Did the Paycheck Protection Program Hit the Target? (Granja, Makridis, Yannelis and Zwick)
Summary
A paper, Did the Paycheck Protection Program Hit the Target?, by João Granja, Christos Makridis, Constantine Yannelis and Eric Zwick, dated July 3rd 2020. The abstract states that the authors use new data on the first round of PPP loans and high-frequency employment data to consider program targeting. It states that they do not find evidence that funds flowed to areas more adversely affected by the pandemic, and that the top-4 banks account for 36% of total pre-policy small business loans but disbursed less than 3% of all PPP loans in the first round. It also states that they do not find evidence that the PPP had a substantial effect on local economic outcomes during the first round. The authors thank the Small Business Administration, Womply and Homebase for providing data.
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
- Issuer
- SBA materials
- Document type
- Zwick Sba V2 Ppp Hit Target
- Case
- Zwick Sba V2 Ppp Hit Target
Cite this document
“Did the Paycheck Protection Program Hit the Target? (Granja, Makridis, Yannelis and Zwick).”
Original at www.ericzwick.com. Archived at Pandemic Darlings, https://pandemicdarlings.com/source-documents/zwick-sba-v2-ppp-hit-target/
, SHA-256 a7dd7da0b2ada269cded15508b8fa67214a750d5d9f6730364fea46da2bcc143.Full text
Did the Paycheck Protection Program Hit the Target?∗
João Granja Christos Makridis
Chicago Booth MIT Sloan
joaogranja@chicagobooth.edu makridis@mit.edu
Constantine Yannelis Eric Zwick
Chicago Booth and NBER Chicago Booth and NBER
constantine.yannelis@chicagobooth.edu ezwick@chicagobooth.edu
July 3rd 2020
Abstract
This paper takes an early look at the Paycheck Protection Program (PPP), a large and
novel small business support program that was part of the initial policy response to the
COVID-19 pandemic. We use new data on the distribution of the first round of PPP loans
and high-frequency micro-level employment data to consider two dimensions of program
targeting. First, we do not find evidence that funds flowed to areas more adversely af-
fected by the economic effects of the pandemic, as measured by declines in hours worked
or business shutdowns. If anything, funds flowed to areas less hard hit. Second, we find
significant heterogeneity across banks in terms of disbursing PPP funds, which does not
only reflect differences in underlying loan demand. The top-4 banks alone account for
36% of total pre-policy small business loans, but disbursed less than 3% of all PPP loans
in the first round. Areas that were significantly more exposed to low-PPP banks received
much lower loan allocations. We do not find evidence that the PPP had a substantial effect
on local economic outcomes—including declines in hours worked, business shutdowns, ini-
tial unemployment insurance claims, and small business revenues—during the first round
of the program. Firms appear to use first round funds to build up savings and meet loan
and other commitments, which points to possible medium-run impacts. As data become
available, we will continue to study employment and establishment responses to the pro-
gram and the impact of PPP support on the economic recovery. Measuring these responses
is critical for evaluating the social insurance value of the PPP and similar policies.
∗
We thank seminar participants at the University of Chicago Booth School of Business, the Stigler Center
Economic Effects of COVID-19 Workshop, the Federal Reserve Bank of New York, the Federal Reserve Bank of
Philadelphia and the Bank of Portugal as well as Scott Baker, Sylvain Catherine, Jediphi Cabal, Raj Chetty, Mike
Faulkender, Sam Hanson, Steve Kaplan, Mike Minnis, Tiago Pinheiro, Larry Schmidt, Adi Sunderam, and Luigi
Zingales for comments. Laurence O’Brien and Igor Kuznetsov provided excellent research assistance. João Granja
gratefully acknowledges support from the Jane and Basil Vasiliou Faculty Scholarship and from the Booth School
of Business at the University of Chicago. Yannelis and Zwick gratefully acknowledges financial support from the
Booth School of Business at the University of Chicago. We are grateful to the Small Business Administration,
Womply, and Homebase for providing data. This draft is preliminary and comments are welcome.
1
1 Introduction
The COVID-19 pandemic triggered an unprecedented economic freeze and a massive immedi-
ate policy response. Among the firms most affected by the freeze were millions of small busi-
nesses without access to public financial markets or other ways to manage short-term costs.
Without an existing system of social insurance to support these firms, policymakers rushed to
develop new programs to help contain the damage, culminating in the CARES Act.
This paper takes an early look at a large and novel small business support program that
was part of the initial crisis response package, the Paycheck Protection Program (PPP). The
PPP offered guaranteed loans to small businesses through the Small Business Administration
(SBA) in order to stimulate lending to liquidity constrained firms. The loans are forgivable if
firms do not lay off workers or change their compensation. Our first goal is to describe the
targeting of the first round of PPP funding. We then build on our targeting results to evaluate
the short-term economic impacts of the first round (i.e., from April 2nd to May 2nd). As
data become available, we will continue to expand this work by studying the second round of
funding, the medium-term economic impacts, and ultimately the social insurance value of the
program.
We bring new data from two sources to study the PPP. First, we obtained data from the SBA
on the number and amount of PPP loans disbursed by each participating lender during the first
round of the program. The data offer a clear look at which lenders were most active in disburs-
ing loans and at the geographic distribution of PPP lending across the U.S. economy. Second,
we obtained high-frequency employment data from Homebase, a software company that pro-
vides free scheduling, payroll reporting and other services to small businesses, primarily in the
retail and hospitality sectors. The granularity of the data, coupled with the focus on sectors
most adversely affected by the pandemic, allows us to trace out the response of employment,
wages, hours worked, and business closures in almost real-time and evaluate the effects of PPP
support.
In the first part of the paper, we consider two dimensions of program targeting. First,
did the funds flow to where the economic shock was greatest? A central policy goal is to
prevent unnecessary mass layoffs and firm bankruptcies by injecting liquidity into firms. These
potential benefits are likely greatest in areas with more pre-policy economic dislocation and
disease spread. We find no evidence that funds flowed to areas that were more adversely
2
affected by the economic effects of the pandemic, as measured by declines in hours worked
or business shutdowns. If anything, we find some suggestive evidence that funds flowed to
areas less hard hit. The fraction of establishments receiving PPP loans is greater in areas with
better employment outcomes, fewer COVID-19 related infections and deaths, and less social
distancing.
Second, given that the PPP used the banking system as a conduit to access firms, what role
did the banks play in mediating policy targeting? Anecdotal evidence suggests some banks
were eager to participate in the program, while others were unable or unwilling to process large
numbers of loans in the short program window.1 Given the magnitude and pace of the evolving
pandemic and the resulting disruptions in the marketplace, it is important for policymakers to
understand whether banks of different sizes and lending strategies had equal access to the
lending program. In particular, we ask whether there are systematic differences in program
targeting at the aggregate level driven by bank behavior and then quantify the resulting bank
allocation effect on the labor market.
Lender heterogeneity in PPP participation appears to be one reason why we find a weak
correlation between economic declines and PPP lending. We find significant heterogeneity
across banks in terms of disbursing PPP funds, which reflects more than mere differences in
underlying loan demand. For example, because of an asset cap restriction in place since 2018,
Wells Fargo disbursed a significantly smaller portion of PPP loans relative to their market share
of small business loans. We construct a measure of geographic exposure to bank performance
in the PPP using the distribution of deposits across geographic regions. The measure exploits
the fact that most small business lending is local (Brevoort, Holmes and Wolken, 2010; Granja,
Leuz and Rajan, 2018), comparing lenders that did more PPP lending, relative to other small
business lending, versus their counterparts who did less. We find that areas that were signifi-
cantly more exposed to banks whose PPP lending shares exceeded their small business lending
market shares received disproportionately larger allocations of PPP loans.
Overall, our findings shed light on the nature of and mechanism for how the first round
of PPP loans were distributed. PPP loans were disproportionately allocated to areas least af-
fected by the crisis: fifteen percent of establishments in the regions most affected by declines in
1
For example an article from Forbes notes that in the days preceding the launch of the program, Fifth-Third
Bank did not initially participate in the program, while Old National only processed loans for customers that
had an existing account. Bank of America was the first bank to process PPP loans, and they only took loans
from customers with “pre-existing business lending and business deposit relationship with Bank of America, as of
February 15, 2020.”
3
hours worked and business shutdowns received PPP funding; in contrast, thirty percent of all
establishments received PPP funding in the least affected regions. A major factor behind this
pattern was the significant heterogeneity in the intensity of PPP participation across lenders.
Our findings indicate that underperforming banks—whose participation in the PPP underper-
formed their share of the small business lending market—account for two-thirds of the small
business lending market, but only twenty percent of total PPP disbursements in the first round.
The top-4 banks in the U.S. economy (JPMorgan Chase, Bank of America, Wells Fargo, and
Citibank) alone accounted for 36% of the total number of small business loans but disbursed
less than 3% of all PPP loans. These banks were disproportionately located in areas that re-
ceived less PPP funding. We also find that, while the regional funding gap narrowed during
the second round, a large gap persisted across states for at least another month after the initial
funds were exhausted.
In the second part of the paper, we study the effect of the PPP on economic outcomes during
the first round of the program. Our results on bank participation motivate a research design
to evaluate the PPP using bank-driven differences in regional exposure to the program. This
variation across regions allows us to isolate the effect of the PPP from differences in loan de-
mand or confounding correlations between PPP funding and local economic outcomes. Our
research design relies on the assumption that pre-policy bank deposit shares in particular re-
gions are not correlated with the various outcomes we study. This assumption holds once we
condition on relevant observables, such as the relationship between PPP funding and the initial
severity of the crisis. We use this research design to study business shutdowns, reductions in
hours worked, initial unemployment insurance (UI) claims, and small business revenues at the
congressional district and county levels.
We present two findings. First, we do not find evidence that the PPP had a substantial
effect on local economic outcomes during the first round of the program. We establish this
finding using weekly firm-level employment and shutdown data from Homebase, which covers
small retail establishments especially hard hit by the crisis. We confirm the firm-level evidence
using initial unemployment insurance claims at the county level. Given that one motivation
for the program was to provide an escape valve for congested state UI systems, the absence
of a significant effect on UI claims during the initial weeks of the program is striking. Our
confidence intervals on employment outcomes are wide enough to permit modest effects of
the program, but precise enough to reject large effects. The fact that the program disbursed
4
significant funds, yet had little effect on employment, leads to the natural question of what
firms did with additional cash on hand.
Our second finding draws on the Census Small Business Pulse Survey to show that PPP
funds during the first round allowed firms to build up liquidity and to meet loan and other
non-payroll spending commitments. For these firms, the PPP may have strengthened balance
sheets at a time when shelter-in-place orders prevented workers from doing work and when
unemployment insurance was more generous than wages for a large share of workers. This
finding is important because it implies that, while employment effects are small in the short run,
they may well be positive in the medium run because firms are less likely to close permanently.
At the same time, because program eligibility was defined broadly, many less affected firms
received PPP funding and may have continued as they would have in the absence of the funds,
either by spending less out of retained earnings or by borrowing less from other sources. Our
results likely reflect a mix of these factors.
This paper is part of a broader research program to evaluate the impacts of COVID-19
economic policy responses that target private firms and households. As data become available,
we will build on these findings to evaluate the employment responses, to look at firm closures,
and to study the speed and nature of the economic recovery. Measuring these responses is
critical for evaluating the social insurance value of the PPP and similar policies.
Related Literature. This paper joins a literature focusing on how government interventions
following crises impact recovery and the broader economy. Agarwal, Amromin, Ben-David,
Chomsisengphet, Piskorski and Seru (2017) and Ganong and Noel (2018) study the impact of
mortgage modifications following the Great Recession. House and Shapiro (2008) and Zwick
and Mahon (2017) study the effect of fiscal stimulus in the form of temporary tax incentives
for business investment, and Zwick (Forthcoming) documents the role of delegated agents in
mediating take-up of tax-based liquidity support for small firms. Mian and Sufi (2012), Parker,
Souleles, Johnson and McClelland (2013), Kaplan and Violante (2014) and Baker, Farrokhnia,
Meyer, Pagel and Yannelis (2020b) study how stimulus payments following recessions affect
household consumption. This paper evaluates a very large stimulus program aimed at provid-
ing liquidity and support to small firms.
Specifically, this paper also joins a rapidly growing literature studying the impact of the
2020 COVID-19 pandemic on the economy. Jones, Philippon and Venkateswaran (2020),
5
Barro, Ursua and Weng (2020), Eichenbaum, Rebelo and Trabandt (2020), and Berger, Herken-
hoff and Mongey (2020) study the macroeconomics of infectious disease outbreaks and Gorm-
sen and Koijen (2020) use asset prices to back out growth expectations.2 Baker, Bloom, Davis
and Terry (2020) study changes in risk expectations induced by the COVID-19 pandemic.
Barrios and Hochberg (2020), Allcott, Boxell, Conway, Gentzkow, Thaler and Yang (2020),
and Makridis and Rothwell (2020) study how partisan affiliation impacts the response to the
pandemic, and Coibion, Gorodnichenko and Weber (2020), Cajner, Crane, Decker, Grigsby,
Hamins-Puertolas, Hurst, Kurz and Yildirmaz (2020), and Bartik, Betrand, Lin, Rothstein and
Unrath (2020) use a combination of administrative and survey data to study short term la-
bor market impacts. Baker, Farrokhnia, Meyer, Pagel and Yannelis (2020a), Chetty, Friedman,
Hendren and Stepner (2020), and Cox, Ganong, Noel, Vavra, Wong, Farrell and Greig (2020)
study consumption during the COVID-19 pandemic using high-frequency household transac-
tion data, documenting a deeper decline in and slower recovery of consumption among higher
income areas and earners. Taking a more aggregate approach, Mulligan (2020) and Makridis
and Hartley (2020) estimate baseline annual GDP effects of COVID-19 of $7 and $2.14 trillion,
respectively. Guerrieri, Lorenzoni, Straub and Werning (2020) show how supply-side shocks
can generate substantial shocks to demand and aggregate output. Similarly, Papanikolaou and
Schmidt (2020) explore how industry differences in remote work can explain revenue fore-
casts and other supply-side disruptions, building on approaches to measuring the task content
of occupations as in Dingel and Neiman (2020) and Gallipoli and Makridis (2018). Humphries,
Neilson and Ulyssea (2020) take a different approach using a social media survey on nearly
3,000 small and medium size businesses to document large differences in expectations and
awareness of the PPP. We join this emerging literature by providing early microeconomic ev-
idence on how firms and employees were affected as a function of credit supply in the first
stages of the pandemic.
Finally, the paper joins work studying loan guarantees, an important and widely used form
of government intervention in credit markets. Classic work such as Smith (1983), Gale (1990)
and Gale (1991) focus on modeling government credit interventions such as loan guarantees.
2
A related emerging empirical literature also explores the role of non-pecuniary factors, like social capital, as
mediating factors during the pandemic. For example, Makridis and Wu (2020) show how counties with greater
social capital have had fewer infections and a slower spread of infection, conditional on local demographic factors.
Similarly, Barrios, Benmelech, Hochberg and Zingales (2020) and Ding, Levine, Lin and Xie (2020) both show
how areas with greater social capital were more likely to engage in disease-mitigating activities because their
residents have greater trust in institutions and collective efficacy.
6
Early empirical work focused on loan guarantee programs in France (Lelarge, Sraer and Thes-
mar, 2010). Recent theoretical work has focused on government guarantees to banks (Atke-
son, d’Avernas, Eisfeldt and Weill, 2018; Kelly, Lustig and Van Nieuwerburgh, 2016), economic
stimulus (Lucas, 2016) and a burgeoning empirical literature examines the effects of loan guar-
antees on credit supply, employment and small business outcomes (Bachas, Kim and Yannelis,
2020; Barrot, Martin, Sauvagnat and Vallee, 2019; Mullins and Toro, 2017; Gonzalez-Uribe
and Wang, 2019). Barrios, Minnis, Minnis and Sijthoff (2020) provide an empirical frame-
work and Elenev, Landvoigt and Van Nieuwerburgh (2020) focus on theoretical considerations
for assessing the optimal targeting of PPP loans during the pandemic and its role in expanding
credit supply to small versus larger firms that differ in their liquidity constraints. Cororaton and
Rosen (2020) examine the firm characteristics of public firms that have received PPP loans, un-
derscoring the importance of targeting loans towards the firms that need liquidity most. These
patterns contrast in some ways with the rollout of a similar lending program in Italy where
smaller firms and those in more adversely affected areas were more likely to receive support
(De Marco, 2020). Along these lines, we provide a comprehensive assessment of an impor-
tant and large loan guarantee program, evaluating its impacts during a period of economic
contraction and uncertainty.
The remainder of this draft is organized as follows. Section 2 describes the PPP. Section
3 discusses the main data sources used. Section 4 describes how the distribution of relative
performance in the PPP is correlated with bank and other characteristics. Section 5 documents
how differences across banks in PPP activity imply geographic differences in PPP exposure.
Section 6 explores the implications for PPP targeting to different geographic areas. Section
7 analyzes the early effects of the PPP on employment, and section 8 explores mechanisms.
Section 9 concludes.
2 The Paycheck Protection Program (PPP)
The Paycheck Protection Program (PPP) began on April 3rd, 2020 as part of the 2020 CARES
Act as a temporary source of liquidity for small businesses, authorizing $349 billion in forgiv-
able loans to help small businesses pay their employees and additional fixed expenses during
the COVID-19 pandemic. Firms apply for support through banks and the Small Business Ad-
ministration (SBA) is responsible for overseeing the program and processing loan guarantees
7
and forgiveness. An advantage of using the banking system (including FinTech) as a conduit
for providing liquidity to firms is that, because nearly all small businesses have pre-existing
relationships with banks, this connection could be used to ensure timely transmission of funds.
The lending program is generally targeted toward small businesses of 500 or fewer employ-
ees.3 Although the initial round of funding was exhausted on April 16th, funds were drawn
from the Economic Injury Disaster Loan Program (EIDL) in the interim to continue funding
small businesses until the second round of $310 billion in PPP funding was passed by Congress
as part of the fourth COVID-19 aid bill.4 Small businesses were eligible as of April 3rd and
independent contractors and self-employed workers were eligible as of April 10th.
The terms of the loan are the same for all businesses. The maximum amount of a PPP
loan is the lesser of 2.5 times the average monthly payroll costs or $10 million. The average
monthly payroll is based on prior year’s payroll after subtracting the portion of compensation
to individual employees that exceeds $100,000.5 The interest rate on all loans is 1% and
their maturity is two years. The loans will be forgiven if two conditions are met. First, the
loan proceeds must be used to cover payroll costs, mortgage interest, rent, and utility costs
over the eight-week period following the provision of the loan, but not more than 25 percent
of the loan forgiveness amount may be attributable to non-payroll costs. Second, employee
counts and compensation levels must be maintained. If companies cut pay or employment
levels, loans may not be forgiven.6 However, if companies lay off workers or cut compensation
between February 15th and April 26th, but subsequently restore their employment levels and
employee compensation, their standing can be restored.
An important feature of the program is that the SBA waived its standard “credit elsewhere”
test used to grant regular SBA 7(a) loans. This test determines whether the borrower has the
ability to obtain the requested loan funds from alternative sources and amounts to a significant
barrier in the access to regular SBA loans. Instead, in the PPP, applicants were only required to
3
A notable exception was made for firms operating in NAICS Code 72 (accommodations and food services),
which are eligible to apply insofar as they employ under 500 employees per physical location. Firms whose
maximum tangible net worth is not more than $15 million and average net income after Federal income taxes
(excluding any carry-over losses) of the business for the two full fiscal years before the date of the application is
not more than $5 million can also apply. See the SBA for further information about the program.
4
Recipients of an EIDL loan can receive a $10,000 loan advance that does not need to be paid back. The EIDL
loan itself is capped at a maximum of $2 million, is not forgivable, and the funds can be used flexibly for operating
expenses.
5
Payroll costs include wages and salaries but also payments for vacation, family and medical leave, healthcare
coverage, retirement benefits, and state and local taxes.
6
Loan payments on the remainder of the loan can be deferred for six months and interest accrues at 1%.
8
provide documentation of their payroll and other expenses, together with a simple two-page
application process where they certify that the documents are true and that current economic
uncertainty makes this loan request necessary to support ongoing operations. In sum, the PPP
program was designed to be a “first-come-first-served” program with eligibility guidelines that
allowed it to reach a broad spectrum of small businesses.
During the first weeks of April, demand for PPP loans outstripped supply, which was limited
by statute. Between April 3 and 16 all of the initial $349 billion was disbursed, and the program
stopped issuing loans for a period of time. The House and Senate passed a bill to add an
additional $320 billion in funding on April 21 and 23 respectively, which was signed into law
on the 24th. The PPP began accepting applications on April 27 for the second round of funding.
Under the second round, funds were disbursed much less quickly, with unallocated PPP funds
being available in late June.
3 Data
We obtained confidential data on the number of approved PPP loans and approved PPP amounts
from the Small Business Administration during the first round through a Freedom of Informa-
tion Act request. The data set contains information on the amounts and number of loans
approved by each lender, amounts and number of loans received by small businesses in each
state, and total amounts and number of PPP loans received by small businesses in each con-
gressional district as of April 15, 2020. The PPP loan amounts in our records account for 336
billion of the 349 billion allocated to the program under the first round of the CARES Act.
We hand-match this data set with the Reports of Condition and Income (Call Reports) filed
by all active commercial banks as of the fourth quarter of 2019. We are able to match 4,228
out of 4,980 distinct participants in the PPP program to the Call Reports data set. We did not
match 1,031 commercial and savings banks that filed a Call Report in the fourth quarter of
2019. We assume that these banks did not participate in the PPP program and made no PPP
loans. We further classified 631 PPP program participants as credit unions and the remaining
121 participants as non-bank PPP lenders. This group includes small community development
funds, as well as some large non-bank Fintech lenders. The commercial banks in the PPP
sample that we matched to the Call Report account for 93.7% of all PPP loans and 96.8% of
the total amount of loans disbursed under the PPP. By contrast, the group of PPP lenders that
9
we classified as credit unions accounted for 3.3% of all loans and 1.4% of the total PPP lending
amounts. Non-bank PPP lenders accounted for 3% of all loans and 1.8% of total PPP lending
amounts.
We obtain financial characteristics of all banks from the Call Report filed by commercial
and savings banks in the fourth quarter of 2019. The Call Report provides detailed data on the
size, capital structure, and asset composition of each commercial and savings bank operating
in the United States. Importantly, we obtain information on the number and amount of small
business loans outstanding of each commercial and savings bank from the Loans to Small
Business and Small Farms Schedule of the Call Reports. Using this information, we benchmark
the participation of all commercial and savings banks in the PPP program relative to their share
of the small business lending market.
To compute measures of exposure of each state, congressional district, and county to PPP
lenders, we match the matched-PPP-call reports data set with Summary of Deposits data con-
taining the location of all branches and respective deposit amounts of all depository institutions
operating in the United States as of June 30th, 2019. A significant number of depository insti-
tutions merged in the second half of 2019, which means that some branches are assigned to
commercial and savings banks that no longer exist as stand-alone institutions. Notably, Sun-
Trust Banks, Inc. merged with Branch Banking and Trust Company (BB&T) to create the sixth
largest financial institutions in the United States. We use the bank mergers file from the Na-
tional Information Center to adjust the branch network of merged institutions and account for
these mergers. We take advantage of the idea that most small business lending is mostly local
(e.g. Granja, Leuz and Rajan (2018)) and use the distribution of deposits across geographic
regions to create a measure of exposure of these regions to lenders that did more versus less
PPP lending than the expected small business lending share benchmark.
To evaluate whether PPP amounts were allocated to areas that were hardest-hit by the
COVID-19 crisis, we use data from multiple available sources on the employment, social dis-
tancing, and health impact of the crisis. We obtained detailed data on hours worked among
employees of firms that use Homebase to manage their scheduling and time clock. Homebase
processes exact hours worked by the employees of a large number of businesses in the United
States. We use information obtained from Homebase to track employment indicators at a daily
frequency across different states and congressional districts. The Homebase data set dispro-
portionately covers small firms in food service, retail, and other sectors (Bartik, Bertrand, Lin,
10
Rothstein, and Unrath, 2020).7 We complement the Homebase data set with official weekly
state unemployment insurance filings from the Department of Labor. We conduct our main
analyses at the congressional district level because that is the finest level of geographic disag-
gregation for which we have aggregate PPP lending. We use data from the County Business
Patterns dataset to approximate the number of establishments in the congressional district and
compute measures of the average amount of PPP lending per establishment and the fraction
of establishments receiving PPP loans in the region.
Because Homebase primarily covers small retail, beverage, and food service establishments,
it is not representative of aggregate employment. Nevertheless, the Homebase data are quite
useful for evaluating the employment impacts of the PPP specifically, as many hard-hit firms
are in the industries Homebase does cover and much of the early employment losses came from
these firms.8
To broaden our picture of the employment effects, we supplement the Homebase data with
two additional data sources. First, we obtain county-by-week initial unemployment insurance
claims from state web sites or by contacting state employment offices for data. We utilize
initial unemployment insurance claims as a measure of flows into unemployment. We addi-
tionally supplement the Homebase data with data from Womply, a company that aggregates
data from credit card processors. The Womply data includes card spending at small businesses
at the county level, defined by the location where a transaction occurred. Small businesses are
defined as businesses with revenues below SBA thresholds.
To understand the mechanisms underlying our results, we also draw on newly-available
data from the Census Bureau’s Small Business Pulse Survey (SBPS), launched within seven
weeks of the national emergency declaration in March (Buffington, Dennis, Dinlersoz, Foster
and Klimek, 2020). To obtain real-time information tailored towards small businesses, the
SBPS is run weekly and entirely online that takes less than 10 minutes to complete and relies
largely on checkboxes, rather than numerical responses. Rather than contacting businesses
with an official letter, they were reached via email based on the Census Bureau’s Business
Register that has been populated using responses to the Economic Census across the 50 states
7
More information about Homebase can be found at www.joinhomebase.com.
8
We note that we were made aware of an error in the current raw formulation of the HomeBase data. This
resulted in doubling the number of hours three sectors: retail, professional services and transportation. The error
only affects the level, not the relative difference, across industries or states and hence should not change our two
primary outcomes, business shutdowns and the ratio of hours worked. However, these results should be treated
with some caution and the issue is scheduled to be corrected in the next data release.
11
(and D.C. and Puerto Rico). Furthermore, the SBPS focuses on businesses with receipts that
are greater than or equal to $1,000, but retain 500 employees or fewer. Each week, the sample
weights are adjusted to maintain representativeness.
While we focus our analysis on aggregations of small businesses across states and metropoli-
tan areas over time, we observe a wide array of information. For example, we observe changes
in revenues (increase, decrease, and no change) and the actual level of revenues. We also
observe how firms are changing their labor force through temporary business closures, em-
ployment, or hours worked. Finally, and most importantly for our purposes, we draw upon
responses to questions about liquidity (e.g., cash on hand), loans, defaults, and applications
for financial assistance. We observe separate responses to questions about use of the PPP, EIDL,
and SBA loan forgiveness programs. We note that Humphries, Neilson and Ulyssea (2020) and
Bartik, Bertrand, Cullen, Glaeser, Luca and Stanton (2020) have pursued similar approaches
through other channels aimed at gathering information about small businesses.
Finally, we obtain counts of COVID-19 cases by county and state from the Center for Dis-
ease Control and use data on the effectiveness of social distancing from Unacast. Unacast
provides a social distancing scoreboard that describes daily changes in average mobility. Un-
acast measures the change in average distance travelled using individual’s GPS signals. The
data is availably on a daily basis, at the county level. We obtain information on the effective
dates of statewide shelter-in-place orders from the New York Times.9
4 PPPE and Bank Characteristics
4.1 Paycheck Protection Program Exposure
Table 1 shows summary statistics for the 20 largest financial institutions in the United States,
as measured by total assets. The left-most column gives the institution’s name, while the first
column of the table gives total assets as of the fourth quarter of 2019. The second and third
columns, respectively, show the share of total PPP volume and the share of the small business
loan (SBL) market of each institution. The fifth column presents relative bank performance
9
The New York Times presents some aggregates on shelter-in-place orders.
12
which is measured as
Shar eVol.P P P − Shar eSB LM ar ket
P P P Eb = × 0.5 (1)
(Shar eVol.P P P + Shar eSB LM ar ket)
where Shar eVol.P P P the share of PPP volume held by bank b, and Shar eSB LM ar ket is their
total small business loan volume market share. The next three columns present similar infor-
mation to columns (2) through (4), using the market share of total number of loans rather
than their volume, where
Shar eN br.P P P − Shar eN br.SB LM ar ket
P P P Eb = × 0.5. (2)
(Shar eN br.P P P + Shar eN br.SB LM ar ket)
Here Shar eN br.P P P the share of the number of PPP loans held by bank b, and the term
Shar eN br.SB LM ar ket is their total small business loan market share, based on the number
of loans outstanding in each bank’s balance sheet as of the fourth quarter of 2019.
