Lessons Learned From The US Economic Policy Response To Covid-10

RECESSION
REMEDIES
Lessons Learned from the
U.S. Economic Policy Response to
COVID-19
Edited by
Wendy Edelberg, Louise Sheiner,
and David Wessel
Recession Remedies
Lessons Learned from the
U.S. Economic Policy Response
to COVID-19
Edited by
Wendy Edelberg, Louise Sheiner, and
David Wessel
Chapter 4
Lessons Learned from Support to

Business during COVID-19
Gabriel Chodorow-Reich, Ben Iverson, and
Adi Sunderam1
Introduction
The United States responded to the recession caused by the COVID-19 pandemic
with massive and unprecedented support for businesses. New federal business
subsidies during the first year of the pandemic, 2020Q2–2021Q1, including the
Paycheck Protection Program (PPP), Economic Injury Disaster Loan (EIDL)
Advances, and targeted aid for sectors such as airlines and restaurants, totaled
$600 billion, or about 2.7 percent of potential GDP, while expanded EIDL Loans
added an additional $200 billion of support. The Federal Reserve authorized
purchases of up to $750 billion in corporate bonds through the newly created
Corporate Credit Facilities (CCFs) and up to $600 billion in long-term, low
interest rate loans to midsize corporations through the new Main Street Lend-
ing Program (MSLP).
At the same time, the business sector overall fared much better during
the COVID-19 recession and recovery than had been expected at the outset.
Indeed, this resilience was different from previous downturns. Business bank-
ruptcy filings declined during a recession year for the first time since 1980 and
remained below their pre-pandemic level into 2021. After peaking in April
2020, the unemployment rate fell faster than in any other post–World War II
recovery period, and job vacancies in 2021 reached their highest level on record.
We critically evaluate the business aid programs and their role in cush-
ioning the downturn and spurring the economic recovery. We do so especially
with an eye toward future non-pandemic-related downturns, during which
1. The authors are grateful to Eric Milstein and Madeline Kitch for providing excellent research
assistance. The authors thank Beverly Hirtle, Owen Zidar, participants in the October authors’
conference, and the editors of this volume for their insightful feedback.
123
124 | Recession Remedies
policymakers may be tempted to return to these programs in the hope of

achieving a similarly rapid recovery. However, the variety of other policy sup-
port enacted during the COVID-19 recession and reviewed elsewhere in this
book—as well as the unusual course of a lockdown-driven recession—pose
serious confounders to immediately concluding a causal link between the
business aid programs and the economic trajectory. Our task will be to evaluate
the role played by the business programs specifically and to highlight where
uncertainties remain.
Our evaluation starts by setting out a framework for assessing business aid.
If financial markets functioned frictionlessly and there were no externalities,
there would be no rationale for government intervention on efficiency grounds.
We identify two plausible deviations from this benchmark: (a) market failures
that prevent long-run solvent firms from obtaining temporary liquidity and
(b) externalities from worker layoffs or firm failure. Accordingly, business
support should focus on alleviating financial frictions or avoiding labor market
congestion, bankruptcy court congestion, and aggregate demand externalities
that result when firms contract. We then review the impact of policies enacted
during the pandemic period and reach the following conclusions.
First, policies to support small businesses likely could have achieved their
objectives with much smaller budgetary cost by focusing on smaller firms and
featuring a smaller subsidy component. The PPP made 5.1 million potentially
forgivable loans between April and August 2020 with a total face value of
$522 billion. More than 50 percent of these loans were under $25,000 and 80
percent were less than $100,000, yet loans greater than $500,000 that went to
larger recipients account for half the budgetary cost. We survey the academic
literature evaluating PPP and find no credible evidence that the largest PPP
loans had a substantial positive employment effect in the short or medium
run. The evidence for the efficacy of loans to the smallest firms is more mixed.
The closely related EIDL program, which gave nonforgivable, long-term
loans to small businesses, also had extraordinarily high take-up, with 3.6 mil-
lion loans totaling $194 billion through November 2020 and an additional $124
billion over the following year. Relative to PPP, these loans have the benefit of
providing immediate liquidity but at much lower cost to taxpayers. In addition,
EIDL loans were potentially better targeted, as only businesses with an expec-
tation of long-term viability could apply. However, lending to already indebted
firms may leave them overleveraged, creating debt-overhang problems that
impede the recovery. Open questions for the small business support policies
include their long-term impact on firm survival and employment and whether
loans or grants are better tools from a cost-benefit perspective.
Second, the academic literature has largely neglected many of the other
business subsidy programs. Two of the largest were the Employee Retention
Credit and grants to air carriers. While both had features designed to link dis-
bursements to payroll, the fungibility of funds raises the possibility that they
may instead have benefited shareholders. Such concerns may be particularly
Support to Business | 125
significant for the grants to air carriers, which mostly went to large, publicly
traded firms, many of which had previously undergone successful bankruptcy
restructuring, albeit not all simultaneously.
Third, Federal Reserve (Fed) interventions into the corporate bond market
clearly can play a stabilizing role. Indeed, despite the fact that the CCFs used
only approximately $15 billion of their $750 billion capacity, both informal
event study analysis and more rigorous academic studies find that they sig-
nificantly lowered bond yields in the spring of 2020. The key open question
is whether doing so is desirable. In the COVID-19 crisis, large benefits were
obtained even with low take-up, but those outcomes were in part due to the
rapid macroeconomic recovery. Had the pandemic more strongly affected the
economy in late 2020 and early 2021, the costs of intervention may have been
significantly higher.
Fourth, the Fed’s direct support for bank lending had little direct impact.
A key design feature of the MSLP was that banks offloaded 95 percent of each
loan to the Fed but retained a 5 percent slice, meaning that banks would only
make loans that offered similar returns as the rest of their balance sheet. If
banks had been balance-sheet constrained as they were during the 2007–09
recession, such a policy could have proven very useful. As it turned out, banks
remained in relatively good health, and only $18 billion of the $600 billion
facility was used.
Finally, given that our reading of the literature suggests that one should
be skeptical of a crucial role for much of the business aid in supporting the
recovery, we review other explanations for the performance of the business
sector. Using Compustat financials data, we show that large firms initially
reacted by raising substantial external financing from private markets. These
firms raised debt by drawing down existing credit lines and increasing bond
issuance and conserved equity largely by pausing share repurchase programs.
This increase in financing allowed these firms to withstand the initial decline
in net income. We then show that sales recovered much faster during the pan-
demic than during the 2007–09 downturn. Since our Compustat data covers
only public firms, it is possible that small- and medium-sized private firms
reacted quite differently to the pandemic. Further research is needed to shed
light on the behavior of such firms.
We end by articulating four main lessons for the prospects of business aid
programs to support employment and business survival in a non-pandem-
ic-related recession. First, policymakers should not blindly redeploy the 2020
tool kit despite the positive trajectory of the current recovery, as other factors,
including the nature of recovery from a temporary lockdown and general sup-
port for households, likely played a more important role. Second, if necessary,
support for small businesses could likely achieve a similar objective with much
smaller budgetary cost than PPP by focusing on smaller firms and providing
a smaller subsidy component. Third, the fungibility of funds given to large
firms, such as publicly traded airlines, and the history of successful bankruptcy
resolution for these firms suggest caution in the granting of such aid in the
future. Finally, while the Fed clearly has the ability to intervene successfully
in corporate credit markets, the question of whether it should do so involves
careful consideration of the reason for a decline in bond prices. In addition,
while not a significant element of the COVID-19 response, a policy such as the
MSLP could prove useful in a future recession when banks are constrained.
Background on Business Performance

The economic recovery that began in the summer of 2020 was much faster than
expected at the time or than historical experience would have predicted. To set
the stage for our subsequent analysis, in this section we put the macroeconomic
and business sector performance into context.
Macroeconomic Context
Figure 4.1 shows the paths of actual GDP (left panel) and the unemployment rate
(right panel) against the May 2020 median forecast in the Survey of Professional
Forecasters and the July 2020 forecast of the Congressional Budget Office. Despite
making their forecasts after the CARES Act had passed, both sets of forecasters
proved far too pessimistic about the depth of the downturn and the speed of
the recovery. Mostly notably, the rebound in 2020Q3 far exceeded expectations.
Figure 4.2 shows the historically rapid nature of the recovery, focusing on
the labor market. The top panel replicates and extends the finding of Hall and
Kudlyak (2021) that the unemployment rate has historically fallen by roughly
0.1 log point per year during recoveries and expansions. Against this backdrop,
the more-than halving of the unemployment rate from the high of almost 15
percent in April 2020 to about 6 percent in April 2021 is unprecedented. The
bottom panel plots total job vacancies, perhaps the best high-frequency mea-
sure of business demand. After falling sharply during the lockdown period,
vacancies rebounded and reached a series high by early 2021 before skyrocketing
during the summer and fall.
Business Bankruptcies
Along with the overall better-than-expected macroeconomic performance,
business survival fared much better than feared at the recession’s onset. We will
focus on business bankruptcy rates as a proxy for the health of businesses gen-
erally. Historically, business bankruptcy rates have been highly correlated with
economic conditions: in quarterly data from 1980–2019, a 1 percentage point rise
in the U.S. unemployment rate coincides with an increase of about 600 business
bankruptcies filings in the same quarter. The relationship between unemploy-
ment and bankruptcies was especially strong during the global financial crisis
Figure 4.1
Actual and Projected Macroeconomic

