still-waiting-your-stimulus-check-you-have-until-12pm-wednesday-give-irs-your-bank-information/
4 For a full analysis of the costs, see: https://www.jct.gov/publications.html?func=startdown&id=5255. We apply
the CARES Act Economic Impact Payment formulas to the 2018 U.S. population using the American Community
Survey and independently estimate that the individual payments will total $296 billion, so that is the estimated cost
that we use throughout this paper.
5 This assumes a population in 2020 of 332.6 million, following the Census Bureau’s projection for the 2020 U.S.
population: https://www.census.gov/content/dam/Census/library/publications/2020/demo/p25-1144.pdf
6 For more details about the implementation of this second round of payments, see:
https://www.irs.gov/newsroom/treasury-and-irs-begin-delivering-second-round-of-economic-impact-payments-tomillions-of-americans
7 We describe this dataset in further detail in the Data section. Facteus is a private company that works with debit
and payroll-card issuers to aggregate and standardize anonymized transaction-level information. Data are available at
a one-day lag for the set of accounts in their data.

spending by $546 in the two weeks after the deposit. Consumers spent an additional $122 to pay off debt.
Larger stimulus payments increased the consumer spending of recipients proportionally. We find no evidence of anticipatory increases in spending in the days leading up to the stimulus payments. The increase in
spending benefits many merchants, with Walmart capturing the largest amount of stimulus-driven spending.
Spending at Walmart increased by $94 per person in the two weeks following stimulus receipt, or 16% of
the overall increase in consumption following the stimulus payment. But increases in spending were spread
evenly over major sectors. We then analyze 37,474 stimulus recipients from the second round of payments
in January, 2021, and we find similar spending responses—stimulus recipients spend 39% of their payments
in the two weeks following receipt and they use an additional 14% of their payment to pay off debt.
While the average MPC in our sample from the April 2020 stimulus checks is 46%, there is significant
heterogeneity in the MPC across our sample of stimulus recipients. In the two weeks following a stimulus
payment, 12% of the recipients decrease spending, 9% of the recipients do not change their spending from
the prior two weeks, and 10% of recipients spend $1,200 or more in the two weeks following the stimulus
payment (relative to the two weeks prior to the payment). The remaining 69% of our sample have spending
changes in the two weeks following the stimulus payment that are distributed roughly uniformly between 0
and 1.
We conclude by using the American Community Survey to re-weight our sample of stimulus recipients
to be representative of the U.S. as a whole. We use this re-weighted sample to estimate the immediate
effect of the stimulus payments on consumer spending. Ignoring equilibrium effects and assuming that each
stimulus recipient has a constant MPC, we estimate that consumers used $144 billion of the $296 billion
of stimulus payments within two weeks of receiving their payments ($130 billion in increased consumer
spending and $14 billion in increased debt payments). A stimulus program of the same size directed at
low-income individuals with the highest marginal propensity to consume could increase consumer spending
by $158 billion (or 53% of the program cost). Lastly, we show that policymakers could have obtained the
same aggregate increase in total consumer spending and debt payments at a cost of $246 billion—instead
of the program’s actual cost of $296—by offering stimulus payments only to lower-income individuals and
households.
A large literature explores the marginal propensity to consume from unanticipated income shocks. For

example, Kan, Peng, and Wang (2017) analyze a $2.6 billion shopping voucher program in Taiwan and find
that each dollar of vouchers leads to $0.24 of increased spending. Fagereng, Holm, and Natvik (2019) use
lottery winners in Norway to measure the MPC for lottery prizes, which range from around 50% for highliquidity winners of large prizes to 100% for low-liquidity lottery winners who win small prizes. Gross,
Notowidigdo, and Wang (2020) use the removal of bankruptcy flags from credit reports to argue that the
MPC from sharp increases in credit card limits is 37%. Agarwal, Liu, and Souleles (2007) use an event
study framework to measure how consumer spending responds to the 2001 federal tax rebates. And Ganong
et al. (2020) use firm-wide variation in monthly pay to estimate that a $1 increase in income leads to a $0.23
increase in consumption, with significantly lower spending responses for high-liquidity households. In Table
A1, we describe results from a selection of papers that measure the MPC in response to anticipated and
unanticipated changes in income. Much of the prior literature measures MPCs using monthly or quarterly
consumer spending data. Our paper adds to this literature by using high-frequency transaction-level data to
measure the effect of stimulus payments on daily consumer spending.
Our estimates of the MPC using transaction-level data closely match survey results. In two recent
surveys, Coibion, Gorodnichenko, and Weber (2020), and Drescher, Fessler, and Lindner (2020) survey
potential Covid-19 stimulus recipients in the U.S. and Europe and find that respondents have spent or expect
to spend 40% of stimulus payments in the U.S. and expect to spend between between 33% and 57% of
stimulus payments in Europe, with MPCs decreasing in income. And Kubota, Onishi, and Toyama (2020)
find that consumers in Japan spend 49% of Covid-19 stimulus payments within six weeks. There is a large
body of literature, typified by Sahm, Shapiro, and Slemrod (2010), that uses data from household surveys
in a similar context. Sahm, Shapiro, and Slemrod estimate an MPC of 33% in response to the 2008 tax
rebates. In a related paper, Parker and Souleles (2017) compare self-reported MPCs to realized MPCs after
Federal stimulus payments in 2008. They find that households spend roughly 50% of the $910 stimulus
payment, closely matching our findings. Fuster, Kaplan, and Zafar (2018), survey consumers to identify
MPCs in response to hypothetical windfalls. They find that respondents report hypothetical MPCs of only
8%, but among those who expect to spend some of their hypothetical windfall, the average reported MPC
is 54%. And in related work, Canbary and Grant (2019) use a survey of households to measure the MPC
for households with different socioeconomic statuses, arguing that the MPC ranges from 0.53 for high-

SES households to 0.94 for low-SES households. Chetty, Friedman, and Stepner (2021) use aggregated
consumer spending data from Affinity Solutions to estimate that consumers spent around $450 of their
$1,200 first round stimulus checks (38%) with minimal heterogeneity by income; and consumers spent $140
of their $600 second round checks (23%) with MPCs ranging from 8% to 23% for zipcodes with different
per-household incomes.8 Lastly, Misra, Singh, and Zhang (2021) use zipcode-level aggregated data from
Facteus to estimate an MPC of 50% in response to Covid-19 stimulus payments, focusing on geographic
heterogeneity in the MPC.
In the paper that is most similar to ours, Baker et al. (2020) identify a set of 3,197 people who received
Covid-19 Economic Impact Payments. These recipients all use a financial app called SaverLife, which
encourages users to save money. They find that in the first ten days after a stimulus payment, consumers
spent $0.25-$0.35 per dollar of stimulus. The three main advantages of our paper are representativeness,
precision, and a comparison of consumer spending responses from multiple rounds of stimulus. SaverLife
is an app that encourages saving. Our dataset is also a convenience sample of stimulus recipients from
card-issuers who provide data to Facteus, but our sample is not explicitly selected on savings behavior. we
also have a broad enough sample of stimulus recipients to re-weight our data to match the U.S. population
and investigate the representativeness of our sample. We find a precise average MPC of 0.44 in our panel
of individuals who use the debit and payroll cards in our sample and the same MPC when we re-weight our
data to match the income, location, and age distribution of individuals in the 2018 ACS. Among individuals
with a high savings rate in our sample, we estimate an MPC of 24%, closely matching the results from
the SaverLife users in Baker et al (2020). Also, because we track more than 6-times as many stimulus
recipients, we have the ability to more precisely estimate the day-by-day effects of stimulus payments on
overall consumption and consumption at individual firms.
Our paper is one of many that uses high-frequency data to measure the economic effects of Covid19-associated policies on consumers. For example, a growing set of papers measure the sharp decline in
consumer spending, mobility, employment, and business activity in March and April of 2020. These papers
use a variety of alternative data sources from companies like Unacast, Second Measure, Womply, Safegraph,
8 The $450 and $140 numbers come from the estimated MPC for zipcodes with average household income of

$46,000 to $59,000.

ADP, and Burning Glass to track high-frequency measures of consumer behavior. For several relevant examples, see Aaronson et al. (2020); Alexander and Karger (2020); Baker et al. (2020); Carvalho et al. (2020);
Chetty et al. (2020); Gupta et al. (2020); and Lewis, Mertens, and Stock (2020). And our evaluation of
the short-run effect of Covid-19 Economic Impact Payments on consumer spending complements Granja et
al. (2020) and Ganong, Noel, and Vavra (2020) who present descriptive statistics and policy counterfactuals
related to the short-run effects of the Paycheck Protection Program and the unemployment insurance component of the CARES Act (respectively). Together, the Economic Impact Payments, the Paycheck Protection
Program, and the expansion of unemployment insurance comprise some of the largest-scale federal policy
responses to Covid-19.

Data
We use data from a company called Facteus that standardizes transaction-level data from dozens of banks
and card-providers from 2012 through 2020. Facteus works with hundreds of card-issuers to aggregate,
standardize, and anonymize this information, perturbing transaction amounts, demographic information, and
transaction timing by randomly chosen values.9 Facteus’s data describes millions of bank accounts, debit
cards, payroll cards, and load cards between 2012 and 2020. We begin by identifying 238,407 accounts in
the Facteus panel that received a deposit anytime in April 2020 from a government agency, including the
IRS, SSA, or state unemployment insurance offices.
In Figure A1 (Panel A), we overlay the distribution of government payment amounts from two groups of
accounts: those receiving a government payment from April 1—April 9, and those receiving a government
payment from April 10—April 15. In this figure, we can clearly see the stimulus payments. Before April 10,
government payments to the accounts in our sample were distributed smoothly (in value) between $0 and
$2,800 with a long right tail. But on and after April 10, the distribution of government payments reflects the
lumpiness of the CARES Act’s payment amounts. Recall that individuals earning under $75,000 received
9 For example, transaction values are perturbed by adding a random number chosen uniformly from a small range

surrounding that number. Birth date information is perturbed by up to 1-2 years in either direction, and transaction
time is perturbed by several hours to avoid identification of individuals. We ignore these perturbations when estimating
our event study results.

a $1,200 payment and joint filers received a $2,400 payment. Adults also received a $500 payment for
each child in their household (up to a four child limit). Consistent with this payment algorithm, we see
large spikes in payment frequency at $1,200, $1,700, $2,200, $2,400, and $2,700. The payment amounts
are perturbed in Facteus’s data, so the transaction amounts are not exact. But we will use the large mass of
accounts receiving a $1,200 payment from the IRS between April 10th and April 15th as our main analysis
sample going forward.10 In Panel B of Figure A1, we show the distribution of government deposits from
the second round of stimulus checks in January, 2021.
We filter our main analysis sample in seven steps to ensure that we focus on primary bank accounts
for a set of consumers. For a complete description of the filtering process, see the Data Appendix. To
summarize the data appendix, we require that accounts record at least ten transactions in January 2020 and
meet minimal thresholds for spending and deposit activity. We also exclude accounts with multiple payments
from the IRS after April 10th to avoid confounding our estimate of the MPC with consumption responses
to contemporaneous tax refunds. These filters leave us with a primary dataset of 20,635 consumers: 13,054
who received a $1,200 payment between April 10th and April 15th, and 7,581 who did not receive any
stimulus payment from the IRS after April 10th. Borusyak and Jaravel (2017) recommend the inclusion of a
never-treated control group of non-recipients in event studies. This never-treated control group helps to pin
down the values of unit and time fixed effects while allowing us to separately estimate days-since-event fixed
effects. Our full sample consists of 30,402 consumers: with an additional 5,442 consumers who received
a $1,700 payment, 3,333 consumers who received a $2,200 payment, and 632 consumers who received a
$2,400 payment from the IRS between April 10th and April 15th. After analyzing the first round of stimulus
payments, we then use identical models to measure the effect of the second round of stimulus payments on
consumer spending for 37,474 consumers in the Facteus data who use these accounts as primary accounts.
In our final analysis sample, we see all bank transactions for these consumers including checks received,
fund transfers, ATM withdrawals, and debit card payments.
In Table A3, we present summary statistics describing the 13,054 recipients of $1,200 stimulus payment
10 We drop accounts in our sample that receive the stimulus payment after April 15th as this date coincides with
the release of the online payment tracker tool by IRS. Thus, the payments could be precisely tracked and their arrival
can not be assumed to be unanticipated. See https://www.irs.gov/newsroom/treasury-irs-unveil-online-application-tohelp-with-economic-impact-payments.