Figure 1 shows the distribution of relative performance in the PPP comparing PPP market
share against the overall stock of small business loans. The top panel shows the distribution
of relative bank performance in the PPP, based on the total volume of PPP loans granted. The
bottom panel shows the distribution of relative bank performance in the PPP, based on the
number of PPP loans granted. Both figures show a wide dispersion of relative performance,
with the distribution of performance based on the total number of loans showing greater mass
at the tails.
Figure 2 plots the bank measure of relative performance in the PPP by percentile of bank
size. The top panel shows relative performance based on the total volume of PPP loans granted,
the middle panel shows relative performance based on the number of PPP loans issued, while
the bottom panel shows average PPP loan size. All three panels indicate a similar pattern—
larger banks issued more PPP loans than expected given their share of the small business mar-
ket, whether by volume, number of loans, or average loan size. This pattern could be consistent
with larger banks being better suited to take advantage of the PPP program as it was rolled
out. This pattern reverses at the very top of the bank size distribution. The very largest banks,
those in the top percentile of bank size, significantly underperformed in PPP lending relative
to their pre-policy share of small business lending. This underperformance is clear both in
terms of lending volume (Panel A) and number of loans (Panel B). Panel C suggests that the
underperformance of the top percentile of banks occurs despite their making the largest PPP
13
loans in the sample in terms of average loan amount.
Figure 3 provides evidence concerning the significant dislocations between the share of PPP
lending of underperforming banks and the share of PPP that we would expect had these banks
issued PPP loans in proportion to their share of the small business lending market. The blue
hollow triangles and red hollow circles represent, respectively, the cumulative share of the PPP
and small business lending of banks whose PPPE is below a certain threshold. The figure shows
that commercial and savings banks, representing 20% of the small business lending market,
simply did not participate in the PPP lending program, (P P P E = −0.5). The plot further shows
that the group of banks whose PPP share is below their share of the small business lending
market, (P P P E < 0), made less than 20% of the PPP loans, but account for approximately
two-thirds of the entire small business lending market. Overall, the evidence is consistent with
substantial heterogeneity across lenders in their responses to the program’s rollout.
The fact that lenders were significantly heterogeneous in accepting and processing PPP
loans would not necessarily result in aggregate differences in PPP lending across geographic
areas if small businesses could easily substitute and place their PPP applications to lenders
that were willing to accept and quickly expedite them. If many lenders, however, prioritize
their existing business relationships in the processing of PPP applications, firms’ pre-existing
relationships might determine to a large extent whether they are able to tap into PPP funds. In
this case, the exposure of geographic areas to banks that over/underperformed in the deploy-
ment of the PPP might significantly determine the aggregate PPP amounts received by small
businesses located in these areas. Next, we examine if geographic areas that were exposed to
banks with weak PPP performance received less PPP lending overall.
5 Geographic Exposure to Bank PPP Performance
We next explore how the geography of the PPPE is related to PPP lending outcomes. Figure 4
presents a map of county level exposure to PPPE based on the share of deposits of each bank
in the county. Exposure varies across the United States with some Western areas containing
a large Wells Fargo presence that exhibit lower levels of PPPE, suggesting greater exposure
to lenders that underperformed in the PPP program relative to their small business lending
benchmark. By contrast, the counties with lower median household income (ρ = −0.13), a
lower share of college educated individuals (ρ = −0.16), and a smaller COVID-19 shock were
14
more likely to be exposed to lenders that overperformed in the PPP roll-out.
Figure 5 explores the relationship between PPPE exposure and PPP lending. The top panel
of Figure 5 plots aggregate PPP volume per small business by exposure to PPPE for each state,
while the bottom panel shows the fraction of all small businesses receiving PPP loans in the
state. Both panels present a similar pattern—there is a strong positive relationship between
PPP lending and PPPE exposure at the state level. States with higher exposure to banks that
performed well in terms of PPPE also saw greater PPP lending. Figure 7 presents a similar
pattern at the congressional district level, and a similar correlation emerges.
Table 2 makes this graphical evidence explicit. The top panel shows the relationship be-
tween PPPE and aggregate lending, at the congressional district level. Column (1) shows
the correlation between aggregate PPP lending and PPPE at the congressional district level.
The relationship is highly statistically significant, with an F-statistic of approximately 45. A
one-standard deviation increase in the congressional district exposure to PPPE based on total
amounts of outstanding PPP and small business loans and weighted by the share of deposits
of each bank leads to a 16.1 percent increase in PPP lending. Column (2) adds in aggregate
employment and payroll controls, and column (3) adds in industry shares. The results remain
highly statistically significant at the 1 percent level. Column (4) shows that the correlation
holds even within states when we add state fixed effects, although the coefficient remains
significant at only the 10 percent level.
The bottom panel repeats the analysis, replacing aggregate lending per business with the
fraction of establishments receiving PPP loans in each congressional district. This panel also
indicates a very strong relationship between our PPPE measure and lending. In column (1),
the first stage F-statistic is now above 200, and even with state fixed effects in column (4) the
relationship is highly statistically significant at the 1 percent level. A one-standard deviation
increase in our measure of congressional district exposure to bank PPP performance is associ-
ated with an increase of 4.3 percentage points in the fraction of establishments receiving loans
in a congressional district. These results suggest that businesses were much more likely to re-
ceive a PPP loan simply because they were located closer to banks that processed a large share
of PPP loans relative to their benchmark share of small business loans.
A potential concern with the above results is that the causality runs reverse. That is, banks
do relatively better where demand for PPP loans is abundant. To address this concern, we
15
compare survey measures on firm applications and PPP receipt.10 The Census Small Business
Pulse survey measures small businesses conditions in relation to the COVID-19 crisis at the
state by sector level, including PPP application and receipt. We compare the difference between
application and receipt to our measure of bank exposure.
The top panel of Figure 6 shows volume-based PPPE exposure and the difference between
the percentage of businesses reporting having applied to PPP and percentage of business that
received PPP in each state. The bottom panel shows PPPE based on the number of loans and
the difference between the percentage of businesses reporting having applied to PPP and the
percentage of business that received PPP in each state. Both panels show a very similar pattern,
the difference between PPP application and receipt is much lower in states with higher exposure
to banks that allocated more PPP funds. In other words, conditional on applying, businesses
were more likely to receive PPP funds in states where there were more banks allocating funds.
Appendix A presents a case study, which demonstrates the importance of financial interme-
diation in the allocation of PPP funds. Wells Fargo was severely constrained from expanding
its balance sheet as a result of an asset cap imposed by the Fed in the aftermath of the fake
accounts scandal. This asset cap was only lifted on April 10 when the Fed excluded PPP loans
from the formula it uses to restrict Wells Fargo’s growth. The asset cap limited Wells Fargo’s
ability to lend under the PPP in the early days for the first phase of the program. Wells Fargo
performed poorly in terms of allocating PPP funds during the first wave and firms located in
areas with higher Wells Fargo market share see lower PPP allocations, both in terms of overall
aggregate loan volume per business and in the fraction of businesses receiving PPP loans.
The collection of results in this section suggests that exposure to bank-specific heterogeneity
in their willingness and ability to extend PPP loans was a significant determinant of the allo-
cation of PPP loans in the economy. Next, we examine how the PPP allocation and exposure
to over/underperforming banks correlated with the local magnitude of the epidemic.
6 Are PPP Allocations Targeted to the Hardest Hit Regions?
Were PPP funds disbursed to geographic areas that were most affected by the epidemic? Figure
8 shows the relationship between PPP allocations, exposure to Bank PPP performance, and the
fraction of businesses in each congressional district that shut down during the week of March
10
Census Small Business Pulse survey data can be found here.
16
29th to April 4th, just before PPP funds were disbursed.11 We estimate business shutdowns
in the congressional district using the high-frequency data set obtained from Homebase. The
figure indicates little if any correlation between PPP allocation or relative bank performance
and hours worked or business shutdowns. In Figure 9, we follow Bartik, Bertrand, Cullen,
Glaeser, Luca and Stanton (2020) and repeat the analysis using the ratio of hours worked on
March 31st, 2020 relative to a baseline of the average hours worked in the same weekday of
the last two weeks of January. Again, we find that PPP allocations across congressional districts
are very weakly correlated to the impact of the epidemic crisis on labor markets and aggregate
firm outcomes. We have also investigated the robustness of these results to alternative cutoffs
when measuring the decline in hours worked and employment (e.g., as of April).12
To better illustrate the relation between firm and employment outcomes at the congres-
sional district level and PPP allocations, we stratify congressional districts into 20 bins based
on the impact of the COVID-19 epidemic on the fraction of businesses that shut down and on
the average decline in hours worked in the congressional district. Figure 10 plots the average
fraction of business receiving PPP loans in each business shutdown bin (top panel) and hours
worked bin (bottom panel). The plots suggest that approximately 15% percent of businesses
located in the most affected congressional districts were able to obtain PPP funding until April
15th, 2020. By contrast, more than 30% of all businesses operating in the least affected con-
gressional districts were able to tap into PPP funding.
The results suggest that PPP funds were not targeted towards geographic areas that were
most affected by the pandemic, at least in terms of small business employment drops. This fact
could be a result of the pre-existing bank relationships across counties, rather than a problem
with implementation: banks were caught off guard by the pandemic and the corresponding
actions taken to social distance. A related factor likely influencing these geographic patterns is
differential loan demand in harder hit areas. Because PPP support is more generous for firms
that maintain their payroll, the program likely appealed more to firms with smaller reductions
in their business. To the extent these geographic patterns reflect such differences in loan de-
11
Following Bartik, Betrand, Lin, Rothstein and Unrath (2020), we define a business shutdown as businesses
that report zero hours worked during a week.
12
In Tables A.2 and A.3, we employ cross-sectional specifications at the congressional district to further support
the idea that the PPP funding did not flow to areas with largest pre-PPP declines in employment and ratios of
shutdown businesses. The tables report the results of ordinary least squares (OLS) regressions examining the
relation between the allocation of PPP funds and the share of businesses that shut down operations in the last
week of March, and the decline in hours worked between January and the last week of March.
17
mand, the evidence suggests the PPP functioned less as social insurance to support the hardest
hit areas and more as liquidity support for less affected firms. Nevertheless, our bank-level
results point to an important loan supply channel distorting the distribution of PPP loans.
The appendix presents suggestive evidence that, if anything, funds were disproportionately
allocated to geographic areas that were less hard hit by the virus. Figures A.2 and A.3 repeat
the analyses of Figures 8 and 9 at the state-level. Figures A.4 and A.5 show that there is a
slight negative correlation between loans and PPPE with COVID-19 confirmed cases and deaths.
This fact is consistent with Figure A.6, which indicates that states with earlier shelter-in-place
orders—which were presumably harder hit by the epidemic—saw lower fund allocations. Fig-
ure A.7 shows that there is little correlation between the magnitude of social distancing at the
state level and PPP allocation and bank exposure. Finally, Figure A.8 confirms our findings
using the Homebase data with another public data source—we find no consistent relationship
between PPP allocation and bank exposure with state UI claims. The totality of the evidence
suggests that there was little targeting of funds to geographic areas that were harder bit by
the epidemic and, if anything, areas hit harder by the virus and subsequent economic impacts
received smaller portions of PPP funds.
7 The Early Effects of PPP on Employment and Local Eco-
nomic Activity
7.1 Research Design
Our results on PPP performance differences across banks motivate a research design for evalu-
ating the PPP. The basic idea is to use differences in local area PPP exposure (PPPE) to partition
geographies and compare the evolution of local outcomes for high versus low PPPE regions.
By exploiting differential exposure to banks that performed poorly in distributing PPP funding
during the first round of the program, we can isolate the effect of the PPP from other differ-
ences across regions that may drive differences in PPP loan demand. As described above, we
map bank level aggregates for PPP lending from the SBA data onto local geographies using
measures of local bank branch presence. The research design is akin to a Bartik instrument
and therefore relies on the assumption that pre-policy bank deposit shares in particular regions
are not correlated with the various outcomes we study, conditional on observables.
18
We focus our analysis in the time period between the third week of January and the last
week of April to study the short-term effects of the PPP in the immediate aftermath of the
pandemic when the injection of liquidity was thought to matter the most. We choose this time
window for three reasons. First, starting our sample period in January allows to establish a
baseline period prior to the full economic impact of the virus hitting US firms. Second, the PPP
began disbursing funds on April 3 and all of the initial $349 billion was allocated by April 16.
During this period, banks played a key role in allocating limited funds, creating the variation
we exploit to identify the effects of the program. Third, beginning on April 27 a second round
of PPP funding began. Partly because PPP funding limits were no longer binding, the second
wave closed some of the gap between initially high and low PPPE exposure regions. Thus, were
we to use PPPE to study the effect of the program later in May and June, we would need to
interpret the research design as assigning some firms funding with a lag, instead of as assigning
some firms no funding at all.13
Given the rapid nature and size of the economic shock, we highlight an important consid-
eration when analyzing data from this time period. Our targeting analysis shows that regions
receiving more PPP funding were less hard hit by the initial shock, in part due to the banking
channel we emphasize. Thus, it is important to properly condition on this non-random assign-
ment of PPP funding. For example, if one does not break out the data finely enough or con-
dition properly for targeting differences—for example by treating all of March as a pre-period
benchmark or by looking at the correlation between PPP receipt and employment outcomes—
then one might detect a spurious effect of the program. We show this issue is very clear in
week-by-week outcomes around the policy window.
To account for these targeting differences, we estimate the effects of the program using
the period prior to large scale lockdowns as a base period and separately report “effects” for
the “Pre-Lockdown Covid” period, “Post-Lockdown Pre-PPP” period and “Post-PPP” period. We
then compare the post-PPP period effects to the ‘Post-Lockdown Pre-PPP” period to estimate
the causal effect of the program. We also include time-varying controls and state-by-time-
by-industry fixed effects to estimate treatment effects under weaker versions of the Bartik as-
sumption. Once we adjust for targeting differences, including these more restrictive controls
has little effect on our estimates.
In our main analysis, we present reduced form regressions of local economic outcomes on
13
We plan to extend our analysis to study the second round as new data become available.
19
PPPE. We choose this approach because we only have PPP funding for certain geographic levels,
namely, state and congressional district.14 In particular, we estimate variants of the following
specification
Eisnt = αi + δsnt + β1 1[Pre-Lockdown] × P P P Ec
+ β2 1[Post-Lockdown & Pre-PPP] × P P P Ec
+ β3 1[Post-PPP] × P P P Ec + γX c + "isnt ,
where Eisnt is an outcome, business shutdowns or the decline in hours worked in our main
specification for firm i operating in state s and industry n in week t. The term αi captures firm
fixed effects, while δsnt are state-by-industry-by-week fixed effects, P P P Ec is PPP exposure at
the congressional district level measured using the number of firms exposed, and "isnt is an
error term. 1[Pre-Lockdown] is an indicator variable that takes the value of one between the
weeks of March 1st–March 7th and March 15th–March21st, 1[Post-Lockdown, Pre-PPP] is an
indicator variable that takes the value of one during the weeks of March 22nd–March 28th
and March 29th–April 4th, and 1[Post-PPP] is an indicator variable that takes the value of one
following the week of April 5th–April 12th.
The coefficients β1 , β2 , and β3 capture the differential effect of PPP exposure on the out-
come of interest during each time period relative to the baseline period. The identifying as-
sumption is that the firms differentially exposed would have trended similarly in the absence
of the PPP, conditional on fixed effects and controls. Due to the inclusion of state-by-industry-
by-week fixed effects, we are comparing trajectories for firms within state-by-industry groups
and allowing general time trends within these groups. The coefficient β1 captures differences
between firms after the initial spread of COVID-19 but prior to subsequent lockdowns. The co-
efficient β2 captures targeting effects—differences in outcomes related to PPP exposure prior
to the disbursement of funds—while the coefficient β3 captures the differential effect of PPP
exposure on the outcome of interest after the PPP launched. The difference between the coef-
ficients β3 − β2 identifies the effect of PPP exposure conditional on targeting.
14
Appendix Table A.4 reports instrumental variable regressions to study Homebase outcomes at the congres-
sional district level, and we briefly interpret these results below. We also use instrumental variable regressions to
study Census survey outcomes at the state level.
20
7.2 Small Business Employment
A significant portion of the policy and media attention regarding the PPP focused on the pro-
gram’s potential employment effects. We examine business shutdowns (i.e., hours worked
reduced to zero during the entire week), declines in hours worked, and unemployment insur-
ance claims and find no evidence of substantial effects of the PPP on these outcomes.
Recall that our targeting analysis revealed that if anything a larger fraction of businesses
located in areas that were less hard hit received PPP funding during the first round. Specif-
ically, the plots of Figure 10 indicate that in congressional districts that saw least decline in
hours worked and the lowest rates of business shutdowns saw a greater fraction of their small
businesses receiving small business loans.
Figure 11 presents simple difference-in-difference graphs for each of our employment and
local activity outcomes. We divide all firms in the sample based on whether they are located
in congressional districts with above- or below-median high and low PPPEs. We use vertical
markers to demarcate the post-lockdown, pre-PPP period; the post-PPP launch; when the first
round of PPP funds are exhausted; and when the second round of PPP funding begins. For both
measures of small business employment, we see a dramatic decline in employment starting in
the week prior to the lockdowns. Consistent with our targeting results, this decline is modestly
larger for regions with low PPPE. Importantly, during the first round of PPP, the gap between
high and low PPPE areas does not widen further, indicating little incremental impact of PPP
during this time.
Figure 12 plots weekly coefficients and standard error bands for regressions of each out-
come on congressional district PPPE. Specifically, we estimate regressions of the form
Eisnt = αi + δsnt + β t P P P Ec × 1[Week=t] + "isnt
where Ei t is business shutdowns or the decline in hours worked for firm i operating in state
s and industry n in week t, αi are firm fixed effects, αsnt are state-by-industry-by-week fixed
effects, P P P Ec is PPP exposure at the congressional district level measured using the number of
firms exposed, 1[Week=t] is an indicator variable that takes the value of one during week t and
"isnt is an error term. The coefficients β t capture the effect of PPP exposure on the outcome of
interest under the identifying assumption the firms differentially exposed would have trended
similarly in the absence of the PPP. The baseline week is the first week of the sample period
21
(January 19th–January 25th). The state-by-industry-by-week fixed effects imply that we are
comparing trajectories for firms within state-by-industry groups and allowing general time
trends within these groups.
The results align with the raw differences across high and low PPPE regions in Figure 11.
Common trends hold until the week of the declaration of the national emergency and lock-
downs take effect. At this point, higher PPPE is associated with better employment outcomes
in terms of fewer business shutdowns and smaller reductions in hours worked. When the PPP
is launched and funds start to reach firms, we see no incremental effects of the program on
these outcomes relative to pre-PPP levels.15
Table 3 presents our difference-in-difference estimates, in which we pool the weekly effects
into four time intervals reflecting the pre-Covid-19 crisis period; the pre-lockdown - post-Covid-
19 period; the post-lockdown, pre-PPP period; and the post-PPP period. We measure the effect
of the PPP by comparing the coefficients for the latter two intervals. In the top panel, the
outcome of interest is business shutdowns, while in the bottom panel it is the decline in hours
worked. The first two columns and second two columns exclude and include firm fixed ef-
fects, respectively. The first and third columns use the fraction of firms receiving PPP as the
program exposure variable. The second and fourth columns use congressional district PPPE as
the program exposure variable, and thus can be thought of as reduced form regressions.16
The table confirms the finding of no statistically or economically significant relationship
between PPP loans or PPP bank exposure and these employment outcomes. Moreover, our
least squares estimates are not simply statistically insignificant with large confidence intervals;
rather, they are precise zeros. To address potential concerns about the strength of the research
design, Appendix Table A.4 instruments the fraction of firms receiving PPP loans with the PPPE
measure. Again, we find nearly identical results. While one concern is that our first-stage effect
is weak, we show that our F -statistic is above the rule-of-thumb and highly significant: a 10
percentage point increase in PPPE is associated with a 3.4 to 3.8 percentage point increase in
the share of firms receiving PPP.
As another way of interpreting our magnitudes, consider the following comparison. The
15
Appendix Figure A.9 shows that our instrument improves the reliability of inferences by mitigating pre-trends,
likely driven by differences in loan demand. However, the results when we use PPP funding instead of PPPE as
the right-hand-side variable are consistent with our preferred specification: we see little evidence of substantial
effects of PPP on these outcomes relative to pre-PPP outcomes.
16
Appendix Table A.4 estimates 2SLS regressions in first differences comparing the post-PPP period to the post-
lockdown, pre-PPP period. The effects are qualitatively identical but with modestly larger standard errors.
22
difference between PPPE for top versus bottom quartile congressional districts is 0.42. This
difference implies an increase in the share of establishments receiving PPP funding of 14.3
percentage points.17 Using the reduced form estimates in Table 3, column (4), this change
in funding implies a change in the probability of firm shutdown of -0.08 percentage points
(= (−0.042 − 0.040) × 0.42). The lower bound of the 95% confidence interval is well below
a one percentage point effect. An analogous calculation for the decline in hours worked gives
similarly small effect sizes. Thus, relative to the aggregate patterns in Figure 11—a 40 per-
centage point increase in the probability of firm shutdown and 60 percentage point reduction
in the ratio of hours worked relative to January—we can reject modest effect sizes during this
time period.
Table 4 presents similar results using another separate measure of employment outcomes:
UI claims. The dependent variable is the ratio between the weekly initial unemployment in-
surance filings hand-collected from several state labor departments and county employment
measured using the Bureau of Labor and Statistics January County Employment data. The re-
sults paint a very similar picture to those in Table 3. Again, there is no statistically or econom-
ically significant relationship between PPP loans or PPP bank exposure and UI claims. Again,
the first and third columns use the fraction of firms receiving PPP as the program exposure
variable while the second and fourth columns use the respective county PPPE as the program
exposure variable. While the results are slightly less precise than those using outcomes from
the Homebase data, the coefficient on the interaction post the rollout of the PPP program indi-
cates no discernable changes in the number of UI filings, relative to the post-lockdown period.
7.3 Local Economic Activity
While much of the focus of the PPP was in countering a surge in unemployment and busi-
ness closures, we now explore the potential effects of PPP on the evolution of small business
revenues as a proxy for local consumption activity. Table 5 presents analogous results. For ex-
ample, in columns (1) and (3), we find no statistically significant association between county
PPPE and consumption expenditures. Here, we control for state × industry × week fixed ef-
fects, exploiting differences in the exposure of firms to PPP across counties within the same
state and industry. This rules out potential concerns about the passage of stay-at-home and
17
This calculation comes from 0.42×0.34 from Appendix Table A.4, Panel A, establishment-level results, column
(1).
23
nonessential business closure laws at the state-level, which could otherwise confound our esti-
mates. Columns (2) and (4) subsequently add county × industry fixed effects, thus absorbing
invariant characteristics of small businesses at the county×industry level. Again, we find no ev-
idence of a positive association with consumption expenditures. Moreover, column (4) controls
for weekly infections and deaths per capita, as well as an index of social distancing (measured
by distance traveled among residents) to control for other time-varying county-specific shocks
present during these weeks.
8 Interpretation and Mechanisms
Additional funds from the PPP do not have significant effects on employment or local economic
activity during the first round of the program. If firms did not maintain or increase employment,
what did they do with these funds?
There are several non-mutually exclusive channels through which businesses may have ab-
sorbed the funds without immediate employment effects. First, program eligibility was defined
broadly, so many less affected firms likely received funds and continued as they would have
in the absence of the funds. In these cases, the program’s benefits accrue to the firm’s owners.
Related to this idea, it is possible that PPP funds crowded out private market borrowing, as
firms or banks may substitute publicly guaranteed PPP loans for other lending that would have
happened otherwise (Holmstrom and Tirole, 1998).
Second, despite restrictions on the share of funds ultimately forgiven that needed to be
spent on payroll, firms retained significant flexibility in how they could use the funds over
time. Firms could elect to use funds for required payments initially, to leave the funds in the
bank, or to use only the share that they are able to use when they can. For these firms, the
PPP may have strengthened balance sheets at a time when shelter-in-place orders prevented
workers from doing work and when unemployment insurance was more generous than wages
for a large share of workers. An implication of this channel is that, while employment effects
are zero in the short run, they may well be positive in the medium run because firms are less
likely to close permanently.
There was also likely considerable uncertainty about how to use the funds, either because
of evolving regulatory guidance or because small businesses struggled to understand the rules
in the first place. Moreover, firms may respond to periods of high uncertainty, such as the
24
ongoing pandemic (Baker, Bloom, Davis and Terry, 2020), by holding cash as a precautionary
buffer.
Third, some firms may have increased employment or called back workers. Small effects
for these firms are indeed consistent with our null overall result. These effects may be par-
ticularly muted given historic expansion of UI benefits, causing former employees to prefer
unemployment over a return to work (Chetty, 2008). Nevertheless, it is unlikely that the PPP
was responsible for large changes in employment at the end of April and into the first week in
May. However, we caution that more data are likely needed to better support this claim.
Overall, we find evidence that PPP funds during the time period we study are used as
liquidity support and to meet loan and other spending commitments.
8.1 Payments
While we do not find short term effects on employment and firm revenue, the PPP may have
had significant effects promoting financial stability and the long-term recovery by extending
a lifeline to small businesses and preventing a rapid wave of small business bankruptcies and
liquidations. In Table 6 and Figure 13, we use information from the Census Small Business
Pulse Survey to relate the state PPPE with the percentage of businesses reporting having re-
ceived PPP in each state-industry group. The results suggest that between April 26th and June
6th, the period covered by the weekly surveys, state PPPE is strongly and positively related to
the percentage of businesses reporting receiving PPPE within a state and industry. Over the
entire sample period, a one-unit increase in state PPPE is associated with a 35% increase in the
share of businesses who report receiving PPP. Consistent with the idea that the second round of
funding lessened some of the geographic imbalance in the allocation of PPP disbursements, we
also see that the effect of state PPPE on the percentage of businesses reporting receiving PPP is
more pronounced in the early waves of the survey. Between April 26 to May 2, a comparable
unit increase in PPPE is associated with an 85% increase in PPP received. By May 31 to June
6, the coefficient declines to 24%, although it remains statistically significant.
We repeat this exercise using the percentage of businesses reporting requesting, but not
receiving PPP in each state and industry. As discussed in Section 5, this exercise also assuages
concerns that the relation between state PPPE and the percentage of businesses receiving PPP
could be driven by demand effects whereby firms and industries that did not request PPP funds
25
were disproportionately located in areas whose local banks have low PPPE. We report the
results of this analysis in Table A.5 and Appendix Figure A.10. The results indicate that states
with low PPPE had a greater proportion of businesses applying for but not receiving funds
under the program thus suggesting that the lower fraction of firms receiving PPP in those
areas is driven not by lower demand for funds from the program, but rather by their lower
access to banks that processed applications during the first round.18
Having confirmed that PPPE is significantly related with access to PPP funds, we further
explore the information in the Small Business Pulse Survey to examine whether receipt of PPP
allowed firms to avoid becoming delinquent on scheduled payments (either loan or non-loan).
In Table 8, we show regressions examining the relationship between the allocation of PPP
funds and the percentage of firms reporting missing loan payments at the state-industry level.
In Table 7, we show similar regressions using the percentage of firms missing other scheduled
payments such as rent, utilities, supplier payments, and payroll.19 In light of our targeting
results, these regressions add controls for pre-PPP measures of crisis severity, including the
pre-PPP decline in hours worked from Homebase, the pre-PPP counts of COVID cases and
deaths per capita, and the pre-PPP social distancing index.
The results in Table 8, column (1), indicate that a percentage point increase in the share
of firms reporting receiving PPP is not significantly associated with a decline in the percentage
of firms missing loan payments. This result, however, could indicate that areas and industries
with a lower percentage of businesses receiving PPP had a larger fraction of businesses that
did not apply because they were not eligible to apply or decided they did not need the funds.
To address this issue, we use state PPPE to capture geographic differences in access to the
supply of PPP funds resulting from differences across regions in their exposure to bank PPP
performance. These differences are plausibly unrelated to demand factors and therefore less
likely to be confounded by them. In column (2), we run a similar specification, but instead
focus on the relation between percentage of firms reporting missing payments with state PPPE.
In this case, an increase in state PPPE is associated with a significantly lower percentage firms
18
In the Appendix, we further repeat these empirical exercises at the level of the Metropolitan Statistical Area
(MSA) rather than at the State-Industry level and we find similar results in that MSAs with lower exposure to PPP
see a lower percentage of small businesses receiving PPP.
19
Unfortunately, the Pulse survey does not separate this category of other payments into payroll versus non-
payroll components. However, it does focus on “required” payments, which firms may interpret as referring to
payments for past labor rather than discretionary payments based on retaining workers going forward. Results
for this measure should be interpreted with some uncertainty about respondents’ interpretation of the question.
26
reporting missing loans payments. In columns (3) and (4), we present results of an IV strategy
whereby we instrument for the percentage of firms receiving PPP using state PPPE. Using this
strategy, we find that a ten percentage point increase in firms receiving PPP is associated with
1.7 percentage point decline in missing loan payments.