Trajectories, 2020Q1–2021Q2
A. Actual and Projected GDP, B. Actual and Projected
Indexed to 2020Q1 Unemployment Rate
104 16
Actual
GDP indexed to 2020Q1
100 CBO 12
(July 2020)
Percent
96 8
SPF
(May 2020)
92 4
88 0
Q1 Q2 Q3 Q4 Q1 Q2 Q1 Q2 Q3 Q4 Q1 Q2
2020 2021 2020 2021
Source: Bureau of Economic Analysis n.d.; Bureau of Labor Statis-

tics n.d.; Federal Reserve Bank of Philadelphia 2020; Congressional
Budget Office (CBO) 2020.
Note: GDP amounts are indexed to 100 in 2020Q1, such that subse-
quent observations can be interpreted as percent changes from this
initial level. SPF May 2020 is the release date of the second quarter Survey of Professional
Forecasters (SPF) report by the Federal Reserve Bank of Philadelphia.
of 2008, as can be seen in Figure 4.3, which plots the unemployment rate and
bankruptcy filings over time.
Given this context, the sharp increases in unemployment in March and
April 2020 were cause for concern. If historical relationships had held, the
10-percentage-point increase in the unemployment rate would have led to the
prediction of an additional 6,000 business bankruptcies in the second quarter
of 2020 alone, doubling the 5,952 business bankruptcies in 2020Q1.
These fears did not materialize. Instead, bankruptcies fell with the onset of
the COVID-19 pandemic. As shown in Wang et al. (2021), business bankruptcies
fell 17 percent in 2020 relative to 2019, and filing rates in 2021 were similar to
those in 2020. The decline in bankruptcy filings is striking given that there had
not been a decline in bankruptcies during a recession since official bankruptcy
statistics began being collected in 1980. Further, bankruptcy rates were already
quite low in 2019, making a further decline unlikely ex ante.
Figure 4.2
Unemployment Rate and Job Vacancies

A. Unemployment Rate, by Business Cycle, 1948–2021
15
12
Percent (log scale)
1948 1960 1970 1980 1990 2000 2010 2021
B. Total Vacancies, 2000–2021

12
10
8
Millions
2
2000 2004 2008 2012 2016 2021
Source: Bureau of Labor Statistics n.d.; authors’ calculations.

Note: In Figure 4.2a, slope lines have been added to show the
average rate of decline in the unemployment rate following each
recession. Steeper lines indicate more rapid recoveries. Recessions
are demarcated by grey bars.
The timing and breakdown of business bankruptcies can give some indi-
cation of what precipitated the overall decline. Figure 4.4, provided by Wang
et al. (2021), shows how weekly bankruptcy filing rates evolved for small and
large businesses throughout 2020 relative to 2019. Small businesses, defined as
those with less than $10 million in assets, saw filing rates fall dramatically at the
Figure 4.3
Business Bankruptcies and Unemployment,

1980Q2–2021Q2
40,000
13
30,000
11
Unemployment
Number of bankruptcies
rate
Unemployment rate
Business 9
20,000 bankruptcies
10,000
5
0 3
1980 1990 2000 2010 2020
Source: Bureau of Labor Statistics n.d.; U.S. Courts Bankruptcy

Filings Statistics n.d.
Note: Recessions are demarcated by grey bars.
immediate onset of the pandemic, well before government support programs

were put in place to support these businesses.2 As the year wore on, small
business filings rebounded somewhat from the initial drop but still stabilized
around 20 percent lower than 2019 levels. Meanwhile, large business filings saw
a short-lived decline in late March 2020 but for the most part remained close
to 2019 levels throughout 2020.
In this section, we have focused on business bankruptcy rates because
official statistics on overall business failure have not yet been released.3 Given
that the smallest businesses in the economy are unlikely to use bankruptcy
(Greenwood, Iverson, and Thesmar 2020), it is possible that business exit rates
increased even while bankruptcy rates declined. Crane et al. (2021) leverage
2. This decline was not due to physical court closures, as Wang et al. (2021) show that filings
declined at the same rate in bankruptcy districts where courts were never closed.
3. The U.S. Census Bureau’s (n.d.) Business Dynamics Statistics provide measures of firm
startups and shutdowns, but the most recent release as of this writing is for 2019.
Figure 4.4
Year-Over-Year Change in Business Bankruptcy

Filings, by Week, 2020
A. Small Businesses
2
Change in log weekly bankruptcies
National EIP 1
emergency
1
-1
-2
Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sep. Oct. Nov. Dec.
B. Large Businesses
2
Change in log weekly bankruptcies
National EIP 1
emergency
1
-1
-2
Jan. Feb. Mar. Apr. May. Jun. Jul. Aug. Sep. Oct. Nov. Dec.
Source: U.S. Courts Bankruptcy Filings Statistics n.d.

Note: Small businesses are those with less than $10 million in assets
and large businesses are those with more than $10 million in assets.
95% confidence intervals shown in dotted gray lines.
alternative indicators of business exit (e.g., paycheck issuance and phone-track-

ing data) to estimate business exit rates in the first year of the pandemic. Using
these sources, they estimate that the business exit rate was about 25 percent
higher than baseline in the first year of the pandemic, but they note that these
alternative data sources have limitations that could lead to overstating or under-
stating the true exit rate. A key difficulty, which was particularly exacerbated
by the COVID-19 pandemic, is determining whether business closures are
temporary or permanent. Given these difficulties, it will likely be necessary to

wait for administrative data to fully understand the pattern of business closures.
In summary, both the overall macroeconomy and business survival, specif-
ically, fared much better during the pandemic than initially feared or historical
experience would have predicted. Against this backdrop, we next evaluate
the role of direct government aid to businesses. However, it is important to
recognize that these programs came on top of several other policies and fac-
tors specific to COVID-19 that likely aided the rapid recovery and interacted
in important ways with business aid. On the policy front, fiscal support to
households played an important role in supporting consumer demand and thus
indirectly helping businesses. This support included three separate rounds of
direct payments to households in April 2020, January 2021, and March 2021,
totaling over $850 billion, and extended and enhanced Unemployment Insur-
ance (UI). Enhanced generosity of UI in particular makes it less important to
incentivize businesses to maintain employment (as the PPP and other programs
did) since these workers are supported in other ways, especially during a period
such as summer 2020 when public health conditions warranted having many
people remain at home anyway. Bolstered by government support, total house-
hold income rose in 2020 despite the recession, and households increased their
liquid assets, especially lower income households.4 The increase in household
income and wealth created the conditions for the sharp rebound in consumer
demand as the economy exited the recession, with additional fuel coming from
widespread availability of COVID-19 vaccines in early 2021.
Similarly, the evolution of business practices and the course of the pan-
demic itself played an important part in determining business conditions.
Widespread use of videoconferencing technologies allowed many to work
from home. Widespread testing protocols allowed some workers to return
to work. And the relatively quick development of vaccines meant that many
businesses were able to partially or fully reopen sooner than might have been
anticipated. All these factors together created a quick economic rebound in late
2020 and early 2021, which meant that many businesses only faced short-term
cash flow shortfalls rather than fundamental insolvency. The totality of these
circumstances make it important to try to isolate the role for and effectiveness
of direct government support for businesses.
Framework for Evaluation