in our data. The average recipient has total monthly deposits of $1,740 from January to March, 2020 and
total monthly spending of $1,250 over that time period. Multiplying by twelve, this implies an average
annual income of $20,880 and average consumer spending of $15,036. In the summary table, we separate
out different transactions marked as ATM Withdrawals, deposits, government deposits, and loads (onto
payroll cards), but in the main analysis we combine all types of deposits into an aggregate deposit measure
and all types of spending into an aggregate spending measure. In Table A4, we show the same summary
statistics, but for our combined group of $1,200 stimulus recipients and the control group described above.
The average consumer in our overall sample received payroll and other deposits totaling $1,691 in January—
March, 2020, implying an average annual income of roughly $20,300.11 In Tables A5 - A7, we show the
same summary statistics for recipients of $1,700, $2,200, and $2,400 stimulus payment (respectively). And
in Table A2 we show summary statistics for non-recipients. In Tables A8 - A11 we show summary statistics
for recipients of the second round of stimulus payments.
In Figure A3, we plot changes over time in account deposits, spending, and 2020 savings for all the
accounts in our sample. We define ‘savings’ as cumulative deposits minus cumulative spending and debt
payments since January 1st, 2020. We see a linear increase in deposits and spending through the end of
March with the exception of a large increase in spending at the end of February, caused by EITC refunds.
The time series of spending is smoother than the lumpy time series of deposits because of regular weekly
and biweekly direct deposits from employers. On April 15th, when we see the largest number of stimulus
payments, we see a sharp increase in aggregate deposits (because of the stimulus payments) and consumer
spending. Figure A4 shows a similar pattern when we plot the time-series of aggregate deposits and spending
in calendar time.
The data from Facteus has several advantages for our analyses in this paper: first, we can see daily
transactions for a large set of accounts, allowing for precise estimation of daily consumer spending. Second,
we can disentangle consumer spending, payroll deposits, government deposits, and ATM withdrawals. And
third, we have fine-grained geographic features for each transaction.
The one major concern about the Facteus data is representativeness. The only demographic information
we have for each consumer is their age and geographic location. We do not see information describing each
11 Based on their incomes, we expect that these consumers will receive a stimulus payment at a later date.

consumer’s gender, household structure, or secondary and tertiary accounts in this data.12 If a consumer
has a credit card, we see when they pay off the credit card, but we do not see the individual credit card
transactions. If the consumer has a secondary debit or payroll card, we cannot see deposits or spending from
that secondary account if the consumer uses their second account as a main source of deposits and spending.
That being said, the consumers in our sample chose the account in Facteus’s data to receive a direct deposit
from the IRS. And aggregate changes in deposits and spending in Figure A3 imply that consumers use these
accounts for a large share of deposits and spending. We view this as evidence that many of these consumers
are using this account as a primary bank account. If these stimulus recipients have secondary bank accounts
or credit cards from which they spend additional money immediately after receiving a stimulus payment,
then our estimate of the average MPC (0.44) is likely a lower bound.

Empirical Strategy
Our main calculation of the average MPC relies on a basic event study framework to measure the effect
of stimulus payments on daily consumer spending, in the style of Agarwal, Lui, and Souleles (2007). We
begin with a dataset of all consumers described above in the data section. We collapse the transaction-level
data to the individual-by-day level. We then analyze four event studies, focusing on the simplest model
(Model A) throughout the paper. Models B and C are presented as robustness checks.
Model A. We focus on consumers who receive a stimulus payment from April 10th—April 15th and
we require that our individual-by-day panel be balanced by further subsetting our dataset to the two weeks
before and after each consumer’s stimulus deposit. We estimate our model separately for consumers who
receive $1,200, $1,700, $2,220, and $2,400 payments between April 10th - April 15th. Recall that $1,200
recipients represent single adults with no dependents, $1,700 recipients represent adults with a single dependent child, $2,200 recipients represent adults with two dependent children, and $2,400 recipients represent
12 Although, as we show later, the MPC is not measurably different for adults with no dependent children and adults

with one or two children. We identify these three groups of consumers using the exact stimulus payment amount.

married adults with a joint account. We analyze this linear regression:
13

Yi,t =

∑ βs 1(stimulus received)i,t+s + εi,t

s=−14

In our analysis, Yi,t is either:
(1) Individual i’s total spending on day t.
(2) Individual i’s total deposits on day t.
The coefficients of interest, βs , represent the days-since-event fixed effects for the two weeks before and
after the stimulus payment. The indicator variable 1(stimulus received)i,t+s is 1 if individual i had received
a stimulus payment on date t + s and 0 otherwise. Because we do not include additional covariates, this
model measures the average level of spending on each day surrounding the stimulus payment relative to an
omitted day (in our case, t − 5). As illustrated in the bottom left panel of Figure 1, the four different groups
of recipients have virtually identical trends in pre-stimulus spending, but experience different post-stimulus
increases in spending ordered according to the stimulus amount received by each group.
Model B. In our second model, we add a series of controls to our baseline event study framework (Model
A). We regress:
13

Yi,t =

∑ βs 1(stimulus received)i,t+s + δi,t + αi + εi,t

s=−14

In this model δi,t and αi represents state-by-date and individual fixed effects, respectively. We include
the δi,t fixed effects as covariates in this event study to absorb regional time-varying features of our data.
One worry is that our baseline estimates may be confounded by systematic variation in stimulus timing for
different types of individuals. Second, there is significant evidence that the passage of stay-at-home orders
causes sharp changes in consumer spending (see Alexander and Karger, 2020). Because of this, we want to
ensure that our estimates of the MPC are not confounded by time-varying state policies (like stay-at-home
orders) that affect business closures and consumer spending.
We include in our model the never-treated control group of consumers. This never-treated control group
helps to pin down the values of unit and time fixed effects while allowing us to separately estimate event-

time fixed effects (as is recommended by Borusyak and Jaravel (2017)). This type of event study can suffer
from bias if treatment effects are time-varying. For more information, see Goodman-Bacon (2018) and Sun
and Abraham (2020). But the results from Model B are indistinguishable from the results of the more simple
averaging exercise in Model A.
Model C. In our third and final model, we address concerns about time-varying treatment effects by estimating an event study model in two steps. First, we regress Yi,t on person-by-day of the week fixed effects
during the pre-pandemic period (January 1st, 2020 through March 15th, 2020). We use this regression to
calculate a daily residual spending measure for each consumer in our final analysis sample of stimulus recipients. Then, we regress this residual spending measure on days-since-stimulus-receipt fixed effects. This
method is similar to one proposed by Goodman-Bacon (2019) who suggests estimating residual outcomes
in the pre-period and using those residual outcomes in the main difference-in-difference specification in the
post-period. As we show later, our main results are unaffected by our use of Model A (as compared to Model
B and Model C).

Results
In Figure 1, we use our baseline event study framework (Model A) to measure the immediate effect of
stimulus payments on consumers. Confidence intervals rely on standard errors that are clustered two-way at
the state and date level. In Panel A, we plot the βs values where the outcome is daily aggregate deposits into
each account.13 In the two weeks immediately preceding and following the receipt of the stimulus payment,
we see no measurable variation in aggregate deposits. But on the exact date when we identify the stimulus
payments, we can see the stimulus payment amounts deposited into each account. In Panel B, we plot the
spending response to this unexpected $1,200 payment. Spending increases sharply in the two days following
the payment, before slowly returning to baseline levels after two weeks. Overall, stimulus recipients increase
spending by $546 in the two weeks following receipt (46% of the stimulus amount). In Panel C, we show
that consumers also sharply increase spending on debt payments (bill pay, utility, and rental payments). In
13 In Figure A7, we show the level change in non-stimulus deposits (relative to day t − 5) for each group of stimulus
recipients.

Panel C we plot daily consumer use of the stimulus payments (spending and debt). In Panel E, we show
the cumulative daily spending (relative to t − 5) from two weeks before the stimulus payments to two weeks
after. And in Panel F, we plot these cumulative daily spending amounts as a fraction of the stimulus payment
made to individuals in each group. In Panel F, we see that different types of households report similar MPCs
in the two weeks after the stimulus payment, with MPCs ranging from 0.40 on average for $2,200 recipients
to 0.50 on average for $2,400 recipients.14
In Figure 2 we report the same six panels for recipients of the second round of stimulus payments.
The most remarkable pattern when comparing Figures 1 and 2 is the striking similarities between consumer
spending responses to the first round of stimulus payments in April, 2020 and the second round of payments
in January, 2021. Although the second round of payments was ninth months later and included smaller
payments, the average MPC ranged from an average of 39% for $600 recipients to an average of 47% for
$2,400 recipients.
We can use the same event study framework to calculate each consumer’s individual increase in spending
in response to the stimulus payment. To do this, we calculate each consumer’s abnormal consumption
following the stimulus payment as

14
14
Cia = Ci,r(i)+14
−Ci,r(i)
14 is consumer i’s total spending in the 14 days preceding date t and r(i) is the date when
where Ci,t

consumer i received their stimulus payment. The individual-level change in consumer spending in the two
Ca

i
weeks following the stimulus payment is then δi = 1,200
. We calculate these changes in spending using the

difference in consumption over a two-week period to account for any constant day-of-the-week effects or
biweekly payroll-related spending decisions.
In Panel A of Figure A2, we plot the distribution of spending changes in our sample of $1,200 stimulus recipients, winsorized at -2 and 2. We see significant variation across stimulus recipients. 69% of our
stimulus recipients have spending changes ranging uniformly between 0 and 1. 9% of recipients spent ap14 In Figure A8, we plot the main event studies over a 28 day pre- and post-event window to confirm the flat pre-

trend and tapering effect of the stimulus payment on consumer spending. The top panel of Figure A9 shows the sharp
increase in consumer spending after stimulus receipt is robust to the addition of thee controls in Model B. And the
bottom panel of Figure A9 shows the main results from Model C.