In Table 7, we find that a ten percentage point increase in the share of firms receiving PPP
is associated with an even larger effect on missed non-loan payments. This result might not be
surprising when we consider that a large fraction of small businesses do not necessarily have
loans. We find that, across all specifications, an increase in the percentage of firms receiving
PPP or in the access to PPP funds is associated with a lower percentage of firms reporting
missing these types of payments. Specifically, the results of the IV strategy in columns (3) and
(4) suggest that a ten percentage point increase in firms receiving PPP is associated with a
5.5 percentage point decline in the number of firms reporting missing any type of scheduled
payments. The relationship is highly statistically significant at the 1 percent level.
Finally, we show in Appendix Table A.7 that the share of firms reporting missing payments
is positively related to the fraction of small businesses reporting having applied, but not having
received PPP funds. The percentage of firms that applied, but not did not receive, PPP funding is
plausibly less related to differences in demand for PPP funds across geographies and industries
and likely better captures timely access of small businesses to the supply of PPP funds. These
results further suggest that PPP might have been crucial in allowing small businesses to make
scheduled payments and survive the economic crisis without permanently closing.20
8.2 Liquidity Support and Precautionary Savings
We additionally find that the PPP funds increased firms’ cash on hand. We use the Census Small
Business Pulse Survey to investigate whether access to PPP funds is associated with a greater
percentage of firms reporting having greater liquidity support. Table A.16 shows regressions
results examining the relation between access to PPP funds during the first round and the
percentage of businesses reporting having at least two months of cash-on-hand. This exercise
also offers a useful sanity check of the informativeness of the survey data. As above, these
regressions include controls for pre-PPP measures of crisis severity.
Similar to results presented above, the coefficients reported in column (1) of Table 9 do
20
We also show in the appendix that these results are economically and statistically similar when we use MSA-
level observations.
27
not indicate an economically or statistically significant relation between cash-on-hand and the
percentage of firms in that state and industry that reported receiving PPP. However, when
we turn our attention to the relation between state PPPE, which better isolates access to the
supply of PPP funds, access to PPP is economically and significantly related to the share of
firms reporting significant liquidity. In the reduced form specification presented in column (2),
we find that a unit increase in state PPPE is associated with a ten percentage point increase in
the share of firms reporting significant liquidity. In the IV regression, which uses state PPPE
to instrument for the share of firms receiving PPP, a ten percentage point increase in the share
of firms receiving PPP is associated with a 2.9 percentage point increase in the share of firms
reporting at least two months of cash to cover business operations. Moreover, in the appendix
we report similar results when we use the share of small businesses that applied, but did not
receive, PPP as our main right-hand-side variable of interest.
Overall, these results are consistent with the idea that the PPP provided firms with an
important liquidity cushion that they used to navigate the turmoil of the initial months of
the pandemic. Another possible interpretation, given our evidence that PPP did not induce
increases in employment or declines in initial unemployment filings, is that these businesses
are maintaining the loaned PPP funds in bank accounts as precautionary savings until they are
ready to resume activities, perhaps as demand for their goods and services return to normal or
as relaxed shelter-in-place orders permit them to reopen for business. Generally, the results are
not consistent with the idea that the PPP served as a large-scale alternative to unemployment
insurance for delivering funds directly to affected workers.
9 Concluding Remarks and Next Steps
This paper takes an early look at a large and novel small business support program that was
part of the initial crisis response package, the Paycheck Protection Program (PPP). We consider
two dimensions of program targeting. First, did the funds flow to where the economic shock
was greatest? Second, given the PPP used the banking system as a conduit to access firms,
we ask what role did the banks play in mediating policy targeting? We also explore effects on
employment and other firm outcomes.
We find little evidence that funds were targeted towards geographic regions more severely
affected by the pandemic. If anything, preliminary evidence indicates that the opposite is
28
true and funds were targeted towards areas less severely affected by the virus. We do find
that bank heterogeneity played an important role in mediating funds. We construct a new
measure of geographic exposure of regions to banks that over or underperformed in terms of
PPP allocation relative to their share of small business lending. States with higher exposure
to banks that performed well in terms of bank PPP exposure also saw higher levels of PPP
lending.21
Using a number of data sources, we do not find evidence that the PPP had a substantial
effect on employment during the first round of the program. Our estimates are precise enough
to rule out modest employment effects. While there appears to be little effect of the program on
employment, the program may have played an important role in promoting financial stability.
Firms with greater exposure to the PPP hold more cash on hand, and are more likely to make
loan and other scheduled payments. Measuring these responses is critical for evaluating the
social insurance value of the PPP and similar policies and designing them effectively. As data
become available, we will continue to examine whether the program merely delayed firms’
inability to meet commitments or whether there are other medium-term and long-term effects.
21
The analysis here focuses on ex ante targeting of the PPP, that is, the distribution of funding provided at
the start of the program. Ultimate targeting will depend on the extent of loan forgiveness and defaults, as well
as subsequent changes to the PPP, including conditions for recoupment based on ex post economic hardship
and changes to program eligibility criteria going forward. See Hanson, Stein, Sunderam and Zwick (2020) for a
discussion of these dynamic policy considerations in the design of business liquidity support during the pandemic.
29
References
Agarwal, Sumit, Gene Amromin, Itzhak Ben-David, Souphala Chomsisengphet, Tomasz
Piskorski, and Amit Seru. 2017. “Policy Intervention in Debt Renegotiation: Evidence from
the Home Affordable Modification Program.” Journal of Political Economy, 125(3): 654–712.
Allcott, Hunt, Levi Boxell, Jacob Conway, Matthew Gentzkow, Michael Thaler, and David Y
Yang. 2020. “Polarization and Public Health: Partisan Differences in Social Distancing during
the Coronavirus Pandemic.” NBER Working Paper, , (w26946).
Atkeson, Andrew, Adrien d’Avernas, Andrea Eisfeldt, and Pierre-Olivier Weill. 2018. “Gov-
ernment Guarantees and the Valuation of American Banks.” NBER Macroeconomics Annual.
Bachas, Natalie, Olivia Kim, and Constantine Yannelis. 2020. “Loan Guarantees and Credit
Supply.” Journal of Financial Economics.
Baker, Scott R, Nicholas Bloom, Steven J Davis, and Stephen J Terry. 2020. “Covid-Induced
Economic Uncertainty.” National Bureau of Economic Research.
Baker, Scott R, Robert A Farrokhnia, Steffen Meyer, Michaela Pagel, and Constantine Yan-
nelis. 2020a. “How Does Household Spending Respond to an Epidemic? Consumption Dur-
ing the 2020 COVID-19 Pandemic.” National Bureau of Economic Research.
Baker, Scott R, Robert A Farrokhnia, Steffen Meyer, Michaela Pagel, and Constantine Yan-
nelis. 2020b. “Income, Liquidity, and the Consumption Response to the 2020 Economic Stim-
ulus Payments.” National Bureau of Economic Research.
Barrios, John, and Yael Hochberg. 2020. “Risk Perception Through the Lens Of Politics in the
Time of the COVID-19 Pandemic.” Working Paper.
Barrios, John M., Efraim Benmelech, Yael V. Hochberg, and Luigi Zingales. 2020. “Civic
Capital and Social Distancing During the COVID-19 Pandemic.” NBER working paper.
Barrios, John, Michael Minnis, William Minnis, and Joost Sijthoff. 2020. “Assessing the
Payroll Protection Program: A Framework and Preliminary Results.” Working paper.
Barro, Robert J, José F Ursua, and Joanna Weng. 2020. “The Coronavirus and the Great
Influenza Epidemic.”
30
Barrot, Jean-Noel, Thorsten Martin, Julien Sauvagnat, and Boris Vallee. 2019. “Employ-
ment Effects of Alleviating Financing Frictions: Worker-level Evidence from a Loan Guaran-
tee Program.” Working Paper.
Bartik, Alexander W., Marianne Bertrand, Zoe B. Cullen, Edward L. Glaeser, Michael Luca,
and Christopher T. Stanton. 2020. “How Are Small Businesses Adjusting to COVID-19?
Early Evidence from a Survey.” Proceedings of the National Academy of Sciences.
Bartik, Alexander W., Marianne Betrand, Feng Lin, Jesse Rothstein, and Matt Unrath.
2020. “Measuring the labor market at the onset of the COVID-19 crisis.” Brookings Papers on
Economic Activity.
Berger, David W, Kyle F Herkenhoff, and Simon Mongey. 2020. “An SEIR Infectious Disease
Model with Testing and Conditional Quarantine.” National Bureau of Economic Research
Working Paper 26901.
Brevoort, Kenneth P., John A. Holmes, and John D. Wolken. 2010. “Distance Still Matters:
The Information Revolution in Small Business Lending and the Persistent Role of Location,
1993-2003.” Board of Governors of the Federal Reserve System.
Buffington, Catherine, Carrie Dennis, Emin Dinlersoz, Lucia Foster, and Shawn Klimek.
2020. “Measuring the Effect of COVID-19 on U.S. Small Businesses: The Small Business
Pulse Survey.” Census for Economic Studies, working paper.
Cajner, Tomaz, Leland Crane, Ryan A. Decker, John Grigsby, Adrian Hamins-Puertolas,
Erik Hurst, Christopher Kurz, and Ahu Yildirmaz. 2020. “The U.S. Labor Market During
the Beginning of the Pandemic Recession.” NBER working paper.
Chetty, Raj. 2008. “Moral Hazard Versus Liquidity and Optimal Unemployment Insurance.”
Journal of Political Economy, 116(2): 173–234.
Chetty, Raj, John N. Friedman, Nathaniel Hendren, and Michael Stepner. 2020. “How did
COVID-19 and Stabilization Policies Affect Spending and Employment? A New Real-time
Economic Tracker Based on Private Sector Data.” NBER working paper.
Coibion, Olivier, Yuriy Gorodnichenko, and Michael Weber. 2020. “Labor Markets During
the COVID-19 Crisis: A Preliminary View.” Fama-Miller Working Paper.
Cororaton, Anna, and Samuel Rosen. 2020. “Public Firm Borrowers of the US Paycheck Pro-
tection Program.” SSRN working paper.
31
Cox, Natalie, Peter Ganong, Pascal Noel, Joseph Vavra, Arlene Wong, Diana Farrell, and
Fiona Greig. 2020. “Initial Impacts of the Pandemic on Consumer Behavior: Evidence from
Linked Income, Spending, and Savings Data.” Brookings Papers on Economic Activity.
De Marco, Filippo. 2020. “Public Guarantees for Small Businesses in Italy during COVID-19.”
SSRN working paper.
Dingel, Jonathan I., and Brent Neiman. 2020. “How Many Jobs Can Be Done at Home.”
Journal of Public Economics.
Ding, Wenzhi, Ross Levine, Chen Lin, and Wensi Xie. 2020. “Social Distancing and Social
Capital: Why U.S. Counties Respond Differently to COVID-19.” NBER working paper.
Eichenbaum, Martin S, Sergio Rebelo, and Mathias Trabandt. 2020. “The Macroeconomics
of Epidemics.” National Bureau of Economic Research.
Elenev, Vadim, Tim Landvoigt, and Stijn Van Nieuwerburgh. 2020. “Can the Covid Bailouts
Save the Economy?” National Bureau of Economic Research.
Gale, William. 1990. “Federal Lending and the Market for Credit.” Journal of Public Economics,
, (42): 177–93.
Gale, William. 1991. “Economic Effects of Federal Credit Programs.” American Economic Re-
view, 81(1): 133–52.
Gallipoli, Giovanni, and Christos Makridis. 2018. “Structural Transformation and the Rise
of Information Technology.” Journal of Monetary Economics, 97: 91–110.
Ganong, Peter, and Pascal Noel. 2018. “Liquidity vs. Wealth in Household Debt Obligations:
Evidence from Housing Policy in the Great Recession.” National Bureau of Economic Re-
search.
Gonzalez-Uribe, Juanita, and Su Wang. 2019. “Dissecting the Effect of Financial Constraints
on Small Firms.” Working Paper.
Gormsen, Niels Joachim, and Ralph SJ Koijen. 2020. “Coronavirus: Impact on Stock Prices
and Growth Expectations.” University of Chicago, Becker Friedman Institute for Economics
Working Paper, , (2020-22).
Granja, João, Christian Leuz, and Raghuram Rajan. 2018. “Going the Extra Mile: Distant
Lending and Credit Cycles.” National Bureau of Economic Research.
32
Guerrieri, Veronica, Guido Lorenzoni, Ludwig Straub, and Ivan Werning. 2020. “Macroe-
conomic Implications of COVID-19: Can Negative Supply Shocks Cause Demand Shortages.”
Working paper.
Hanson, Samuel, Jeremy Stein, Adi Sunderam, and Eric Zwick. 2020. “Business Conti-
nuity Insurance and Business Continuity Loans: Keeping America’s Lights on During the
Pandemic.” Policy Brief.
Holmstrom, Bengt, and Jean Tirole. 1998. “Private and Public Supply of Liquidity.” Journal
of Political Economy, 106(1): 1–40.
House, Christopher, and Matthew Shapiro. 2008. “Temporary Investment Tax Incentives:
Theory with Evidence from Bonus Depreciation.” American Economic Review, 98(3): 737–
768.
Humphries, John Eric, Christopher Neilson, and Gabriel Ulyssea. 2020. “Information Fric-
tions and Access to the Paycheck Protection Program.”
Jones, Callum, Thomas Philippon, and Venky Venkateswaran. 2020. “Optimal Mitigation
Policies in a Pandemic.” Working Paper.
Kaplan, Greg, and Gianluca Violante. 2014. “A Model of the Consumption Response to Fiscal
Stimulus Payments.” Econometrica, 82(4): 1199–1239.
Kelly, Bryan, Hanno Lustig, and Stijn Van Nieuwerburgh. 2016. “Too-Systemic-to-Fail: What
Option Markets Imply about Sector-Wide Government Guarantees.” American Economic Re-
view, 106(6): 1278–1319.
Lelarge, Clair, David Sraer, and David Thesmar. 2010. “Entrepreneurship and Credit Con-
straints: Evidence From a French Loan Guarantee Program.” International Differences in En-
trepreneurship, 243–273.
Lucas, Deborah. 2016. “Credit Policy as Fiscal Policy.” Brookings Papers on Economic Activity,
, (Spring): 1–57.
Makridis, Christos A., and Cary Wu. 2020. “Ties that Bind (and Social Distance): How Social
Capital Helps Communities Weather the COVID-19 Pandemic.” SSRN working paper.
Makridis, Christos A., and Jonathan Hartley. 2020. “The Cost of COVID-19: A Rough Esti-
mate of the 2020 GDP Impact.” Mercatus Center, Policy Brief Special Edition.
33
Makridis, Christos A., and Jonathan Rothwell. 2020. “The Real Cost of Political Polarization:
Evidence from the COVID-19 Pandemic.” SSRN working paper.
Mian, Atif, and Amir Sufi. 2012. “The Effects of Fiscal Stimulus: Evidence from the 2009 Cash
for Clunkers Program.” Quarterly Journal of Economics, 127: 1107–1142.
Mulligan, Casey B. 2020. “Economic Activity and the Value of Medical Innovation During a
Pandemic.” NBER Working paper.
Mullins, William, and Patricio Toro. 2017. “Credit Guarantees and New Bank Relationships.”
Policy Research Working Paper No, 8241.
Papanikolaou, Dimitris, and Lawrence D. W. Schmidt. 2020. “Working Remotely and the
Supply-side Impact of COVID-19.” Working paper.
Parker, Jonathan A, Nicholas S Souleles, David S Johnson, and Robert McClelland. 2013.
“Consumer Spending and the Economic Stimulus Payments of 2008.” American Economic
Review, 103(6): 2530–53.
Smith, Bruce. 1983. “Limited Information, Credit Rationing, and Optimal Government Lend-
ing Policy.” American Economic Review, 73(3): 305–18.
Zwick, Eric. Forthcoming. “The Costs of Corporate Tax Complexity.” American Economic Jour-
nal: Economic Policy.
Zwick, Eric, and James Mahon. 2017. “Tax Policy and Heterogeneous Investment Behavior.”
American Economic Review, 107(1): 217–248.
34
Figure 1: Histogram of Bank Paycheck Protection Program Exposure (PPPE)
Panel A of Figure 1 plots the distribution of the measure of relative bank performance in the PPP based on the volume of PPP loans granted and
Shar eVol.P P P−Shar eSB LM ar ket
stock of small business loans at the bank as of fourth quarter of 2019. We compute this measure as: P P P E b = (Shar eVol.P P P+Shar eSB LM ar ket)×0.5
.
Panel B of Figure 1 plots the distribution of the measure of relative bank performance in the PPP based on the number of PPP loans
granted and number of small business loans held by the bank as of fourth quarter of 2019. We compute this measure as: P P P E b =
Shar eN br.P P P−Shar eN br.SB LM ar ket
(Shar eN br.P P P+Shar eN br.SB LM ar ket)×0.5
. Data is obtained from the SBA and call reports.
Panel A: Histogram of Volume-based PPPE
50
45
40
35
Percent
25 30
20
15
10
5
0
-.5 -.4 -.3 -.2 -.1 0 .1 .2 .3 .4 .5
PPPE (Volume of Loans)
Panel B: Histogram of Number Loans-based PPPE
50
45
40
35
Percent
25 30
20
15
10
5
0
-.5 -.4 -.3 -.2 -.1 0 .1 .2 .3 .4 .5
PPPE (Number of Loans)
35
Figure 2: PPPE and Commercial Bank Size
Figure 2 plots average PPPE based on volume of PPP loans (Panel A) number of PPP loans (Panel B) and average amount of PPP loan (Panel
C) in each percentile size bin. The size bins stratify all commercial banks operating as of the fourth quarter of 2019 based on their total assets.
Data is obtained from the SBA and call reports.
Panel A: Volume-based PPPE and Size
Bank PPPE (Volume of Lending)
-.2 -.175-.15-.125 -.1 -.075-.05-.025 0 .025 .05 .075 .1
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100
Size Bin (Percentile)
Panel B: Number of Loans-based PPPE and Size
.35
.3
Bank PPPE (Number of Loans)
.05 .1 .15 .2 .25
0
-.05
-.1
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100
Size Bin (Percentile)
Panel C: Average Amount of Loan and Size
500
450
400
Av. Loan Amount (in $000)
200 250 300 350
150
100
50
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100
Size Bin (Percentile)
36
Figure 3: PPPE and PPP Allocation
Figure 3 plots the cumulative share of PPP and SBL lending by all banks whose PPPE is below x, where x ∈ (−0.5, 0.5). Data is obtained from the SBA and call reports.
1
Cumulative Share of PPP Lending
Cumulative Share of SBL Lending
Cumulative Shares of PPP Lending and SBL Lending
37
.2 .4 .6 .8
0
-.5 0 .5
Bank PPPE
Figure 4: Map of County Exposure to PPPE
Figure 4 plots the average exposure of each county to the volume-based PPPE. County exposure to PPPE is computed as the average of the PPPE of each bank with a branch presence in the county.
The PPPE of each bank is weighted by the share of deposits of the bank in the county as of June 30th, 2019. Data is from the SBA, Call Reports, and FDIC’s Summary of Deposits.
38
Figure 5: State Exposure to PPPE and PPP per Establishment
Figure 5 are scatterplots of the total PPP allocation per small business establishment in the state and the state exposure to the volume-based
PPPE (Panel A) and fraction of small business establishments receiving a PPP loan and the state exposure to the PPPE based on the number
of loans (Panel B). Data comes from SBA, Call Reports, Summary of Deposits, and County Business Patterns.
Panel A: State Exposure to Volume-Based PPPE and PPP Allocation per Small Business
70000
PPP allocation per Small Business Establishment (State)
ND
60000
HI
MN
WI
KS
MA
NE
OH
IA
DC
50000
SD NH
MO PA
IL INOK
AL
TX LA ME VT
MD TN
UT CT RI MI KY
VA CO
AK DE
40000
AR MS
GA NJ WY MT
ID
WV
NYWA
CA SC NC
AZ NM
30000
OR FL
NV
-.3 -.2 -.1 0 .1 .2
State PPPE
Panel B: State Exposure to Number-Based PPPE and Fraction of Small Businesses receiving PPP
.5
Fraction of Small Businesses receiving PPP (State)
ND
NE
SD
.4
OK
WY IA
MT ME
MSKS
HI
AR VT
.3 MO
MNWI
ID NH
AL
UT KY
CO LATNRI MA IN
TX OH
AK
IL PA WV
.2 SC CT
GA VADE
MD NM MI
NC
OR FL WA
NY
NJ
NV AZ DC
CA
.1
-.2 0 .2 .4
State PPPE
39
Figure 6: State Exposure to PPPE and Unmet PPP Demand
Figure 5 are scatterplots of the state exposure to the volume-based PPPE and the difference between percentage of businesses reporting having
applied to PPP and percentage of business that received PPP in each state (Panel A) and the state exposure to the PPPE based on the number
of loans and the difference between percentage of businesses reporting having applied to PPP and percentage of business that received PPP
in each state (Panel B). Data comes from the Census Bureau, SBA, Call Reports, Summary of Deposits, and County Business Patterns.
Panel A: State Exposure to Volume-Based PPPE and Difference between % Businesses Applying and Receiving PPP
CA
NJ
50
Diff. between % of businesses applying and receiving PPP
DC FL
NV GA
AZ OR
NY
CT
40
WA
NM NC
MDMA RI
VA
TX OH MI
PA
SC HI
IL CO
TN
30 UT LA
AKWV DE IN
NH
MS VT
KY
WI
20 MN AL
ID
IA SD MO ME
KS
AR OK
WY ND
10 MT
NE
-.2 -.1 0 .1 .2
State PPPE (Volume)
Panel B: State Exposure to Numer-Based PPPE and Difference between % Businesses Applying and Receiving PPP
CA
NJ
50
Diff. between % of businesses applying and receiving PPP
DC FL
NV OR GA
AZ
NY
CT
40
WA
NC NM
MD RI MA
VA TX OHPA MI
SC HI
CO IL
30 UT LA TN
DE
AK IN
WV
NH
MS VT
KY
WI
20 AL MN
ID
MO SD
IA ME
KS
AR OK
WY ND
10 MT
NE
-.2 0 .2 .4
State PPPE (Nbr. Loans)
40
Figure 7: Congressional District Exposure to PPPE and PPP per Establishment
Figure 7 are scatterplots of the total PPP allocation per establishment in the congressional district and the congressional district exposure to
the volume-based PPPE (Panel A) and fraction of establishments receiving a PPP loan and the congressional district exposure to the PPPE
based on the number of loans (Panel B). Data comes from the SBA, Call Reports, Summary of Deposits, and County Business Patterns.
Panel A: Congressional District Exposure to Volume-Based PPPE and PPP Allocation per Small Business
PPP allocation per Establishment (Congressional District)
0 50000 100000 150000
-.4 -.2 0 .2 .4
Congressional District PPPE
Panel B: Congressional District Exposure to Number-Based PPPE and Fraction of Small Businesses receiving PPP
Fraction of Establishments receiving PPP (Congressional District)
0 .1 .2 .3 .4 .5
-.4 -.2 0 .2 .4
Congressional District PPPE
41
Figure 8: Business Shutdowns and PPP Allocation by Congressional District
Figure 8 presents four scatterplots of the share of businesses in each state that shutdown that in the week of March 29th–April 4th and four alternative measures of allocation of PPP funds across
states. The figure on the top left plots the amount of PPP loans received by small businesses in each state divided by the total number of small businesses in the state. The figure on the top right
corner plots the fraction of small businesses in each state that received a PPP loan. The figure on the bottom left corner plots the fraction of small businesses in each state that received a PPP loan
and the state exposure to the PPPE measured in terms of the total volume of loans. The figure on the bottom right corner plots the fraction of small businesses in each state that received a PPP
loan and the state exposure to the PPPE measured in terms of the total number of loans.
Share of Business Shutdowns (Cong. Dist.)
Share of Business Shutdowns (Cong. Dist.)
(Date: 31/03/2020) (Date: 31/03/2020)
.8 .8
.6
.6
.4
.4
.2
.2
42
0
0 0 00 00 0
00 00 00
50 10 15 0 .1 .2 .3 .4 .5
PPP amount per establishment Fraction of Businesses receiving PPP
Share of Business Shutdowns (Cong. Dist.) Share of Business Shutdowns (Cong. Dist.)
(Date: 31/03/2020) (Date: 31/03/2020)
.8 .8
.6 .6
.4 .4
.2 .2
0 0
-.4 -.2 0 .2 .4 -.4 -.2 0 .2 .4
Congressional District PPPE (Volume) Congressional District PPPE (Nbr. Loans)
Figure 9: Decline in Hours Worked and PPP Allocation by Congressional District
Figure 9 presents four scatterplots of the decline in hours worked in each congressional district relative to a January baseline and four alternative measures of allocation of PPP funds across states.
The figure on the top left plots the amount of PPP loans received by small businesses in each state divided by the total number of small businesses in the state. The figure on the top right corner
plots the fraction of small businesses in each state that received a PPP loan. The figure on the bottom left corner plots the fraction of small businesses in each state that received a PPP loan and the
state exposure to the PPPE measured in terms of the total volume of loans. The figure on the bottom right corner plots the fraction of small businesses in each state that received a PPP loan and
the state exposure to the PPPE measured in terms of the total number of loans.
Ratio of Hours Worked relative to January Baseline
(Date: 31/03/2020) (Date: 31/03/2020)
Ratio of Hours Worked relative to January Baseline
0 .25 .5 .75 1
0 .25 .5 .75 1
43
0 0
00 00 00
50 00 00
10 15 0 .1 .2 .3 .4 .5
PPP amount per establishment Fraction of Businesses receiving PPP
Ratio of Hours Worked relative to January Baseline Ratio of Hours Worked relative to January Baseline
(Date: 31/03/2020) (Date: 31/03/2020)
0 .25 .5 .75 1 0 .25 .5 .75 1
-.4 -.2 0 .2 .4 -.4 -.2 0 .2 .4
Congressional District PPPE (Volume) Congressional District PPPE (Nbr. Loans)
Figure 10: PPP Allocation by Employment Shock Bin
Figure 10 stratifies congressional districts on 20 bins based on the share of Homebase businesses that shutdown in the week of March29th–
April 4th (Panel A) and on their decline in hours worked relative to a January baseline. The y-axis represents the fraction of businesses
receiving PPP funds in each bin computed as total number of PPP loans in that bin divided by total number of establishments of congressional
districts in that bin. Data is from SBA, Homebase, and County Business Patterns.
Fraction of Businesses receiving PPP
.15 .18 .21 .24 .27 .3
tric 10 12 13 14 15 16 17 18 19 tric
2 3 4 5 6 7 8 9 11
ts ts
Dis Dis
ffec ffec
ted ted
st A st A
Mo Lea
Business Shutdown Shock
Fraction of Businesses receiving PPP
.18 .21 .24 .27 .3
.15
tric 10 12 13 14 15 16 17 18 19 tric
2 3 4 5 6 7 8 9 11
ts ts
Dis Dis
ffec ffec
ted ted
st A st A
Mo Lea
Decline in Hours Worked Shock
44
45
W
W ee
ee k1
k1 Initial Unemployment Insurance Claims (as % of Jan 2020 BLS County Emp.) :J
an % Decline in Hours Worked relative to January
:J 19 0 .2 .4 .6 .8 1
an 0 1% 2% 3% 4% -J
19 W an
W -Ja
n2 ee 25
ee 5 k2
k2 :J
:J an
an 26
26 W -F
W -F
eb ee eb
1
ee 1 k3
k3 :F
:F eb
eb 02
02 W -F
W -F
eb ee eb
8
ee 8 k4
k4 :F
:F eb
eb 9-
W 9- W Fe
Fe ee b
ee
k5 b 15 k5 15
:F :F
eb eb
16 16
W -F
W -F ee eb
Low PPPE County
ee eb 22
High PPPE County
k6 22 k6
Low PPPE Cong. Dist.
:F
High PPPE Cong. Dist.