Given that the COVID-19 pandemic and associated recession were quite
unusual, we cannot simply use outcomes to assess the success of business
support programs and the suitability of such programs for future recessions.
4. The JPMorgan Chase Institute (2022) found that median cash balances were 65 percent
higher than 2019 levels at the end of 2021 among low-income families. Cash balances for
high-income families were about 35 percent higher at the end of 2021.
We instead start by highlighting conditions under which governments should

provide support to businesses during a recession. We then assess business
support programs in part by asking how well they address the rationales we
highlight. Following Hanson et al. (2020), we suggest two main rationales,
focusing on efficiency concerns: (a) market failures that prevent long-run solvent
firms from obtaining temporary liquidity and (b) externalities from worker
layoffs or firm failure.
If financial markets functioned frictionlessly and there were no externali-
ties, there would be no rationale for government intervention. In this case, firms
that were solvent in the long run could simply raise capital by issuing equity
or borrowing against their future cash flows from banks or financial markets.
The availability of private financing would allow firms to weather temporary
revenue shocks, like the COVID-19 pandemic and its accompanying public
health interventions. For instance, consider a restaurant that faces temporar-
ily low cash flows due to the pandemic but will ultimately be viable (i.e., have
post-pandemic profits that exceed the costs of surviving the pandemic). If
financial markets functioned perfectly, the restaurant would be able to borrow
enough to survive. This argument applies even in the face of the extreme mac-
roeconomic uncertainty created by the pandemic. In uncertain environments,
firms retain option value by deferring the decision to shut down until there is
more clarity on the path of the economy. In frictionless markets, lenders and
investors recognize that this option is valuable and are willing to contribute
funding immediately in exchange for the possibility of a future payoff.
Thus, deviations from this frictionless benchmark are necessary for gov-
ernment interventions to be warranted on efficiency grounds. The first rationale
for intervention we consider arises because credit markets may not function
well enough to enable firms with viable long-run business prospects to raise
enough financing to meet temporary liquidity needs. The lockdown conditions
that prevailed in the spring of 2020 and caused revenue at many firms to fall
precipitously provide an example par excellence of when fixed costs such as rent
or debt obligations could cause firms to fail if they cannot arrange temporary
financing, but such circumstances arise in all recessions. Again, in a first-
best world with perfect credit markets, full enforcement of contracts, and no
asymmetric information, long-run solvent firms could obtain such financing
from private sources and government intervention would not be necessary.
However, these conditions may fail in a variety of ways, particularly in a
crisis. For instance, lending may become constrained because banks take losses
on their existing loans at the onset of a crisis, reducing their capital buffers
and creating debt overhang. In early 2020, there was significant concern that
bank capital buffers would be rapidly depleted during the COVID-19 pandemic
(Feldman and Schmidt 2021). Alternatively, the nature of firm cash flows may
change in a way that makes it difficult for banks to continue lending. For
instance, it may become more difficult for lenders to discriminate between
long-run solvent and insolvent borrowers, causing them to exit credit markets
completely. In this case, even solvent firms may not be able to borrow. A third
potential financial friction involves changes in the nature of cash flows that
make it difficult for solvent firms to fully pledge future cash flows to lenders.
For instance, banks may have an advantage in holding low-risk assets i.e.,
in making relatively safe loans (Diamond 2020). If an economic downturn
increases uncertainty about future cash flows, as the COVID-19 pandemic
did, new loans will be riskier, even if they are made to firms that will be viable
in the long run, on average. Banks with a preference for relatively safe lending
may not be well-suited to provide such incremental financing to firms. Firms
that have access to financing outside of banks could then turn to other capital
providers, but finding new financing is costly for all firms and may be impos-
sible for many small and medium firms (Fazzari, Hubbard, and Peterson 1988).
Finally, credit markets may suffer from fire sales (Shleifer and Vishny 1992;
Stein 2012) or market freezes (Diamond and Rajan 2011), which can impede
the ability of healthy firms to raise financing. In the presence of such frictions,
government interventions may be helpful. These interventions can take the form
of direct assistance, supplements to bank financing, or central bank policies,
such as asset purchases, that help to ensure well-functioning financial markets.
The second rationale for government intervention involves negative exter-
nalities from firm shrinkage or exit. The idea is that there are benefits to keeping
firms alive that accrue to neither the firms themselves nor their lenders. In
such cases, government intervention can be valuable even if financial markets
function well. For instance, if too many firms simultaneously seek bankruptcy
protection, the resulting congestion in bankruptcy courts can lead to inefficient
liquidations (Iverson 2018; Greenwood, Iverson, and Thesmar 2020). Existing
research suggests that the deadweight loss from such congestion can be large.
For instance, Iverson (2018) found that a 6 percent increase in bankruptcy
caseloads increases the loss given default on commercial and industrial bank
loans by 3.9 percentage points (relative to a mean loss given default of 36 per-
cent). In a typical recession, caseloads rise 25 to 50 percent, suggesting scope
for significant losses from congestion.
Labor market congestion is a second type of externality that can justify
government intervention. If too many laid-off workers simultaneously search
for new jobs, they can impede the employer–employee matching process,
resulting in fewer hires and lower quality matches (Blank and Maghzian 2021).
More broadly, such separations risk destroying firm-specific human capital,
slowing down the eventual recovery. The widespread use during the pandemic
of temporary layoffs, in which workers expect to be recalled to their previous
employer, mitigates such concerns but may not eliminate them.
A third type of externality occurs when lower consumption by laid-off
workers contributes to lower aggregate demand, leading output to fall further
(Chodorow-Reich and Karabarbounis 2016; Farhi and Werning 2016). Concerns
about aggregate demand externalities loom particularly large when interest
rates are stuck at the zero lower bound. While other policies—notably, generous
UI—can alternatively target the decline in consumption by laid-off workers,

such considerations nonetheless strengthen the rationale for employment
subsidy policies that also have this effect.
Social insurance (i.e., subsidies to business that rise in bad times) for busi-
ness owners is a third rationale for government intervention that is sometimes
proposed. There may be social benefits to encouraging entrepreneurship, and
since entrepreneurs bear a large amount of uninsurable, undiversifiable risk,
supporting small businesses could be valuable. This is particularly true given
that small-business owners are typically not eligible for other forms of social
insurance, like UI. Moreover, to the extent a pandemic-type shock was com-
pletely unforeseen, ex post transfers to business owners could correct for the
absence of pandemic insurance ex ante (Romer and Romer forthcoming). On
the other hand, as pointed out by Hanson, Sunderam, and Zwick (2021), busi-
ness owners are on average relatively wealthy, so the social insurance benefits
of supporting them are likely small.
While these rationales provide a case for supporting businesses in a generic
recession, it is worth noting that they may have provided an especially strong
case in the recession caused by the COVID-19 pandemic. Three features of the
pandemic-related recession made it different from most others. First, during
the early months of the COVID-19 pandemic, the correlation between firms’
short-run cash flows and their longer-run solvency was likely much weaker
than in a typical recession. The pandemic and associated public health inter-
ventions caused precipitous revenue declines for many fundamentally healthy
firms. Against this backdrop, the risk that government support would prop
up insolvent firms through so-called zombie lending was weaker than usual.
Second, the turmoil in bond markets in March 2020, while not completely
unprecedented, was significantly more severe than market dislocations in a
typical recession. In other words, financing frictions in bond markets were
larger than usual, again strengthening the case for government intervention.
Third, macroeconomic uncertainty was significantly higher than normal
in the COVID-19 recession (Altig et al. 2020). This both exacerbated standard
financial frictions and increased the option value inherent in keeping firms
alive, relative to typical recessions.
The rationales outlined above also have implications for the types of inter-
ventions that are likely to be most effective. For instance, if financial market
frictions are the rationale for intervention, it may be beneficial to target the firms
and sectors most affected by such frictions. Because small firms typically face
greater financial constraints than larger firms (Fazzari, Hubbard, and Peterson
1988; Zwick and Mahon 2016) and have access to fewer sources of financing, the
case for targeting government support toward small firms may be stronger than
the case for unconditional support. Similarly, for firms that depend on particular
banks for financing, these relationships make it difficult to seek funds from other
sources (Rajan 1992; Darmouni 2020). Thus, steps to encourage bank lending may
be particularly impactful. In contrast, large firms typically have many sources
of financing, including public debt and equity markets, and multiple banks with
which they maintain relationships. These characteristics suggest that the gains
from government support of large firms may be relatively smaller.
It is also worth noting that while the types of externalities discussed pro-
vide rationales for government intervention, it is not clear whether they justify
direct aid to businesses specifically. For instance, aid to businesses may reduce
the congestion of bankruptcy courts in an unexpected recession, but outside
of crisis times simply hiring more bankruptcy judges is a more direct policy
intervention. Similarly, aid to businesses may prevent them from firing workers
and reduce labor market congestion. However, job retention subsidies may be
a better-targeted policy response to the problem.
Finally, the stated purpose of a policy may not equate to its ultimate effect,
because money is fungible. Policies requiring that aid be used to support payroll
provide a leading example. If the recipient would have met the required payroll
target even absent the aid, then the policy has in effect provided unrestricted
support to the owners of the business. Evaluating specific programs therefore
requires determining how the funds were actually used.
Summary of Major Programs