proximately none of their stimulus payment. 12% of recipients reduce their spending in the weeks following
a stimulus payment (δi <0), and 10% of recipients increase their spending by more than $1,200 in the two
weeks following the stimulus payment (MPC>1). These negative or large changes in spending represent
abnormal spending that is not explained by person or day-of-the-week effects. For example, if an April
15th stimulus recipient decided in March to buy a car on April 20th, that would dramatically increase their
measurable spending after the stimulus payment. But this is not in and of itself due to the stimulus payment.
Panel B of figure A2 plots the distribution of changes in spending from two weeks before to two weeks after
a stimulus payment in our sample of $1,200 recipients in response to a placebo stimulus date, identified as
21 days prior to actual date of stimulus receipt. The symmetric distribution of the MPCs in this panel confirms that our estimated changes in spending in Panel A capture the response to the unanticipated receipt of
stimulus payment rather than just the heterogeneity in consumption patterns in the cross-section. In Figure
A11 we plot the distribution of changes in spending at the individual-level for recipients of $1,700, $2,200,
and $2,400 payments as well. The distributions are quite similar.
To explore sources of this individual-level heterogeneity in the MPC, in Figure 3 we plot the average
individual-level MPC as a function of total spending, total deposits, and total savings in January—March
2020, as well as consumer age. We define total savings as the difference between deposits and spending plus
debt payments in January—March 2020. We do not see account balances in Facteus’s data, so this measure
of savings (a 3-month difference in income and spending flows) is the closest we can come to approximating
pre-Covid-19 liquidity. We see significant variation in the MPC as a function of the pre-Covid-19 savings
rate. Consumers with the highest pre-pandemic savings rate spend only 24% of the stimulus payment on
average in the two weeks following receipt. Consumers with low pre-pandemic propensities to save spend
60% of the first stimulus payment in the two weeks following receipt. Age and aggregate spending levels are
largely uncorrelated with individual marginal propensities to consume. But aggregate deposits in January—
March 2020 (a measure of total income) are also highly correlated with MPCs. There are inimal differences
between heterogenetiy in the MPC in the first and second round of payments. Although the second round
of payments saw a slightly lower MPC of 39% (vs. an MPC of 46% in the first round), we see similar
differences in the MPC for high-income and low-income consumers.
We can further decompose spending for our treatment groups into spending in specific sectors of the

economy using the Merchant Category Code (MCC) attached to each transaction.15 In Figure 4, we plot
spending at merchants in five specific sectors and an ‘other’ category containing all other transactions. See
Table A13 for basic summary statistics about typical monthly spending levels at merchants in each of these
categories. Focusing on the $1,200 recipients, we see that spending at groceries, utilities, restaurants, discount stores, and other merchants jump in the two weeks following a stimulus payment. ATM withdrawals
also sharply increase. This increase is quite similar across sectors: the increase in spending at grocery stores
reflects 43% of typical monthly spending on grocery stores—recipients spend $102 in the two weeks following the stimulus receipt relative to the prior two weeks, and in a typical month, $1,200 recipients spend
$236 on groceries. Similarly, withdrawals from ATMs and spending on utilities, restaurants, discount stores,
and all other merchants see increases equal to 43%, 41%, 33%, 58%, and 44% of typical monthly spending
at merchants in these sectors.
In Figure A13, we go to a more granular level and plot the spending response for eight companies that
might be especially salient during the Covid-19 pandemic: Walmart, Amazon, Dollar General, 7-Eleven,
AT&T, Verizon, Sprint, and Comcast. In Panel A, we see that Walmart captures a full 16% of the increased
spending in our sample due to stimulus payments. The other seven firms also see sharp increases in spending,
although those increases are of significantly smaller magnitudes. In Table A12, we show the increase in
spending following stimulus payments for all merchants in our data with more than $500,000 of spending
from our panel in January–March 202016 . Besides for Walmart, Amazon, 7-Eleven, and Dollar General, no
other merchant captures more than $5 of each stimulus payment.
To ensure that our results are not being driven by cyclical increases in deposits or spending due to regular
paychecks or other bi-weekly or monthly government deposits, in Figure A17, we present our Model A event
study where we assume that each stimulus payment was received 21 days before it was actually received.
Here, we see that those placebo event studies show no pre-trends or post-receipt increase in deposits or
spending, consistent with our argument that the stimulus payments were an unanticipated shock to income
that drove changes in personal spending.
15 MCCs are standardized codes assigned by card issuers to merchants in order to give credit card rewards to cardholders and to meet governmental reporting requirements for specific transactions.
16 This is approximately equivalent to an average expenditure of $4.5 per individual at a company.

Policy Counterfactuals
One concern with the results presented above is that the Facteus data may not be representative of the
U.S. as a whole, which would affect the generalizability of our MPC estimate. We attempt to address this
concern by re-weighting our sample of first-round stimulus recipients to match stimulus recipients in the
U.S. To perform this re-weighting exercise, we rely on the 2018 individual-level ACS data from IPUMS
(Ruggles et al., 2020). For each household, we calculate the number of children aged 16 or younger. We
assign the household head and his/her spouse an equal fraction of the household’s children. In cases where
there is no spouse, we assign children to the household head. And in cases where additional adults live in a
given household, we treat each of those adults as an independent household.
Then, we assign each adult member of each household (over 16 years of age) a stimulus payment based
on their total income and their assigned dependent children using the income eligibility criteria defined by
IRS. We assign all individuals with a personal income of up to $75,000 a stimulus payment of $1,200. For
adults with personal income between $75,000 and $99,000, we assign a stimulus payment that is reduced
by $5 for every $100 earned over $75,000. We add to each adult’s stimulus payment $500 for each child in
their household (including ‘partial’ children, as assigned above). We do not assign any stimulus payment to
those individuals whose personal income might be low enough to receive stimulus payment but who have
spouses earning more than $150,000. After assigning these stimulus payments to each adult, we estimate
that the total cost of Covid-19 Economic Impact Payments will be $296 billion. This is in-line with the Joint
Committee on Taxation’s estimate of $293 billion,17 and small discrepancies are to be expected because we
rely on data from 2018 to estimate the cost of this 2020 program.
We merge our dataset of individuals from the 2018 ACS onto our dataset of MPCs from the Facteus
panel. We estimate MPCs for each adult in the 2018 ACS by regressing the Facteus-based MPC on age
fixed effects, state fixed effects, and income ventiles in our Facteus panel. We then use the coefficients from
this regression to predict the MPC for each adult in the 2018 ACS. In Table 1, we use this matched dataset
to explore the effect of five policies on consumer spending. We first estimate the immediate increase in
consumption in response to the first round of Economic Impact Payments. We estimate that consumers used
17 For a full analysis of the JCT’s estimated costs, see: https://www.jct.gov/publications.html?func=startdown&id=5255.

$130 billion of the $296 billion individual stimulus payments. This ignores any equilibrium effect of the
stimulus payments on prices, and it also ignores any spillover effects from the initial spending increases.
In rows 2-4 of Table 1, we explore predicted spending responses to more progressive versions of the
individual stimulus payments. We keep the overall value of the policy constant, but target lower-income
individuals and households. At the most extreme, we evaluate a policy that gives a payment of $3,700 to any
individual earning less than $10,000 annually (or $20,000 jointly). A back-of-the-envelope calculation based
on the Facteus MPCs implies that this would increase consumer spending by $176 billion. An important
assumption here is that the MPC is constant for each person and does not change with the size of the
individual payment. In the last row of Panel A of Table 1, we evaluate a policy that would give approximately
$1,100 payment to each adult in the United States, independent of income. We argue that this policy would
increase spending roughly the same ($136 billion) as the actual CARES Act payments to individuals and
households. In other words, while the stimulus bill was means-tested, it will have almost the same effect
on consumer spending as a policy that sends a payment to each adult in the United States, irrespective of
income.18 In Panel B of Table 1, we hold total consumer spending (instead of the cost of the stimulus bill)
fixed and show that the same increase in consumer spending of $130 billion could have been achieved with
a $246 billion bill by distributing the stimulus payment to only low-income individuals.

Conclusion
In this paper, we analyze the short-run effects of Covid-19 Economic Impact Payments on consumer
spending. We show that a $1,200 payment from the IRS in April, 2020 caused consumers to spend an
additional $546 on average in the two weeks following stimulus receipt and caused consumers to increase
debt payments by 46%. These patterns are quite similar to consumer responses to the second round of
stimulus payments in January, 2021, where consumers spent 39% of their stimluus payments. Our estimated
MPC of 46% masks significant heterogeneity. Consumers who live paycheck-to-paycheck, spending all of
the income they receive each month, have an average MPC of 60% while high-income consumers and
18 Importantly, this last counterfactual policy evaluation relies on the assumption that the highest-income stimulus
recipients have MPCs that are representative of non-stimulus recipients—those individuals with individual incomes
and household incomes above the CARES Act cutoff.

consumers who generally save a significant fraction of their income have an MPC closer to 24%; and.
We show that consumer age, income, and location are only marginally correlated with individual MPCs
after controlling for each individual’s pre-pandemic savings behavior. Walmart captures much (16%) of the
increase in consumer spending due to Covid-19 Economic Impact Payments.
Ignoring equilibrium effects and assuming a constant MPC for each person, we estimate that the $296
billion of payments to individuals from the CARES Act will increase consumer spending by $130 billion
(44% of total outlays). A stimulus bill of the same overall size targeted at lower-income individuals earning
under $10,000 would have instead increased consumer spending by $158 billion. Consumer spending is not
the main goal of most stimulus programs. Instead, governments use stimulus programs to keep households
afloat during recessions and times of economic uncertainty. Nonetheless, we hope that our findings provide
a precise estimate of how government disbursements (with no strings attached) affect consumer spending
during a time of economic uncertainty.

References
[1] Daniel Aaronson, Scott A. Brave, R. Andrew Butters, and Michael Fogarty. “The stay-at-home labor
market: Google searches, unemployment insurance, and public health orders.” Fed Letter forthcoming,
2020.
[2] Sumit Agarwal, Chunlin Liu, and Nicholas S. Souleles. “The Reaction of Consumer Spending and Debt
to Tax Rebates—Evidence from Consumer Credit Data.” Journal of Political Economy Vol. 115, No. 6,
pp. 986-1,019, 2007.
[3] Sumit Agarwal, and Wenlan Qian.“Consumption and debt response to unanticipated income shocks:
Evidence from a natural experiment in Singapore.“ American Economic Review 104, no. 12 (2014):
4205-30.
[4] Diane Alexander and Ezra Karger. “Do stay-at-home orders cause people to stay at home? Effects
of stay-at-home orders on consumer behavior.” Working Paper Federal Reserve Bank of Chicago, No.
2020-12, 2020.
[5] Scott R. Baker, R.A. Farrokhnia, Steffen Meyer, Michaela Pagel, Constantine Yannelis. “How Does
Household Spending Respond to an Epidemic? Consumption During the 2020 COVID-19 Pandemic.”
NBER Working Paper No. 26949, 2020.
[6] Scott R. Baker, Robert A. Farrokhnia, Steffen Meyer, Michaela Pagel, and Constantine Yannelis. ‘Income, liquidity, and the consumption response to the 2020 economic stimulus payments.“ NBER Working Paper No. w27097. 2020.
[7] Ronald Bodkin. “Windfall income and consumption.“ The American Economic Review 49, no. 4 (1959):
602-614.
[8] Kirill Borusyak and Xavier Jaravel. “Revisiting Event Study Designs, with an Application to the Estimation of the Marginal Propensity to Consume.” Working Paper 2017.