:F eb
eb 23
23 -F
-F W eb
W eb
29 ee 29
ee k7
k7 :M
:M ar
ar W 1-
W 1-
Ma ee Ma
ee r7 k8 r7
k8 : M
: M
ar ar
W 8-
W 8-
Ma ee Ma
ee r1
k9 r1
4
k9
: M 4
: M
ar ar
W 15
W 15
-M ee -M
ee ar k1 ar
k1 21 0: 21
0:
Ma Ma
r2 r2
Initial Lockdown Orders
W 2-
Initial Lockdown Orders
W 2- Ma
Ma ee
ee r2 r2
k1 8 k1 8
1: 1:
Ma Ma
r2 r2
PPP Launch
W 9-
PPP Launch
W 9- Ap
Ap
ee
k1 r4
ee
k1 r 4
2: 2:
Ap Ap
r5 W r5
W -A ee -A
ee pr pr
k1 11 k1 11
3: 3:
Ap Ap
Initial Unemployment Filings and County PPPE (Nbr.-Based)
r1 PPP Runs out of Funds
r1 PPP Runs out of Funds
W 2 -A W 2-
Ap
ee pr ee r1
k1 18 k1 8
4: 4:
Ap Ap
r1 r1
Ratio Hours Worked over Time and Congressional District PPPE (Nbr-Based)
Figure 11 shows the ratio of hours worked over time, Data is from SBA, the percent of businesses shut down, initial unemployment filings and small business revenue by above and below PPPE
W 9 -A W 9-
Ap
ee pr ee r2
k1 25 k1 5
5: 5:
Ap PPP 2nd Round begins Ap PPP 2nd Round begins
r2 r2
6 -M 6-
ay Ma
2 y2
W W
ee ee
k1 Decline in Y/Y Change in Total Consumer Spending relative to January k1
:J % Businesses Shutdown
:J an
an -.6 -.4 -.2 0 .2 19 0 .1 .2 .3 .4
19 -J
W -Ja W an
ee n2 ee 25
k2 5 k2
:J :J
an an
26
W -F 26
ee eb W -F
eb
k3 1 ee
k3 1
:F
eb :F
02 eb
W -F 02
ee
k4
eb
8 W -F
eb
ee 8
:F k4
eb :F
W 9-
Fe eb
ee b 9-
k5 15 W Fe
:F ee
k5 b 15
eb :F
16
W -F eb
ee eb 16
k6 22 W -F
Low PPPE County
eb
Figure 11: PPPE and Post-PPP Outcomes
:F ee
High PPPE County
eb k6 22
23 :F
High PPPE Cong. Dist. Low PPPE Cong. Dist.
-F eb
W eb 23
ee 29 -F
k7
:M W eb
29
ee
ar k7
W 1- :M
ee Ma
k8 r7 ar
: M W 1-
Ma
ar ee r7
W 8- k8
ee Ma : M
k9 r1 ar
: M 4 W 8-
ee Ma
ar r1
W 15 k9 4
ee -M : M
k1 ar ar
0: 21
Ma W 15
-M
ee
r2 k1 ar
W 2-
Ma Initial Lockdown Orders 0: 21
ee
k1 r2 Ma
1: 8 r2
Initial Lockdown Orders
Ma W 2-
ee Ma
r2 r2
W 9-
Ap PPP Launch
k1 8
ee 1:
k1 r4 Ma
2: r2
Ap W 9- PPP Launch
W r5
-A ee Ap
r
ee pr k1 4
k1 11 2:
3: Ap
Ap r5
r1 W -A
W 2 -A PPP Runs out of Funds ee pr
ee pr k1 11
Small Business Revenue and County PPPE (Nbr.-Based)
k1 18 3:
4: Ap
Ap r1 PPP Runs out of Funds
r1 W 2-
W 9 -A ee Ap
ee r1
k1 pr
25 k1 8
% Businesses Shutdown and Congressional District PPPE (Nbr-Based)
5: 4:
Ap
PPP 2nd Round begins
Ap r1
W r2 9-
ee 6 -M W Ap
ay ee r2
k1
6: 2 k1 5
5:
PPP 2nd Round begins
Ma Ap
y3 r2
-M 6-
ay Ma
10 y2
exposure. Homebase, and County Business Patterns.
Figure 12: Weekly Elasticities of Employment and Firm Outcomes to PPPE
Figure 12 plots coefficients and respective standard errors of regression analyses investigating the week-by-week elasticities between employment and firm outcomes and exposure to PPPE. The
top-left figure plots the coefficients and respective standard errors of an empirical specification that augments the specification of Panel A of Table 5 column (4) to include a full set of interactions
between the Congressional District PPPE and week dummy variables. The top-left figure plots the coefficients and respective standard errors of an empirical specification that augments the
specification of Panel B of Table 5 column (4) to include a full set of interactions between the Congressional District PPPE and week dummy variables. The bottom-left figure plots the coefficients
and respective standard errors of an empirical specification that augments the specification of Table 6 column (4) to include a full set of interactions between the County PPPE and week dummy
variables. The bottom-right figure plots the coefficients and respective standard errors of an empirical specification that augments the specification of Table 7 column (4) to include a full set of
interactions between the County PPPE and week dummy variables.
Weekly Elasticities of Shutdown and Congressional District PPPE Weekly Elasticities of Weekly Hours Worked and Congressional District PPPE
State-Industry-Week and Firm Fixed Effects State-Industry-Week and Firm Fixed Effects
.06
0 Vertical bands represent +/- 1.96 * St. Error of each point estimate
Initial Lockdown Orders PPP Runs out of Funds PPP 2nd Round begins
PPP Launch
.04
-.02
Beta Beta
Initial Lockdown Orders PPP Runs out of Funds PPP 2nd Round begins
.02
PPP Launch
-.04
46 0
Vertical bands represent +/- 1.96 * St. Error of each point estimate
-.06
-.02
b1 b8 15 2
b2 b2
9 ar7 ar1
4
ar2
1
ar2
8 r4 Ap r18 r25 2
ay b1 b8 15 2
b2 b2
9 ar7 ar1
4
ar2
1
ar2
8 r4 Ap r18 r25 2
ay
-Fe -Fe 9-F -M -Ap r11 -Ap -Ap -Fe -Fe 9-F -M -Ap r11 -Ap -Ap
-Fe -Fe -M 5 -M -M -M -Fe -Fe -M 5 -M -M -M
26 02 eb ar1 r29 r5- r26 26 02 eb ar1 r29 r5- r26
16 23 ar8 r22 r12 r19 16 23 ar8 r22 r12 r19
an :F :F :M ar1 Ma Ap Ap an :F :F :M ar1 Ma Ap Ap
:J eb eb :F :F : M Ma Ap Ap :J eb eb :F :F : M Ma Ap Ap
eb eb k7 : M 2: 5: eb eb k7 : M 2: 5:
k2 k3 k4 k8 k1 k1 3: 4:
k1
k2 k3 k4 k8 k1 k1 3: 4:
k1
k5 k6 We k9 0: 1: k1 k1 k5 k6 We k9 0: 1: k1 k1
We We We e We k1 We We We We e We k1 We
e e e We We We We e We We We e e e We We We We e We We We
e e e e We e e e e e e e e We e e e e
e e
Weekly Elasticities of Initial County UI Claims and County PPPE Weekly Elasticities of Small Business Revenue and County PPPE
State-by-Week and County Fixed Effects State-Industry-Week and County-Industry Fixed Effects
PPP 2nd Round begins
.2 .08
Initial Lockdown Orders PPP Runs out of Funds
Initial Lockdown Orders PPP Runs out of Funds PPP 2nd Round begins
Vertical bands represent +/- 1.96 * St. Error of each point estimate
PPP Launch
PPP Launch
.06
Vertical bands represent +/- 1.96 * St. Error of each point estimate
0
.04
Beta Beta
.02
-.2
0
-.4 -.02
8 15 22 29 r7 r14 1 r28 r4 r11 r18 r25 y2 b1 8 15 22 29 r7 r14 1 r28 r4 r11 r18 r25 y2
-Fe -Fe -Fe -M ar2 -Ap -M -Fe -Fe -Fe -Fe -M ar2 -Ap -M
b 9-F a -M -M Ap -Ap -Ap a b 9-F a -M -M Ap -Ap -Ap a
02 eb b b ar1 a 5-M a r29 r5- 26 02 eb b b ar1 a 5-M a r29 r5-
16 23 ar8 r22 r12 r19 r26 16 23 ar8 r22 r12 r19 r26
:F :F :M ar1 Ma Ap Ap an :F :F :M ar1 Ma Ap Ap
eb eb :F :F : M Ma Ap Ap :J eb eb :F :F : M Ma Ap Ap
eb eb k7 : M 2: 5: eb eb k7 : M 2: 5:
k3 k4 k8 k1 k1 3: 4:
k1
k2 k3 k4 k8 k1 k1 3: 4:
k1
k5 k6 We k9 0: 1: k1 k1 k5 k6 We k9 0: 1: k1 k1
We We e We k1 We We We We e We k1 We
e e We We We We e We We We e e e We We We We e We We We
e e e e We e e e e e e e e We e e e e
e e
Year Year
Figure 13: % Receiving PPP and State Exposure to PPPE (Industry×State Level)
Figure 13 are scatterplots of the state exposure to the number-based PPPE and the percentage of businesses reporting receiving PPP at the
state-by-industry level. The plots represent the evolution of the two variables for each survey week. Data comes from the Census Bureau,
SBA, Call Reports, Summary of Deposits, and County Business Patterns.
Week 15: April26 - May2nd Week 16: May3rd - May9
100 100
90 90 MSKSOK
OK SC NEMS
MO
MO NHOK
80 MO 80 ALLA MO AR MT
NEIA OK
MT OK
IA LA TN MSKS
WI
MO IAKS
MT ID IL LA OH MNWI IN KS OK
OK NC IL TN MA KY IAKS
IAKSOK COVA SC
% Businesses reporting receiving PPP % Businesses reporting receiving PPP
WI TN
CT AR IA
70 ID WI AR
ARNEIA
MS
70 NY
UT GA
ID
CO
MD
ILALLA TNOH
TX
AL MA
PAMN
MAMI WI IN MS
NJ
WA
NY
UT MD
VA ILALTX TN
TNOH
PAMN
MA KYMO
MO
WI IN
INE
NH
HI ME
AL MN ARNE KS
NE NHOK NC
GA VA SCILAL
TX TNOH
PAMN WI AR EIAKSOK
60 VA IL LA MN
MNMO IN IA ME 60
AZ
OR
FL
FL NJ
NY
FL WA
NY ID
CO
ID TX
ILALNM
LA
TX
CT OHMA
PA KY
MI
MN IN NH KSOK VT
ME
TX TN MA MN AZ
AZ OR NJ
NY IDMD SC
IL
IL TX
TX TN
TN PA MI
MN WI
MO IN NH
UT VA ILAL LA OH MN MI
MNWI IN MSKS NV
AZ OR UT
FL NJ
NY NC
GACO
NC MD SC
OH
IL CT PA MI KY
MI
MNWI KS
NH ME
COVA IL TX TNOH MN MO IN OR UT
NY GACO
NC PA
LACT OH MNMO
MAMI
MI IN IA OK
AR
NY
UT NC
GACO TX
ILAL PA
CT OH
PAMNKY
MI
WI
WI
MO KSOK NV CA
WA
UT
NJ
VA SC
IL
ILTXNM PA
OH
PA MI WI
MN
MO IN
50 COVA LA MA KYWI IA 50 OR FL WA
NJ GACO
MD CT
TNOH
LACT
TX PA MI
OR WA VA SC NM
TX PA
OHMAMI WI CA
CA FL WA
NJ
WA
NY NC
GA
GA MD
MD SC OHMA WI AR OK
UT NC
GACO SC IL CT
TN PA
MAMI MO IN CA
AZ OR FL
FL NY
WA
UT MD
GACOVA ILTXCT IN
WA NC
GACO
NC ILTX
IL LA OH
TX PA
OHMA
PA KYWI
WI IN
IN
NV
NV CA FL
FL GA
NC VA CT MNMO
WI
40 AZ NC MD SC
ILAL
AL
TXCTTNOH MN 40 CA
AZ
AZ OR NC VA MA
CA WA
NJ
WA GACO
NC MD IL MI IN NV CA FL
AZ OR UT
NY VA SCTX MI MO IN CA
AZ NJ OH
FL NJ
NY
NY GACO TX MA
PA
PA AZ
CA OR
WA MD
VA
FL NY
NY
FL WA NC MD PA
OH CA DC WA
30 FL VA SCTX
MD
NCCO MA
PA 30 NJ
WA
NY MI
FL NJ
NY
WA MI OR CO
NV AZ
AZ
CA OR FL NY
NJ
FL WA NC VA ILTX OH MI VA
CA NJ NY
CA FL WA
NJ NC
GA
20 CA
CA FL NY
FL WA 20
CA
FL
10 CA 10
0 0
-.2 0 .2 .4 -.2 0 .2 .4
State PPPE (Nbr. Loans) State PPPE (Nbr. Loans)
Week 17: May10 - May16 Week 18: May17 - May23
100 MT 100 AL AR
LA AR
AL UT TN
CT
UT ID MS ME ID AR IA
90 TN IA 90 SCTX
VA ALNM MS OK
MN MO ARNEIA OK NV FL UT LACT
NM TN MN
KY IN KS ME
NC
MD ILAL LA TN
TX MNKY IN NH OR UT GA MD
GA VA AL LA TNOH
TX
TX PA
MI
MN WI
WI IA OK
NJ VA SC MI WI NH CA NY MD SCIL TNOH
LACT PA
OH MO IN MS
80 OR WA
UT
NY
GA MD
VA ILALTXCT
CT MI
OH MN WI NH
NEIA
KS 80 AZ OR
OR FL NJ GACO
NCCO IL CT OH
IL PAMNWI
MN
MA
MAMI MO
WI NE NH OK
KS ME
AZ
AZ FL
NJ
NY GA VA SC TX
ALLA
NM OHOH MNWIMO
WI IN IAKS
NH OK NV NY
NJ NC VA IL TNOH
CT OHMA
MN
MAMI MO ARNEIA
NHOK
OK
AZ ILAL
TX TN MAMA
MN MT OK
IA
MS VT FL UT
NY GA MD
CO IL LACT
SC TNOH
PA MO IAKS
KS
NV GACOVA SC
MD TX IN
AR
IN KSOK ME AZ FL NJ
NY ID
COVA IL MN
OH MI WI
MA IN E KS
CA OR FL NC MD ILTXLACT
TN PA
MA
PA MI
KYMO IN MS CA MD IL
VA SC WI
MO
WI IN
% Businesses reporting receiving PPP % Businesses reporting receiving PPP
NJ
UT
NY
FL WA CO LA
LA TN MI WA GACO
NC IL TX PA
MAMI IN
70 AZ NY
NY
UT
WA NC VA ILTX TN
TN
TN PA
OH
PA
OH MNKY
MI
MO
PR
WI
MO IN KS OK ME 70
NV CA WA
NJ
UT NC
GACO SC TX
ILAL
IL NM
TX
TN
TNOH
PA MI
PAMN
MNKYWI IN IANHOK
IN VT
CA
AZ NJ
NY ID LA
TX CT PA
MA PR
MI WI OK CA OR FL NJ ALLACT
LA OH MN
MAMI
KY MT
NV
CA OR FL WA
NY
NY
WA
UT GA
NC
GA
CO
COVA SC
ID
AL
ILTXLA
NM PA
OH
CT OHMA
PA
MAMI
MN
MO
MO
WI ARNE NH
AR IA KS CA
CA OR FL WA
FL WA
FL NY
NJ
NY
GA
NC
GA MD ILTX
COVA
NC MD
MD TX PA
MAMI
KY INE
OK
OR NY
NJ GACOVA IL NV CA WA
NV
NV CA
CA
AZ OR
OR FL WA
FL WA NC MD ILTX
OH
OH MN
MA
MN PR CA
CA NJ ID TX
NM
TX CT PA
KY
TN PA MI WI AR MS
60 CA OR FL WA
NY
NJ
NY OHMA
PA MI WI
KY
MI
IN
ARE OK
60 AZ OR NY
FL WA ID
NCCO SC
CO
AL PA PR IN IAKS
CA DC UT
WA
NJ
FL WA
NC
GA
ID VA IL CT
GA VA TX TN PAMNWIMO
MO CA
AZ
OR
OR FL WA
UT GA VA SC LA OH MO
PA MI WI
CA
AZ OR NC VA SCTX KY AZ
AZ OR FL GA MD
NC TX MAMI WI IA CA DC FL NJ
NY GA MD
NC MD
50 CA
CA OR FL UT NC
ID NM MA WI 50 AZ
AZ
CA
OR FL UT
FL UT
NY ID
CO
COVA
VA TX PAMNMO
FL NJ SC AZ GA
ID PR
NV AZ IL
WA
NY NY NC IL MA
40 40 AZ FL
CA MD SC
PA WA
CO
30 30
20 20
10 10
0 0
-.2 0 .2 .4 -.2 0 .2 .4
State PPPE (Nbr. Loans) State PPPE (Nbr. Loans)
Week 19: May24 - May30 Week 20: May31 - Jun6
100 CT OH NE 100 MN KS
LA
LA MS OH MO NEMS
ID FL GA VA SC NH ME
VT
MD
CO WI MS OK ID VA ALLA CT
TN EMS
90 ILAL CT
SC TNOH 90
FL
TN MA
LA TN PAMNKY AR IAKS
IN OK
GA VA SC LACT
TN MI INEIA KS ILTX CT OH
AZ
AZ FL NJ NC MD
CO TX
ILAL
VA IL TX TN
LACT MA
TN PAMNMO
MA IN MS
AR
OK AZ
CA FL NCCO
IL LACT OHMAMI
MNKY
MI
WI INE KS
AR OK
NV
CA
AZ OR
UT
NY
FL WA NC
GA
GACO
NC ILALLACT OH
IL TX PA MI
MN
MO
WI
PR AR KS
NH NY GACO
IDMD
CT
ILAL TNOH PA MO
LA OHMA MI MO AR MT OK
80 CA OR FL NJ LACT
TX PA
PA
MA
TN PA MI
MI INEMT KS
NH 80 AZ NY
UT
NJ
NJ
IL
VA SC
MD TX
ILAL
SC MN
MNKYWI IAKSOK
CA OR FL WA
UT
NY
NJ
NY
WA
UT
GACO
IDMD
VA SC TX
ILAL TNOHMA
LACT KY
MI
MN
MNKY
KYWI
MO
WI
IN
AR E
IA
IN NH
HI
OK
VT
VT NV AZ
CA
CA OR FL NY
FL WA
NJ
NY
NC
NCCO
MD
VA SC
TX
LA
NM
TX
TX
CT
TN PA
CT MA
PA
MA
MA
PA
MI
KY
MI
WI
MO
WI
MO INEIA
IN
MSNHOK
KS
OK
ME
CA
AZ
AZ
OR
OR FL WA
NJ
NY
NJ
UT GA
ID VA
VA
MD ILAL
CO TX OH
NMTNOH
TX MA
PA
KYWI
MO
MO IN
IN NH
OK
KS
ME CA
AZ OR NJ
FL WA
NY
UT NC
GA
NCCO
MD
IDMD ILILAL
TX
LACT
TNOH MN
MA
PA KY
MI
MO
PR IN IAKS
IN
OK
% Businesses reporting receiving PPP % Businesses reporting receiving PPP
70 NV CA FL WA
NJ
NY GA
NC SC MN
MNWI
MA IN KS 70 NV FL WA
NY COVA ILTX TN MA
PA WI INE NH KS ME
CA UT
NY
FL NJ GACO
NC ILAL
TX TNOH
TX TN MAKY
MA
PA
WI NEIANHOK CA OR NY
FL WA
NY
NJ
NY GACOVA SC
SC
PAMN
LA TNOH
TX MN WI AR
IN
IN IA OK
OR FL NJ
WA MD OH MO IA KS CA OR UT
FL NJ GA VA TX OH MI PR
MO
WI IANHHI
OR FL
FL NJ
NY NC
GA
NCCOVA IL
ID VA SCTX CT OH MI
MN WI
MO
MO IN MS NV CA
AZ OR VA ILAL
IDMD MO
PR
60 NV
CA
CA
AZ
CA
FL WA
FL NCCOVA
MD
VA IL TX TNOHMAMI
MN PR IN IA
AR ME 60
NV CA
AZ OR NY
NJ
UT
WA NCCO
NC ILTXCT
SC MNMO
MA
TN PA
MI IANH
MS OK
NV CA OR NY
NC MD
VA TX TN PA
PAMN
MI
WI
MT
KS CA
AZ OR
OR
OR
NJ
NY
UT
WA GACO
NC MD ILTX
CO
TX MN
AZ SC TX MN AZ
CA OR NJ
FL WA IL OH
WA
NY
NY NC VA MAKY AZ
CA WA
NY ID
CO TX TN PA MI
NM IA FL WA IL
VA SC
50 CA
AZ OR FL TX 50 DC WA GA MD
NC
CA DC NY IL NM FL NC NM
NJ
WA OH VA OHMA
NC
GACO
MD PAMN IN
40 40
30 30
20 20
10 10
0 0
-.2 0 .2 .4 -.2 0 .2 .4
State PPPE (Nbr. Loans) State PPPE (Nbr. Loans)
47
Table 1: Top 20 Banks by Total Assets and PPPE
Table 1 reports individual bank statistics and the PPPE index for the 20 largest financial institutions in the United States. Total Assets is computed using information from fourth quarter 2019 call
reports. Share of total vol. PPP program is the total amount disbursed by each financial institution relative to the total amount disbursed under the first wave of the PPP. Share of SBL market is the
share of the total outstanding amount of small business loans held by each financial institution relative to the total outstanding amount of small business loans as of 2019:Q4. PPPE (volume) is
the volume-based bank PPP index. Total assets are in $ millions. Share of loans in PPP program is the total number of loans processed by each financial institution relative to the total number of
loans processed in the first wave of the PPP. Share of loans in SBL market is the share of the total number of outstanding small business loans held by each financial institution relative to the total
outstanding number of small business loans as of 2019:Q4. PPPE (Nbr. Loans) is the number-based bank PPP index.
(1) (2) (3) (4) (5) (6) (7)
Financial Institution Name Total Share Share PPPE Share Share PPPE
Assets of total of SBL (vol.) of loans of loans (Nbr.
vol. Market in PPP in SBL Loans)
PPP Market
JPMORGAN CHASE BANK, NATIONAL ASSOCIATION 2,337,707 3.892% 6.547% -0.127 1.432% 10.47% -0.380
BANK OF AMERICA, NATIONAL ASSOCIATION 1,866,841 1.199% 9.510% -0.388 .5565% 11.86% -0.455
WELLS FARGO BANK, NATIONAL ASSOCIATION 1,736,928 .0362% 6.502% -0.494 .0664% 4.308% -0.485
48 CITIBANK, N.A. 1,453,998 .3348% 2.121% -0.364 .4431% 9.729% -0.456
U.S. BANK NATIONAL ASSOCIATION 486,004 .6953% 3.327% -0.327 1.120% 5.643% -0.334
TRUIST BANK 461,256 3.160% 2.011% 0.111 2.078% 1.732% 0.045
CAPITAL ONE, NATIONAL ASSOCIATION 453,626 .0212% 2.822% -0.493 .0134% 10.38% -0.499
PNC BANK, NATIONAL ASSOCIATION 397,703 2.765% 1.124% 0.211 1.235% 1.373% -0.027
BANK OF NEW YORK MELLON, THE 342,225 0% .0024% -0.500 0% .0001% -0.500
TD BANK, N.A. 338,272 1.837% .6876% 0.228 1.698% .5698% 0.249
STATE STREET BANK AND TRUST COMPANY 242,148 0% 0% 0.000 0% 4.493% -0.500
CHARLES SCHWAB BANK 236,995 0% .0745% -0.500 0% .0039% -0.500
MORGAN STANLEY BANK, N.A. 229,681 0% .1445% -0.500 0% .0089% -0.500
GOLDMAN SACHS BANK USA 228,836 0% .0032% -0.500 0% .0001% -0.500
HSBC BANK USA, NATIONAL ASSOCIATION 172,888 .1411% .0845% 0.125 .0697% .0140% 0.332
FIFTH THIRD BANK, NATIONAL ASSOCIATION 167,845 .9991% .4589% 0.185 .5948% .1920% 0.256
ALLY BANK 167,492 .2639% 2.118% -0.389 .0631% 1.382% -0.456
CITIZENS BANK, NATIONAL ASSOCIATION 165,742 1.072% .8077% 0.070 1.148% .5274% 0.185
KEYBANK NATIONAL ASSOCIATION 143,390 2.370% .7295% 0.265 2.236% .2743% 0.391
BMO HARRIS BANK NATIONAL ASSOCIATION 137,588 1.385% 1.959% -0.086 .7133% .5413% 0.069
ALL OTHER BANKS 6,889,908 79.82% 58.96% -0.048 86.52% 40.97% 0.218
Table 2: PPPE and PPP Allocation
Table 2 reports the results of ordinary least squares (OLS) regressions examining the impact of the congressional district exposure to PPPE
on the cross-sectional allocation of PPP funds to congressional districts. The dependent variable of the specifications, Ln(Total PPP Allocation
per establishment), is the natural logarithm total amount of PPP funds disbursed to small businesses in each congressional district divided by
the number of establishments in the congressional district. The dependent variable of the specifications in Panel B, Fraction of Establishments
receiving PPP is the total number of PPP loans made to small businesses in each congressional district divided by the number of establishments
in the congressional district. Cong. Dist. Exposure to PPPE (Vol) is the congressional district average of the PPPE based on total amounts of
outstanding PPP and small business loans, weighted by the share of deposits of each bank in each congressional district. Cong. Dist. Exposure
to PPPE (Nbr.) is the congressional district average of the PPPE based on the number of outstanding loans, weighted by the share of deposits
of each bank in each congressional district. Ln(Payroll) is the natural logarithm of the sum of payroll of all establishments in the congressional
district. Ln(Employment) is the natural logarithm of total employment in the congressional district. Industry Shares are additional controls
for the share of establishments in each two-digit NAICS code industry. The specification of column (4) includes state fixed effects. Standard
errors are presented in parentheses, and are clustered at the level of the state. ***, **, and *, represent statistical significance at 1%, 5%, and
10% levels, respectively.
Panel A: Total PPP Allocation per Establishment at the Congressional District
(1) (2) (3) (4)
Ln(Total PPP Allocation per Establishment)
Cong. Dist. PPPE (Vol) 0.160∗∗∗ 0.174∗∗∗ 0.126∗∗∗ 0.148∗∗
(0.023) (0.023) (0.038) (0.067)
Ln(Total Payroll) 0.290 0.279 0.191
(0.223) (0.233) (0.242)
Ln(Employment) -0.089 -0.113 0.112
(0.397) (0.351) (0.376)
Observations 436 436 436 436
Adjusted R2 0.071 0.109 0.135 0.102
Industry Shares No No Yes Yes
State Fixed Effects No No No Yes
Panel B: Fraction of Establishment receiving PPP at the Congressional District
(1) (2) (3) (4)
Fraction of Establishments receiving PPP
Cong. Dist. PPPE (Nbr.) 0.068∗∗∗ 0.067∗∗∗ 0.056∗∗∗ 0.047∗∗∗
(0.005) (0.005) (0.007) (0.009)
Ln(Total Payroll) -0.017 -0.045 -0.029
(0.038) (0.036) (0.037)
Ln(Employment) 0.028 0.051 0.050
(0.062) (0.059) (0.066)
Observations 436 436 436 436
Adjusted R2 0.414 0.412 0.449 0.482
Industry Shares No No Yes Yes
State Fixed Effects No No No Yes
49
Table 3: Evolution of Homebase Employment Outcomes and Exposure to PPP
Table 3 reports the results of ordinary least squares (OLS) regressions examining the relation between the geographic allocation of PPP funds
during the first round and weekly firm employment outcomes measured using the Homebase data set between the second week of March until
the last week of April. The dependent variable in Panel A, Bus. Shutdowns, is an indicator variable that takes the value of one if the business
reported zero hours worked over the course of the week. The dependent variable in Panel B, Decline Hours Worked, is the decline in hours
worked in each establishment relative to the average hours worked in the same establishment during the last two weeks of January. Pre-Covid
is an indicator variable that takes the value of prior to the first week of March. This period represents the base period in the specifications
that include firm fixed effects and as a result this variable is omitted in such specifications. Pre-Lockdown is an indicator variable that takes
the value of one prior to the week of March 22nd – March 28th (exclusive) Post-Lockdown & Pre-PPP is an indicator variable that takes the
value of one during the weeks of March 22nd – March 28th and March 29th – April 4th. Post-PPP is an indicator variable that takes the value
of one following the week of April 5th – April 12th (inclusive), Fraction receiving PPP is the percentage of businesses that received PPP loans
during the first round in the congressional district where the firm is located. Congressional District PPPE (Nbr.) is the congressional district
average of the PPPE based on the number of outstanding loans, weighted by the share of branches of each bank in each congressional district.
Standard errors are presented in parentheses, and are double-clustered at the level of the state and week. ***, **, and *, represent statistical
significance at 1%, 5%, and 10% levels, respectively.