Table 4.1 lists the major business aid programs, the amount authorized, the
amount utilized during the mostly pre-vaccine year stretching from 2020Q2
to 2021Q1, and the amount in 2021Q2–2021Q4. Several of these programs were
administered by the Small Business Administration (SBA).5 The largest single
program measured by dollars utilized was the PPP, whose size exceeds all the
other federal subsidy programs combined. Other programs administered by
the SBA include Economic Injury Disaster Loans (EIDL), EIDL advances, and
SBA loan forbearance. These non-PPP SBA programs provided in aggregate
$344 billion in liquidity to small businesses. The new Federal Reserve programs
had even larger authorizations but much lower utilization. Moreover, these
programs involved asset purchases, making the subsidy amount far smaller
than the authorized purchases. Finally, many state and local governments
enacted business support policies.
We now discuss each of these programs in greater detail, with emphasis on
evaluation of their effectiveness and the lessons learned for future downturns.
5. Our focus is on programs aimed at general business survival that were active during 2020. In
addition to the programs listed in Table 4.1, businesses also received subsidies through the
Provider Relief fund ($64 billion allocated thus far) and tax credits to support paid sick leave
($113 billion). In 2021, restaurants received support through the Restaurant Revitalization
Fund ($28 billion.)
Table 4.1
Distribution of Major Business Aid Programs,

Billions of Dollars Authorized and Utilized
Utilized
Authorized 2020Q2–2021Q1 2021Q2–2021Q4
Federal government subsidies 604 252
SBA programs
Paycheck Protection Program 814 457 180
Economic Injury Disaster Loan advances 35 20 7
SBA forbearance 7 7
Other programs
Employee Retention Tax Credit 71 47
Grants to air carriers 58 29 12
Food Assistance Program 30 21 6
Federal government loans 941 169
Paycheck Protection Program 814 735 58
EIDL loans 206 111
Federal Reserve programs 1,350 33
Corporate Credit Facility 750 15
Main Street Lending Facility 600 18
State and local programs 15
Source: Bureau of Economic Analysis n.d.; Federal Reserve n.d.;

Small Business Administration n.d.; authors' calculations.
Note: Authorized refers to cumulative authorizations across bills and
is blank for mandatory spending. Dollar values in 2020Q2–2021Q1
and 2021Q2–2021Q4 refer to the amount of business subsidies,
loans, or purchases actually made. Federal government loans include loans to all recipients
including those not in the business sector.
SBA Programs
Paycheck Protection Program
The PPP was the largest and most visible of the federal subsidy programs.
Initially enacted at the end of March 2020 under the CARES Act with an
authorization of $350 billion, the program was extended and modified several
times and eventually made nearly 12 million loans totaling $800 billion before
expiring at the end of May 2021. The first round of PPP funding lasted from
April to August 2020 and offered term loans of an amount equal to 2.5 times
average monthly payroll with a cap of $10 million. Firms were eligible if they
had fewer than 500 employees or operated in the Accommodation and Food
Services Sector with fewer than 500 employees per location. The Coronavirus
Response and Relief Supplemental Appropriations Act of 2021, signed at the
end of December 2020, replenished the funding for new PPP loans. It also
allowed firms with fewer than 300 employees and at least a 25 percent reduction
in gross receipts between comparable quarters in 2019 and 2020 to receive a
second PPP loan, again based on 2.5 times monthly payroll but with a cap of
$2 million. The first and second loans were forgivable if the borrower main-
tained employee and compensation levels for a specified 8- to 24-week period
following the disbursement and used at least 60 percent of the proceeds on
payroll costs. As of December 2021, 80 percent of the total PPP loan amount,
or $634 billion, had been forgiven.
In terms of the rationales articulated for government intervention in the
Framework for Evaluation section above, PPP can be thought of as serving
two purposes. First, the loan aspect of the program may be thought of as an
attempt to overcome financial frictions for small firms by directly supplying
them with funds. Second, the grant aspect of the program can be thought
of as an attempt to reduce labor market congestion or to generate aggregate
demand externalities more broadly. We now review evidence that suggests
that to the extent the program achieved these goals at all, it could have done
so on a far smaller scale.
We begin our analysis of the PPP by highlighting the sharp disparities in
the dollar amount allocated to smaller and larger firms. Figure 4.5 shows the
number and dollar value of loans by loan size for the first PPP round (covering the
period April–August 2020) using data from SBA on the universe of PPP loans.6
Because of the statutory link between loan amount and payroll, the distribution
of loan sizes closely approximates the distribution of firm sizes of loan recipients.
While half of the loans were under $25,000, in total these loans account for only
6 percent of the dollar cost. At the other extreme, just 1.6 percent of the loans
exceeded $1 million, but these loans account for one-third of the dollar cost.
The academic literature has taken several approaches to evaluating the
PPP. Perhaps the simplest is to ask how recipients adjust their balance sheets
after receiving the funds. Using administrative bank supervisory data on firms
with credit line commitments of at least $1 million matched to their PPP loan,
Chodorow-Reich et al. (forthcoming) found that by the end of June 2020 these
firms had reduced their non-PPP borrowing from banks by $0.95 for every $1 of
PPP funds. While not a causal estimate of the use of PPP funds, this adjustment
suggests that for these larger PPP recipients (i.e., the mean PPP loan in their
data is about $1 million) the PPP loan might have partially or mostly replaced
private financing.
6. Firms that received their first PPP loan in the tranche starting in January 2021 skewed much
smaller than in the initial allocation, with 96 percent of the loans and 72 percent of the dollars
in loans of less than $25,000. The distribution of second PPP loans was much closer to the
initial tranche.
Figure 4.5
Receipt of Paycheck Protection Program, by

Loan Size
3.0 Number of 100
Dollars
PPP recipients
loaned
2.5
Number of recipients (millions)
Dollars
forgiven
Dollar amount (billions)

2.0
1.5 50
1.0
0.5
0.0 0
$24,999 $25– $100– $250– $500– $1–2.49 $2.5–10
or less 99k 249k 499k 999k million million
Loan amount
Source: Small Business Administration n.d.a.

Note: Left axis and purple bars represent the number of loans given,
while the right axis and dark and light green bars represent the
amount of dollars loaned and forgiven.
A second approach uses the 500-employee threshold as a natural experi-

ment that separates eligible firms just below the threshold from ineligible firms
just above it. Autor et al. (2020), Chetty et al. (2020), and Hubbard and Strain
(2020) all pursue this methodology, with Autor et al. and Chetty et al. finding
that eligible firms increased their relative employment by 2 to 3 percent in the
summer of 2020 and Hubbard and Strain finding no effect in the neighbor-
hood of the cutoff. Even the upper bound of these effects is modest relative to
the size of the program, consistent with the evidence from Chodorow-Reich
et al. (forthcoming) that larger recipients may have used a large portion of the
funds to pay down other debt.
A third approach exploits the haphazard nature of the initial rollout period,
when demand for PPP loans exceeded the CARES Act appropriation. Specifically,
during the first weeks of the program, banks prioritized existing customers in
processing PPP applications, and some banks had more efficient PPP operations
than others. The CARES Act appropriation ran out on April 16, freezing new
loan activity until Congress appropriated an additional $310 billion on April 24