[9] Christian Broda, and Jonathan A. Parker. “The economic stimulus payments of 2008 and the aggregate
demand for consumption.“ Journal of Monetary Economics 68 (2014): S20-S36.
[10] Zara Canbary, Charles Grant. “The Marginal Propensity to Consume for Different Socio-economic
Groups.” Economics and Finance Working Paper Series No. 1916, 2019.
[11] Vasco M. Carvalho, Stephen Hansen, Alvaro Ortiz, Juan Ramon Garcia, Tomasa Rodrigo, Sevi Rodriguez Mora, Jose Ruiz. “Tracking the Covid-19 crisis with high-resolution transaction data.” CEPR
Working Paper No. 14642, 2020.
[12] Raj Chetty, John N. Friedman, Nathaniel Hendren, Michael Stepner, and the Opportunity Insights
Team. “Real-Time Economics: A New Platform to Track the Impacts of COVID-19 on People, Businesses, and Communities Using Private Sector Data.” Working Paper 2020.
[13] Raj Chetty, John N. Friedman, and Michael Stepner. “Effects of January 2021 Stimulus Payments on
Consumer Spending.” Summary Memorandum 2021.
[14] Olivier Coibion, Yuriy Gorodnichenko, Michael Weber. “How Did U.S. Consumers Use Their Stimulus Payments?” NBER Working Paper 2020.
[15] Katharina Drescher, Pirmin Fessler, Peter Lindner. “Helicopter money in Europe: New evidence on
the marginal propensity to consume across European households.” Economics Letters Vol. 195, October
2020.
[16] Andreas Fagereng, Martin B. Holm, and Gisle J. Natvik. “MPC heterogeneity and household balance
sheets.” Working Paper 2019.
[17] Andreas Fuster, Greg Kaplan, and Basit Zafar. “What would you do with $500? Spending Responses
to Gains, Losses, News and Loans.” NBER Working Paper No. 24386, 2018.
[18] Peter Ganong, Damon Jones, Pascal Noel, Diana Farrell, Fiona Greig, Chris Wheat. “Wealth, Race,
and Consumption Smoothing of Typical Income Shocks.” Working Paper 2020.

[19] Peter Ganong, Pascal Noel, Joseph Vavra. “US Unemployment Insurance Replacement Rates During
the Pandemic.” Working Paper 2020.
[20] Andrew Goodman-Bacon. “Difference-in-Differences with Variation in Treatment Timing.” NBER
Working Paper No. 25018, 2018.
[21] Andrew Goodman-Bacon. “So You’ve been Told to do my Difference-in-Differences Thing: A Guide.”
Working Paper 2019.
[22] João Granja, Christos Makridis, Constantine Yannelis, Eric Zwick. “Did the Paycheck Protection Program Hit the Target?” NBER Working Paper No. 27095, 2020.
[23] Tal Gross, Matthew J. Notodiwigdo, and Jialan Wang. “The Marginal Propensity to Consume over the
Business Cycle.” American Economic Journal: Macroeconomics Vol. 12, No. 2, pp. 351-384, 2020.
[24] Sumedha Gupta, Thuy D. Nguyen, Felipe Lozano Rojas, Shyam Raman, Byungkyu Lee, Ana Bento,
Kosali I. Simon, and Coady Wing. “Tracking Public and Private Response to the COVID-19 Epidemic:
Evidence from State and Local Government Actions.” NBER Working Paper No. 27027, 2020.
[25] Joshua K. Hausman.“Fiscal policy and economic recovery: The case of the 1936 veterans’ bonus.“
American Economic Review 106, no. 4 (2016): 1100-1143.
[26] Chang-Tai Hsieh. “Do consumers react to anticipated income changes? Evidence from the Alaska
permanent fund.“ American Economic Review 93, no. 1 (2003): 397-405.
[27] David S. Johnson, Jonathan A. Parker, and Nicholas S. Souleles. “Household expenditure and the
income tax rebates of 2001.“ American Economic Review 96, no. 5 (2006): 1589-1610.
[28] Kamhon Kan, Sin-Kun Peng, and Ping Wang. “Understanding Consumption Behavior: Evidence from
Consumers’ Reaction to Shopping Vouchers.” American Economic Journal: Economic Policy Vol. 9,
No. 1, pp. 137-153, 2017.
[29] So Kubota, Koichiro Onishi, and Yuta Toyama. “Consumption Responses to COVID-19 Payments:
Evidence from a Natural Experiment and Bank Account Data.” Working Paper 2021.

[30] Mordechai E. Krenin. “Windfall income and consumption: Additional evidence.“ The American Economic Review (1961): 388-390
[31] Daniel Lewis, Karel Mertens, James H. Stock. “U.S. Economic Activity During the Early Weeks of
the SARS-Cov-2 Outbreak.” NBER Working Paper No. 26954, 2020.
[32] Kanishka Misra, Vishal Singh, and Qianyun Poppy Zhang. “Impact of Stay-at-home-orders and Costof-living on Stimulus Response: Evidence from the Cares Act.” Working Paper 2021.
[33] Jonathan A. Parker. “The reaction of household consumption to predictable changes in social security
taxes.“ American Economic Review 89, no. 4 (1999): 959-973.
[34] Jonathan A. Parker, Nicholas S. Souleles. “Reported Preference vs. Revealed Preference: Evidence
from the Propensity to Spend Tax Rebates.” Working Paper 2017.
[35] Jonathan A. Parker, Nicholas S. Souleles, David S. Johnson, and Robert McClelland. “Consumer
spending and the economic stimulus payments of 2008.“ American Economic Review 103, no. 6 (2013):
2530-53.
[36] Steven Ruggles, Sarah Flood, Ronald Goeken, Josiah Grover, Erin Meyer, Jose Pacas and Matthew
Sobek. “IPUMS USA: Version 10.0 [dataset].” Minneapolis, MN: IPUMS, 2020.
[37] Claudia R. Sahm, Matthew D. Shapiro, and Joel Slemrod,. “Household Response to the 2008 Tax
Rebate: Survey Evidence and Aggregate Implications.” Tax Policy and the Economy Vol. 24, No. 1, pp.
69-110, 2010.
[38] Nicholas S. Souleles.“The response of household consumption to income tax refunds.“ American
Economic Review 89, no. 4 (1999): 947-958.
[39] Nicholas S. Souleles. “Consumer response to the Reagan tax cuts.“ Journal of Public Economics 85,
no. 1 (2002): 99-120.

[40] elvin Stephens Jr. “The consumption response to predictable changes in discretionary income: Evidence from the repayment of vehicle loans.“ The Review of Economics and Statistics 90, no. 2 (2008):
241-252.
[41] Liyang Sun and Sarah Abraham. “Estimating Dynamic Treatment Effects in Event Studies with Heterogeneous Treatment Effects.” Working Paper 2020.

Data Appendix
We use these seven steps to subset our data to a set of consumers who are likely using the bank account in
Facteus as their primary bank account:
1. We subset to accounts (cards) that recorded at least ten transactions in January 2020 to ensure that the
consumers in our sample are using their accounts actively before the pandemic.
2. We subset to accounts that recorded at least $1,000 of aggregate spending and $1,000 of aggregate
deposits across January—March 2020.
3. We subset to accounts that received at least one deposit from a government agency any time between
January 1st and June 8th 2020. This could include a federal tax refund, unemployment insurance,
Social Security payments, or an Economic Impact Payment. For the second round of stimulus payments, we subset to accounts that received at least one deposit from a government agency between
December 17th, 2020 and January 29th, 2021.
4. We exclude accounts that received multiple IRS deposits between April 10 and June 8 for 1st round
of stimulus payment recipients. Similarly, for the second round of stimulus recipients, we exclude
accounts that received multiple IRS deposits between December 17th, 2020 and January 29th, 2021.
We do this to remove a handful of accounts that received tax refunds and stimulus payments in close
proximity and for whom we cannot identify a distinct value of stimulus payment.
5. We exclude accounts that we identify as Walmart employees. These are accounts that receive at least
one deposit of value greater than $500 from Walmart in January–March 2020.

6. We subset to accounts for which we can identify the resident state of the account holder. This is done
in two ways. First, for each account we look for a known zip code associated with the account at
the time that the IRS or a state government made a deposit into the account in 2020. Second, if the
state cannot be identified in this way, we assign each account to the most frequently occurring state
associated with all other transactions in their account.
7. We subset to two sets of accounts:
• Accounts that received an IRS payment of (i) $1,200, (ii) $1,700, (iii) $2,200, (iv) $2,400
between April 10th and April 15th. We consider the accounts that received an IRS payment of
$1,200 as our main sample of ‘treated’ units. We focus on stimulus payments made between
April 10th and April 15th because early payments were less likely to be anticipated. In midand late-April, the IRS heavily publicized a website where consumers could check the expected
timing of their upcoming stimulus payment. For the 2nd round of stimulus payment, we identify
accounts that received (i) $600, (ii) $1,200, (iii) $1,800, (iv) $2,400 between December 28th,
2020 and January 4th, 2021. The IRS started sending out the 2nd round of stimulus payments
in the last week of December.
• Accounts that did not receive any IRS payment worth more than $500 after April 10th. This is
our main sample of ‘control’ units who we will include for visual and regression-based comparisons. Based on the reported deposits into these accounts, we expect that most of these people
will receive a stimulus payment over the summer.

Figure 1: Effect of 1st Stimulus Payment on Deposits, Spending & Debt: Event Time (Model A)
Aggregate Deposits
Aggregate Spending
Average Daily Value (dollars)

Average Daily Value (dollars)
400
Coefficient on event time dummy

Coefficient on event time dummy

2500
2000
1500
1000
500
0
-14

-7
0
7
Days relative to receipt of stimulus payment

300

200

100

-14

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

Aggregate Debt

Aggregate Debt and Spending

Average Daily Value (dollars)

Average Daily Value (dollars)
600
Coefficient on event time dummy

150
Coefficient on event time dummy

-7
0
7
Days relative to receipt of stimulus payment

100

400

200

-50
-14

-7
0
7
Days relative to receipt of stimulus payment

-14

-7
0
7
Days relative to receipt of stimulus payment

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

Cumulative Aggregate Spending

Daily Value (dollars)

% Share of Stimulus Value

1500

% Share

Value (dollars)

.6
1000

500

.4
.2
0

-.2
-14

-7
0
7
Days relative to receipt of stimulus payment

-14

-7
0
7
Days relative to receipt of stimulus payment

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

Notes: Data at the account-day level. Panels A & B show plotted coefficients on event time dummies from regression of aggregate deposits and
spending on day-since-event time fixed effects (Model A). Standard errors are clustered two-way by state and calendar date. Time 0 in event time
is defined as the date on which the account received a stimulus payment. The shaded regions are 95% confidence intervals. Panels C & D plot the
same days-since event fixed effects for debt payments and aggregate spending and debt. Panels E & F plot cumulative point estimates from Model
A for aggregate spending. Aggregate spending includes transactions labeled as fees, spending and ATM withdrawals. Aggregate Debt includes
transactions labeled as funds transfer, bill pay, or spending transactions with MCC codes associated with utility payments or rental payments.
Aggregate Deposits includes transactions labeld as deposits, loads, government deposits and stimulus payment.