Panel A: Business Shutdowns
(1) (2) (3) (4)
Bus. Shutdown
Pre-Covid × Fraction receiving PPP -0.005
(0.004)
Pre-Lockdown × Fraction receiving PPP -0.032∗∗∗ -0.027∗∗∗
(0.010) (0.002)
Post-Lockdown & Pre-PPP × Fraction receiving PPP -0.104∗∗ -0.098∗∗
(0.042) (0.043)
Post-PPP × Fraction receiving PPP -0.103∗ -0.097∗
(0.049) (0.052)
Pre-Lockdown × Congressional District PPPE (Nbr.) -0.008∗∗ -0.007∗∗∗
(0.003) (0.002)
Post-Lockdown & Pre-PPP × Congressional District PPPE (Nbr.) -0.041∗∗∗ -0.040∗∗∗
(0.007) (0.007)
Post-PPP × Congressional District PPPE (Nbr.) -0.043∗∗∗ -0.042∗∗∗
(0.009) (0.010)
Observations 726868 726868 726868 726868
Adjusted R2 0.261 0.262 0.522 0.522
P-Value Diff. in Coefficients .8864 .5061 .9106 .6163
Firm Fixed Effects No No Yes Yes
State×Industry×Week Fixed Effects Yes Yes Yes Yes
Panel B: Decline in Hours Worked
(1) (2) (3) (4)
Decline Hours Worked
Pre-Covid × Fraction receiving PPP -0.007
(0.009)
Pre-Lockdown × Fraction receiving PPP 0.062∗∗ 0.068∗∗∗
(0.025) (0.024)
Post-Lockdown & Pre-PPP × Fraction receiving PPP 0.109∗∗ 0.113∗∗
(0.044) (0.047)
Post-PPP × Fraction receiving PPP 0.112∗∗ 0.116∗∗
(0.046) (0.049)
Pre-Covid × Congressional District PPPE (Nbr.) 0.000
(0.002)
Pre-Lockdown × Congressional District PPPE (Nbr.) 0.026∗∗∗ 0.026∗∗∗
(0.004) (0.003)
Post-Lockdown & Pre-PPP × Congressional District PPPE (Nbr.) 0.040∗∗∗ 0.040∗∗∗
(0.009) (0.010)
Post-PPP × Congressional District PPPE (Nbr.) 0.044∗∗∗ 0.043∗∗∗
(0.009) (0.009)
Observations 726868 726868 726868 726868
Adjusted R2 0.419 0.420 0.659 0.660
P-Value Diff. in Coefficients .813 .1661 .7928 .1718
Firm Fixed Effects No No Yes Yes
State×Industry×Week Fixed Effects Yes Yes Yes Yes
50
Table 4: Evolution of County Unemployment Outcomes and Exposure to PPP
Table 4 reports the results of ordinary least squares (OLS) regressions examining the relation between the geographic allocation of PPP funds
during the first round and weekly initial unemployment insurance filings at the county level between the second week of March until the last
week of April. The dependent variable, County Initial UI Filings Ratios, is the ratio between the weekly initial unemployment insurance filings
hand-collected from several state labor departments and county employment measured using the Bureau of Labor and Statistics January
County Employment data. Pre-Covid is an indicator variable that takes the value of prior to the first week of March. This period represents the
base period in the specifications that include county fixed effects and as a result this variable is omitted in such specifications. Pre-Lockdown
is an indicator variable that takes the value of one prior to the week of March 22nd – March 28th (exclusive) Post-Lockdown & Pre-PPP is
an indicator variable that takes the value of one during the weeks of March 22nd – March 28th and March 29th – April 4th. Post-PPP is
an indicator variable that takes the value of one following the week of April 5th – April 12th (inclusive), County PPPE (Nbr.) is the county
average of the PPPE based on the number of outstanding loans, weighted by the share of branches of each bank in each county. Weekly County
Covid-19 Cases (per capita) are the weekly per capita number of reported Covid-19 cases. Weekly County Covid-19 Deaths (per capita) are the
weekly per capita number of reported Covid-19 deaths. Social Distancing Index is the change in average distance travelled in the county using
individuals’ GPS signals. Standard errors are presented in parentheses, and are double-clustered at the level of the state and week. ***, **,
and *, represent statistical significance at 1%, 5%, and 10% levels, respectively.
(1) (2) (3) (4)
County Initial UI Filings Ratio
Pre-Covid × County PPPE (Nbr.) 0.009∗ 0.009∗
(0.005) (0.005)
Pre-Lockdown × County PPPE (Nbr.) -0.021 -0.019 -0.032∗∗∗ -0.029∗∗∗
(0.014) (0.011) (0.004) (0.006)
Post-Lockdown & Pre-PPP × County PPPE (Nbr.) -0.089∗ -0.077 -0.100 -0.096
(0.049) (0.052) (0.063) (0.064)
Post-PPP × County PPPE (Nbr.) -0.076∗∗ -0.050 -0.085∗ -0.070
(0.034) (0.037) (0.048) (0.048)
Weekly County Covid-19 Cases (per capita) 3.389 3.324∗∗
(2.050) (1.274)
Weekly County Covid-19 Deaths (per capita) 8.186 -32.817∗∗
(20.726) (11.587)
Social Distancing Index -0.787∗∗∗ -0.654∗∗∗
(0.199) (0.145)
Observations 23971 23889 23971 23887
Adjusted R2 0.758 0.762 0.830 0.832
P-Value Diff. in Coefficients .3909 .2305 .4621 .2591
County Fixed Effects No No Yes Yes
State×Week Fixed Effects Yes Yes Yes Yes
51
Table 5: Evolution of Small Business Revenues and Exposure to PPP
Table 5 reports the results of ordinary least squares (OLS) regressions examining the relation between the geographic allocation of PPP funds
during the first round and weekly small business revenues at the county×industry level between the second week of March until the last
week of April collected from Womply, a company specializing in processing revenue for small businesses. The dependent variable, Y/Y Change
in Total Consumer Spending, is the average year-on-year change in small business revenues in a given county and industry. Pre-Covid is an
indicator variable that takes the value of prior to the first week of March. This period represents the base period in the specifications that
include county×industry fixed effects and as a result this variable is omitted in such specifications. Pre-Lockdown is an indicator variable that
takes the value of one prior to the week of March 22nd – March 28th (exclusive) Post-Lockdown & Pre-PPP is an indicator variable that takes
the value of one during the weeks of March 22nd – March 28th and March 29th – April 4th. Post-PPP is an indicator variable that takes the
value of one following the week of April 5th – April 12th (inclusive), County PPPE (Nbr.) is the county average of the PPPE based on the
number of outstanding loans, weighted by the share of branches of each bank in each county. Weekly County Covid-19 Cases (per capita) are
the weekly per capita number of reported Covid-19 cases. Weekly County Covid-19 Deaths (per capita) are the weekly per capita number of
reported Covid-19 deaths. Social Distancing Index is the change in average distance travelled in the county using individuals’ GPS signals.
Standard errors are presented in parentheses, and are double-clustered at the level of the state and week. ***, **, and *, represent statistical
significance at 1%, 5%, and 10% levels, respectively.
(1) (2) (3) (4)
Y/Y Change in Total Consumer Spending
Pre-Covid × County PPPE (Nbr.) -0.068∗∗∗ -0.075∗∗∗
(0.009) (0.009)
Pre-Lockdown × County PPPE (Nbr.) -0.032∗∗ -0.045∗∗∗ 0.033∗∗∗ 0.028∗∗∗
(0.012) (0.011) (0.010) (0.009)
Post-Lockdown & Pre-PPP × County PPPE (Nbr.) -0.019∗∗∗ -0.035∗∗∗ 0.042∗∗∗ 0.033∗∗∗
(0.006) (0.007) (0.006) (0.007)
Post-PPP × County PPPE (Nbr.) -0.014∗∗ -0.033∗∗∗ 0.046∗∗∗ 0.029∗∗
(0.005) (0.008) (0.008) (0.010)
Weekly County Covid-19 Cases (per capita) -0.017 0.163
(0.306) (0.349)
Weekly County Covid-19 Deaths (per capita) 4.020 -7.032
(3.781) (8.209)
Social Distancing Index 0.255∗∗∗ 0.288∗∗∗
(0.071) (0.063)
Observations 210076 193328 209950 193216
Adjusted R2 0.362 0.361 0.665 0.668
P-Value Diff. in Coefficients .1093 .6282 .3683 .3787
County×Industry Fixed Effects No No Yes Yes
State×Industry×Week Fixed Effects Yes Yes Yes Yes
52
Table 6: % Receiving PPP and Exposure to State PPPE
Table 6 reports the results of ordinary least squares (OLS) regressions examining the relation between the geographic allocation of PPP funds during the first round and the percentage of firms
reporting having received PPP funds at the state-by-industry level in each week of the survey collected from the Census Small Business Pulse Survey. The dependent variable is the percentage
of businesses that received PPP funds in a state-by-industry group during each week of the survey. State PPPE (Nbr.) is the state average of the PPPE based on the number of outstanding loans.
Standard errors are presented in parentheses, and are double-clustered at the level of the state and week. ***, **, and *, represent statistical significance at 1%, 5%, and 10% levels, respectively.
53 (1) (2) (3) (4) (5) (6) (7)
% PPP Received
State PPPE (Nbr.) 35.865∗∗∗ 89.549∗∗∗ 51.603∗∗∗ 23.899∗∗∗ 28.381∗∗∗ 17.061∗∗∗ 24.315∗∗∗
(3.201) (8.169) (4.298) (3.594) (5.024) (3.610) (3.879)
Observations 2230 277 386 390 390 393 394
Adjusted R2 0.653 0.629 0.529 0.397 0.416 0.387 0.516
Industry Fixed Effects No Yes Yes Yes Yes Yes Yes
Industry×Week Fixed Effects Yes No No No No No No
Week: Full Sample Apr26–May2 May3–May9 May10–May16 May17–May23 May24–May30 May31–Jun6
Table 7: % Receiving PPP and Missed Loans Payments
Table 7 reports the results of ordinary least squares (OLS) and instrumental variables (IV) regressions examining the relation between the geographic allocation of PPP funds during the first round
and the percentage of firms reporting missing payments at the state-by-industry level collected from the Census Small Business Pulse Survey. The dependent variable is the percentage of firms
reporting a missed scheduled loan payment. % PPP Received is the percentage of businesses reporting having received PPP funds in a state-by-industry group. State PPPE (Nbr.) is the state average
of the PPPE based on the number of outstanding loans. Pre-PPP Decline Hours Worked is the average decline in hours worked in each state between January and the last week of March. Pre-PPP
State Covid-19 Cases (per capita) are per capita number of reported Covid-19 cases in the state. Pre-PPP State Covid-19 Deaths (per capita) are the weekly per capita number of reported Covid-19
deaths in the state. Pre-PPP State Social Distancing Index is the change in average distance travelled in the state until the end of March using individuals’ GPS signals. All specifications include
industry×week fixed-effects. Standard errors are presented in parentheses, and are clustered at the level of the state. ***, **, and *, represent statistical significance at 1%, 5%, and 10% levels,
respectively.
(1) (2) (3) (4)
OLS Reduced Form IV
First-Stage Second-Stage
Dep. Variable: % Missing Loan Payments % Missing Loan Payments % PPP Received % Missing Loan Payments
54 %PPP Received -0.005 -0.149∗∗
(0.014) (0.045)
State PPPE (Nbr.) -4.401∗∗∗ 34.402∗∗∗
(1.380) (3.055)
Pre-PPP Decline Hours Worked -3.536∗∗ -3.242∗∗∗ 8.034∗∗ -1.740
(1.330) (0.977) (3.041) (0.904)
Pre-PPP State Covid-19 Cases (per capita) 0.853∗ 0.801∗ 1.639 1.001∗
(0.466) (0.409) (0.995) (0.473)
Pre-PPP State Covid-19 Deaths (per capita) -26.308∗ -26.047∗ -7.200 -23.355
(13.770) (13.396) (39.721) (15.393)
Pre-PPP Social Distancing Index -5.795∗∗ -6.787∗∗∗ 14.786 -2.501
(2.803) (2.363) (8.984) (2.421)
Observations 2218 2487 2218 2218
Adjusted R2 0.573 0.554 0.664 -0.064
FStat 17.386
Industry×Week Fixed Effects Yes Yes Yes Yes
Table 8: % Receiving PPP and Missed Other Scheduled Payments
Table 8 reports the results of ordinary least squares (OLS) and instrumental variables (IV) regressions examining the relation between the geographic allocation of PPP funds during the first round
and the percentage of firms reporting missing payments at the state-by-industry level collected from the Census Small Business Pulse Survey. The dependent variable is the percentage of firms
reporting a missed other scheduled payment such as rent, utilities, and payroll. % PPP Received is the percentage of businesses reporting having received PPP funds in a state-by-industry group.
State PPPE (Nbr.) is the state average of the PPPE based on the number of outstanding loans. Pre-PPP Decline Hours Worked is the average decline in hours worked in each state between January
and the last week of March. Pre-PPP State Covid-19 Cases (per capita) are per capita number of reported Covid-19 cases in the state. Pre-PPP State Covid-19 Deaths (per capita) are the weekly
per capita number of reported Covid-19 deaths in the state. Pre-PPP State Social Distancing Index is the change in average distance travelled in the state until the end of March using individuals’
GPS signals. All specifications include industry×week fixed-effects. Standard errors are presented in parentheses, and are clustered at the level of the state. ***, **, and *, represent statistical
significance at 1%, 5%, and 10% levels, respectively.
(1) (2) (3) (4)
OLS Reduced Form IV
First-Stage Second-Stage
Dep. Variable: % Missing Schd. Payments % Missing Schd. Payments % PPP Received % Missing Schd. Payments
55 %PPP Received -0.091∗∗∗ -0.435∗∗∗
(0.022) (0.067)
State PPPE (Nbr.) -14.807∗∗∗ 34.662∗∗∗
(2.651) (3.053)
Pre-PPP Decline Hours Worked -12.380∗∗∗ -12.176∗∗∗ 7.824∗∗ -8.192∗
(3.587) (4.094) (3.097) (4.094)
Pre-PPP State Covid-19 Cases (per capita) 4.390∗∗∗ 4.189∗∗∗ 1.674∗ 4.754∗∗∗
(0.736) (0.577) (0.988) (0.462)
Pre-PPP State Covid-19 Deaths (per capita) -107.446∗∗∗ -101.315∗∗∗ -8.416 -100.872∗∗∗
(25.205) (20.723) (39.435) (17.170)
Pre-PPP Social Distancing Index -15.797∗∗ -13.084∗∗ 14.833 -7.857
(6.450) (6.317) (9.059) (5.592)
Observations 2206 2448 2206 2206
Adjusted R2 0.708 0.698 0.663 -0.028
FStat 128.871
Industry×Week Fixed Effects Yes Yes Yes Yes
Table 9: % Receiving PPP and Cash-on-Hand
Table 9 reports the results of ordinary least squares (OLS) and instrumental variables (IV) regressions examining the relation between the geographic allocation of PPP funds during the first round
and the percentage of businesses at the state-by-industry level reporting at least two months of cash on hand to sustain operations as collected from the Census Small Business Pulse Survey. The
dependent variable is the fraction of businesses with cash on hand to sustain operations for two months or more. % PPP Received is the percentage of businesses reporting having received PPP
funds in a state-by-industry group. State PPPE (Nbr.) is the state average of the PPPE based on the number of outstanding loans. Pre-PPP Decline Hours Worked is the average decline in hours
worked in each state between January and the last week of March. Pre-PPP State Covid-19 Cases (per capita) are per capita number of reported Covid-19 cases in the state. Pre-PPP State Covid-19
Deaths (per capita) are the weekly per capita number of reported Covid-19 deaths in the state. Pre-PPP State Social Distancing Index is the change in average distance travelled in the state until the
end of March using individuals’ GPS signals. All specifications include industry×week fixed-effects. Standard errors are presented in parentheses, and are clustered at the level of the state. ***,
**, and *, represent statistical significance at 1%, 5%, and 10% levels, respectively.
(1) (2) (3) (4)
OLS Reduced Form IV
First-Stage Second-Stage
Dep. Variable: % Cash 3 months % Cash 3 months % PPP Received % Cash 3 months
%PPP Received -0.000 0.270∗
56 (0.035) (0.125)
State PPPE (Nbr.) 8.920∗∗ 32.502∗∗∗
(3.768) (3.116)
Pre-PPP Decline Hours Worked 11.966∗ 15.861∗∗∗ 4.354 11.007∗
(6.170) (5.378) (5.855) (4.844)
Pre-PPP State Covid-19 Cases (per capita) -1.979∗∗ -2.003∗∗ 2.095∗∗ -2.476∗∗
(0.733) (0.818) (0.937) (0.726)
Pre-PPP State Covid-19 Deaths (per capita) 75.244∗∗ 76.722∗∗ -26.761 76.491∗
(28.941) (31.278) (44.659) (35.062)
Pre-PPP Social Distancing Index -6.850 -12.334 13.236 -12.138
(9.277) (11.326) (11.661) (9.223)
Observations 903 918 903 903
Adjusted R2 0.629 0.640 0.797 -0.089
FStat 19.527
Industry×Week Fixed Effects Yes Yes Yes Yes
A Case Study: Wells Fargo
This section explores a case study, illlustrating the importance of financial intermediation in
allocating PPP funds. Figure A.1 and Table A.1 present a case study of a particular bank with
a very low share of PPP loans relative to overall market share—Wells Fargo. Wells Fargo was
severely constrained from expanding its balance sheet as a result of an asset cap imposed by
the Fed in the aftermath of the fake accounts scandal. This asset cap was only lifted on April
10, when the Fed excluded PPP loans from the formula it uses to restrict Wells Fargo’s growth.
The asset cap limited Wells Fargo’s ability to lend under the PPP in the early days for the first
phase of the program. Table 1 shows that Wells Fargo, the third largest bank in the nation
by total assets, held a 6.5% share of the total outstanding small business loans but processed
only 0.04% of the total volume of loans in the PPP program until April 15. Figure A.1 shows
PPP allocations by the market share of Wells Fargo. The top panel shows volume, while the
bottom panel shows the number of loans. Both figures show a similar pattern—firms located in
areas with higher Wells Fargo market share see lower PPP allocations, both in terms of overall
aggregate loan volume per business and in the fraction of businesses receiving PPP loans.
Table A.1 presents similar information, regressing the log total volume and number of loans
on the share of Wells Fargo branches in congressional districts. The relationship between PPP
allocations and Wells Fargo branches echoes the results seen in Table 2. Areas with higher Wells
Fargo exposure (and lower PPPE exposure) see lower PPP allocations. The effect is statistically
significant at the 5 percent level or higher in columns (1) through (3), but loses significance
when state fixed effects are included. More broadly, Figure A.1 and Table A.1 illustrate an
example of the variation underlying overall PPPE exposure.
57
Figure A.1: Wells Fargo Exposure and PPP per Establishment
Figure A.1 are scatterplots of the total PPP allocation per establishment in the congressional district and the share of branches of Wells Fargo
in the Congressional District (Panel A) and fraction of establishments receiving a PPP loan and the share of Branches of Wells Fargo in the
Congressional District (Panel B). Data comes from the SBA, Summary of Deposits, and County Business Patterns.
PPP allocation per Establishment (Congressional District)
50000 100000 150000
0
0 .1 .2 .3 .4
Share of Wells Fargo Branches
Fraction of Establishments receiving PPP (Congressional District)
0 .1 .2 .3 .4 .5
0 .1 .2 .3 .4
Share of Wells Fargo Branches
58
Figure A.2: Decline in Hours Worked and PPP Allocation by State
Figure A.3 presents four scatterplots of the decline in hours worked relative to a January baseline and four alternative measures of allocation of PPP funds across states. The figure on the top left
plots the amount of PPP loans received by small businesses in each state divided by the total number of small businesses in the state. The figure on the top right corner plots the fraction of small
businesses in each state that received a PPP loan. The figure on the bottom left corner plots the fraction of small businesses in each state that received a PPP loan and the state exposure to the
PPPE measured in terms of the total volume of loans. The figure on the bottom right corner plots the fraction of small businesses in each state that received a PPP loan and the state exposure to
the PPPE measured in terms of the total number of loans.
(Date: 31/03/2020) (Date: 31/03/2020)
Ratio of Hours Worked relative to January Baseline
SD Ratio of Hours Worked relative to January Baseline SD
UT NE UT
OK ND NE
ID MS KS OK ND
AZSC GAAKVARI TXALMO IA
AR WI ID MO MSKS
MT
ORNC WVWY DE
0 .25 .5 .75 1
FL COKYTN
ME IN MN AZ SCAK AL WI IA
NV NMCAWA GATX RI AR MT
0 .25 .5 .75 1
MD
LA VA
DE CO WY
NJ CT NH OH
ILPA CA FL
ORNCNM TN
IN
KY MN ME
NY MI MA NVNJWA MD
CTWV
IL OHLA
NH
VT DC HI PA
NY MI MA
59 DC VTHI
0 0 0 0 0
00 00 00 00 00
30 40 50 60 70 .1 .2 .3 .4 .5
PPP Amount per Small Business in State Fraction of Businesses receiving PPP
Ratio of Hours Worked relative to January Baseline Ratio of Hours Worked relative to January Baseline
(Date: 31/03/2020) (Date: 31/03/2020)
SD SD
UT NE UT NE
ND OK ND
OK
KS
ID KSMS AZ SC
AKID MO IA MS
SC
AZ AKGA IAMOAL
WIRI
AR VAGA
TX ALRIWI AR
MT
WY
VATX MT DECO
0 .25 .5 .75 1
DE WY ORFL TNKY
0 .25 .5 .75 1
OR
MN FL MDCO
NC TNWV KY CA NCMN ME
NV CA WA
NM
CT NJLA
IL INOHME NV WAMD LA
NJ ILM
CT INWV
OHNH
PA NH PA
NY MAMI NY MIMA HIVT
DC VT HI DC
0 .1 .2 -.4 -.2 0 .2 .4
-.3 -.2 -.1
State PPPE (Volume) State PPPE (Nbr. Loans)
Figure A.3: Business Shutdowns and PPP Allocation by State
Figure A.3 presents four scatterplots of the share of businesses in each state that shutdown that in the week of March29th–April 4th and four alternative measures of allocation of PPP funds across
states. The figure on the top left plots the amount of PPP loans received by small businesses in each state divided by the total number of small businesses in the state. The figure on the top right
corner plots the fraction of small businesses in each state that received a PPP loan. The figure on the bottom left corner plots the fraction of small businesses in each state that received a PPP loan
and the state exposure to the PPPE measured in terms of the total volume of loans. The figure on the bottom right corner plots the fraction of small businesses in each state that received a PPP
loan and the state exposure to the PPPE measured in terms of the total number of loans.
(Date: 31/03/2020) (Date: 31/03/2020)
Share of Business Shutdowns (State)
Share of Business Shutdowns (State)
.8 .8
.6 .6
NY MI
PADC MA
HI NY MI HI
WV ME
VT DC PA MA
NVOR WA NJ MD ILIN NH OH WV ME
CA DE RI
CT
KY MN NV WA MDCTILOH
VT
RIKY NHMN
.4
FLNMNC GA CO LA TXALMO IA WI .4 CA NJOR DE IN
LA
SC MTMS AKVA TN KS FL
NCNMGATXCO AL WI
VA MO MS IA
AZ AR
IDWY OK SCAKTN MT
KS
NE AZ ID AR WYOK
.2
UT SD ND
.2
UT SDNEND
60
0
0
0 0 0 0 0
00 00 00 00 00
30 40 50 60 70 .1 .2 .3 .4 .5
PPP Amount per Small Business in State Fraction of Businesses receiving PPP in State
Share of Business Shutdowns (State) Share of Business Shutdowns (State)
(Date: 31/03/2020) (Date: 31/03/2020)
.8 .8
.6 .6
NY MI HI NY MI HI
DC PA MA WV ME DC PA
MAWV ME
VT
.4
NV
CA MD
ORWADE MNLA
NJ ILCT RI OHNH
IN
KY NV MN
OR CT
CA WADEMDIL
NJ
LA
RI KY NH
INOH VT
CO
GA
FL NCTXNM WI
.4
FLGA CO MOAL
NC
VASC
AK ALMO
TN MT
IA MS
KS
NM
VA
AK
TX IA WI
MTKS MS
AZ ID AR SC TN
WY OK AZ ID WYAR OK
.2 SD NE ND NE
UT .2
UT SD ND
0 0
-.4 -.2 0 .2 .4 0 .1 .2
-.3 -.2 -.1
State PPPE (Nbr. Loans) State PPPE (Volume)
Figure A.4: COVID-19 Cases and PPP Allocation by State
Figure A.4 presents four scatterplots of the number of confirmed COVID-19 cases per thousand as of April, 3rd 2020 and four alternative measures of allocation of PPP funds across states. The
figure on the top left plots the amount of PPP loans received by small businesses in each state divided by the total number of small businesses in the state. The figure on the top right corner plots
the fraction of small businesses in each state that received a PPP loan. The figure on the bottom left corner plots the fraction of small businesses in each state that received a PPP loan and the state
exposure to the PPPE measured in terms of the total volume of loans. The figure on the bottom right corner plots the fraction of small businesses in each state that received a PPP loan and the
state exposure to the PPPE measured in terms of the total number of loans. Data comes from the Center for Disease Control, SBA, Call Reports, and FDIC Summary of Deposits.
Covid-19 Cases per Thousand (04/03/2020)
Covid-19 Cases per Thousand (04/03/2020)
3 3
NJ NJ
NY NY
LA LA
2 2
MA
CT
MI MA
1 MICT
WA 1
CO VTILPA WA
ID GA RIMD
NVFL DEUT
MS TN IN NH IL
PA CO ID VT
61 SC
CA
AZ
NM WYAKVA KYME
MTAR ALMO
SDIAOH
TX OK
WI
HI ND NV FL MD GA IN
DE RI MS
OR NC WV NEKSMN CAAZ SC
TNUT ALNHWI
MO ME
0 VAAK
NCNM
OR OH
TX MT
ARHIKSWYOK SD ND
IA
WV KY MN NE
0 0 0 0 0
0
00 00 00 00 00
30 40 50 60 70 .1 .2 .3 .4 .5
PPP volume per Small Business Establishment Fraction of Small Businesses Receiving PPP loan
Covid-19 Cases per Thousand (04/03/2020) Covid-19 Cases per Thousand (04/03/2020)
3 3
NJ NJ
NY NY
LA LA
2 2
MA MA
CT MI CT MI
1 1
WA WA
ILCO PA VT CO IL PA VT
ID FLGA ID
FLUT GA
NV
UT TN MDMSERI IN NH
NV MDDE TNRI INMS
NH
CA ALSC
MOWI OK ME CA SC
AL WYWI
MOARMT
AZOR WY
NE
MN IANMNCR
VA
ND
SD
TX MT
AK
WV KSOH
KY HI AZ OR NC VAAKNMOH
TX KY WV
MN NEIA
KS
SD ND ME
OK
HI
0 0
-.2 -.1 0 .1 .2 -.2 0 .2 .4
State PPPE (Volume) State PPPE (Nbr. Loans)
Figure A.5: COVID-19 Deaths and PPP Allocation by State
Figure A.5 presents four scatterplots of the number of COVID-19 deaths per thousand as of April, 3rd 2020 and four alternative measures of allocation of PPP funds across states. The figure on the
top left plots the amount of PPP loans received by small businesses in each state divided by the total number of small businesses in the state. The figure on the top right corner plots the fraction of
small businesses in each state that received a PPP loan. The figure on the bottom left corner plots the fraction of small businesses in each state that received a PPP loan and the state exposure to
the PPPE measured in terms of the total volume of loans. The figure on the bottom right corner plots the fraction of small businesses in each state that received a PPP loan and the state exposure
to the PPPE measured in terms of the total number of loans. Data comes from the Center for Disease Control, SBA, Call Reports, and FDIC Summary of Deposits.
Covid-19 Deaths per Thousand (04/03/2020)
Covid-19 Deaths per Thousand (04/03/2020)
.08 LA .08 LA
NJ NJ
.06 .06
MI
MI
.04 WA CT .04 WA CT
MA
NY
.02 GA CO NY
MA
NV DE IL
IN
VT OK
.02 GA CO
62 FL CA MS KY
MD AL
ME MO PA OH KS WI NV DEIL IN
ORAZSC ID
NM MTAR VARI TN
TX SD NH
IA NE MN HI ND
VT
WI MS OK
0 NC WV WYAK UT PA
FL MDSC
CAAZ OR NM
OHRIKYAL
TN ID
NHMO ME
KS
MT
NC VA TX
AK MNAR IA SDNEND
0 WV UT HI WY
0 0 0 0 0
00 00 00 00 00
30 40 50 60 70 .1 .2 .3 .4 .5
PPP volume per Small Business Establishment Fraction of Small Businesses Receiving PPP loan
Covid-19 Deaths per Thousand (04/03/2020) Covid-19 Deaths per Thousand (04/03/2020)
.08 LA .08 LA
NJ NJ
.06 .06
MI MI
.04 WA CT .04 WA CT
MA MA
NY NY
.02 CO
ILGA
.02 GACO IL
NV DE IN NV DE IN
MS VT OK VT
CA FLALSC WI
KY
MD
OK
PA
OH ME CA FL MD AL PA
OH WI MS
KY KS ME
AZOR ID
MNIA
NE TN
NM VA
TXMO
NDR
AKMTKSRI NH AZ OR SC
IDVA TXTN MN
NMRI MOARIA
NE NH
MT ND
0 UT
WY SD
NC WV HI 0 UT NC AK WY WVSDHI
-.2 -.1 0 .1 .2 -.2 0 .2 .4
State PPPE (Volume) State PPPE (Nbr. Loans)
Figure A.6: Shelter-in-Place Orders and PPP Allocation by State
Figure A.6 presents four scatterplots of the timing of statewide shelter-in-place orders and four alternative measures of allocation of PPP funds across states. The figure on the top left plots the
amount of PPP loans received by small businesses in each state divided by the total number of small businesses in the state. The figure on the top right corner plots the fraction of small businesses
in each state that received a PPP loan. The figure on the bottom left corner plots the fraction of small businesses in each state that received a PPP loan and the state exposure to the PPPE measured
in terms of the total volume of loans. The figure on the bottom right corner plots the fraction of small businesses in each state that received a PPP loan and the state exposure to the PPPE measured
in terms of the total number of loans. Data comes from the New York Times, SBA, Call Reports, and FDIC Summary of Deposits.