and lending resumed on April 27. These delays create an opportunity to com-
pare firms that received their PPP loans earlier and later. Relative to the cutoff
approach, this research design can encompass smaller recipients but only in the
weeks and months immediately following the program’s rollout.
Studies of early versus late recipients produced mixed results. Doniger and
Kay (2021) found sizeable employment effects in areas with more loans pro-
cessed just before the initial CARES Act allotment ran out, especially in smaller
firms. Granja et al. (2020) applied a similar approach to firm and local area
outcomes and found much smaller immediate employment effects, a difference
they attribute to the variation in lending before the replenishment not being
fully random. Faulkender, Jackman, and Miran (2020) attempted to resolve
the nonrandom distribution by using county-level variation in the density
of community banks, which processed loans relatively efficiently, and found
large effects that they interpret as local to the small firms most likely to borrow
from a community bank. However, their main results also display “pre-trends”
wherein counties with higher community bank density had smaller increases
in UI claims even before the PPP went into effect, highlighting the difficulty
of obtaining causal estimates. Bartlett and Morse (2020) compared businesses
in Oakland, CA, that applied for and received or did not receive a PPP loan as
of the beginning of June and found that recipients had a self-reported 20 per-
centage point higher subjective probability of survival if lockdown conditions
persisted for an additional six months, but this effect disappears for firms with
more than 20 employees. Using data from a nationwide survey of small firms,
Bartik et al. (2021) found similar effect sizes for small firms. In addition, they
found that effects of receiving a loan were similar across small firms, suggesting
that the choice to distribute PPP through banks which favored certain clients
did not substantially reduce the program’s overall impact. Moreover, banks
were likely better able to distribute the funds quickly than a program directly
administered by the government. This may have raised the overall impact of
the program by allowing firms that were very cash constrained early in the
pandemic to survive.7
A fourth approach attempts to match firms that received PPP loans to
other firms that have similar characteristics but did not receive PPP funds or
got them later. Wheat and Mac (2021) used deidentified administrative data on
customers of JPMorgan Chase and compared outcomes at firms that received
PPP loans in 2020 to those that received their first PPP loan in 2021. They found
that the 2020 recipients increased total expenses by 42 percent in the month of
receipt relative to the control group, with larger effects for smaller firms, but the
difference almost fully dissipates within three months. Dalton (2021) merges
7. It is worth noting that the banks’ ability to rapidly deliver PPP funds was supported by the
Federal Reserve through the Paycheck Protection Program Liquidity Facility (Anbil, Carlson,
and Stycznski 2021).
the PPP loan–level data with monthly administrative employment records for
all establishments in the Quarterly Census of Employment and Wages. Using
a dynamic event study design that compares recipients to observationally
similar firms that received a loan later or never received a loan, Dalton found
employment effects in the neighborhood of 4 to 6 percent, with larger effects
for smaller establishments. Dalton went on to find positive employment effects
at the end of his sample (seven months after receipt), suggesting the longer-run
average cost per job could be lower than his headline range of $20,000–34,000
per employee-month retained.8 While these studies are the most optimistic for
the efficacy of PPP, they rely on the crucial assumption that 2020 PPP recipi-
ents would have evolved similarly to 2021 recipients or to non-PPP recipients
absent the program. This assumption could fail if, for example, the firms that
did not apply during the summer of 2020 did not expect to meet the payroll
criteria for loan forgiveness, perhaps because they did not expect to reopen.9
Taking stock, three main lessons emerge. First, across research designs,
evidence on both the use of funds and employment effects suggest very limited
impact of the PPP on employment at larger firms in the months following
receipt. This suggests the program could have accomplished its employment
objectives at a much lower cost, for example by capping the maximum loan
size at well below $1 million.10 Second, some studies find evidence of an impact
on smaller businesses in the months immediately following receipt, although
nothing in the range of the statutory requirement that 60 percent of the funds
be spent on payroll. This highlights the lesson that—because money is fungi-
ble—even programs with strict employment requirements such as the PPP may
not have large effects on employment. In this case, businesses used much of
the PPP funds for items other than payroll, such as paying down debt. Third,
there is as yet no evidence of a positive effect of PPP on employment or firm
survival in the medium to long run. This will be an especially important area
for future research.11
8. This cost per job applies only to jobs directly impacted by the PPP. In other contexts total
employment effects tend to be larger than the direct effects (Chodorow-Reich 2019).
9. In the extreme, suppose that all firms that applied for and received PPP in the summer of
2020 did so knowing that they would meet the payroll requirement irrespective of whether
they received a loan and that nonapplicant firms did not apply because they knew they would
have to reduce their payroll irrespective of loan receipt. Then a comparison of these groups
of firms would indicate a positive effect of PPP receipt on employment even though PPP had
no causal impact and all of the employment at recipients was inframarginal.
10.Notably, smaller loans account for an even higher share of loans and loan amount to self-iden-
tified Black or African American recipients. Of total PPP loans to this group, 96 percent
by number and 75 percent by amount were for less than $25,000, and only 7 percent of the
amount was made in loans of more than $500,000.
11. Autor et al. (2022) extended the 500-employee cutoff design through December 2020 and
found the employment differential had fully disappeared by the end of that month.
SBA Forbearance and the EIDL Program

Beyond the PPP, the SBA has been involved in two main programs that directly
support small businesses in response to COVID-19. The first was a forbearance
program in which the SBA was authorized to pay six months of principal,
interest, and fees for all 7(a), 504, and microloans. This relief was provided
automatically to all SBA loans that were fully disbursed prior to September 27,
2020, and were in regular servicing status. SBA loan forbearance was originally
provided as part of the CARES Act in March 2020, with a total of $17 billion
available for relief. Initially, it was uncertain how many businesses would
seek and obtain new SBA loans prior to the September 27 deadline; hence, it
was unclear how much of the $17 billion allocated would be used to provide
forbearance. By the end of 2020 it was clear that not all $17 billion would be
needed, and as part of the Coronavirus Response and Relief Act, passed on
December 27, 2020, $11.5 billion of this amount was rescinded, reducing the
total assistance from the CARES Act to $3.6 billion. At the same time, the act
allocated an additional $3.5 billion in available funds for automatic loan relief,
available for all 7(a), 504, and microloans approved before September 27, 2020,
and fully disbursed after this date. Any additional funds were made available
to pay the first three months of payments for loans approved after September
27, 2020, subject to availability of funds. Thus, in total, about $7 billion in
direct loan payments were made by the SBA to cover payments that would
have normally been made by small businesses. The SBA forbearance program
fulfills two rationales for government support outlined in the Framework for
Evaluation section. First, it provided short-term liquidity to small businesses.
Second, it supported bank balance sheets by providing consistent loan pay-
ments at a time when many small businesses may not have had the revenue to
make payments on their own.12 However, because forbearance was provided
automatically, it likely went to many businesses that did not need liquidity and
would have made payments regardless.
In addition to automatic forbearance, the SBA also offered EIDL loans to
small businesses in need of liquidity during the pandemic. While the EIDL
program existed prior to COVID-19, the program was expanded considerably
in response to the COVID-19 pandemic. To give a sense of the size of this
expansion, in 2019 total EIDL loans to businesses amounted to $98 million.
The COVID-19 EIDL program was several orders of magnitude larger, with
a total of $317 billion in loans approved across 3.9 million loans as of the end
of December 2021. EIDL loans are designed to provide working capital or to
repay other business debt, allowing small businesses to refinance at favorable
rates. Originally, loans were available up to $150,000, but this cap was raised
to $500,000 in March 2021 and to $2 million in September 2021. To be eligible
12. Typically, the SBA guarantees 50 to 85 percent of an SBA loan, while the SBA forbearance
program provided an effective 100 percent guarantee for the six-month period.
Figure 4.6
Economic Injury Disaster Loans, by Loan Size

2.0 120
Number of recipients (millions)
1.5 90
Dollar amount (billions)