Figure 2: Effect of 2nd Stimulus Payment on Deposits, Spending & Debt: Event Time (Model A)
Aggregate Deposits
Aggregate Spending
Average Daily Value (dollars)

Average Daily Value (dollars)
400
Coefficient on event time dummy

Coefficient on event time dummy

2500
2000
1500
1000
500

300

200

100

0
-14

-7
0
7
Days relative to receipt of stimulus payment

-14

$600 Stimulus Recipients

$1200 Stimulus Recipients

$600 Stimulus Recipients

$1200 Stimulus Recipients

$1800 Stimulus Recipients

$2400 Stimulus Recipients

$1800 Stimulus Recipients

$2400 Stimulus Recipients

Aggregate Debt

Aggregate Debt and Spending

Average Daily Value (dollars)

Average Daily Value (dollars)
600
Coefficient on event time dummy

150
Coefficient on event time dummy

-7
0
7
Days relative to receipt of stimulus payment

100

400

200

0
-14

-7
0
7
Days relative to receipt of stimulus payment

-14

-7
0
7
Days relative to receipt of stimulus payment

$600 Stimulus Recipients

$1200 Stimulus Recipients

$600 Stimulus Recipients

$1200 Stimulus Recipients

$1800 Stimulus Recipients

$2400 Stimulus Recipients

$1800 Stimulus Recipients

$2400 Stimulus Recipients

Cumulative Aggregate Spending

Daily Value (dollars)

% Share of Stimulus Value

1500

% Share

Value (dollars)

.6
1000
.4
.2

500
0
0

-.2
-14

-7
0
7
Days relative to receipt of stimulus payment

-14

-7
0
7
Days relative to receipt of stimulus payment

$600 Stimulus Recipients

$1200 Stimulus Recipients

$600 Stimulus Recipients

$1200 Stimulus Recipients

$1800 Stimulus Recipients

$2400 Stimulus Recipients

$1800 Stimulus Recipients

$2400 Stimulus Recipients

MPC

Figure 3: MPC Heterogeneity: 1st and 2nd Stimulus Payment
Aggregate Deposits
Aggregate Spending

.4
.4
.2
.2
0

1000

2000

3000

Average Monthly Deposits in Jan - Mar

4000

5000

1000

2000

1st Stimulus Payment - $1200 Recipients
2nd Stimulus Payment - $600 Recipients

4000

1st Stimulus Payment - $1200 Recipients
2nd Stimulus Payment - $600 Recipients

Aggregate Savings

Age

MPC

3000

Average Monthly Spending in Jan - Mar

.4
.4
.2
.2
-1000

1000

2000

Average Monthly Savings in Jan - Mar
1st Stimulus Payment - $1200 Recipients
2nd Stimulus Payment - $600 Recipients

3000

Age

1st Stimulus Payment - $1200 Recipients
2nd Stimulus Payment - $600 Recipients

Notes: Plotted are binned scatterplots showing the distribution of individual level MPCs by spending levels, deposit levels, savings levels, and
consumer age for the 1st round of $1,200 stimulus recipients and 2nd round of $600 stimulus recipients in the sample. The bands show the 95%
confidence intervals Aggregate spending includes the sum of all transactions labeled as fees, spending and ATM withdrawals in January—March
2020 for each account. Aggregate Deposits includes all transactions labeled as deposits, loads, and government deposits (including stimulus
payments) in January—March 2020 for each account. The level of savings is calculated as the difference between aggregate deposits and aggregate
spending in January—March 2020.

Figure 4: Effect of 1st Stimulus Payment on Spending/Debt by Merchant Category Code
Grocery Stores & Supermarkets
Automated Cash Disbursements
Average Daily Value (dollars)

Average Daily Value (dollars)
100
Coefficient on event time dummy

Coefficient on event time dummy

200

150

100

0
-14

-7
0
7
Days relative to receipt of stimulus payment

-14

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

Note: Representative firms include Walmart, Amazon, Costco, Kroger, Target.

Discount Stores

Average Daily Value (dollars)

Average Daily Value (dollars)
20
Coefficient on event time dummy

10
Coefficient on event time dummy

Note: Representative firms include 7/11, Bank of America, Cardtronics PLC, Walmart.

Fast Food Restaurants & Bars

-5

15
10
5
0
-5

-14

-7
0
7
Days relative to receipt of stimulus payment

-14

-7
0
7
Days relative to receipt of stimulus payment

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

Note: Representative firms include Dunkin Brands, McDonalds, Wendys, Uber Eats, Starbucks.

Note: Representative firms include Dollar General, Big Lots, Dollar Tree, Ross Stores.

Automobile Transactions

Utilities

Average Daily Value (dollars)

Average Daily Value (dollars)
60
Coefficient on event time dummy

20
Coefficient on event time dummy

-7
0
7
Days relative to receipt of stimulus payment

15
10
5
0
-5
-14

-7
0
7
Days relative to receipt of stimulus payment

0
-14

-7
0
7
Days relative to receipt of stimulus payment

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

Note: Representative firms include Automotive Repair and maintenance

Note: Representative firms include Netflix, Amazon, Comcast, Sprint, T-Mobile, America Movil.

Notes: Data at the account-day level. Plotted are coefficients on event time dummies from regression of aggregate spending or deposits on daysince-event time fixed effects (Model A). Standard errors are clustered two-way by state and calendar date. Time 0 in event time is defined as the
date on which the account received a stimulus payment. The shaded regions are 95% confidence intervals. The aggregate spending in each category
(except utilties which we code as aggregate debt) for each account-date observation is identified using Facteus’s pre-processed merchant category
codes.

Table 1: 1st Stimulus Payment Counterfactuals
Counterfactual

Stimulus
Payment

Cost of Stimulus Bill (USD
Billions)

Recipients
(Millions)

Fraction of
Stimulus
Spent

Total Consumer
Spending
(USD Billions)

Panel A : Policy Counterfactual for different groups of recipients
Actual stimulus bill

1,288

296

230

0.44

130

30K individual income or
60K household income

1,790

296

165

0.48

143

20K individual income or
40K household income

2,309

296

128

0.50

149

10K individual income or
20K household income

3,696

296

0.53

158

All adults receive same
amount

1,146

296

258

0.42

124

Panel B : Policy Counterfactual for same aggregate consumption effect
Actual stimulus bill

1,288

296

230

0.44

130

30K individual income or
60K household income

1,650

273

165

0.48

130

20K individual income or
40K household income

2,034

261

128

0.50

130

10K individual income or
20K household income

3,073

246

0.53

130

All adults receive same
amount

1,215

313

258

0.42

130

Notes: Stimulus payment refers to the payment received by the average adult who received a stimulus payment in each scenario. The cost of the stimulus bill (in USD
billions) is the total amount of the stimulus payments distributed amongst the recipient population. The “Recipients” columns records the total number of adults (in
millions) in the U.S. who would receive a stimulus payment under each scenario, as per the population weights in 2018 ACS. The ‘Fraction of stimulus payment’ is the
weighted average of the share of stimulus payments that the recipients are expected to spend using the MPC distribution estimated from Facteus data. Total consumer
spending is a weighted sum of the stimulus payment multiplied by the fraction of the payment spent by each recipient.

Appendix
Table A1: MPCs in the Literature

Citation

Event

Data

MPC

Life Insurance Dividends to WWII
Veterans
German Restitution Payments

BLS Urban Consumption survey

0.7–0.97
0.17

2001 Federal Income Tax Rebates

Israeli Survey of Family Savings
1957/58
Consume Expenditure Survey

0.33

2001 Federal Income Tax Rebates

Credit Card Data

0.4

2008 Economic Stimulus Payments
2011 Growth Dividend Program in
Singapore
2008 Economic Stimulus Payments
1936 Veterans’ Bonus
Lottery Prizes in Norway

Consumer Expenditure Survey

0.52

Properitary Dataset from a bank
in Singapore
Nielsen Consumer Panel

0.8

Unpredictable Windfalls
1

Bodkin (1959)

Kreinin (1961)

Johnson, Parker, and Souleles (2006)
Agarwal, Liu, and Souleles
(2007)
Parker et al. (2013)

Agarwal and Qian (2014)

Broda and Parker (2014)

8
9

Hausman (2016)
Fagereng, Hold and Natvik
(2019)
Baker et al. (2020)
Ganong et al. (2020)

10
11

Covid-19 stimulus payments
Firm changes in Pay

0.5–0.75

BLS Expenditure Survey, 1936
Norwegian Administrative Panel
Data
Transaction Data from SaverLife
JP Morgan Chase Institute Data

0.7
0.5
0.25–0.35
0.22

Consumer Expenditure Survey

0.2

Consumer Expenditure Survey
Consumer Expenditure Survey
Consumer Expenditure Survey

0.64
0.7
0

Consumer Expenditure Survey

0.2

Predictable Income Shocks
12

Parker (1999)

13
14
15

Souleles (1999)
Souleles (2002)
Hseih (2003)

Stephens (2008)

Changes in Social Security Tax
Witholding
Income Tax Refunds
Reagan Tax Cuts
Payments from Alaska Permanent
Fund
Repayment of Vehicle Loans

Table A2: Summary statistics:1st Stimulus Payment, Non- Recipients
Transaction Type
Bill Pay
Funds Transfer
Utilites/Rent
Deposit
Government Deposit
Load
Atm Withdrawal
Fee
Spend
Aggregate Deposits
Aggregate Spending
Aggregate Debt

No. of Ac- Avg.
counts
7,581
7,581
7,581
7,581
7,581
7,581
7,581
7,581
7,581
7,581
7,581
7,581

Med.

2.4
0.0
79.3
0.0
136.3
53.7
171.5
0.0
763.3
508.3
680.8
119.6
220.6
0.0
7.8
4.3
860.4
632.4
1,615.6 1,140.8
1,088.8 822.0
218.1
66.1

10th % 90th % S.D.

Aggregate Category

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
88.3
237.5
186.6
0.0

Aggregate Debt
Aggregate Debt
Aggregate Debt
Aggregate Deposits
Aggregate Deposits
Aggregate Deposits
Aggregate Spending
Aggregate Spending
Aggregate Spending

0.0
0.0
376.8
509.5
1,546.1
1,962.3
706.5
19.8
1,811.7
3,158.8
2,189.5
510.0

40.1
504.1
227.4
622.8
1,384.1
1,064.6
520.8
14.0
915.2
1,812.8
1,068.6
559.9

Notes: Data at account-month level for January–March 2020. The columns report total no. of accounts and monthly mean, median, 10th/90th percentile, standard deviation of transaction values. Deposits and loads refer to income inflows in the form of paycheck deposits or value loads on payroll cards. Government Spending includes
any account deposit received from the IRS, SSA, or state unemployment insurance offices. Utilities/Rent refer to spend transactions that are associated with MCC codes
4800 - 4999 or 6531. Funds Transfer refers to transactions associated with ACH outgoing, card to card/bank to card transfer, and visa money transfer.

Table A3: Summary statistics: 1st Stimulus Payment, $1,200 Stimulus Recipients
Transaction Type
Bill Pay
Funds Transfer
Utilites/Rent
Deposit
Government Deposit
Load
Atm Withdrawal
Fee
Spend
Aggregate Deposits
Aggregate Spending
Aggregate Debt

No. of Ac- Avg.
counts
13,054
13,054
13,054
13,054
13,054
13,054
13,054
13,054
13,054
13,054
13,054
13,054

Med.

2.7
0.0
52.4
0.0
145.3
63.4
429.5
0.0
221.4
0.0
1,084.4 831.1
286.5
0.0
7.3
3.5
959.2
735.0
1,735.3 1,498.3
1,253.1 1,052.0
200.4
77.5

10th % 90th % S.D.

Aggregate Category

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
124.0
332.3
260.5
0.0

Aggregate Debt
Aggregate Debt
Aggregate Debt
Aggregate Deposits
Aggregate Deposits
Aggregate Deposits
Aggregate Spending
Aggregate Spending
Aggregate Spending

0.0
36.1
0.0
281.0
380.7
237.3
1,734.7 933.4
933.0
627.1
2,603.3 1,256.5
925.5
510.0
19.4
11.0
2,054.9 878.6
3,245.5 1,354.0
2,473.3 970.8
508.1
376.2

Table A4: Summary statistics: 1st Stimulus Payment, Non-Recipients & $1,200 Stimulus Recipients
Transaction Type
Bill Pay
Funds Transfer
Utilites/Rent
Deposit
Government Deposit
Load
Atm Withdrawal
Fee
Spend
Aggregate Deposits
Aggregate Spending
Aggregate Debt

No. of Ac- Avg.
counts
20,635
20,635
20,635
20,635
20,635
20,635
20,635
20,635
20,635
20,635
20,635
20,635

Med.