PPP allocation per Small Business Establishment Fraction of Small Businesses Receiving PPP Loans
30000 40000 50000 60000 70000 18 ma .1 .2 .3 .4 .5
18 ma
19 r20 19 r20
m 20
m
20 ar20
20 20 ar20
ma 20
ma 20 21 20 r
21 20 r m 20
m
22 ar20
20
ND 22 ar20
ma 20 ND
ma 20
23 20 r HI MN 23 20
m
r
20
NE
SD
m
24 ar20
20 WI KS 24 ar20
ma 20 OK
ma 20 MA NE 25 20 r WY
IA
25 20
m
r
20 OH IA
m
26 ar20
20
HI MT KS MEMS
26 ar20 NH DC ma 20 VT AR
PA SD
AL MO
ma 20 MO
27 20 r IL IN OK 27 r20
m 20 WI MN
m
28 ar20
20
LA VT MDTN TX
ME 28 ar20 ID NH AL
ma 20 CTMI KY RI UT
ma 20
29 r20 MA KY UT
OHWV CO RI
LA
r
29 20
m 20 DE CO AK VA m
30 ar20
20 IN TN TX
30 ar20 MS
GA AR ma 20 IL AK SCPA GA
ma 20
31 r20 NJ ID MT WY 31 r20 CTDE
MI
NM VA
ma 20 WAWV ma 20 MD
01 r202
ap 0 CANY NCSC
AZ
01 r202
ap 0 NYNJ WA
OR NC FL
02 r202
ap 0 ORNM FL 02 r202
ap 0 CA AZDC
NV
03 r202
ap 0 NV 03 202
ap
r
0
04 r202
63 04 r 2
ap 020
05 202r r
ap
05 202r 0
r
ap 0 te ap 0 te
06 r202
ap Or 06 r202
ap 0 Or
0
07 202
ap
r de 07 202
ap
r de
r20 0 Sta r20 0 Sta
20 20
No No
Date Shelter Order Date Shelter Order
.2 .4
HI VT ME
VT
State PPPE (Nbr. Loans)
ME HI NHMT KS ND
OK
State PPPE (Volume)
NH IN MS IA
SD
NE
WV AR
IN
MI MI WIKYMN MO
OH RI PA OK OHMA RI PA WY
MAWIKY
DE KS
MD MS CT TN
WV MT
AK VASC MO ND
AR IL LANM
DE AK VASC TX AL
NC TX SD
-.1 0 .1
0 .2
CTNM CO TN GAAL ID CO MD
NJ LA
IL WA ID MN
FL IA
WY
NE
UT NYNJ WA
NC GA
NY UT
FL
OR DC OR
CA AZ DC
CA AZ
-.2 NV -.2 NV
18 m
19 ar20
m 20 r 18 m
19 ar20
m 20 r
te te
20 ar20
ma 2 0
21 20
m r
22 ar20
ma 2 0 20 20 ar20
ma 2 0
21 20
m r
22 ar20
ma 2 0 20
r r
Or Or
23 20
m
24 ar20
ma 2 0
25 20
m r 20
20 23 20
m
24 ar20
ma 2 0
25 20
m r 20
20
26 ar20 26 ar20
de de
ma 2 0
27 20
m r
28 ar20
ma 2 0 20 ma 2 0
27 r20
m
28 ar20
ma 2 0 20
29 20
m r
30 ar20
ma 2 0
31 r20
ma 2 0 20 29 r20
m
30 ar20
ma 2 0
31 r20
ma 2 0 20
01 r202
ap
0 2 r2 0 2
ap
0 3 r2 0 2 0
0 01 r202
ap
0 2 r2 0 2
ap
03 202r 0
0
Sta Sta
04 ap r 2
ap 020
05 202
ap r 0
0 ap
0 4 r2 0 2
ap
05 202
ap r 0
0
0
0 6 r2 0 2
ap
07 202
ap r 0
r2 0 0
20 0 6 r2 0 2
ap
07 202
ap r 0
r2 0 0
20
No No
Date Shelter Order Date Shelter Order
Figure A.7: Social Distancing Index and PPP Allocation by State
Figure A.7 presents four scatterplots of and four alternative measures of allocation of PPP funds across states. The figure on the top left plots the amount of PPP loans received by small businesses
in each state divided by the total number of small businesses in the state. The figure on the top right corner plots the fraction of small businesses in each state that received a PPP loan. The figure
on the bottom left corner plots the fraction of small businesses in each state that received a PPP loan and the state exposure to the PPPE measured in terms of the total volume of loans. The figure
on the bottom right corner plots the fraction of small businesses in each state that received a PPP loan and the state exposure to the PPPE measured in terms of the total number of loans. Data
comes from the SBA, Call Reports, and FDIC Summary of Deposits.
-.2 WY
-.2
WY
Social Distance Index (03/28/2020)
Social Distance Index (03/28/2020)
AR AL AR
-.3 SC MTMS TN -.3 SC TN
AL
MT
NC KY OK IA NC KY
MSIA OK
ORAZ WV
NM
ID GA UT SD NE
MO ND AZ OR WV
GA ID
-.4 VA LATX IN
ME OH KS
WI NM INUT MO SDNEND
FL CA
DE IL
RIMD PA MN -.4 VA TX
DEIL OH WI KS
ME
CT NH CA FL MD PA LARI MN
WA CT NH
NJAKCO WA
-.5 MA AK CO
NV NY
MI -.5 NJ
MA
VT NVNY MI
64 -.6 HI
VT
DC -.6 HI
0 0 0 0 0 DC
00 00 00 00 00
30 40 50 60 70 .1 .2 .3 .4 .5
PPP volume per Small Business Establishment Fraction of Small Businesses Receiving PPP loan
-.2 -.2
WY WY
Social Distance Index (03/28/2020) Social Distance Index (03/28/2020)
ALAR AL AR
-.3 TN
IA NCSCMTMS -.3 SC TN MT
IA
NC MS
KY OK KY WV OK
AZOR ID GA WV
NENM AZ OR UT GAID NE
UT ND
SD MO IN NM SD ND
MO IN
-.4 ILFL
TX WIKSOH
VA DEPA ME -.4 VA
DE
TX OH WI KS ME
CA MN LA MD RI NH
CA FL MD ILLA RI
PAMN
NH
CT CT
WA WA
-.5 NJCO AK -.5 NJ CO AK
MA MI MA
MI
NV NY NV NY
VT VT
-.6 HI -.6 HI
DC DC
-.2 -.1 0 .1 .2 -.2 0 .2 .4
State PPPE (Volume) State PPPE (Nbr. Loans)
Figure A.8: State Unemployment Insurance and PPP Allocation by State
Figure A.8 presents four scatterplots of the ratio of state unemployment insurance claims to employment covered by unemployment insurance and four alternative measures of allocation of PPP
funds across states. State unemployment insurance claims are the sum of filed claim in the weeks ended March 21st, March 28th, and April 4th, 2020. The figure on the top left plots the amount
of PPP loans received by small businesses in each state divided by the total number of small businesses in the state. The figure on the top right corner plots the fraction of small businesses in each
state that received a PPP loan. The figure on the bottom left corner plots the fraction of small businesses in each state that received a PPP loan and the state exposure to the PPPE measured in
terms of the total volume of loans. The figure on the bottom right corner plots the fraction of small businesses in each state that received a PPP loan and the state exposure to the PPPE measured
in terms of the total number of loans. Data comes from the Department of Labor, SBA, Call Reports, and FDIC Summary of Deposits.
State Unemployment Claims by State (% Cov. Employment)
State Unemployment Claims by State (% Cov. Employment)
.2 RI
.2
MI MI RI
NV PA HI PA
NV HI
.15 WA KYLA NH
.15 KY NH
WA LA
NJ OHMA
CA MT GAAK ME NJ OHMA
VT MN CA GAAK ME
MT
.1 ID DE ALIN IA DE IN
MNVT
MD WI
KS ID
AL IA
NMNC
AZSC MO .1
NY AR VA IL
OK ND NCMD
NM WI KS
OR MS TN AZNY SC MO
65 NE VAIL AR OK ND
WY TX OR TN MS
UTCT NE
.05 FL
WV CO
CT
TX
UT
WY
FL
SD .05 WV CO
0 0 0 0 0 SD
00 00 00 00 00
30 40 50 60 70 .1 .2 .3 .4 .5
PPP volume per Small Business Establishment Fraction of Small Businesses Receiving PPP loan
State Unemployment Claims by State (% Cov. Employment) State Unemployment Claims by State (% Cov. Employment)
.2 .2
RI MI RI MI
PA HI PA HI
NV NV
.15 WA LA KY NH .15 WA LA KY NH
NJ MA NJ MA
CA GA AKMT OH ME CA GA
OH MT ME
MN VT AK MN VT
IDIA AL DE IN DE
ID AL INIA
.1 .1
NMNC
SC WI
KS
MD NCMD NM WI KS
AZ NY IL VA MO OK
AR
ND
AZ NY VASC IL MOAR OK
OR NE TN MS OR TN NEMSND
WY TX TX WY
UTFLCT FLUT CT
.05 CO WV .05 CO WV
SD SD
-.2 -.1 0 .1 .2 -.2 0 .2 .4
State PPPE (Volume) State PPPE (Nbr. Loans)
Figure A.9: Weekly Elasticities of Employment Outcomes to % Businesses Receiving PPP
Figure A.9 plots coefficients and respective standard errors of regression analyses investigating the week-by-week elasticities between employment and firm outcomes and exposure to PPPE. The
left figure plots the coefficients and respective standard errors of an empirical specification that augments the specification of Panel A of Table 5 column (3) to include a full set of interactions
between the fraction of businesses receiving PPP in the Congressional District during the first round and week dummy variables. The figure on the right plots the coefficients and respective standard
errors of an empirical specification that augments the specification of Panel B of Table 5 column (3) to include a full set of interactions between the Congressional District PPPE and week dummy
variables.
Weekly Elasticities of Shutdown and Fraction Receiving PPP Weekly Elasticities of Shutdown and Fraction Receiving PPP
State-Industry-Week and Firm Fixed Effects State-Industry-Week and Firm Fixed Effects
.05 .2
0 .15 Vertical bands represent +/- 1.96 * St. Error of each point estimate
Initial Lockdown Orders PPP Runs out of Funds PPP 2nd Round begins
66
PPP Launch
.1
-.05
Initial Lockdown Orders PPP Runs out of Funds PPP 2nd Round begins
Vertical bands represent +/- 1.96 * St. Error of each point estimate
PPP Launch
Beta Beta .05
-.1
0
-.15
-.05
-.2
-.1
b1 b8 15 2
b2 b2
9 ar7 r14 1
ar2 ar2
8 r4 pr1 r18 r25 y2 b1 b8 15 2
b2 b2
9 ar7 r14 1
ar2 ar2
8 r4 pr1 r18 r25 y2
-Fe -Fe 9-F -M -Ap 1 -M -Fe -Fe 9-F -M -Ap 1 -M
-Fe -Fe -M -Ap -Ap a -Fe -Fe -M -Ap -Ap a
26 02 eb ar1 a 5-M -M r29 r5- 26 02 eb ar1 a 5-M -M r29 r5-
16 23 ar8 r22 A r12 r19 r26 16 23 ar8 r22 A r12 r19 r26
an :F :F :M ar1 Ma Ap Ap an :F :F :M ar1 Ma Ap Ap
:J eb eb :F :F : M Ma Ap Ap :J eb eb :F :F : M Ma Ap Ap
eb eb k7 : M 2: 5: eb eb k7 : M 2: 5:
k2 k3 k4 k8 k1 k1 3: 4:
k1
k2 k3 k4 k8 k1 k1 3: 4:
k1
k5 k6 We k9 0: 1: k1 k1 k5 k6 We k9 0: 1: k1 k1
We We We e We k1 We We We We e We k1 We
e e e We We We We e We We We e e e We We We We e We We We
e e e e We e e e e e e e e We e e e e
e e
Figure A.10: Unmet PPP Demand and Exposure to State PPPE (Industry×State)
Figure A.10 are scatterplots of the state exposure to the number-based PPPE and the percentage of businesses in each industry within a state
that applied but did not receive PPP. The plots represent the evolution of the two variables for each survey week. Data comes from the Census
Bureau, SBA, Call Reports, Summary of Deposits, and County Business Patterns.
Week 15: April26 - May2nd Week 16: May3rd - May9
CA 60
CA
CA FL
60
AZ NJ GA NJ
NJ NC 50
Diff. between % of businesses applying and receiving PPP Diff. between % of businesses applying and receiving PPP
CA
CA OR FL VA
FL NJ
NY
50 CA FL
FL MD MI
CA
CA OR FL NY
NY GA MD OH
WA TX
NMCT
TN 40 NY
WA
NJ
UT
NY MD OH MI NJ VA
NJ
NY NC
GA IL OH
CA GA
NC PA
40
AZ
CA OR
OR
NJ
WA GA VA SC
NC IL
SC TX TN PA MI
CA
AZ
CA
CA GA
FL WA
FL NY GACO LA OH PA
PA WI NV CA FL
FL WA VA
NY
FL WA VA ILTX
MD TX PA
MAMI AZ FL SC MI MO
CA NCCO OH
PA MI 30 OR FL GA
AZ NY
UT COVA TX TN
TNOHPA
MAMI MO IN NH CA FL NJ
NY
WA VA
VA ALLACT CT
WA
WA GACO ILTXCT MA IN AZ
CA DC OR UT GA MD
NC MD IL MN NV OR
OR FL WA
NJ NC MD
30 NV NY
WA
NY NCCO
NC IL LA TN MAMA IA NV CA
AZ
AZ
CA FL WA
UT
FL WA
NJ
NY NC
GA VA
MD TX
CT OHMAMI
MI
FL UT COVA ILTX
NC MD LA OH KYWI IN MS
NH OR NY GA
NC MI IANH
NY ILTXCT
GACOVA SC
ILAL TN PAMA
PAMN
MA
WI
IN CA OR FL WA
NJ
UT
NY
NJ GACO
GA
NC
NC MD IL
VA SC
NM
LACT
TX
ILTX TNOHMA KS
20 CO
IDMD IL CT PA MI WI
MO IN
TX OH MI IN
IN NV AZ OR NJ
FL WA
NY
WA NC VA
MD
VA ILAL
NC MD NM
LACT
TX PAMN
PA
MNWI
MN IN ME
WA OH MN
MN INE KS OK AZ OR WA
NY
WA GA MD TX
LA TNOH
OH
PA MI
MN IN MSNH
VA SC
VA TX
ILAL
AL PAMNWI IN
MO AR OK NJ
WA
UT
NY NC
ID SC TX CT OH
OH
TXCT PA MA
PAMN
MA WI
MO
IN
IN MS
AR KS
20 TX OH MO
WI OR NY CO
CO
MD IL
IL MA KYWI
WI
MO INE NH HI
KS
CO LA OH MN MO ME IDMD SC TX TNOH MNWI OK
VA ILAL
AL
TX KYMO IA NY
NY
NY NCCO
ID ILAL
ILAL OHMAMI
KY
MI MO AR IAKSOK
UT TN MN
WI
WI
MO AR KS 10 UT
WA GA
NCCO SC TX
ILALLACT
TX TN MA
TN PA
MAMI
MN MO IN IA
AR
IN ME
VT
MI
MN WI OK AZ UT
UT CO ILTX TN PA MN
MA KY
MI WI
MO MS OK
IN IA
MNKYWI MT
MSKS ID TN
TN MNKYWI
MO OK
KS
AR IA OK AL NEMTNHOK
OK
10
ID OK WA LA TN PA KY
LA
TX AR
IN IANH
KS
ME
MNMO IN IAKS OK
OK
PA MO
NE
NEIAKS
0 ID MNWI ARNEMS
MTKS
OK
0 MO
-.2 0 .2 .4 -.2 0 .2 .4
State PPPE (Nbr. Loans) State PPPE (Nbr. Loans)
Week 17: May10 - May16 Week 18: May17 - May23
60 60
50 50
Diff. between % of businesses applying and receiving PPP Diff. between % of businesses applying and receiving PPP
40 40
30 30
MD
NC
SC
CA
NJ NC 20 VA
20 NV OR NY
WA AZ FL NY LA
NV SC OH
OR FL MD AZ NY
CA FL NJ
NY GA VA SCIL CT IN NV OR NM PA MI
KY
CA
CA DC
FL
FL WA
NY NC
NC
ID TX TNOH AR
NE
NJ GA MD IL
NC PR
UT PA PR NV CA OR SC
NV CA
AZ
CA OR FL NY
FL WA
UT
NJ
GA
NCCOVA SCTX
MD LA
TN MA
MO
MO
PR
INE
E 10 NV
NV CA
AZ
CA
WA
NY
NJ
WA
UT
NCCO
NC
MD
VA LA
TX MA MO
10 CA OR FL WA
NJ
NY NC
GACOVA
MD
VA TX
LA OH
PAMNKYMO ME CA
AZ OR NJ
NJ
UT NC
GACO SC
VA IL TN PA WI
AZ
CA
AZ OR NJ
FL WA
NY
NJ
NY NC
GACO
COVA SC
MD
VA ILAL
TX
LA
NM
LA TN PA MI
CT OH MN
MA WI
WI IN CA DC
CA FL NJ
NY
NY
NJ GACO
MD TN MA
MAMI INEIA
NV CA
AZ
AZ
UT
WA
NJ
NY
UT
WA
GA MD
GACO
CO
MD IL
IL TX
AL
TX
TN
LA OH
NM PA
OHMA
MA
MI
KY
MI MO IN IAKS
WI AZ
CA
CA OR FL WA
FL
FL UT
NY
NJ
WA
NJ
NCCO
MD
GACO
NC VA ILAL
VA TX
TX
TX
CT
CT
PA
OH
TNOH
PAMN
MN
MAMI
MO
MO
PR AREMS KS
NH OK
OK VT
NV AZ OR NY
UT NC MD SC
CO
ID IL
SC LA
TX
LACT
TX PA
TNOH
MN
MI
KYWI AR MS KS
NHOK
KS AZ
CA WA
NY
UT
WA
NJ COVA SC
MD ILALLA
TX CT PAMN
MN
NJ
WA
NY GA ILAL
TX CT
CT OHMA
PA
OHMA KYMO
WI
MO IN
AR MS
MT
MSNH
NHOK ME
VT NV AZ FL WA
FL NY
UT
NY NCCO
MD
GACO SC
IL TXCT
TX OH KY
PA MI
MI WI IN IAKS
AR OK
AZ OR
OR
FL NJ
UT
WA
ID
NCCOVA ILAL
CO
TX
ILTX
NM PA
OH
PA
OH
MA
MN
MAMI
MNWIWI
WI INEIA
MO IN IAKS
OK
OK OR
OR
OR
NY
UT
FL WA
UT
NY GACOVA SC
NC
NC
IL
ILAL
IL CT
TNOHMA
PA
OH
MN
MI
MNKYMO
WI
WI
MO
WI
IN
IN
KS
MI WI 0 OR WA
UT
NY GA
IDMD
VA SC
ILAL
TX
NM
LA PA
TNOH
CT MA
PA
MA
MNWI
MI MO AR
INEIA
MS
MT
NH
KSOK ME
0 NV AZ UT NC
GA
ID
CO
MD
VA ILALNMTN MA MNKYPR
WI
MO
WI MS
IA
EMT
NH
KS
OK ME MA
MN WA VA LA OH MN KY IN IA
AR
ID TNOH MN NH
-.2 0 .2 .4 -.2 0 .2 .4
State PPPE (Nbr. Loans) State PPPE (Nbr. Loans)
Week 19: May24 - May30 Week 20: May31 - Jun6
60 60
50 50
Diff. between % of businesses applying and receiving PPP Diff. between % of businesses applying and receiving PPP
40 40
30 30
20 UT 20
NM WA
NV SC CA VA
NC IL
OR LA OH
CA
UT GA VA AL IN
INE FL WA COVA SCTX PA MI
10 OR NJ NC MI
KYMO 10 FL WA LA
NJ
FL WA
NY
NY GA SCTX LA TN MO AR IA
MSKSOK NV AZ NY
NJ
FL WA GA VA
VA LACT PA MI MT
IA
CA WA
NY NC
GACOVA
NC VA SC
ILTXCT OH PAMNKY AZ NY
UT
NY GA MD
MD SC TNOH MO
MI WI
PAMN
NV
NV CA DC
AZ
AZ
CA OR
OR FL
FL
FL
NJ
WA
NY
NJ
WA
GA
NC
ID VA
IDMD
GA
GA VA
VA IL TX
TX
PA
PA
OHMA
PA
MN
MN
MAMI WI
PR
PR
KSOK
VT NV
NV
NV AZ
CA
OR
OR FL WA
FL NJ
NY
WA
UT
NJ
WA
GA
NC
GA
GA
NCCO
IDMD
COVA
ILAL
TX
ILALNM
TX TN
NMTNOH
TX
LA MNPR
MI
MO
WI
MO
AR
INE NH
AR KS
OK
OK
NV CA FL WA NC AL OH
PA NV CA
AZ OR NC IL TX PA
AZ
CA
AZ
CA
OR
OR
OR FL NY
FL WA
NY
NJ
NY
WA
NJ
GA
IDMD
CO
NCCO
VA IL
MD
VA ILAL
IL
SC LA
TX
TX TN
LACT
TX CT
OH
PA
MA
MI
MI
MI
MNKY
KY
MI
WI
PR IN IAKS
IN IAKSOK
CA
CA
AZ
CA
OR
OR
FL
FL
FL
NJ
NY
WA
NY
NJ GACO
NC
NC
MD
CO
CO
VA SC
IL
ILAL
VA SC
MD
MD IL
LA
TX
LA
TX CT OH
OHMA
PAMNWI
MA
PA
MN
MI
MI
MO
MO
WI
IN IAKS
IN
AZ MD
COVA ILTX TN PA
MA
LA OH MIMNWIMO
WI FL NJ
NY NC TX MI
0 AZ OR UT
NJ
NY
FL WA NC
GA
NC
ID
CO
MD
VA SC
ILALLA
NMCT
TNOH
PAMN
MA KY
MI WI
MO AR
INEIA
MS
MT
NH
KS
HI
OK ME
VT
0 AZ DC OR FL NY
UT
NJ ID
NC
GACO
MD
VA SC
ILAL
TX
LACT
TNOH
PA
OH KY
MI
MN
MA MO
WI
PR AR
INEIA
MS
NH
KS
HI
OK ME
VT
MA GA CT PA
CO IL WA OH PR
WI OK
OR AL MO AR
WI MA
MD TX AR
-.2 0 .2 .4 -.2 0 .2 .4
State PPPE (Nbr. Loans) State PPPE (Nbr. Loans)
67
Figure A.11: % Receiving PPP and Exposure to MSA PPPE
Figure A.11 are scatterplots of the MSA exposure to the number-based PPPE and the percentage of businesses in each MSA that applied but
did not receive PPP in each week. The plots represent the evolution of the two variables for each survey week. Data comes from the Census
Bureau, SBA, Call Reports, Summary of Deposits, and County Business Patterns.
Week 15: April26 - May2nd Week 16: May3rd - May9
100 100
90 90
80 80
% of Businesses reporting receiving PPP % of Businesses reporting receiving PPP
70 70 Hartford-East Hartford-Middletown, CT
Columbus, OH Oklahoma City, OK
Nashville-Davidson--Murfreesboro--Franklin, TN
Oklahoma City, OK Birmingham-Hoover, AL Memphis,
St. Louis, TN-MS-AR
MO-IL Buffalo-Cheektowaga, NY
Minneapolis-St. Paul-Bloomington, MN-WI
Chicago-Naperville-Elgin, IL-IN-WI Cleveland-Elyria, OH
60 Milwaukee-Waukesha, WI 60 Jacksonville, FL
Austin-Round
Houston-The Rock-Georgetown,
Woodlands-Sugar
New Orleans-Metairie,
Dallas-Fort LA
Worth-Arlington,
TX
Cincinnati,
Land,
San
TX OH-KY-IN Indianapolis-Carmel-Anderson,
Boston-Cambridge-Newton,
Providence-Warwick,
TXAntonio-New Braunfels, TX
MA-NH
RI-MA
IN
Baltimore-Columbia-Towson, MD Kansas City, MO-KS
Philadelphia-Camden-Wilmington,
Milwaukee-Waukesha, PA-NJ-DE-MDCounty, KY-IN
WI
Louisville/Jefferson
Pittsburgh,
Birmingham-Hoover, AL St. Louis, MO-IL Salt Lake City, UT Richmond,Virginia
Raleigh-Cary, NC VA Beach-Norfolk-Newport News, PA
VA-NC
New York-Newark-Jersey City, NY-NJ-PA
Virginia Minneapolis-St. Paul-Bloomington, MN-WI Detroit-Warren-Dearborn, MI
50 Beach-Norfolk-Newport
KansasNews,
City,VA-NC
MO-KS
Indianapolis-Carmel-Anderson, IN
Memphis, TN-MS-AR Buffalo-Cheektowaga, NY 50 Portland-Vancouver-Hillsboro,
Miami-Fort OR-WA
Denver-Aurora-Lakewood,
Tampa-St. Petersburg-Clearwater,
Atlanta-Sandy
FL
Orlando-Kissimmee-Sanford,
Springs-Alpharetta, GA
Charlotte-Concord-Gastonia,
Lauderdale-Pompano Beach, FL FL
NC-SC
CO
Louisville/Jefferson County, KY-IN Las Vegas-Henderson-Paradise, NV
Phoenix-Mesa-Chandler, AZ
Sacramento-Roseville-Folsom,
Los Angeles-Long Seattle-Tacoma-Bellevue, WA
Beach-Anaheim, CACA
Richmond, VA Providence-Warwick, RI-MA Riverside-San Bernardino-Ontario, CA Washington-Arlington-Alexandria, DC-VA-MD-WV
Baltimore-Columbia-Towson, MD
Dallas-Fort Worth-Arlington, TX San San Francisco-Oakland-Berkeley,
Diego-Chula Vista-Carlsbad, CA CA
40 Chicago-Naperville-Elgin, IL-IN-WI
Raleigh-Cary, NC
Charlotte-Concord-Gastonia,
New
Columbus, OH
Cincinnati,Boston-Cambridge-Newton,
OH-KY-IN
Cleveland-Elyria, OH
Nashville-Davidson--Murfreesboro--Franklin,
Orleans-Metairie, LA
NC-SC Pittsburgh,
TN MA-NH 40
Salt Lake City, UT Austin-Round Rock-Georgetown,San TX
Antonio-New Braunfels,
Philadelphia-Camden-Wilmington, TXPA
PA-NJ-DE-MD San Jose-Sunnyvale-Santa Clara, CA
Houston-The Woodlands-Sugar Land, TX
Denver-Aurora-Lakewood, CO
Phoenix-Mesa-Chandler, AZ
30 Jacksonville, FLAtlanta-Sandy Springs-Alpharetta, GA
Seattle-Tacoma-Bellevue,
Miami-Fort Lauderdale-Pompano
Portland-Vancouver-Hillsboro, OR-WA
WA
Beach, FL
Washington-Arlington-Alexandria,
Detroit-Warren-Dearborn, MI
DC-VA-MD-WV
Hartford-East Hartford-Middletown, CT 30
Sacramento-Roseville-Folsom, CA
Orlando-Kissimmee-Sanford, FL
Las Vegas-Henderson-Paradise, NV New York-Newark-Jersey City, NY-NJ-PA
San San
Diego-Chula Vista-Carlsbad, CA CA
Francisco-Oakland-Berkeley, Tampa-St. Petersburg-Clearwater, FL
20 Los Angeles-Long Beach-Anaheim, CA 20
Riverside-San
San Bernardino-Ontario,
Jose-Sunnyvale-SantaCA
Clara, CA
10 10
0 0
-.2 -.1 0 .1 .2 .3 -.2 -.1 0 .1 .2 .3
MSA PPPE (Nbr. Loans) MSA PPPE (Nbr. Loans)
Week 17: May10 - May16 Week 18: May17 - May23
100 100
90 90
Birmingham-Hoover, AL Buffalo-Cheektowaga, NY
Richmond, VA
80 Columbus, OH
Memphis, TN-MS-AR
80 Austin-Round Rock-Georgetown, TX
Columbus, OH
Hartford-East Hartford-Middletown, CT
Birmingham-Hoover, AL Cleveland-Elyria, OH Oklahoma City, OK
Austin-Round Rock-Georgetown, TX Kansas City, MO-KS Virginia Milwaukee-Waukesha, WI
Beach-Norfolk-Newport News, VA-NC
Nashville-Davidson--Murfreesboro--Franklin, TN
Minneapolis-St. Paul-Bloomington, MN-WI
% of Businesses reporting receiving PPP % of Businesses reporting receiving PPP
Raleigh-Cary,
NewNC
Orleans-Metairie, LA St. Louis, MO-IL
Boston-Cambridge-Newton, MA-NH
Indianapolis-Carmel-Anderson, Pittsburgh,
City,PA
St. Louis, MO-IL
Nashville-Davidson--Murfreesboro--Franklin,
Cincinnati, OH-KY-IN
Hartford-East
Milwaukee-Waukesha, WI Hartford-Middletown, IN
TN CT Oklahoma City, OK Orlando-Kissimmee-Sanford, FL Kansas MO-KS
Indianapolis-Carmel-Anderson, IN
70 Las Vegas-Henderson-Paradise, NV
Phoenix-Mesa-Chandler, AZ
New York-Newark-Jersey
Houston-TheRichmond,
Petersburg-Clearwater, FL
Tampa-St. City, NY-NJ-PA
Woodlands-Sugar Land, TX
Baltimore-Columbia-Towson,
VA MD
Minneapolis-St. Paul-Bloomington,
Cleveland-Elyria,
Detroit-Warren-Dearborn,
Pittsburgh,
MI
OH PA MN-WI 70 Las Vegas-Henderson-Paradise, NV NewSaltCA
Sacramento-Roseville-Folsom,York-Newark-Jersey City, NY-NJ-PA
Lake
Atlanta-Sandy Chicago-Naperville-Elgin,
Springs-Alpharetta, GA
Dallas-Fort Detroit-Warren-Dearborn,
IL-IN-WI
Philadelphia-Camden-Wilmington,
DC-VA-MD-WV TX
Washington-Arlington-Alexandria,Worth-Arlington,
Raleigh-Cary, NCWoodlands-Sugar
Cincinnati, Memphis,
OH-KY-IN
MI TN-MS-AR
PA-NJ-DE-MD
Philadelphia-Camden-Wilmington,
Chicago-Naperville-Elgin, IL-IN-WI
San Antonio-New PA-NJ-DE-MD Buffalo-Cheektowaga, NY
Braunfels, TX Los Angeles-Long Jacksonville,
Beach-Anaheim, CA FLCity, UT Houston-The
Denver-Aurora-Lakewood, CO San
Land, TX
Antonio-New Braunfels, TX
Atlanta-Sandy Springs-Alpharetta,
Seattle-Tacoma-Bellevue, WA GA
Denver-Aurora-Lakewood, Boston-Cambridge-Newton, MA-NH San San San Francisco-Oakland-Berkeley,
Diego-Chula Vista-Carlsbad,
Clara, CA CA CA Charlotte-Concord-Gastonia, NC-SC
San Francisco-Oakland-Berkeley, CA
Portland-Vancouver-Hillsboro, OR-WA
Salt Lake City,
Miami-Fort UT Dallas-Fort NC-SCO
Charlotte-Concord-Gastonia,
Lauderdale-Pompano Beach, FL Worth-Arlington, TX Jose-Sunnyvale-Santa Portland-Vancouver-Hillsboro, OR-WA
Tampa-St. Petersburg-Clearwater, FL
Seattle-Tacoma-Bellevue,
Virginia Beach-Norfolk-Newport News, VA-NC Miami-Fort Lauderdale-PompanoWA Beach, FL Baltimore-Columbia-Towson, MD
Providence-Warwick, RI-MA
Riverside-SanSan Diego-Chula Vista-Carlsbad, CA
Bernardino-Ontario,
Los CA
Angeles-Long Beach-Anaheim, CA FL Riverside-San Bernardino-Ontario, CA
60 San Jose-Sunnyvale-Santa Clara, CA Jacksonville, Washington-Arlington-Alexandria, DC-VA-MD-WV
Orlando-Kissimmee-Sanford, FL
Louisville/Jefferson
Providence-Warwick, RI-MA County, KY-IN
60 Phoenix-Mesa-Chandler, AZ
New Orleans-Metairie, LA Louisville/Jefferson County, KY-IN
Sacramento-Roseville-Folsom, CA
50 50
40 40
30 30
20 20
10 10
0 0
-.2 -.1 0 .1 .2 .3 -.2 -.1 0 .1 .2 .3
MSA PPPE (Nbr. Loans) MSA PPPE (Nbr. Loans)
Week 19: May24 - May30 Week 20: May31 - Jun6
100 100
90 90
Memphis, TN-MS-AR
Hartford-East Hartford-Middletown, CT
San Antonio-New Braunfels, TX
Birmingham-Hoover, AL
80 Cincinnati, OH-KY-IN
Hartford-East
Milwaukee-Waukesha, WI Hartford-Middletown, CT
Buffalo-Cheektowaga, NY
80 Birmingham-Hoover, AL
New Orleans-Metairie, LA
Pittsburgh, PA
Cleveland-Elyria,
Milwaukee-Waukesha,
Louis, WI
Nashville-Davidson--Murfreesboro--Franklin,
St.