Number of
recipients
Total dollars
1.0 loaned 60
0.5 30
0 0
$24,999 or less $25–99k $100–249k
Loan amount
Source: Small Business Administration n.d.b.
for a loan, a business must have fewer than 500 employees and demonstrate
that it suffered working capital losses due to COVID-19. Figure 4.6 shows the
number and size distribution of the first round of EIDL loans made through
December 2020.13 Compared to PPP, the EIDL program disbursed a larger
share of funds in smaller amounts, with about 96 percent by number and 40
percent of the dollar value of loans being less than $100,000.
Importantly, the EIDL program is distinct from PPP loans, as there is
no loan forgiveness expected. Businesses that obtain these loans must meet
certain credit score requirements,14 post collateral for loans above $25,000 and
provide a personal guaranty for loans over $200,000. Thus, the subsidy from the
government comes in the form of a relatively low interest rate of 3.75 percent
combined with long, 30-year maturities and a two-year grace period in which
no loan payments are required. Given expected repayments, the Committee
13. The most recent data released by the SBA ends in December 2020, before the cap was raised
above $150,000.
14. The requirements are a credit score above 570 for loans up to $500,000 and above 625 for
loans larger than $500,000.
for a Responsible Federal Budget (n.d.) expects losses to total only $36.5 billion
even though the program has supported $317 billion in total loans. To the
extent that the COVID-19 pandemic was a short-term liquidity event for many
firms, the EIDL program was well-suited to help businesses bridge a funding
gap until revenue streams could be reestablished.
In addition to the EIDL program, the SBA administered the Targeted
EIDL Advance program, which provided funds to businesses in the most
need. EIDL Advances have no expectation of repayment; they are essentially
a no-strings-attached grant from the SBA. To qualify, a business must operate
in a low-income area, have fewer than 300 employees, and demonstrate that it
has lost at least 30 percent of its revenue over an eight-week period. Businesses
that qualify for an EIDL Advance can receive grants of up to $15,000 with no
repayment requirement. By the middle of July 2020, EIDL Advances totaled
$20 billion across 5.8 million grants disbursed.
Combined, the SBA provided substantial aid to small businesses beyond
PPP in the form of loan forbearance ($7 billion), subsidized lending ($317 billion
in loans), and direct grants ($20 billion). Despite the size of these programs,
they have received much less attention than the PPP program in academic stud-
ies. One exception is Li (2021), who used the Census Bureau’s Small Business
Pulse Survey to show that the local severity of the COVID-19 pandemic was
unrelated to the probability that a small business applied for or received an
EIDL loan or SBA loan forgiveness, suggesting that the programs were poorly
targeted. However, Li (2021) also found that firms that received SBA support
were less likely to report revenue and employee hour decreases in subsequent
weeks. Nonetheless, these are simply correlations seen in the data and should be
interpreted with caution. It is likely that the savviest businesses were the ones
that applied for SBA assistance, and they may have weathered the COVID-19
pandemic better than other firms even if they had not received SBA assistance.
Fairlie and Fossen (forthcoming) also studied the allocation of SBA assis-
tance, with a focus on whether the PPP and EIDL programs effectively reached
minority communities. They found that take-up of the PPP program was slow
in many minority communities and that loan amounts were negatively cor-
related with the minority share across communities. Meanwhile, they found
that the EIDL program was more effective in its reach, with loan numbers and
amounts both positively correlated with minority communities.
Aside from the allocation of assistance, some concern has been raised
about fraud in applying for SBA assistance. The Government Accountability
Office (2021) found that at least $156 million in EIDL loans had been approved
for ineligible businesses, such as real estate developers and multilevel market-
ers. In addition, U.S. financial institutions filed more than 20,000 reports of
suspicious activity related to the EIDL program. The SBA’s Office of Inspector
General released a report in October 2020 finding that about 46 percent of total
EIDL funding through July 2020 had been released to potentially fraudulent
borrowers, many of whom submitted duplicate applications from the same IP
address or email address (SBA 2020). Similarly, Griffin, Kruger, and Mahajan
(2021) argue that a large number of PPP loans were released to potentially
fraudulent borrowers. Given the speed and size of the programs, it is perhaps
inevitable that the SBA could not put in place tight controls—at least initially.
In preparation for future small business assistance, care should be given to
thinking about how to scale up programs quickly without lowering the guard-
rails so dramatically.
We are unaware of any academic study that clearly identifies the effect
of EIDL or SBA loan forgiveness on small business performance. Nonethe-
less, some conclusions can be drawn. First, demand for EIDL loans was very
strong, showing that the program’s subsidized terms were attractive to many
small-business owners. Many small businesses were willing to take on addi-
tional debt despite the uncertainty at the beginning of the pandemic, signifying
at least some expectation of an ability to repay after the two-year grace period.
Their demand for EIDL loans was likely affected also by the long maturity of
these loans. Recent work has shown that many individuals focus on monthly
payment amounts rather than interest rates or overall loan amounts when
considering new credit (Argyle, Nadauld, and Palmer 2020). By stretching
payments over 30 years, EIDL loans have low required monthly payments,
which likely enhanced their attractiveness. As opposed to the PPP, EIDL loans
have the benefit of providing liquidity now but at lower cost to the government
after repayment of the loans.
Another benefit of EIDL loans is their ability to be somewhat targeted
towards long-term viable firms. As laid out in the Framework for Evaluation
section, one argument for government involvement in business support is
that during downturns it can be difficult to separate viable from nonviable
firms, leading capital providers to stop providing capital entirely. During the
pandemic, government-provided liquidity via grant programs, including the
PPP, targeted firms that were hard-hit by the pandemic but not necessarily
those firms that also expected to be viable long term. Indeed, to the extent
that the pandemic fundamentally altered some aspects of the economy (e.g.,
moving more commerce online), the hardest-hit firms in the short run could
also be those that cannot survive in the long run. On the other hand, subsidized
lending programs that force business owners to consider their ability to repay
(e.g., the EIDL) or that force lenders to keep some “skin in the game” (e.g., the
Main Street Lending Program [MSLP], discussed below) can provide needed
liquidity while still attempting to provide capital to firms with better prospects.
Of course, the downside of providing loans to struggling businesses instead
of grants is that it leaves them with more debt, which could slow economic
recovery due to debt overhang. Relative to providing grants, loans create at
least some debt overhang as small businesses use cash flows to repay debt
instead of other potential investments during the recovery phase. The amount
of debt overhang in the aftermath of COVID-19 is still unknown, but the quick
recovery in the economy suggests it has not been overly severe to this point.
Clearly, the non-PPP SBA small-business support programs merit closer study
in the future than they have received to date.
Other Federal Subsidies

The CARES Act and subsequent legislation contained several other provisions
to aid businesses. Two of the largest were the Employee Retention Credit and
grants to air carriers. The Employee Retention Credit was a refundable tax credit
against employment taxes equal to 50 percent of the qualified wages paid by
an employer after March 12, 2020. To be eligible, employers had to experience
either a full or partial suspension of operations due to a government order in
response to COVID-19 or demonstrate a significant decline in gross receipts.
More than $70 billion was claimed for wages paid through 2021Q1 and a further
$31 billion after that date.
In recognition of the immediate disruption to travel, the CARES Act pro-
vided grants to air carriers based on their total payroll and required the funds
to be used exclusively for employee compensation. The program disbursed
$28.6 billion to 611 passenger carriers, cargo carriers, and support contractors
between April and October 2020. Strikingly, $22 billion of this total went to
just six large airlines: American ($6.0 billion), Delta ($5.6 billion), United ($5.1
billion), Southwest ($3.4 billion), Alaska ($1.0 billion), and JetBlue ($1.0 billion).
The academic literature has thus far paid little attention to these other pro-
grams. While both had features designed to link disbursements to payroll, the
fungibility of funds raises the possibility that they may instead have benefited
shareholders. Such concerns may be particularly significant for the grants to air
carriers, which mostly went to large, publicly traded firms that have access to
a variety of capital markets where they may have been able to access liquidity.
Alternatively, the airlines could have renegotiated with their creditors either
out of court or via Chapter 11 bankruptcy. Indeed, most major air carriers
have previously undergone successful bankruptcy restructuring, albeit not all
simultaneously. Finding a suitable counterfactual for large passenger airlines
is difficult. Careful case studies of how these firms used the funds would help
in assessing these programs.
Federal Reserve Programs

The Federal Reserve responded to the COVID-19 pandemic by taking unprece-
dented actions at unprecedented speed. It began by deploying many of the tools
it used during the 2008–09 financial crisis. Specifically, on March 15, 2020, it
cut the federal funds rate to a range of 0 to 0.25 percent and began large-scale
asset purchases, or quantitative easing, in Treasury securities and agency
mortgage-backed securities. On March 17, 2020, the Fed announced several
measures to support market liquidity, including reopening many facilities first
used in the financial crisis: the Primary Dealer Credit Facility, the Commercial
Paper Funding Facility, and the Money Market Mutual Fund Liquidity Facility.
Through their broader effects on financial markets, these steps all indirectly
supported businesses.
Direct support for business credit began on March 23, 2020, when the
Fed and the Treasury announced their new Corporate Credit Facilities. Under
the original announcement, the Primary Market Corporate Credit Facility
(PMCCF) would buy up to $100 billion of newly issued bonds and loans from
investment-grade U.S. firms. The Secondary Market Corporate Credit Facility
(SMCCF) would buy up to $100 billion of existing investment-grade bonds
and loans as well as exchange-traded funds (ETFs) that held such bonds. On
April 9, 2020, the Fed and the Treasury significantly expanded the scale of the
two programs, increasing their total capacity to $750 billion. It also expanded
their scope, allowing the facilities to buy the bonds and loans of firms that had
been investment grade at onset of the pandemic but had subsequently been
downgraded.
The April 9 announcement also established the MSLP, a $600-billion facil-
ity to make loans to firms. The program was aimed at midsized firms, with
requirements that firm employment, revenue, and leverage not be too high.
Banks made qualifying loans and sold 95 percent to the facility while retain-
ing the remaining 5 percent. Restrictions were placed on uses of funds, and
firms participating in the program were subject to restrictions on executive
compensation, dividends, and share repurchases.15
The Corporate Credit Facilities