2.6
0.0
62.3
0.0
142.0
60.0
334.7
0.0
420.5
0.0
936.1
579.1
262.3
0.0
7.5
3.7
922.9
692.1
1,691.3 1,360.1
1,192.7 954.2
206.9
73.6

10th % 90th % S.D.

Aggregate Category

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
111.9
304.3
232.8
0.0

Aggregate Debt
Aggregate Debt
Aggregate Debt
Aggregate Deposits
Aggregate Deposits
Aggregate Deposits
Aggregate Spending
Aggregate Spending
Aggregate Spending

0.0
0.0
379.8
1,459.2
1,180.1
2,409.1
851.1
19.6
1,981.2
3,221.9
2,397.3
508.9

37.6
378.8
233.8
842.1
1,010.4
1,205.4
515.0
12.2
893.5
1,539.6
1,011.0
452.5

Table A5: Summary statistics: 1st Stimulus Payment, $1,700 Stimulus Recipients
Transaction Type
Bill Pay
Funds Transfer
Utilites/Rent
Deposit
Government Deposit
Load
Atm Withdrawal
Fee
Spend
Aggregate Deposits
Aggregate Spending
Aggregate Debt

No. of Ac- Avg.
counts
5,442
5,442
5,442
5,442
5,442
5,442
5,442
5,442
5,442
5,442
5,442
5,442

3.4
89.3
211.4
506.6
581.8
1,657.0
363.6
8.7
1,533.1
2,745.5
1,905.4
304.1

Med.

10th % 90th % S.D.

Aggregate Category

0.0
0.0
101.4
0.0
0.0
1,001.3
0.0
5.5
1,057.8
1,810.2
1,434.6
125.5

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
165.1
200.3
286.6
0.0

Aggregate Debt
Aggregate Debt
Aggregate Debt
Aggregate Deposits
Aggregate Deposits
Aggregate Deposits
Aggregate Spending
Aggregate Spending
Aggregate Spending

0.0
63.1
538.0
1,832.8
2,652.0
4,817.5
1,100.3
22.1
3,637.7
6,909.7
4,257.8
751.3

44.7
423.8
339.2
1,349.9
1,634.7
2,455.7
775.6
10.9
1,479.6
2,856.4
1,639.7
556.4

Table A6: Summary statistics: 1st Stimulus Payment, $2,200 Stimulus Recipients
Transaction Type

No. of Ac- Avg.
counts

Bill Pay
Funds Transfer
Utilites/Rent
Deposit
Government Deposit
Load
Atm Withdrawal
Fee
Spend
Aggregate Deposits
Aggregate Spending
Aggregate Debt

3,333
3,333
3,333
3,333
3,333
3,333
3,333
3,333
3,333
3,333
3,333
3,333

4.2
134.4
259.1
570.9
761.9
2,150.1
416.2
9.4
1,980.6
3,483.0
2,406.2
397.8

Med.

10th % 90th % S.D.

Aggregate Category

0.0
0.0
127.1
0.0
0.0
1,066.8
0.0
5.8
1,303.3
1,947.2
1,721.8
159.2

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
193.8
160.6
305.1
0.0

Aggregate Debt
Aggregate Debt
Aggregate Debt
Aggregate Deposits
Aggregate Deposits
Aggregate Deposits
Aggregate Spending
Aggregate Spending
Aggregate Spending

0.0
102.3
648.4
1,853.0
1,991.8
7,349.4
1,176.4
24.0
4,842.5
9,626.7
5,616.3
957.9

44.4
656.3
401.1
1,718.4
2,242.7
2,965.0
963.8
12.0
1,954.6
3,585.8
2,164.4
789.7

Table A7: Summary statistics: 1st Stimulus Payment, $2,400 Stimulus Recipients
Transaction Type
Bill Pay
Funds Transfer
Utilites/Rent
Deposit
Government Deposit
Load
Atm Withdrawal
Fee
Spend
Aggregate Deposits
Aggregate Spending
Aggregate Debt

No. of Ac- Avg.
counts
632
632
632
632
632
632
632
632
632
632
632
632

3.7
106.3
225.6
481.6
389.3
1,370.3
277.4
7.0
1,283.0
2,241.2
1,567.4
335.5

Med.
0.0
0.0
138.2
0.0
0.0
983.2
0.0
3.0
967.9
1,818.2
1,269.7
175.8

10th % 90th %
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
175.4
472.7
308.8
0.0

0.0
106.8
556.0
2,088.9
1,322.7
3,257.9
974.4
19.2
2,724.9
4,405.9
3,157.2
768.1

S.D.

Aggregate Category

31.5
540.7
288.5
1,239.3
1,029.0
1,708.2
608.3
9.5
1,255.7
1,996.8
1,358.2
616.9

Aggregate Debt
Aggregate Debt
Aggregate Debt
Aggregate Deposits
Aggregate Deposits
Aggregate Deposits
Aggregate Spending
Aggregate Spending
Aggregate Spending

Table A8: Summary statistics: 2nd Stimulus Payment, $600 Stimulus Recipients
Transaction Type
Bill Pay
Funds Transfer
Utilites/Rent
Deposit
Government Deposit
Load
Atm Withdrawal
Fee
Spend
Aggregate Deposits
Aggregate Spending
Aggregate Debt

No. of Ac- Avg.
counts
25,859
25,859
25,859
25,859
25,859
25,859
25,859
25,859
25,859
25,859
25,859
25,859

2.6
41.2
138.9
461.7
252.2
959.4
300.6
7.1
855.0
1,673.3
1,162.6
182.7

Med.

10th % 90th %

S.D.

Aggregate Category

0.0
0.0
62.8
0.0
0.0
666.6
21.8
3.5
692.3
1,436.3
971.5
73.8

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
142.1
511.9
304.5
0.0

34.9
298.6
223.3
937.8
612.6
1,191.0
530.8
10.1
730.9
1,259.7
870.2
377.5

Aggregate Debt
Aggregate Debt
Aggregate Debt
Aggregate Deposits
Aggregate Deposits
Aggregate Deposits
Aggregate Spending
Aggregate Spending
Aggregate Spending

0.0
0.0
361.4
1,807.6
983.6
2,417.4
972.5
18.4
1,733.0
3,043.6
2,218.6
446.1

Table A9: Summary statistics: 2nd Stimulus Payment, $1,200 Stimulus Recipients
Transaction Type
Bill Pay
Funds Transfer
Utilites/Rent
Deposit
Government Deposit
Load
Atm Withdrawal
Fee
Spend
Aggregate Deposits
Aggregate Spending
Aggregate Debt

No. of Ac- Avg.
counts
6,524
6,524
6,524
6,524
6,524
6,524
6,524
6,524
6,524
6,524
6,524
6,524

Med.

4.4
0.0
73.5
0.0
197.6
104.2
489.6
0.0
751.8
0.0
1,232.3 965.2
369.1
39.6
6.8
3.1
1,220.1 903.4
2,473.6 1,783.4
1,596.0 1,254.5
275.5
121.8

10th % 90th % S.D.

Aggregate Category

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
182.2
462.0
333.7
0.0

Aggregate Debt
Aggregate Debt
Aggregate Debt
Aggregate Deposits
Aggregate Deposits
Aggregate Deposits
Aggregate Spending
Aggregate Spending
Aggregate Spending

0.0
0.0
500.7
1,943.6
4,288.1
2,906.1
1,118.0
17.6
2,637.3
6,103.8
3,318.5
631.0

52.8
464.7
281.4
1,038.8
1,813.7
1,382.7
718.9
11.6
1,137.8
2,229.3
1,329.4
554.1

Table A10: Summary statistics: 2nd Stimulus Payment, $1,800 Stimulus Recipients
Transaction Type
Bill Pay
Funds Transfer
Utilites/Rent
Deposit
Government Deposit
Load
Atm Withdrawal
Fee
Spend
Aggregate Deposits
Aggregate Spending
Aggregate Debt

No. of Ac- Avg.
counts
3,438
3,438
3,438
3,438
3,438
3,438
3,438
3,438
3,438
3,438
3,438
3,438

Med.

4.6
0.0
106.1
0.0
218.3
119.2
490.5
0.0
1,116.4
0.0
1,302.4 980.9
416.8
43.0
7.2
3.6
1,443.4 987.6
2,909.3 1,790.7
1,867.4 1,316.9
329.0
140.5

10th % 90th % S.D.

Aggregate Category

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
198.6
396.9
332.2
0.0

Aggregate Debt
Aggregate Debt
Aggregate Debt
Aggregate Deposits
Aggregate Deposits
Aggregate Deposits
Aggregate Spending
Aggregate Spending
Aggregate Spending

0.0
0.0
538.2
1,940.8
6,528.2
3,041.2
1,185.7
18.2
3,311.6
8,444.7
4,249.2
710.8

55.1
610.6
321.3
1,087.9
2,667.0
1,538.7
893.7
10.7
1,450.5
3,069.3
1,753.2
700.1

Table A11: Summary statistics: 2nd Stimulus Payment, $2,400 Stimulus Recipients
Transaction Type
Bill Pay
Funds Transfer
Utilites/Rent
Deposit
Government Deposit
Load
Atm Withdrawal
Fee
Spend
Aggregate Deposits
Aggregate Spending
Aggregate Debt

No. of Ac- Avg.
counts
1,653
1,653
1,653
1,653
1,653
1,653
1,653
1,653
1,653
1,653
1,653
1,653

4.3
107.1
250.0
499.9
1,035.6
1,410.7
429.0
6.9
1,484.0
2,946.2
1,919.9
361.3

Med.

10th % 90th % S.D.

Aggregate Category

0.0
0.0
138.5
0.0
0.0
1,041.4
42.3
3.6
1,017.0
1,832.6
1,364.8
163.7

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
195.2
421.5
341.4
0.0

Aggregate Debt
Aggregate Debt
Aggregate Debt
Aggregate Deposits
Aggregate Deposits
Aggregate Deposits
Aggregate Spending
Aggregate Spending
Aggregate Spending

0.0
0.0
587.9
2,008.3
6,092.9
3,285.4
1,231.3
17.6
3,347.6
8,497.9
4,278.6
757.6

39.7
735.3
360.3
1,169.9
2,721.7
1,701.7
942.0
9.1
1,503.2
3,194.7
1,824.9
833.2

Table A12: Firm-Specific Summary statistics: 1st Stimulus Payment, $1,200 Stimulus Recipients

Company

Walmart
Kroger
Seven & i Holdings Co.
Amazon.com Inc.
Dollar General Corp.
AT&T
Cardtronics, PLC
Sprint Corporation
Royal Dutch Shell
Dollar Tree Inc
Verizon Communications Inc.
Yum! Brands Inc.
McDonald’s Corp
Bank of America Corp
Apple Inc.
Comcast
Food Delivery
Walgreens Boots Alliance Inc
JPMorgan Chase Co
Alimentation Couche-Tard Inc Class B
Restaurant Brans International Inc
CVS Health Corp.
Chevron Corporation
Starbucks Corp
Other

Average Monthly
Spending

Average No.
of
monthly Transactions

Change
in
Spending after
stimulus
receipt

Change in Spending as a fraction of
stimulus payment

162.51
33.59
26.16
25.92
22.01
12.50
12.17
10.66
9.18
8.66
8.37
7.61
7.35
6.48
5.79
5.48
5.39
4.83
4.50
4.02
3.50
2.62
2.50
1.23
860.04

2.64
0.85
0.77
1.05
1.49
0.15
0.12
0.15
0.72
0.57
0.07
0.56
0.75
0.04
0.51
0.05
0.21
0.22
0.02
0.32
0.34
0.14
0.18
0.12
28.40

93.95
4.38
8.10
20.90
5.12
2.67
4.85
0.60
1.38
2.43
2.25
2.38
1.78
2.27
2.60
0.97
4.91
1.73
1.52
0.59
0.94
0.82
0.41
0.17
379.02

0.08
0.00
0.01
0.02
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.32

Notes: Data at account-month level. Transaction amounts are averages for January—March 2020. Spending includes the sum of all transactions labeled as bill pay, fees,
spending and ATM withdrawals in January—March 2020 for each account.