Indianapolis-Carmel-Anderson, IN
OH
MO-IL TN County, KY-IN
Louisville/Jefferson
Orlando-Kissimmee-Sanford, FL Detroit-Warren-Dearborn,
Chicago-Naperville-Elgin, MI MO-IL
Nashville-Davidson--Murfreesboro--Franklin,
IL-IN-WI St. Louis, TN Orlando-Kissimmee-Sanford,
Dallas-Fort FL Cincinnati,Boston-Cambridge-Newton,
Worth-Arlington, TX OH-KY-IN MA-NH Oklahoma City, OK
Jacksonville, FL
OH Kansas City, MO-KS Oklahoma City, OK Detroit-Warren-Dearborn,
OH MI
Kansas City, MO-KS Buffalo-Cheektowaga, NY
% of Businesses reporting receiving PPP % of Businesses reporting receiving PPP
Tampa-St. Petersburg-Clearwater, FL Philadelphia-Camden-Wilmington,
Virginia Columbus, PA-NJ-DE-MD
Providence-Warwick,
Beach-Norfolk-Newport News, VA-NCRI-MA Miami-Fort Lauderdale-Pompano Chicago-Naperville-Elgin, IL-IN-WI
Beach, FLAustin-Round Columbus,
New Orleans-Metairie, LA Rock-Georgetown,
San TX
Jacksonville, FL Denver-Aurora-Lakewood, CO
Houston-The Woodlands-Sugar Land, TX Boston-Cambridge-Newton, MA-NHKY-IN
Louisville/Jefferson County, Atlanta-Sandy Springs-Alpharetta,
Tampa-St. Petersburg-Clearwater,
GA
FL
Minneapolis-St.
Antonio-New Braunfels,
Paul-Bloomington,
TX MN-WI
70 Las Vegas-Henderson-Paradise, NV
Riverside-San
San
San San
Phoenix-Mesa-Chandler,
Sacramento-Roseville-Folsom,
Francisco-Oakland-Berkeley,
Diego-Chula
Bernardino-Ontario,
Jose-Sunnyvale-Santa
Los Clara, CA
CASalt Lake
Portland-Vancouver-Hillsboro,
Vista-Carlsbad, CA
CA
Angeles-Long
AZ CA
New York-Newark-Jersey
Miami-Fort
Beach-Anaheim,
OR-WA
City, UT City, NY-NJ-PA
Lauderdale-Pompano
Seattle-Tacoma-Bellevue,
CA Beach, FL
WA
Dallas-Fort
Atlanta-Sandy Springs-Alpharetta, GA Richmond,
Austin-Round
Charlotte-Concord-Gastonia,
Worth-Arlington,
VA
Rock-Georgetown, TXTX Pittsburgh, PA
Cleveland-Elyria, OH
Baltimore-Columbia-Towson, MD
Indianapolis-Carmel-Anderson, IN
70 San Francisco-Oakland-Berkeley,New
Las Vegas-Henderson-Paradise, NV Beach-Anaheim,
Sacramento-Roseville-Folsom,
Los Angeles-Long
CASalt Lake
York-Newark-Jersey
CA
CA City, UT Charlotte-Concord-Gastonia,
City, NY-NJ-PA NC-SC VA
Richmond,Philadelphia-Camden-Wilmington,
Houston-The Woodlands-Sugar Land, TX MD
Baltimore-Columbia-Towson, PA-NJ-DE-MD
Raleigh-Cary, NC NC-SC
Washington-Arlington-Alexandria, DC-VA-MD-WV Seattle-Tacoma-Bellevue, WA Providence-Warwick, RI-MA
Minneapolis-St. Paul-Bloomington, MN-WI Phoenix-Mesa-Chandler, AZ
Riverside-San Bernardino-Ontario, CA Raleigh-Cary, NC
Denver-Aurora-Lakewood, CO
Washington-Arlington-Alexandria, DC-VA-MD-WV
San San Diego-Chula Vista-Carlsbad,
Jose-Sunnyvale-Santa Clara, CA CA
Portland-Vancouver-Hillsboro, OR-WA
Virginia Beach-Norfolk-Newport News, VA-NC
60 60 Memphis, TN-MS-AR
50 50
40 40
30 30
20 20
10 10
0 0
-.2 -.1 0 .1 .2 .3 -.2 -.1 0 .1 .2 .3
MSA PPPE (Nbr. Loans) MSA PPPE (Nbr. Loans)
68
Figure A.12: Unmet PPP Demand and Exposure to MSA PPPE
Figure A.12 are scatterplots of the MSA exposure to the number-based PPPE and the percentage of businesses in each MSA that applied but
did not receive PPP in each week. The plots represent the evolution of the two variables for each survey week. Data comes from the Census
Bureau, SBA, Call Reports, Summary of Deposits, and County Business Patterns.
Week 15: April26 - May2nd Week 16: May3rd - May9
60 Riverside-San Bernardino-Ontario, CA 60
San Jose-Sunnyvale-Santa Clara, CA
Los Angeles-Long Beach-Anaheim, CA
San Diego-Chula Vista-Carlsbad, CA
Jacksonville,
San Francisco-Oakland-Berkeley, CA
Sacramento-Roseville-Folsom, CA FLTampa-St. Petersburg-Clearwater, FL
New York-Newark-Jersey City, NY-NJ-PA
Diff. between % of businesses applying and receiving PPP Diff. between % of businesses applying and receiving PPP
50 50
Orlando-Kissimmee-Sanford, FL
Las Vegas-Henderson-Paradise, NV Atlanta-Sandy Springs-Alpharetta, GA Hartford-East
Pittsburgh, PA
Hartford-Middletown, CT
Portland-Vancouver-Hillsboro, OR-WA
Miami-Fort Lauderdale-Pompano Beach, FL
Washington-Arlington-Alexandria,
Phoenix-Mesa-Chandler, AZ Seattle-Tacoma-Bellevue, WA DC-VA-MD-WV Land, TX
Houston-The Woodlands-Sugar
Detroit-Warren-Dearborn, MI
Cleveland-Elyria, OH
Charlotte-Concord-Gastonia, NC-SCNashville-Davidson--Murfreesboro--Franklin, TN
40 Austin-Round Rock-Georgetown,
Denver-Aurora-Lakewood, CO TX
Columbus, OH
40
Dallas-Fort Worth-Arlington, TX
Philadelphia-Camden-Wilmington, PA-NJ-DE-MD
Salt Lake City, UT Boston-Cambridge-Newton, MA-NH
Raleigh-Cary, NC
New Orleans-Metairie, San
Chicago-Naperville-Elgin, LA Antonio-New Braunfels, TX
IL-IN-WI
Cincinnati, OH-KY-IN
Baltimore-Columbia-Towson, MD San Diego-Chula Vista-Carlsbad, CA
Providence-Warwick, RI-MA San Francisco-Oakland-Berkeley, CA
Memphis, TN-MS-AR Miami-Fort
Sacramento-Roseville-Folsom, CA Lauderdale-Pompano Beach, FL
30 Richmond, VA
Buffalo-Cheektowaga, NY
30 Las Vegas-Henderson-Paradise, NV
San Jose-Sunnyvale-Santa Clara, CA
Riverside-San Los Angeles-Long
Bernardino-Ontario, CA Beach-Anaheim, CA Richmond, VA
Washington-Arlington-Alexandria, DC-VA-MD-WV
Virginia Beach-Norfolk-Newport News, VA-NC
St. Louis, MO-IL
Indianapolis-Carmel-Anderson, IN Atlanta-Sandy Springs-Alpharetta, GA
Milwaukee-Waukesha, WI Seattle-Tacoma-Bellevue, WA
Phoenix-Mesa-Chandler, AZ Tampa-St. Petersburg-Clearwater, FL Detroit-Warren-Dearborn, MI
Louisville/Jefferson County, KY-IN Orlando-Kissimmee-Sanford, FL
New York-Newark-Jersey City, NY-NJ-PA
Birmingham-Hoover, AL
Portland-Vancouver-Hillsboro, OR-WA Charlotte-Concord-Gastonia, NC-SC Philadelphia-Camden-Wilmington, PA-NJ-DE-MD
20 Minneapolis-St. Paul-Bloomington, MN-WI
20 Salt Lake City, UT New Orleans-Metairie,
Virginia LACincinnati, OH-KY-IN News, VA-NC
Beach-Norfolk-Newport
Baltimore-Columbia-Towson,
Denver-Aurora-Lakewood, CO
MD
Cleveland-Elyria, OH
Oklahoma City, OK Raleigh-Cary, NC Milwaukee-Waukesha, WI
Hartford-East Hartford-Middletown,
Boston-Cambridge-Newton, MA-NH CT
Jacksonville, FL Dallas-Fort Worth-Arlington, TX
Birmingham-Hoover, AL Pittsburgh, PA
Houston-The Woodlands-Sugar Land, TX
San Antonio-New Braunfels, TX
Minneapolis-St.
Chicago-Naperville-Elgin, IL-IN-WI Paul-Bloomington,
Louisville/Jefferson
St. Louis, MO-IL MN-WI
County, KY-IN
Indianapolis-Carmel-Anderson, IN
Kansas City, MO-KS Austin-Round Rock-Georgetown, TX Kansas City, MO-KS
Memphis, TN-MS-AR
Nashville-Davidson--Murfreesboro--Franklin, TNRI-MA
Providence-Warwick,
Buffalo-Cheektowaga, NY
Columbus, OH Oklahoma City, OK
10 10
0 0
-.2 -.1 0 .1 .2 .3 -.2 -.1 0 .1 .2 .3
MSA PPPE (Nbr. Loans) MSA PPPE (Nbr. Loans)
Week 17: May10 - May16 Week 18: May17 - May23
60 60
50
Diff. between % of businesses applying and receiving PPP Diff. between % of businesses applying and receiving PPP
50
40 40
30
30
20
20
Sacramento-Roseville-Folsom, CA
Los Angeles-Long Beach-Anaheim, CA Lauderdale-Pompano Beach, FL Riverside-San Bernardino-Ontario, CA Memphis, TN-MS-AR
Baltimore-Columbia-Towson, MD
Riverside-San Bernardino-Ontario, CA
Miami-Fort
Richmond,
Washington-Arlington-Alexandria, VA
DC-VA-MD-WV
10 Las San Diego-Chula Vista-Carlsbad, CA
Phoenix-Mesa-Chandler,
Vegas-Henderson-Paradise,
Los NV
Angeles-Long
New York-Newark-Jersey City, NY-NJ-PA
AZ CA
Beach-Anaheim, Raleigh-Cary, NC NC-SC
10 Las San Diego-Chula NV
Vegas-Henderson-Paradise,Vista-Carlsbad, CA
Portland-Vancouver-Hillsboro,
San Jose-Sunnyvale-Santa Clara, CA
Atlanta-Sandy Orlando-Kissimmee-Sanford,
Springs-Alpharetta,
Tampa-St. City, NY-NJ-PA
OR-WA
New York-Newark-Jersey Charlotte-Concord-Gastonia,
Houston-The
Petersburg-Clearwater,
Seattle-Tacoma-Bellevue, WA
New Orleans-Metairie,
GA
FL
FL Philadelphia-Camden-Wilmington, PA-NJ-DE-MD
NC-SC
Woodlands-Sugar
LA
Land, TX
Columbus, OH Buffalo-Cheektowaga, NY
Clara, CA
Sacramento-Roseville-Folsom,
San Jose-Sunnyvale-Santa
San Francisco-Oakland-Berkeley, CA CA
Charlotte-Concord-Gastonia,
Miami-Fort Lauderdale-Pompano Beach, FL Louisville/Jefferson County, KY-IN
Jacksonville, FL Cleveland-Elyria, OH PA
Pittsburgh, New Orleans-Metairie,
Washington-Arlington-Alexandria,
Atlanta-Sandy Houston-The
Springs-Alpharetta, GA Woodlands-Sugar
DC-VA-MD-WV LA
San
Land,
Antonio-New Braunfels,
TX
Philadelphia-Camden-Wilmington, TXPA
PA-NJ-DE-MD
San Francisco-Oakland-Berkeley, CA Dallas-Fort Worth-Arlington,
Chicago-Naperville-Elgin, TX
Cincinnati,
IL-IN-WI OH-KY-IN
Detroit-Warren-Dearborn, MI Seattle-Tacoma-Bellevue,
Portland-Vancouver-Hillsboro, OR-WA WA Denver-Aurora-Lakewood,
Orlando-Kissimmee-Sanford, FLCOD etroit-Warren-Dearborn, MIPittsburgh,
Dallas-Fort Worth-Arlington,
Virginia Beach-Norfolk-Newport
TX News, VA-NC
Phoenix-Mesa-Chandler, AZSalt Lake City, UT Austin-Round Rock-Georgetown, TX Providence-Warwick, RI-MA Cleveland-Elyria, OH
Denver-Aurora-Lakewood, CO San Antonio-NewSt. Louis, MO-IL
Braunfels, TX
Boston-Cambridge-Newton, MA-NH Chicago-Naperville-Elgin, IL-IN-WI
Tampa-St. Petersburg-Clearwater, FL Providence-Warwick,
Hartford-East RI-MA
Hartford-Middletown, CT
Raleigh-Cary, NC Baltimore-Columbia-Towson, MD Indianapolis-Carmel-Anderson,
Memphis, TN-MS-AR
Nashville-Davidson--Murfreesboro--Franklin, TNMN-WI IN Salt Lake
Jacksonville, FLCity, UT Richmond, VA Cincinnati, St. Louis,
OH-KY-IN MO-IL
Kansas City, MO-KS
Indianapolis-Carmel-Anderson, IN Oklahoma City, OK
Virginia Minneapolis-St.
Beach-Norfolk-Newport
Paul-Bloomington,
News, VA-NC Birmingham-Hoover, AL Minneapolis-St.
Columbus, OH Paul-Bloomington, MN-WI
Hartford-East
Louisville/Jefferson
Hartford-Middletown,
County, KY-INCT Boston-Cambridge-Newton,
Nashville-Davidson--Murfreesboro--Franklin, TN MA-NH
Austin-Round Rock-Georgetown, TX
Milwaukee-Waukesha, WI
Birmingham-Hoover, AL
Kansas Oklahoma City, OK
0
Milwaukee-Waukesha, WICity, MO-KS
0 Buffalo-Cheektowaga, NY
-.2 -.1 0 .1 .2 .3 -.2 -.1 0 .1 .2 .3
MSA PPPE (Nbr. Loans) MSA PPPE (Nbr. Loans)
Week 19: May24 - May30 Week 20: May31 - Jun6
60 60
Diff. between % of businesses applying and receiving PPP Diff. between % of businesses applying and receiving PPP
50 50
40 40
30 30
20 20
10 Charlotte-Concord-Gastonia, NC-SC 10
Virginia Beach-Norfolk-Newport News, VA-NC
Las San Jose-Sunnyvale-Santa
Vegas-Henderson-Paradise, NV Clara, CA Las Vegas-Henderson-Paradise, NV Jacksonville, FL New Orleans-Metairie, LA Cincinnati,
Columbus, OH-KY-IN
OH
Dallas-Fort
Tampa-St. Petersburg-Clearwater, FL Worth-Arlington,
New Orleans-Metairie, TX San Antonio-New Braunfels, TX
Riverside-San Bernardino-Ontario,
San SanLos
CA
Francisco-Oakland-Berkeley,
Angeles-Long
Diego-Chula Portland-Vancouver-Hillsboro,
Vista-Carlsbad, CA
Phoenix-Mesa-Chandler,
CA
Beach-Anaheim,
New
AZ CA Atlanta-Sandy
OR-WA Springs-Alpharetta,
Beach,
City, NY-NJ-PA
Lauderdale-Pompano
York-Newark-Jersey
Miami-Fort FL GA Virginia LA
Beach-Norfolk-Newport News, VA-NC Sacramento-Roseville-Folsom,
Tampa-St. Petersburg-Clearwater,
Seattle-Tacoma-Bellevue,
CA WA FL
Dallas-Fort NC-SC
Charlotte-Concord-Gastonia, Philadelphia-Camden-Wilmington,
Detroit-Warren-Dearborn,
Worth-Arlington, TX Memphis,
MI PA-NJ-DE-MD
TN-MS-AR
Raleigh-Cary, NC Rock-Georgetown,
Austin-Round Philadelphia-Camden-Wilmington,
Detroit-Warren-Dearborn,
TX MILouisville/Jefferson
PA-NJ-DE-MDCounty, KY-IN San San Diego-Chula
Jose-Sunnyvale-Santa Clara, CA CA
Vista-Carlsbad, Baltimore-Columbia-Towson, MD
Seattle-Tacoma-Bellevue, WA Houston-The Woodlands-Sugar
Washington-Arlington-Alexandria, DC-VA-MD-WV Land, TX Phoenix-Mesa-Chandler, AZ York-Newark-Jersey
Portland-Vancouver-Hillsboro,
New OR-WA Washington-Arlington-Alexandria,
City, NY-NJ-PA DC-VA-MD-WV
Richmond, VA Cleveland-Elyria, OH Riverside-San Bernardino-Ontario, CA
SanLos Angeles-Long
Francisco-Oakland-Berkeley,
Beach-Anaheim,
CA CA
Miami-Fort Atlanta-Sandy Houston-The
Springs-Alpharetta, GA Richmond,
Woodlands-Sugar
COVA IL-IN-WI
Chicago-Naperville-Elgin,
Denver-Aurora-Lakewood,
Lauderdale-Pompano Beach, FL Land, TX
Chicago-Naperville-Elgin, Columbus,
Milwaukee-Waukesha,
Cincinnati, OH Pittsburgh,
OH-KY-IN
Minneapolis-St.
IL-IN-WI WI PA MN-WI
Paul-Bloomington, Orlando-Kissimmee-Sanford, FL Milwaukee-Waukesha, WI
Orlando-Kissimmee-Sanford,Baltimore-Columbia-Towson,
FL MD
St. Indianapolis-Carmel-Anderson, IN
Memphis,
Louis, MO-IL
TN-MS-AR Minneapolis-St. Hartford-East
Paul-Bloomington, MN-WI
St. Louis, MO-IL Hartford-Middletown, CT
Sacramento-Roseville-Folsom,
SaltCALake City, UT Denver-Aurora-Lakewood, CO Boston-Cambridge-Newton,
Kansas City, MO-KS MA-NH Austin-Round Rock-Georgetown, TX Kansas City,PA
MO-KS MA-NH
Buffalo-Cheektowaga, NY
Jacksonville, FL Nashville-Davidson--Murfreesboro--Franklin,
Hartford-East Hartford-Middletown, CT
TN Oklahoma City, OK Birmingham-Hoover,
Raleigh-Cary, NC AL Boston-Cambridge-Newton,
Pittsburgh,
Cleveland-Elyria, OH
0 0 Providence-Warwick, TNRI-MA
Indianapolis-Carmel-Anderson,
Nashville-Davidson--Murfreesboro--Franklin,
Louisville/Jefferson IN
County, KY-IN Oklahoma City, OK
Birmingham-Hoover, AL San Antonio-New Braunfels,
Providence-Warwick,
TX RI-MABuffalo-Cheektowaga, NY Salt Lake City, UT
-.2 -.1 0 .1 .2 .3 -.2 -.1 0 .1 .2 .3
MSA PPPE (Nbr. Loans) MSA PPPE (Nbr. Loans)
69
Table A.1: Wells Fargo and PPP Allocation
Table A.1 reports the results of ordinary least squares (OLS) regressions examining the impact of the exposure of the Congressional District
to Wells Fargo on the allocation of PPP funds. The dependent variable of the specifications, Ln(Total PPP Allocation per establishment),
is the natural logarithm total amount of PPP funds disbursed to small businesses in each congressional district divided by the number of
establishments in the congressional district. The dependent variable of the specifications in Panel B, Ln(Total PPP Allocation per establishment)
is the total number of PPP loans made to small businesses in each congressional district divided by the number of establishments in the
congressional district. Share of Wells in the Congressional District is share of branches owned by Wells Fargo in the Congressional District.
Ln(Payroll) is the natural logarithm of the sum of payroll of all establishments in the congressional district Ln(Employment) is the natural
logarithm of total employment in the congressional district Industry Shares are additional controls for the share of establishments in each
two-digit NAICS code industry. The specification of column (4) includes state fixed effects Standard errors are presented in parentheses, and
are clustered at the level of the state. ***, **, and *, represent statistical significance at 1%, 5%, and 10% levels, respectively.
Panel A: Total PPP Allocation per Establishment at the Congressional District
(1) (2) (3) (4)
Ln(Total PPP Allocation per Establishment)
Share of Wells Branches in Cong Dist. -1.871∗∗∗ -1.977∗∗∗ -1.277∗∗ -2.317
(0.526) (0.594) (0.629) (1.721)
Ln(Total Payroll) 0.078 0.116 0.127
(0.299) (0.253) (0.251)
Ln(Employment) 0.259 0.105 0.183
(0.521) (0.354) (0.367)
Observations 436 436 436 436
Adjusted R2 0.047 0.088 0.125 0.097
Industry Shares No No Yes Yes
State Fixed Effects No No No Yes
Panel B: Fraction of Establishment receiving PPP at the Congressional District
(1) (2) (3) (4)
Fraction of Establishments receiving PPP
Share of Wells Branches in Cong Dist. -0.588∗∗∗ -0.492∗∗∗ -0.272∗∗ -0.287
(0.148) (0.137) (0.108) (0.276)
Ln(Total Payroll) -0.162∗∗∗ -0.119∗∗∗ -0.071∗
(0.061) (0.043) (0.041)
Ln(Employment) 0.233∗∗ 0.172∗∗∗ 0.099
(0.098) (0.062) (0.070)
Observations 436 436 436 436
Adjusted R2 0.151 0.210 0.376 0.455
Industry Shares No No Yes Yes
State Fixed Effects No No No Yes
70
Table A.2: Business Shutdowns and PPP Allocation
Table A.2 reports the results of ordinary least squares (OLS) regressions examining the relation between the allocation of PPP funds and the share of businesses that shutdown operations in the last
week of March. The dependent variable, Share of Firms Shutdown (March 31st, 2020), is the share of businesses in the congressional district that did not operate in the week of March29th–April 4th.
Ln(Total PPP Allocation per establishment) is the natural logarithm total amount of PPP funds disbursed to small businesses in each congressional district divided by the number of establishments in
the congressional district. Fraction of Establishments receiving PPP is the total number of PPP loans made to small businesses in each congressional district divided by the number of establishments
in the congressional district. Congressional District PPPE (Vol) is the congressional district average of the PPPE based on total amounts of outstanding PPP and small business loans, weighted by the
share of deposits of each bank in each congressional district. Congressional District PPPE (Nbr.) is the congressional district average of the PPPE based on the number of outstanding loans, weighted
by the share of deposits of each bank in each congressional district. Ln(Payroll) is the natural logarithm of the sum of payroll of all establishments in the congressional district Ln(Employment)
is the natural logarithm of total employment in the congressional district Industry Shares are additional controls for the share of establishments in each two-digit NAICS code industry. Standard
errors are presented in parentheses, and are clustered at the level of the state. ***, **, and *, represent statistical significance at 1%, 5%, and 10% levels, respectively.