Any evaluation of the CCFs must wrestle with two facts. First, take-up was
very low. As shown in Table 4.1, the CCFs used only approximately $15 billion
of their $750-billion capacity.
Second, despite this low take-up, the CCFs appear to have had meaningful
announcement effects on bond prices, as shown in Figure 4.7. Investment-grade
credit spreads fell sharply after the initial program announcement on March
23, while high-yield credit spreads were more significantly impacted when
the programs were significantly expanded on April 9. Spreads fell further
after Federal Reserve Chair Jerome Powell’s remarks on May 29: “The Fed is
strongly committed to using our tools to do whatever we can for as long as
it takes to provide some relief and some stability now. … We crossed a lot of
red lines, that had not been crossed before. … This is that situation in which
you do that, and then you figure it out afterward (Smialek 2020).” These price
movements were accompanied by significant bond issuance by firms, which
took advantage of improving market conditions to build up their liquidity
buffers (Halling, Yu, and Zechner 2020).
15. While the Corporate Credit Facilities and the Main Street Lending Program were jointly
designed by the Federal Reserve and the Treasury Department, press reports indicated that
some of the more restrictive elements of the program design were insisted upon by Treasury.
See, for example, Timiraos and Davidson (2020).
Figure 4.7
Corporate Credit Spreads and Select Policy

Changes, 2020
A. A and BBB-Rated Bond Spreads
3/15: Funds rate cut to 0%; $700B UST and MBS purchase
3/23: PMCCF, SMCCF, TALF announced; open-ended UST, MBS purchases
5 3/27: CARES Act passed
4/9: PMCCF, SMCCF expanded; Main Street announced
4 5/29: Powell “red line” comments
Percent
6/15: SMCCF purchases to track index

3
2
BBB
1
A
0
Jan. Feb. Mar. Apr. May Jun. Jul. Aug.
B. B and CCC-Rated Bond Spreads
3/15: Funds rate cut to 0%; $700B UST and MBS purchase
3/23: PMCCF, SMCCF, TALF announced; open-ended UST, MBS purchases
3/27: CARES Act passed
20 4/9: PMCCF, SMCCF expanded; Main Street announced
5/29: Powell “red line” comments
16 6/15: SMCCF purchases to track index
Percent
12 CCC
8
B
4
0
Jan. Feb. Mar. Apr. May Jun. Jul. Aug.
Source: ICE/Bank of America n.d.

Note: Lines show ICE/BofA corporate bond index option-adjust-
ed spreads, in percentage points, for A, BBB, B, and CCC rated
bonds. The acroynms refer to: PMCCF (Primary Market Corporate
Credit Facility); SMCCF (Secondary Market Corporate Credit Facili-
ty); TALF (Term Asset-Backed Securities Loan Facility); MBS (Mortgage-Backed Securities).
Consistent with Figure 4.7, academic studies, including Haddad, Moreira,

and Muir (2021), Gilchrist et al. (2020), and Boyarchenko, Kovner, and Shachar
(2021), find significant effects on credit spreads of the announcement of the
CCFs when taking a simple event study approach. Yet, while event studies find
large effects, they are potentially confounded by other news about the path of
the pandemic and the macroeconomy that were released around the same time.
Thus, the same studies try to achieve more careful identification of the effects
of the CCFs by also taking a second approach: a differences-in-differences
approach that compares spreads on bonds that were eligible for CCFs purchases
and bonds that were not, before and after the key program announcements.
These empirical exercises find that the CCFs lowered credit spreads, but they
generally found smaller magnitudes than the simple event study approach.
Boyarchenko, Kovner, and Shachar (2021) also argued that purchases themselves
had important effects on bond prices, over and above the simple announce-
ments of the programs.
While the differences-in-differences approach offers more careful iden-
tification, it may understate the effects of the CCFs for two reasons. First,
the programs may have had general equilibrium effects that simultaneously
moved all bond prices. Second, investors may have anticipated that the pro-
grams would be expanded if market conditions deteriorated further. Thus,
program announcements may have moved the prices of ineligible bonds.
Haddad, Moreira, and Muir (2021) used prices of options on bond ETFs to
argue that the market did indeed anticipate significant expansions of the CCFs
if markets deteriorated.
In terms of the rationales articulated for government intervention in Sec-
tion III, the CCFs are best rationalized as an attempt to reduce the financial
frictions that prevailed in the corporate bond market early in the pandemic.
Bond price declines in March 2020 were in part driven by fire sale dynamics
(Ma, Xiao, and Zeng 2021; Falato, Goldstein, and Hortaçsu 2021), and the CCFs
may have helped mitigate fire sale problems. Consistent with the idea that the
CCFs reduced financial frictions, O’Hara and Zhou (2021) and Kargar et al.
(2021) show that market liquidity improved significantly for eligible bonds.
We next turn to the potential costs of the CCFs. As discussed in Hanson
et al. (2020), the expected cost of the CCFs depends in part on one’s theory
of disruptions in the corporate bond market. It could be the case that bond
market fire sales are akin to bank runs—that there are multiple equilibria, a
“bad” fire sale equilibrium with low asset prices in which many investors try
to fire sell their bonds and a “good” equilibrium featuring high prices and few
sales. Under this multiple equilibrium view, the CCFs take little risk.
In contrast, it could be the case that there are not multiple equilibria, but
government actions still have benefits. For instance, suppose that losses could
be borne either by the government, in which case they must be financed by
future taxation, or by the private sector, in which case they are amplified by
private sector financial frictions and spillovers. If the distortions associated
with taxation are relatively low and private sector frictions are relatively high,
then government intervention may be warranted, but it is not a free lunch
(Hanson, Scharfstein, and Sunderam 2019).
The low take-up and large price impact of the CCFs are not sufficient to
distinguish between these two views. Under the multiple equilibrium view,
the very existence of the CCFs shifted markets from the bad equilibrium to
the good one, like deposit insurance in the canonical Diamond and Dybvig
(1983) treatment of bank runs. Thus, there are large benefits to government
intervention in terms of prices and market functioning, even though utilization
of the facilities is low.
However, low take-up and large price impact are also consistent with the
idea that there are not multiple equilibria. Instead, the government opened itself
up to significant risk taking through the CCFs, but that risk did not realize due
to the path of the pandemic. Haddad, Moreira, and Muir (2021) used prices of
options on bond ETFs to argue that the market anticipated that the government
could take significant losses on its bond purchases if the economic impact of
the pandemic had been worse.
Given the rationales articulated in the Framework for Evaluation section,
it is worth noting that the CCFs targeted large firms with access to public mar-
kets. While the financial frictions these firms faced were likely more severe
during the initial stages of the pandemic than normal, they were also likely
much less severe than the financial frictions faced by smaller firms. In other
words, the CCFs were not targeted toward firms facing the most significant
financial frictions. Nonetheless, since public firms are large employers with
large macroeconomic impacts, interventions targeted at them may have rela-
tively large benefits.
Taking stock, the key lesson of the CCFs is that it is possible for the gov-
ernment to play a major stabilizing role in bond markets and reducing financial
frictions. The critical open question is whether doing so is desirable. In the
COVID-19 crisis, large benefits were obtained at low cost with low take-up,
but those outcomes were in part due to the path of the pandemic. Had the
pandemic more strongly affected the economy in late 2020 and early 2021, the
costs of intervention may have been significantly higher. The costs and benefits
of such intervention in future market disruptions are uncertain.
The Main Street Lending Program

We next turn to the MSLP, which targeted smaller firms than the CCF. As
shown in Table 4.1, like the CCFs, the MSLP had very low take-up. It used just
over $18 billion of its $600 billion capacity.
A key design feature of the MSLP was the way that banks and the govern-
ment shared risk. Banks sold 95 percent of qualifying loans to the facility while
retaining a 5 percent slice of the loan with the same risk (i.e., a “pari passu”
loan participation) on their balance sheets. This design choice meant that
loans made under the MSLP had to offer similar returns to other loans banks
were willing to make. In other words, the MSLP did not encourage subsidized
lending. Though banks only had to retain a fraction of the loans, they had to
earn a satisfactory return on the retained portions. And since banks and the
government shared risk and repayments proportionately, the overall returns
on MSLP loans were similar to the returns on the bank-retained portions.16
When would banks find a facility with such a design useful? At times when
banks are highly balance-sheet constrained but when there are many loans on
which banks could earn a satisfactory return. At such times, the MSLP would
expand the size of banks’ effective balance sheets. At the onset of the pandemic,
there was a considerable chance that banks would become capital constrained.
As documented by Chodorow-Reich et al. (forthcoming), Greenwald, Krainer,
and Paul (2021), and Kapan and Minoiu (2021), there were significant draw-
downs of bank credit lines in the early stages of the pandemic. Greenwald,
Krainer, and Paul (2021) argued that drawdowns may have changed decisions
about new lending, suggesting that balance-sheet constraints may have entered
banks’ calculus. In addition, Acharya, Engle, and Steffen (2021) showed that
banks with larger drawdowns suffered particularly large stock price declines.
However, as shown in Figure 4.8, neither bank stock prices nor bank capital
ratios declined as significantly during the pandemic as they did during the
global financial crisis, and they recovered from their lows far more quickly.
For instance, bank regulatory capital declined 29 percent peak-to-trough in
the global financial crisis, compared with 7 percent during the pandemic.
In terms of the rationales for intervention outlined earlier in the chapter,
the MSLP is best rationalized as an attempt to reduce the potential financial
frictions in the banking sector. There is a rich body of literature demonstrating
that bank capital supply shocks can negatively impact firm investment and
employment (e.g., Bernanke 1983; Bernanke and Lown 1991; Peek and Rosengren
1997; Ashcraft 2005; Khwaja and Mian 2008; Ivashina and Scharfstein 2010;
Chodorow-Reich 2014). Furthermore, these impacts tend to be particularly
severe for smaller firms without access to public capital markets, and the MSLP
focused on such firms.
Taking stock, the key lesson is that the MSLP could have had a larger
impact if the pandemic’s effect on the macroeconomy and the banking sector
had been more severe and more protracted.17 The key open question is whether
other tools for shoring up bank balance sheets could achieve the same goals
at lower cost. For instance, increasing the amount of bank equity would also
improve the health of their balance sheets and support additional lending. In
16. The returns were not exactly the same, because the banks received origination and servicing
fees, while the government did not.
17. Since market prices are generally not available for bank loans, it is difficult to study
announcement effects of the MSLP in the way that the academic literature has for the
CCFs. Nonetheless, Minoiu, Zarutskie, and Zlate (2021) argue that the MSLP may have
been perceived by banks as a backstop. As such, banks may have lent more at the initial
stages of the pandemic because they understood that future lending would be supported
by the MSLP.
Figure 4.8
Bank Stock Prices and Regulatory Capital