Table A13: Merchant Group Specific Summary statistics: 1st Stimulus Payment, $1,200 Stimulus Recipients
Spending/Debt Category

Grocery Stores & Supermarkets
Automated Cash Disbursements
Utilities
Automobile Transactions
Fast Food, Restaurants & Bars
Discount Stores
Other Spending

Average
Monthly
Spending

Average No. of
monthly Transactions

Change in
Spending
after stimulus receipt

Change
in
Spending as
a
fraction
of stimulus
payment

234.88
224.25
133.07
111.84
102.60
38.49
541.02

5.11
2.88
2.24
7.01
8.40
1.71
15.33

101.18
96.72
57.59
40.08
33.70
22.53
252.51

0.08
0.08
0.05
0.03
0.03
0.02
0.21

Notes: Data at account-month level. Transaction amounts are averages for January—March 2020. Spending in each category is identified using Facteus’s pre-processed
merchant category codes. Spending includes the sum of all transactions labeled as fees, spending and ATM withdrawals in January—March 2020 for each account.

Figure A1: Distribution of Government Deposits
1st Stimulus Payment
40

Percent

0
0

400

800

1200

1600

2000

2400

2800

Value of Government Deposits

April 1st - April 9th

April 10th - April 15th

2nd Stimulus Payment
60

Percent

600

1200

1800

2400

3000

Value of Government Deposits

Notes: The graph in panel A shows the distribution of government deposits made between April 10th and April 15th in our sample of accounts from Facteus overlaid
on the distribution of government deposits made between April 1st and April 9th. The graph in panel B shows the distribution of government deposits made between
December 17th , 2020 and January 29th, 2021. The values shown are trimmed at the 90th percentile of the distribution.

Figure A2: Distribution of Changes in Consumer Spending: 1st Stimulus Payment, $1,200 Recipients
Stimulus Payment
5

Percent

-2

-1.5

-1

-.5

MPC

1.5

Placebo Stimulus Payment
15

Percent

-2

-1.5

-1

-.5

MPC

Notes: Histogram shows distribution of changes in consumer spending in the two weeks after stimulus receipt relative to the two weeks before stimulus receipt after
winsorizing the values to [-2,2]. Panel A shows MPCs calculated as the difference between total spending in the two weeks following stimulus receipt and the two weeks
preceding stimulus receipt for all accounts in our data that received a 1st stimulus payment of $1,200 between April 10th and April 15th 2020. Panel B shows MPCs
calculated similarly but relative to the placebo stimulus date identified as 21 days prior to the date of actual stimulus receipt.

Figure A3: Time-Trend in Account Balance
1st Stimulus Payment Recipients

Average Cumulative Daily Value ($)

12000

9000

6000

3000

0
Jan2020

Feb2020

Mar2020

Apr2020

May2020

Jun2020

Aggregate Deposits
Aggregate Spending + Debt
Savings

2nd Stimulus Payment Recipients

Average Cumulative Daily Value ($)

12000

9000

6000

3000

0
Jan2020

Feb2020

Mar2020

Apr2020

May2020

Jun2020

Aggregate Deposits
Aggregate Spending + Debt
Savings

Notes: The vertical line in figure in panel A marks April 15th, 2020 when a majority of the accounts in the sample receive stimulus payment deposit. Out of all the accounts
that received a $1,200 stimulus payment between April 10th and April 15th, 60% receive the stimulus payment deposit on April 15th. The large increase in deposits,
spending, and savings in February is driven by EITC deposits—the federal PATH Act requires that the IRS wait to deposit EITC tax credits into accounts until February
15th or later, and the deposits take around 10 days to be deposited into bank accounts, so we see large increases in governmental deposits and spending immediately after
that date.

Figure A4: Effect of 1st Stimulus Payment on Spending and Deposits: Calendar Time
Aggregate Deposits
Aggregate Spending
400

Average Daily Value ($)

1500

1000

500

0
11Mar20

300

200

100

0
1Apr20

22Apr20

13May20

3Jun20

11Mar20

13May20

3Jun20

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

Aggregate Spending & Debt

400

Average Daily Value ($)

22Apr20

$1700 Stimulus Recipients

Aggregate Debt

0
11Mar20

1Apr20

$1200 Stimulus Recipients

300

200

100

0
1Apr20

22Apr20

13May20

3Jun20

11Mar20

1Apr20

22Apr20

13May20

3Jun20

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

Notes: Plotted are time-series of aggregate per-consumer spending and deposits for March 15th - June 9th, 2020 . We separately plot spending and deposits for $1,200
, $1,700, $2,200, and $2,400 stimulus recipients. The time-series shows post-March 15th data to avoid irregular spending and deposits patterns seen in the data in late
February and early March due to EITC and tax refund receipt. Aggregate spending includes transactions labeled as fees, spending and ATM withdrawals. Aggregate
Deposits include all transactions labeled as deposits, loads, and government deposits (including stimulus payments).Aggregate debt includes all transactions labeled as
funds transfer, bill pay, or associated with utilities and rental payments.

Figure A5: Effect of 2nd Stimulus Payment on Spending and Deposits: Calendar Time
Aggregate Deposits
Aggregate Spending
250

Average Daily Value ($)

1500

1000

500

0
1Nov20

200

150

100

0
1Dec20

1Jan21

1Feb21

1Nov20

1Feb21

$600 Stimulus Recipients

$1200 Stimulus Recipients

$1800 Stimulus Recipients

$2400 Stimulus Recipients

$1800 Stimulus Recipients

$2400 Stimulus Recipients

Aggregate Spending & Debt

100

400

Average Daily Value ($)

1Jan21

$1200 Stimulus Recipients

Aggregate Debt

0
1Nov20

1Dec20

$600 Stimulus Recipients

300

200

100

0
1Dec20

1Jan21

1Feb21

1Nov20

1Dec20

1Jan21

1Feb21

$600 Stimulus Recipients

$1200 Stimulus Recipients

$600 Stimulus Recipients

$1200 Stimulus Recipients

$1800 Stimulus Recipients

$2400 Stimulus Recipients

$1800 Stimulus Recipients

$2400 Stimulus Recipients

Notes: Plotted are time-series of aggregate per-consumer spending and deposits for November 1st, 2020 - January 31st, 2021. We separately plot spending, deposits,
and debt payments for $600 , $1,200, $1,800, and $2,400 stimulus recipients. Aggregate spending includes transactions labeled as fees, spending and ATM withdrawals.
Aggregate Deposits include all transactions labeled as deposits, loads, and government deposits (including stimulus payments). Aggregate debt includes all transactions
labeled as funds transfer, bill pay, or associated with utilities and rental payments.

MPC

Figure A6: Total Stimulus Use Heterogeneity: 1st and 2nd Stimulus Payment
Aggregate Deposits
Aggregate Spending

.2
0

1000

2000

3000

Average Monthly Deposits in Jan - Mar

4000

5000

1000

2000

1st Stimulus Payment - $1200 Recipients
2nd Stimulus Payment - $600 Recipients

4000

1st Stimulus Payment - $1200 Recipients
2nd Stimulus Payment - $600 Recipients

Aggregate Savings

Age

MPC

3000

Average Monthly Spending in Jan - Mar

.4
.4
.2
.2
-1000

1000

2000

Average Monthly Savings in Jan - Mar
1st Stimulus Payment - $1200 Recipients
2nd Stimulus Payment - $600 Recipients

3000

Age

1st Stimulus Payment - $1200 Recipients
2nd Stimulus Payment - $600 Recipients

Notes: Plotted are binned scatterplots showing the distribution of individual level spending and debt payments by spending levels, deposit levels, savings levels, and
consumer age for the 1st round of $1,200 stimulus recipients and 2nd round of $600 stimulus recipients in the sample. The bands show the 95% confidence intervals
Aggregate spending includes the sum of all transactions labeled as fees, spending and ATM withdrawals in January—March 2020 for each account. Aggregate Deposits
includes all transactions labeled as deposits, loads, and government deposits (including stimulus payments) in January—March 2020 for each account. The level of savings
is calculated as the difference between aggregate deposits and aggregate spending in January—March 2020.

Figure A7: Effect of 1st Stimulus Payment on Deposits, Excluding Stimulus Value : (Model A)

Average Daily Value (dollars)

Coefficient on event time dummy

200

100

-100

-200
-14

-7
0
7
Days relative to receipt of stimulus payment

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

Figure A8: Effect of 1st Stimulus Payment on Spending and Deposits: Extended Event Time
Aggregate Deposits
Average Daily Value (dollars)

Coefficient on event time dummy

2500
2000
1500
1000
500
0
-28

-21

-14
-7
0
7
14
Days relative to receipt of stimulus payment

$1200 Stimulus Recipients
$2200 Stimulus Recipients

$1700 Stimulus Recipients
$2400 Stimulus Recipients

Aggregate Spending
Average Daily Value (dollars)

Coefficient on event time dummy

400

300

200

100

0
-28

-21

-14
-7
0
7
14
Days relative to receipt of stimulus payment

$1200 Stimulus Recipients
$2200 Stimulus Recipients

$1700 Stimulus Recipients
$2400 Stimulus Recipients

Notes: Data at the account-day level. Plotted are coefficients on event time dummies from regression of aggregate spending or deposits on day-since-event time fixed
effects (Model A). The time - window is extended to 28 days before and after stimulus receipt. Standard errors are clustered two-way by state and calendar date. Time
0 in event time is defined as the date on which the account received a stimulus payment. The shaded regions are 95% confidence intervals. Aggregate spending includes
transactions labeled as fees, spending and ATM withdrawals. Aggregate Deposits includes transactions labeld as deposits, loads, government deposits and stimulus
payment.

Figure A9: Effect of 1st Stimulus Payment on Spending and Deposits: Event Time (Models B & C)
Model B
Aggregate Deposits
Aggregate Spending
Average Daily Value (dollars)

Average Daily Value (dollars)
400
Coefficient on event time dummy

Coefficient on event time dummy

2500
2000
1500
1000
500

300
200
100
0

0
-100
-14

-7
0
7
Days relative to receipt of stimulus payment

-14

-7
0
7
Days relative to receipt of stimulus payment

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

Notes: Data at the account-day level. Plotted are coefficients on event time dummies from a regression which also includes individuals and date-by-state fixed effects
(Model B). Standard errors are clustered two-way by state and date. Time 0 in event time is defined as the date on which stimulus payment is received by the account. The
shaded regions are 95% confidence intervals. Aggregate spending includes transactions labeled as fees, spending and ATM withdrawals. Aggregate Deposits includes all
transactions labeled as deposits, loads, and government deposits (including stimulus payment).