(1) (2) (3) (4) (5) (6) (7) (8)
71 Share of Firms Shutdown (March 31st, 2020)
Fraction receiving PPP -0.078 -0.025
(0.049) (0.045)
Ln(PPP loan per establishment) 0.008 0.008
(0.007) (0.006)
Congressional District PPPE (Vol) 0.020∗ -0.008
(0.010) (0.008)
Congressional District PPPE (Nbr.) 0.019 -0.015
(0.012) (0.011)
Ln(Total Payroll) 0.062 0.011 0.071∗ 0.014 0.096∗∗ 0.006 0.095∗∗ -0.004
(0.039) (0.022) (0.041) (0.021) (0.044) (0.025) (0.044) (0.030)
Ln(Employment) 0.023 0.010 -0.000 -0.001 -0.038 0.016 -0.038 0.028
(0.050) (0.043) (0.053) (0.040) (0.064) (0.046) (0.066) (0.051)
Observations 436 436 436 436 436 436 436 436
Adjusted R2 0.317 0.583 0.315 0.584 0.335 0.584 0.325 0.586
Industry Shares Yes Yes Yes Yes Yes Yes Yes Yes
State Fixed Effects No Yes No Yes No Yes No Yes
Table A.3: Decline in Hours Worked and PPP Allocation
Table A.3 reports the results of ordinary least squares (OLS) regressions examining the relation between the allocation of PPP funds and the share of businesses that shutdown operations in the last
week of March. The dependent variable, Decline in Hours Worked (March 31st, 2020), is the decline in hours worked at establishments located in the congressional district related to the average
hours worked in the same weekdays of the last two weeks of January at the same congressional district. Ln(Total PPP Allocation per establishment) is the natural logarithm total amount of PPP
funds disbursed to small businesses in each congressional district divided by the number of establishments in the congressional district. Fraction of Establishments receiving PPP is the total number
of PPP loans made to small businesses in each congressional district divided by the number of establishments in the congressional district. Congressional District PPPE (Vol) is the congressional
district average of the PPPE based on total amounts of outstanding PPP and small business loans, weighted by the share of deposits of each bank in each congressional district. Congressional District
PPPE (Nbr.) is the congressional district average of the PPPE based on the number of outstanding loans, weighted by the share of branches of each bank in each congressional district. Ln(Payroll)
is the natural logarithm of the sum of payroll of all establishments in the congressional district Ln(Employment) is the natural logarithm of total employment in the congressional district Industry
Shares are additional controls for the share of establishments in each two-digit NAICS code industry. Standard errors are presented in parentheses, and are clustered at the level of the state. ***,
**, and *, represent statistical significance at 1%, 5%, and 10% levels, respectively.
72 (1) (2) (3) (4) (5) (6) (7) (8)
Decline in Hours Worked (March 31st, 2020)
Fraction receiving PPP 0.121∗ 0.051
(0.067) (0.068)
Ln(PPP loan per establishment) -0.005 -0.005
(0.007) (0.007)
Congressional District PPPE (Vol) -0.022∗∗ -0.001
(0.010) (0.011)
Congressional District PPPE (Nbr.) -0.024∗ -0.002
(0.014) (0.017)
Ln(Total Payroll) -0.073∗∗ 0.014 -0.087∗∗ 0.009 -0.114∗∗∗ 0.008 -0.117∗∗∗ 0.007
(0.032) (0.043) (0.034) (0.041) (0.038) (0.047) (0.038) (0.052)
Ln(Employment) -0.011 -0.065 0.017 -0.054 0.063 -0.058 0.069 -0.056
(0.043) (0.069) (0.046) (0.068) (0.059) (0.076) (0.063) (0.083)
Observations 436 436 436 436 436 436 436 436
Adjusted R2 0.274 0.462 0.267 0.461 0.286 0.461 0.280 0.461
Industry Shares Yes Yes Yes Yes Yes Yes Yes Yes
State Fixed Effects No Yes No Yes No Yes No Yes
Table A.4: Employment Outcomes and PPPE: 2SLS Estimator
Table A.4 reports the results of 2SLS regression in first differences comparing the Post-PPP period to the post-lockdown period prior to the passage of the PPP. The dependent variable in Panel A, ∆
Bus. Shutdown is either defined as the average percentage change in business shutdowns in the Congressional District between the pre- and post-PPP period (first three columns) or the percentage
change in the fraction of weeks in which a business decided to shutdown between the pre- and post-PPP period (last three columns). The dependent variable in Panel B, ∆ Hours Worked is either
the average change in firm hours worked in the Congressional District between the pre- and post-PPP period (first three columns) or the percentage change in weekly hours worked at the firm-level
between the pre- and post-PPP period (last three columns). Fraction receiving PPP is the percentage of businesses that received PPP loans during the first round in the congressional district where
the firm is located. Congressional District PPPE (Nbr.) is the congressional district average of the PPPE based on the number of outstanding loans, weighted by the share of branches of each bank
in each congressional district. Standard errors are presented in parentheses, and are clustered at the level of the state. ***, **, and *, represent statistical significance at 1%, 5%, and 10% levels,
respectively.
Panel A: Business Shutdowns
(1) (2) (3) (1) (2) (3)
Congressional District Level Establishment Level
First-Stage Reduced Form Second-Stage First-Stage Reduced Form Second-Stage
Dep. Variable: Fraction PPP ∆ Bus. Shutdown ∆ Bus. Shutdown Fraction PPP ∆ Bus. Shutdown ∆ Bus. Shutdown
Congressional District PPPE (Nbr.) 0.378∗∗∗ -0.017 0.335∗∗∗ -0.009
(0.037) (0.031) (0.041) (0.032)
Fraction receiving PPP -0.044 -0.027
(0.083) (0.095)
Observations 428 428 428 48174 48174 48174
73 Adjusted R2 0.466 0.324 -0.012 0.535 0.012 -0.000
F-Stat 103.207 67.367
State Fixed Effects Yes Yes Yes No No No
State×Industry Fixed Effects No No No Yes Yes Yes
Panel B: Decline in Hours Worked
(1) (2) (3) (1) (2) (3)
Congressional District Level Establishment Level
First-Stage Reduced Form Second-Stage First-Stage Reduced Form Second-Stage
Dep. Variable: Fraction PPP ∆ Hours Worked ∆ Hours Worked Fraction PPP ∆ Hours Worked ∆ Hours Worked
Congressional District PPPE (Nbr.) 0.378∗∗∗ 0.014 0.335∗∗∗ 0.024
(0.037) (0.017) (0.041) (0.019)
Fraction receiving PPP 0.038 0.070
(0.046) (0.059)
Observations 428 428 428 48174 48174 48174
Adjusted R2 0.466 0.397 -0.011 0.535 0.024 -0.001
F-Stat 103.207 67.367
State Fixed Effects Yes Yes Yes No No No
State×Industry Fixed Effects No No No Yes Yes Yes
Table A.5: Unmet Demand for Loans and Exposure to PPP
Table A.5 reports the results of ordinary least squares (OLS) regressions examining the relation between the geographic allocation of PPP funds during the first round and the percentage of
businesses reporting applying but not receiving PPP funds at the state-by-industry level. The dependent variable is the difference between the percentage of businesses that applied for PPP funds
and the percentage of businesses that received PPP funds at the state-by-industry level. State PPPE (Nbr.) is the state average of the PPPE based on the number of outstanding loans. All specifications
include industry×week fixed-effects. Standard errors are presented in parentheses, and are clustered at the level of the state. ***, **, and *, represent statistical significance at 1%, 5%, and 10%
levels, respectively.
74
(1) (2) (3) (4) (5) (6) (7)
% PPP Requested - % PPP Received
State PPPE (Nbr.) -24.564∗∗∗ -78.463∗∗∗ -42.087∗∗∗ -16.478∗∗∗ -15.676∗∗∗ -8.184∗∗∗ -6.107∗∗∗
(2.525) (7.743) (3.943) (2.851) (1.975) (1.789) (1.333)
Observations 2229 277 386 390 390 393 393
Adjusted R2 0.757 0.562 0.421 0.247 0.170 0.074 0.096
Industry Fixed Effects No Yes Yes Yes Yes Yes Yes
Industry×Week Fixed Effects Yes No No No No No No
Week: Full Sample Apr26–May2 May3–May9 May10–May16 May17–May23 May24–May30 May31–Jun6
Table A.6: Unmet Demand for PPP and Missed Loan Payments
Table A.6 reports the results of ordinary least squares (OLS) and instrumental variables (IV) regressions examining the relation between the geographic allocation of PPP funds during the first
round and the percentage of businesses reporting missing scheduled payments at the state-by-industry level. The dependent variable is the percentage of firms reporting a missed scheduled loan
payment. % PPP Requested - % PPP Received is the difference between the percentage of businesses that applied for PPP funds and the percentage of businesses that received PPP funds at a
state-by-industry level. State PPPE (Nbr.) is the state average of the PPPE based on the number of outstanding loans. Pre-PPP Decline Hours Worked is the average decline in hours worked in each
state between January and the last week of March. Pre-PPP State Covid-19 Cases (per capita) are per capita number of reported Covid-19 cases in the state. Pre-PPP State Covid-19 Deaths (per
capita) are the weekly per capita number of reported Covid-19 deaths in the state. Pre-PPP State Social Distancing Index is the change in average distance travelled in the state until the end of
March using individuals’ GPS signals. All specifications include industry×week fixed-effects. Standard errors are presented in parentheses, and are clustered at the level of the state. ***, **, and
*, represent statistical significance at 1%, 5%, and 10% levels, respectively.
(1) (2) (3) (4)
OLS Reduced Form IV
First-Stage Second-Stage
75 Dep. Variable: % Missing Loan Payments % Missing Loan Payments % PPP Request - % PPP Received % Missing Loan Payments
% PPP Requested - % PPP Received 0.056∗∗∗ 0.219∗∗∗
(0.020) (0.043)
State PPPE (Nbr.) -4.401∗∗∗ -23.458∗∗∗
(1.380) (2.264)
Pre-PPP Decline Hours Worked -3.084∗∗ -3.242∗∗∗ -6.071∗∗ -1.610∗
(1.204) (0.977) (2.764) (0.922)
Pre-PPP State Covid-19 Cases (per capita) 0.838∗ 0.801∗ -0.248 0.810∗∗
(0.437) (0.409) (0.552) (0.374)
Pre-PPP State Covid-19 Deaths (per capita) -26.712∗∗ -26.047∗ 24.368 -27.610∗∗
(13.179) (13.396) (17.188) (11.926)
Pre-PPP Social Distancing Index -5.507∗∗ -6.787∗∗∗ -1.591 -4.365∗
(2.580) (2.363) (5.317) (2.328)
Observations 2217 2487 2217 2217
Adjusted R2 0.576 0.554 0.762 -0.060
FStat 107.388
Industry×Week Fixed Effects Yes Yes Yes Yes
Table A.7: Unmet Demand for PPP and Missed Non-Loan Payments
Table A.7 reports the results of ordinary least squares (OLS) and instrumental variables (IV) regressions examining the relation between the geographic allocation of PPP funds during the first
round and the percentage of businesses reporting missing scheduled payments at the state-by-industry level. The dependent variable is the percentage of firms reporting a missed other scheduled
payment such as rent, utilities, and payroll. % PPP Requested - % PPP Received is the difference between the percentage of businesses that applied for PPP funds and the percentage of businesses
that received PPP funds at a state-by-industry level. State PPPE (Nbr.) is the state average of the PPPE based on the number of outstanding loans. Pre-PPP Decline Hours Worked is the average
decline in hours worked in each state between January and the last week of March. Pre-PPP State Covid-19 Cases (per capita) are per capita number of reported Covid-19 cases in the state. Pre-PPP
State Covid-19 Deaths (per capita) are the weekly per capita number of reported Covid-19 deaths in the state. Pre-PPP State Social Distancing Index is the change in average distance travelled in
the state until the end of March using individuals’ GPS signals. All specifications include industry×week fixed-effects. Standard errors are presented in parentheses, and are clustered at the level
of the state. ***, **, and *, represent statistical significance at 1%, 5%, and 10% levels, respectively.
(1) (2) (3) (4)
OLS Reduced Form IV
First-Stage Second-Stage
76 Dep. Variable: % Missing Schd. Payments % Missing Schd. Payments % PPP Request - % PPP Received % Missing Schd. Payments
% PPP Requested - % PPP Received 0.229∗∗∗ 0.639∗∗∗
(0.029) (0.068)
State PPPE (Nbr.) -14.807∗∗∗ -23.577∗∗∗
(2.651) (2.211)
Pre-PPP Decline Hours Worked -11.404∗∗∗ -12.176∗∗∗ -6.119∗∗ -7.682∗∗
(3.328) (4.094) (2.770) (3.127)
Pre-PPP State Covid-19 Cases (per capita) 4.261∗∗∗ 4.189∗∗∗ -0.270 4.199∗∗∗
(0.666) (0.577) (0.556) (0.361)
Pre-PPP State Covid-19 Deaths (per capita) -110.471∗∗∗ -101.315∗∗∗ 24.296 -112.755∗∗∗
(25.156) (20.723) (17.605) (18.614)
Pre-PPP Social Distancing Index -16.215∗∗∗ -13.084∗∗ -1.645 -13.250∗∗∗
(5.683) (6.317) (5.344) (4.319)
Observations 2205 2448 2205 2205
Adjusted R2 0.718 0.698 0.762 0.086
FStat 113.656
Industry×Week Fixed Effects Yes Yes Yes Yes
Table A.8: Unmet Demand for Loans and Cash-on-Hand
Table A.8 reports the results of ordinary least squares (OLS) and instrumental variables (IV) regressions examining the relation between the geographic allocation of PPP funds during the first
round and the percentage of businesses reporting at two months or more of cash-on-hand at the state-by-industry level. The dependent variable is the fraction of businesses with cash on hand to
sustain operations for two months or more. % PPP Requested - % PPP Received is the difference between the percentage of businesses that applied for PPP funds and the percentage of businesses
that received PPP funds at a state-by-industry level. State PPPE (Nbr.) is the state average of the PPPE based on the number of outstanding loans. Pre-PPP Decline Hours Worked is the average
decline in hours worked in each state between January and the last week of March. Pre-PPP State Covid-19 Cases (per capita) are per capita number of reported Covid-19 cases in the state. Pre-PPP
State Covid-19 Deaths (per capita) are the weekly per capita number of reported Covid-19 deaths in the state. Pre-PPP State Social Distancing Index is the change in average distance travelled in
the state until the end of March using individuals’ GPS signals. All specifications include industry×week fixed-effects. Standard errors are presented in parentheses, and are clustered at the level
of the state. ***, **, and *, represent statistical significance at 1%, 5%, and 10% levels, respectively.
(1) (2) (3) (4)
OLS Reduced Form IV
First-Stage Second-Stage
Dep. Variable: % Cash 3 months % Cash 3 months % PPP Request - % PPP Received % Cash 3 months
% PPP Requested - % PPP Received -0.172∗∗∗ -0.411∗∗
77 (0.047) (0.152)
State PPPE (Nbr.) 8.920∗∗ -21.380∗∗∗
(3.768) (2.019)
Pre-PPP Decline Hours Worked 10.973∗ 15.861∗∗∗ -6.302 9.596
(5.868) (5.378) (4.006) (5.830)
Pre-PPP State Covid-19 Cases (per capita) -1.912∗∗ -2.003∗∗ 0.219 -1.820∗∗
(0.743) (0.818) (0.381) (0.756)
Pre-PPP State Covid-19 Deaths (per capita) 75.800∗∗ 76.722∗∗ 17.796 76.570∗∗
(29.392) (31.278) (17.170) (31.054)
Pre-PPP Social Distancing Index -6.618 -12.334 5.529 -6.293
(9.077) (11.326) (5.524) (9.132)
Observations 903 918 903 903
Adjusted R2 0.637 0.640 0.846 -0.077
FStat 112.191
Industry×Week Fixed Effects Yes Yes Yes Yes
Table A.9: % Receiving PPP and % Missing Payments: Week-by-Week
Table A.9 reports the results of instrumental variables (IV) regressions split week by week examining the relation between the geographic allocation of PPP funds during the first round and the
percentage of businesses reporting missing scheduled payments at the state-by-industry level. In the top panel, the dependent variable is the percentage of firms reporting a missed scheduled loan
payment. In the bottom panel, the dependent variable is the percentage of firms reporting a missed other scheduled payment such as rent, utilities, and payroll. % PPP Received is the percentage
of businesses reporting having received PPP funds in a state-by-industry group. State PPPE (Nbr.) is the state average of the PPPE based on the number of outstanding loans. All specifications
include industry×week fixed-effects. Standard errors are presented in parentheses, and are clustered at the level of the state. ***, **, and *, represent statistical significance at 1%, 5%, and 10%
levels, respectively.
Panel A: Missing Loan Payments
(1) (2) (3) (4) (5) (6) (7)
% Missing Loan Payments
%PPP Received -0.173∗∗∗ -0.068∗ -0.148∗∗∗ -0.278∗∗∗ -0.210∗∗∗ -0.349∗∗∗ -0.208∗∗
(0.041) (0.034) (0.048) (0.097) (0.075) (0.122) (0.091)
Observations 2229 277 386 390 389 393 394
Adjusted R2 -0.154 -0.021 -0.076 -0.370 -0.336 -0.665 -0.185
78 FStat 126.358 120.162 144.125 44.222 30.384 22.335 39.285
Industry Fixed Effects No Yes Yes Yes Yes Yes Yes
Industry×Week Fixed Effects Yes No No No No No No
Week: Full Sample Apr26–May2 May3–May9 May10–May16 May17–May23 May24–May30 May31–Jun6
Panel B: Missing Other Scheduled Payments (Rent, Utilities, Payroll)
(1) (2) (3) (4) (5) (6) (7)
% Missing Schd. Payments
%PPP Received -0.555∗∗∗ -0.282∗∗∗ -0.468∗∗∗ -0.880∗∗∗ -0.651∗∗∗ -0.998∗∗∗ -0.630∗∗∗
(0.110) (0.060) (0.092) (0.187) (0.226) (0.291) (0.218)
Observations 2216 277 385 387 388 389 390
Adjusted R2 -0.346 0.153 -0.067 -1.089 -0.687 -1.932 -0.403
FStat 131.427 120.162 144.126 44.423 29.598 25.182 39.759
Industry Fixed Effects No Yes Yes Yes Yes Yes Yes
Industry×Week Fixed Effects Yes No No No No No No
Week: Full Sample Apr26–May2 May3–May9 May10–May16 May17–May23 May24–May30 May31–Jun6
Table A.10: % Receiving PPP and Cash-on-Hand: Week-by-Week
Table A.10 reports the results of instrumental variables (IV) regressions split week-by-week examining the relation between the geographic allocation of PPP funds during the first round and the
percentage of businesses reporting at least two months or more of cash-on-hand at the state-by-industry level. The dependent variable is the difference between the percentage of businesses that
applied for PPP fundings, minus the percentage of businesses that received PPP funds. % PPP Received is the percentage of businesses reporting having received PPP funds in a state-by-industry
group. State PPPE (Nbr.) is the state average of the PPPE based on the number of outstanding loans. All specifications include industry×week fixed-effects. Standard errors are presented in
parentheses, and are clustered at the level of the state. ***, **, and *, represent statistical significance at 1%, 5%, and 10% levels, respectively.
79 (1) (2) (3) (4) (5) (6) (7)
% Cash 3 months
%PPP Received 0.283∗∗ 0.091 0.153 0.327∗ 0.285 0.619∗∗ 0.555
(0.121) (0.079) (0.105) (0.188) (0.310) (0.279) (0.348)
Observations 903 95 141 136 177 180 174
Adjusted R2 -0.247 -0.079 -0.166 -0.291 -0.280 -0.678 -0.480
FStat 144.087 50.909 84.021 49.197 16.342 32.496
Industry Fixed Effects No Yes Yes Yes Yes Yes Yes
Industry×Week Fixed Effects Yes No No No No No No
Week: Full Sample Apr26–May2 May3–May9 May10–May16 May17–May23 May24–May30 May31–Jun6
Table A.11: % Receiving PPP and Exposure to PPP: MSA Level
Table A.11 reports the results of ordinary least squares (OLS) regressions examining the relation between the geographic allocation of PPP funds during the first round and the percentage of
businesses reporting applying but not receiving PPP funds at the MSA level. The dependent variable is the difference between the percentage of businesses that applied for PPP funds and the
percentage of businesses that received PPP funds at the MSA level. MSA PPPE (Nbr.) is the MSA average of the PPPE based on the number of outstanding loans. Standard errors are presented in
parentheses, and are clustered at the level of the MSA. ***, **, and *, represent statistical significance at 1%, 5%, and 10% levels, respectively.
80
(1) (2) (3) (4) (5) (6) (7)
% PPP Received
MSA PPPE (Nbr.) 32.714∗∗∗ 68.793∗∗∗ 47.107∗∗∗ 19.004∗∗∗ 24.207∗∗∗ 16.938∗∗∗ 20.232∗∗∗
(6.579) (6.202) (4.106) (5.071) (4.977) (3.356) (4.363)
Observations 300 50 50 50 50 50 50
Adjusted R2 0.083 0.656 0.657 0.171 0.284 0.242 0.241
Week Fixed Effects Yes No No No No No No
Week: Full Sample Apr26–May2 May3–May9 May10–May16 May17–May23 May24–May30 May31–Jun6
Table A.12: % Receiving PPP and Missed Payments: MSA-Level
Table A.12 reports the results of ordinary least squares (OLS) and instrumental variables (IV) regressions examining the relation between the geographic allocation of PPP funds during the first
round and the percentage of firms reporting missing payments at the MSA level collected from the Census Small Business Pulse Survey. In the top panel, the dependent variable is the percentage
of firms at the MSA reporting a missed scheduled loan payment. In the bottom panel, the dependent variable is the percentage of firms reporting a missed other scheduled payment such as rent,
utilities, and payroll. % PPP Received is the percentage of businesses reporting having received PPP funds in a state-by-industry group. MSA PPPE (Nbr.) is the MSA average of the PPPE based on
the number of outstanding loans. All specifications include week fixed-effects. Robust standard errors are presented in parentheses. ***, **, and *, represent statistical significance at 1%, 5%, and
10% levels, respectively.
Panel A: Missing Loan Payments
(1) (2) (3) (4)
OLS Reduced Form IV
First-Stage Second-Stage
Dep. Variable: % Missing Loan Payments % Missing Loan Payments % PPP Received % Missing Loan Payments
% PPP Received -0.154∗∗∗ -0.181∗∗∗
(0.013) (0.042)
81 MSA PPPE (Nbr.) -5.902∗∗∗ 32.652∗∗∗
(1.583) (6.579)
Observations 298 298 298 298
Adjusted R2 0.357 0.039 0.083 0.347
FStat 24.635
Week Fixed Effects Yes Yes Yes Yes
Panel B: Missing Other Scheduled Payments (Rent, Utilities, Payroll)
(1) (2) (3) (4)
OLS Reduced Form IV
First-Stage Second-Stage
Dep. Variable: % Missing Schd. Payments % Missing Schd. Payments % PPP Received % Missing Schd. Payments
% PPP Received -0.235∗∗∗ -0.451∗∗∗
(0.019) (0.085)
MSA PPPE (Nbr.) -14.753∗∗∗ 32.714∗∗∗
(2.438) (6.579)
Observations 300 300 300 300
Adjusted R2 0.320 0.099 0.083 0.046
FStat 24.728
Week Fixed Effects Yes Yes Yes Yes
Table A.13: % Receiving PPP and Cash-on-Hand: MSA-Level
Table A.13 reports the results of ordinary least squares (OLS) and instrumental variables (IV) regressions examining the relation between the geographic allocation of PPP funds during the first
round and the percentage of businesses at the MSA level reporting at least two months of cash on hand to sustain operations as collected from the Census Small Business Pulse Survey. The
dependent variable is the fraction of businesses in the MSA reporting cash on hand to sustain operations for two months or more. % PPP Received is the percentage of businesses reporting having
received PPP funds at the MSA level. MSA PPPE (Nbr.) is the MSA average of the PPPE based on the number of outstanding loans. All specifications include week fixed-effects. Robust standard
errors are presented in parentheses. ***, **, and *, represent statistical significance at 1%, 5%, and 10% levels, respectively.
(1) (2) (3) (4)
OLS Reduced Form IV
82 First-Stage Second-Stage
Dep. Variable: % Cash 3 months % Cash 3 months % PPP Received % Cash 3 months
% PPP Received 0.172∗∗∗ -0.172∗
(0.018) (0.097)
MSA PPPE (Nbr.) -5.628∗∗ 32.714∗∗∗
(2.468) (6.579)
Observations 300 300 300 300
Adjusted R2 0.194 0.014 0.083 -0.595
FStat 24.728
Week Fixed Effects Yes Yes Yes Yes
Table A.14: Unmet Demand and Exposure to PPP: MSA Level
Table A.14 reports the results of ordinary least squares (OLS) regressions examining the relation between the geographic allocation of PPP funds during the first round and the percentage of
businesses reporting applying but not receiving PPP funds at the MSA level. The dependent variable is the difference between the percentage of businesses that applied for PPP funds and the
percentage of businesses that received PPP funds at the MSA level. MSA PPPE (Nbr.) is the MSA average of the PPPE based on the number of outstanding loans. Robust standard errors are
presented in parentheses. ***, **, and *, represent statistical significance at 1%, 5%, and 10% levels, respectively.
83
(1) (2) (3) (4) (5) (6) (7)
% PPP Requested - % PPP Received
MSA PPPE (Nbr.) -22.987∗∗∗ -60.564∗∗∗ -36.165∗∗∗ -13.531∗∗∗ -11.946∗∗∗ -9.384∗∗∗ -6.331∗∗∗
(6.526) (7.053) (3.150) (2.527) (2.570) (1.500) (1.845)
Observations 300 50 50 50 50 50 50
Adjusted R2 0.043 0.585 0.643 0.360 0.270 0.298 0.121
Week Fixed Effects Yes No No No No No No
Week: Full Sample Apr26–May2 May3–May9 May10–May16 May17–May23 May24–May30 May31–Jun6
Table A.15: Unmet Demand and Missed Payments:MSA Level
Table A.15 reports the results of ordinary least squares (OLS) and instrumental variables (IV) regressions examining the relation between the geographic allocation of PPP funds during the first
round and the percentage of businesses reporting missing scheduled payments at the MSA level. In the top panel, the dependent variable is the percentage of firms reporting a missed scheduled
loan payment. In the bottom panel, the dependent variable is the percentage of firms reporting a missed other scheduled payment such as rent, utilities, and payroll. % PPP Requested - % PPP
Received is the difference between the percentage of businesses that applied for PPP funds and the percentage of businesses that received PPP funds at the MSA level. MSA PPPE (Nbr.) is the
MSA average of the PPPE based on the number of outstanding loans. All specifications include week fixed-effects. Robust standard errors are presented in parentheses. ***, **, and *, represent
statistical significance at 1%, 5%, and 10% levels, respectively.
Panel A: Missing Loan Payments
(1) (2) (3) (4)
OLS Reduced Form IV
First-Stage Second-Stage
Dep. Variable: % Missing Loan Payments % Missing Loan Payments % PPP Request - % PPP Received % Missing Loan Payments
% PPP Requested - % PPP Received 0.166∗∗∗ 0.258∗∗∗
(0.013) (0.067)
84 MSA PPPE (Nbr.) -5.902∗∗∗ -22.911∗∗∗
(1.583) (6.526)
Observations 298 298 298 298
Adjusted R2 0.380 0.039 0.043 0.265
FStat 12.325
Week Fixed Effects Yes Yes Yes Yes
Panel B: Missing Other Scheduled Payments (Rent, Utilities, Payroll)
(1) (2) (3) (4)
OLS Reduced Form IV
First-Stage Second-Stage
Dep. Variable: % Missing Schd. Payments % Missing Schd. Payments % PPP Request - % PPP Received % Missing Schd. Payments
% PPP Requested - % PPP Received 0.258∗∗∗ 0.642∗∗∗
(0.020) (0.155)
MSA PPPE (Nbr.) -14.753∗∗∗ -22.987∗∗∗
(2.438) (6.526)
Observations 300 300 300 300
Adjusted R2 0.355 0.099 0.043 -0.434
FStat 12.406
Week Fixed Effects Yes Yes Yes Yes
Table A.16: Unmet Demand and Cash-on-Hand: MSA Level
Table A.16 reports the results of ordinary least squares (OLS) and instrumental variables (IV) regressions examining the relation between the geographic allocation of PPP funds during the first
round and the percentage of businesses reporting at least two months or more of cash-on-hand at the MSA level. The dependent variable is the fraction of businesses with cash on hand to sustain
operations for two months or more. % PPP Requested - % PPP Received is the difference between the percentage of businesses that applied for PPP funds and the percentage of businesses that
received PPP funds at the MSA level. All specifications include week fixed-effects. Robust standard errors are presented in parentheses. ***, **, and *, represent statistical significance at 1%, 5%,
and 10% levels, respectively.
(1) (2) (3) (4)
OLS Reduced Form IV
85 First-Stage Second-Stage
Dep. Variable: % Cash 3 months % Cash 3 months % PPP Request - % PPP Received % Cash 3 months
% PPP Requested - % PPP Received -0.203∗∗∗ 0.245
(0.019) (0.156)
MSA PPPE (Nbr.) -5.628∗∗ -22.987∗∗∗
(2.468) (6.526)
Observations 300 300 300 300
Adjusted R2 0.247 0.014 0.043 -0.972
FStat 12.406
Week Fixed Effects Yes Yes Yes Yes
File and source
- File
- zwick-sba-v2-ppp-hit-target.pdf
- Size
- 2,988,420 bytes
- SHA-256
- a7dd7da0b2ada269cded15508b8fa67214a750d5d9f6730364fea46da2bcc143
- Our copy
- zwick-sba-v2-ppp-hit-target.pdf
- Original
- www.ericzwick.com