Model C
Aggregate Deposits

Aggregate Spending

Average Daily Value (dollars)

Average Daily Value (dollars)
400
Coefficient on event time dummy

Coefficient on event time dummy

2500
2000
1500
1000
500
0

300

200

100

0
-14

-7
0
7
Days relative to receipt of stimulus payment

-14

-7
0
7
Days relative to receipt of stimulus payment

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

Notes: Data at the account-day level. Plotted are coefficients on event time dummies from a regression of residual spending on days-since-event time (Model C). Residual
spending is calculated as the difference between realized consumer spending and predicted consumer spending from a regression of spending on person-by-day-of-theweek fixed effects in January—March, 2020. Standard errors are clustered by state and date. Time 0 in event time is defined as the date on which stimulus payment is
received by the account. Aggregate spending includes transactions labeled as bill pay, fees, spending and ATM withdrawals. Aggregate Deposits includes all transactions
labeled as deposits, loads, and government deposits (including stimulus payments).

Figure A10: Effect of 1st Stimulus Payment on Aggregate Spending: $1,200 Recipients

Average Daily Value (dollars)

Coefficient on event time dummy

200

150

100

-50
-14

-7
0
7
Days relative to receipt of stimulus payment
Model A

Model B

14
Model C

Notes: Data at the account-day level. Plotted are coefficients on event time dummies from regression of aggregate spending on day-since-event time fixed effects from
Models A,B and C. Standard errors are clustered two-way by state and calendar date. Time 0 in event time is defined as the date on which the account received a stimulus
payment. The shaded regions are 95% confidence intervals. Aggregate spending includes transactions labeled as fees, spending and ATM withdrawals.

Figure A11: Distribution of Changes in Consumer Spending - 1st Stimulus Payment
$1200 Stimulus Recepients
$1700 Stimulus Recipients
5

Percent

2
1
1

-2

-1.5

-1

-.5

MPC

1.5

-2

-1.5

$2200 Stimulus Recipients

MPC

1.5

Percent

-.5

$2400 Stimulus Recipients

-1

-2

-1.5

-1

-.5

MPC

1.5

-2

-1.5

-1

-.5

MPC

Notes: Histogram shows distribution of changes in consumer spending in the two weeks after stimulus receipt relative to the two weeks before stimulus receipt after
winsorizing the values to [-2,2]. MPCs are calculated as the difference between total spending in the two weeks following stimulus receipt and the two weeks preceding
stimulus receipt for all accounts in our data that received a stimulus payment between April 10th and April 15th 2020.

Figure A12: Distribution of Changes in Consumer Spending - 2nd Stimulus Payment
$600 Stimulus Recepients
$1200 Stimulus Recipients
15
10

Percent

10
6

-2

-1.5

-1

-.5

MPC

1.5

-2

-1.5

$1800 Stimulus Recipients

-1

-.5

MPC

1.5

$2400 Stimulus Recipients
8

Percent

-2

-1.5

-1

-.5

MPC

1.5

-2

-1.5

-1

-.5

MPC

Notes: Histogram shows distribution of changes in consumer spending in the two weeks after stimulus receipt relative to the two weeks before stimulus receipt after
winsorizing the values to [-2,2]. MPCs are calculated as the difference between total spending in the two weeks following stimulus receipt and the two weeks preceding
stimulus receipt for all accounts in our data that we identify as receiving a 2nd stimulus payment between December 28th, 2020 and January 4th, 2021.

Figure A13: Effect of 1st Stimulus Payment on Spending at Specific Companies: Event Time
Average Daily Value: Walmart (dollars)

Average Daily Value: Amazon.com Inc. (dollars)
15
Coefficient on event time dummy

Coefficient on event time dummy

200

150

100

-5
-14

-7
0
7
Days relative to receipt of stimulus payment

-14

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

Average Daily Value: Seven & i Holdings Co. (dollars)

Average Daily Value: AT&T (dollars)
6
Coefficient on event time dummy

8
Coefficient on event time dummy

-7
0
7
Days relative to receipt of stimulus payment

6
4
2
0
-2

-2
-14

-7
0
7
Days relative to receipt of stimulus payment

-14

-7
0
7
Days relative to receipt of stimulus payment

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

Average Daily Value: Dollar General Corp. (dollars)

Average Daily Value: Sprint Corporation (dollars)
3

Coefficient on event time dummy

-2

-1

-4
-14

-7
0
7
Days relative to receipt of stimulus payment

-14

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

Average Daily Value: Verizon Communications Inc. (dollars)

Average Daily Value: Comcast (dollars)

8
Coefficient on event time dummy

Coefficient on event time dummy

-7
0
7
Days relative to receipt of stimulus payment

-2

6
4
2
0
-2

-14

-7
0
7
Days relative to receipt of stimulus payment

-14

-7
0
7
Days relative to receipt of stimulus payment

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$1200 Stimulus Recipients

$1700 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

$2200 Stimulus Recipients

$2400 Stimulus Recipients

Notes: Data at the account-day level. Plotted are coefficients on event time dummies from regression of aggregate spending or deposits on day-since-event time fixed
effects (Model A). Standard errors are clustered two-way by state and calendar date. Time 0 in event time is defined as the date on which the account received a stimulus
payment. The shaded regions are 95% confidence intervals. The aggregate spending at each merchant for each account-date is identified using Facteus’s pre-processed
merchant tickers. The biggest companies are identified as firms at which the aggregate spending in January–March 2020 by all accounts in the sample exceeds $500,000.

Figure A14: Effect of 2nd Stimulus Payment on Spending/Debt by Merchant Category Code
Grocery Stores & Supermarkets
Automated Cash Disbursements
Average Daily Value (dollars)
60

150

Coefficient on event time dummy

Average Daily Value (dollars)

100

-20
-14

-7
0
7
Days relative to receipt of stimulus payment

-14

$1200 Stimulus Recipients

$600 Stimulus Recipients

$1200 Stimulus Recipients

$1800 Stimulus Recipients

$2400 Stimulus Recipients

$1800 Stimulus Recipients

$2400 Stimulus Recipients

Note: Representative firms include 7/11, Bank of America, Cardtronics PLC, Walmart.

Fast Food Restaurants & Bars

Discount Stores

Average Daily Value (dollars)

Average Daily Value (dollars)
15
Coefficient on event time dummy

15
Coefficient on event time dummy

$600 Stimulus Recipients

Note: Representative firms include Walmart, Amazon, Costco, Kroger, Target.

-5

-5
-14

-7
0
7
Days relative to receipt of stimulus payment

-14

-7
0
7
Days relative to receipt of stimulus payment

$600 Stimulus Recipients

$1200 Stimulus Recipients

$600 Stimulus Recipients

$1200 Stimulus Recipients

$1800 Stimulus Recipients

$2400 Stimulus Recipients

$1800 Stimulus Recipients

$2400 Stimulus Recipients

Note: Representative firms include Dunkin Brands, McDonalds, Wendys, Uber Eats, Starbucks.

Note: Representative firms include Dollar General, Big Lots, Dollar Tree, Ross Stores.

Automobile Transactions

Utilities

Average Daily Value (dollars)

Average Daily Value (dollars)
60
Coefficient on event time dummy

15
Coefficient on event time dummy

-7
0
7
Days relative to receipt of stimulus payment

0
-14

-7
0
7
Days relative to receipt of stimulus payment

0
-14

-7
0
7
Days relative to receipt of stimulus payment

$600 Stimulus Recipients

$1200 Stimulus Recipients

$600 Stimulus Recipients

$1200 Stimulus Recipients

$1800 Stimulus Recipients

$2400 Stimulus Recipients

$1800 Stimulus Recipients

$2400 Stimulus Recipients

Note: Representative firms include Automotive Repair and maintenance

Note: Representative firms include Netflix, Amazon, Comcast, Sprint, T-Mobile, America Movil.

Notes: Data at the account-day level. Plotted are coefficients on event time dummies from regression of aggregate spending or deposits on day-since-event time fixed
effects (Model A). Standard errors are clustered two-way by state and calendar date. Time 0 in event time is defined as the date on which the account received a
stimulus payment. The shaded regions are 95% confidence intervals. The aggregate spending in each category (except utilties which we code as aggregate debt) for each
account-date observation is identified using Facteus’s pre-processed merchant category codes.

Figure A15: Effect of 1st Stimulus Payment on Spending and Deposits: Placebo Event Time
Aggregate Deposits
Average Daily Value (dollars)

Coefficient on event time dummy

2500
2000
1500
1000
500
0
-14

-7
0
7
Days relative to receipt of stimulus payment
$1200 Stimulus Recipients
$2200 Stimulus Recipients

$1700 Stimulus Recipients
$2400 Stimulus Recipients

Aggregate Spending
Average Daily Value (dollars)

Coefficient on event time dummy

400

200

-200
-14

-7
0
7
Days relative to receipt of stimulus payment
$1200 Stimulus Recipients
$2200 Stimulus Recipients

$1700 Stimulus Recipients
$2400 Stimulus Recipients

Notes: Data at the account-day level. Plotted are coefficients on event time dummies from regression of aggregate spending or deposits on day-since-event time fixed
effects (Model A). Standard errors are clustered two-way by state and calendar date. The shaded regions are 95% confidence intervals. Time 0 in placebo event time
is defined as 21 days before date on which the account received a stimulus payment. We avoid using an earlier placebo stimulus date to avoid confounding the results
with spending responses to EITC and tax refund receipt in late February and early March. Aggregate spending includes transactions labeled as fees, spending and ATM
withdrawals. Aggregate Deposits includes transactions labeld as deposits, loads, government deposits and stimulus payment.

Figure A16: Geographic Distribution of 1st Stimulus Payment Recipients in Facteus vs. ACS

% Share of Total Stimulus Recipients in Facteus

IL
PA

0
0

% Share of Total Stimulus Recipients in ACS, 2018
Notes: The scatterplot shows the % of share of total 1st stimulus payment recipients in each state in the Facteus sample vs. ACS, 2018. The stimulus recipients in ACS
are identified using total personal income and the IRS income eligibility criteria for receipt of stimulus payment.

Figure A17: Age & Income Distribution of 1st Stimulus Payment Recipients: Facteus vs. ACS
Income Distribution
Facteus

ACS, 2018

Percent

20000

40000

60000

80000

100000

20000

40000

Annual Income ($)

60000

80000

100000

100

Annual Income ($)

Age Distribution
ACS, 2018
20

Percent

Facteus
20

60
Age

100

60
Age

Notes: The histograms show the distribution of age and income of 1st stimulus payment recipients in Facteus vs. ACS, 2018.The stimulus recipients in ACS are identified
using total personal income and the IRS income eligibility criteria for receipt of stimulus payment. The histogram plots the weighted total personal income of the
identified stimulus recipients in ACS. To calculate the total personal income of stimulus recipients in the Facteus sample, we multiply the total aggregate deposits value
for January–March,2020 for each account by four to get a value for annual income.

File and source

File: heterogeneity-in-the-mpc-evidence-from-covid-19-stimulus-payments.pdf
Size: 1,307,220 bytes
SHA-256: 77ce9c85ca5400a36daad36edce8c2eb8a6adb620020c05dcaf208d89aa05230
Our copy: heterogeneity-in-the-mpc-evidence-from-covid-19-stimulus-payments.pdf
Original: www.chicagofed.org

Heterogeneity in the Marginal Propensity to Consume

Summary

Full text

File and source

Related records