Ratios, 2007–2020
A. Bank Stock Prices
120
KBW Nasdaq bank index
100
80
60
40
20
2007 2009 2011 2013 2015 2017 2020
B. Bank Regulatory Capital Ratios

14
Bank regulatory capital ratio
12
10
4
2007 2009 2011 2013 2015 2017 2020
Source: Nasdaq n.d.; Federal Reserve Board n.d.

Note: Panel A shows the evolution of the KBW bank index, a
capitalization-weighted index of 24 stocks that is designed to
track the performance of U.S. money center and regional banking
firms. (The index does not include traditional brokerage firms like
Goldman Sachs and Morgan Stanley that are now organized as BHCs.) Panel B shows the
risk-based capital ratios of U.S. publicly traded BHCs from 2006Q1 to 2020Q1 using data
from Form FR Y-9C. Specifically, we plot the Tier 1 capital ratio (the ratio of Tier 1 Capital
to Risk-Weighted Assets) and the CET1 ratio (the ratio of Common Equity Tier 1 Capital to
Risk-Weighted Assets). Prior to 2014Q1 (for Advanced Approaches BHCs) or 2015Q1 (for all
other BHCs) when Common Equity Tier 1 Capital is first reported on the FR Y-9C, we con-
structed a proxy for Common Equity Tier 1 Capital (sometimes referred to as Tier 1 Com-
mon Equity) by taking the appropriate deductions from each BHC’s reported Tier 1 Capital.
the next crisis, the government could encourage higher amounts of equity in
the banking system in two ways. First, as argued by Greenwood et al. (2017)
and Blank et al. (2020), it could use the bank stress tests as a regulatory tool to
encourage banks to raise equity from capital markets. Second, in a more severe
crisis, the government could directly inject equity into the banking system, as
it did during the global financial crisis. Encouraging banks to raise equity from
capital markets minimizes the government’s risk exposure and involvement in
bank operations. In contrast, when the government injects equity itself, a host
of governance problems can arise. The MSLP sits between these extremes. The
government is still involved, but it avoids some of the governance problems
involved with direct equity ownership.18
State and Local Programs

While our focus is on the federal support programs, we note that all states and
many counties and cities created programs to provide grants or below-mar-
ket-rate loans to private businesses. Funding for these initiatives came from
the CARES Act Federal Coronavirus Relief Fund, from other federal sources,
and from state and local government tax revenue. We collected information
on state-administered business relief programs through internet searches
and list the total amount disbursed in Table 4.1. State grant and fee offset pro-
grams total $14.7 billion, with more than half of the financing coming from
the CARES Act. Notably, many of these initiatives had caps of $100,000 or
less and employment caps well below 500, making them much more targeted
toward small businesses than the federal PPP. While these programs were much
smaller than the federal programs, they may have offered more “bang for the
buck” by focusing on those businesses that faced the largest financing frictions.
Other Factors Impacting Businesses

The CCFs and the MSLP were designed to support relatively large firms through
the pandemic. Given that these facilities saw relatively little use, the question
arises: how did these firms weather the pandemic? We study this question using
Compustat data on nonfinancial firms. Notably, while Compustat is restricted to
relatively large firms with publicly issued equity or debt, the patterns uncovered
may also help to shed light on the experience of smaller firms.
Figure 4.9 shows that prior to the pandemic, Compustat nonfinancial firms
in total earned large positive net income of over $200 billion per quarter. On
average, they did not accumulate additional cash and they used their profits
18. It does not avoid all such problems, however. For instance, for programs like the Main
Street Lending Program, there are important questions about whether the government or
the originating bank should have control rights if loans default.
Figure 4.9
Quarterly Net Income and Financing for

Nonfinancial Public Firms, 2019Q1–2021Q1
400
300 Net Change

income in cash
200
Billions of dollars
100
-100 Net
financing
-200
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1
2021
2019 2020
Source: Compustat n.d.

Note: Cash is defined as cash and short-term equivalents; net financ-
ing is defined as net stock issuance (issuance minus repurchases)
plus net debt issuance (issuance minus retirements) minus dividends.
to reduce their outstanding net financing; that is, on average, they retired debt
and repurchased equity. Total net income then fell sharply with the onset of
the pandemic in the first quarter of 2020 and remained low in the second
quarter. At the same time, firms increased their issuance of net new financing
and built up their cash buffers.19 This behavior is consistent with the idea that
firms feared a prolonged downturn at the beginning of the pandemic. However,
firms’ fears were not realized, as Figure 4.9 shows that net income recovered
to its pre-pandemic level by the third quarter of 2020.
Why did net income not fall further at the height of the initial pandemic-re-
lated lockdowns? As Figure 4.10 shows, firms were able to reduce operating costs
as their sales fell. A significant portion of this cost adjustment likely occurred
19. The change in cash is larger than the sum of net income and net new financing. The differ-
ence reflects (a) the conversion of noncash assets to cash; (b) depreciation, which shows up
in net income but is not a cash expense; (c) trade credit (i.e., firms stretching their accounts
payable and cutting their accounts receivable); and (d) capital expenditures.
Figure 4.10
Sales and Operating Costs for Nonfinancial

Public Firms, 2019Q1–2021Q1
400
Sales Operating
costs
300
Billions of dollars
200
100
0
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1
2021
2019 2020
through payrolls. This highlights the fact that there are two potential paths for the
government to support households and firms. First, as in the U.S. unemployment
insurance scheme, firms can lay off workers to reduce costs and the government
can then provide direct aid to workers. Second, in schemes like the PPP, firms
can retain workers and the government can help offset the costs of payroll.
How did firms increase their cash and net new financing early in the
pandemic? Figure 4.11 breaks total new financing of nonfinancial firms in
Compustat into three categories: net new equity issuance, net new debt issu-
ance (including interest payments), and dividends paid to equity. Prior to the
pandemic, dividend payments exceeded $100 billion each quarter and equity
repurchases averaged $100 billion per quarter. Net debt issuance was generally
small but positive. Figure 4.11 shows that equity repurchases (negative net equity
issuance) shrank dramatically with the onset of the pandemic while dividend
payments remained stable. Firms raised over $250 billion of new debt financing
in the first quarter of 2020. This debt came from two sources: capital markets
and drawdowns of credit lines. Additional debt financing was raised in the
Figure 4.11
Net Financing for Nonfinancial Public Firms,

2019Q1–2021Q1
300
200
Net
debt
Billions of dollars
100
-100
Net Dividends
equity
-200
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1
2021
2019 2020

Note: Net financing is defined as net stock issuance plus net debt
issuance minus dividends.
second quarter of 2020, and firms started to repay this financing at the end
of 2020 as the economic outlook improved. Importantly, this reflects capital
raising only by firms in the Compustat dataset, which are larger firms that
have access to public capital markets. Smaller firms likely found it somewhat
more difficult to raise capital during this time, which again highlights the
importance of targeting programs such as PPP and EIDL loans to those firms.
Why was the recovery in net income so fast in 2020? Figure 4.12 compares
the evolution of sales (revenues) during the pandemic and the global financial
crisis (GFC). The figure shows that the aggregate drop in sales for nonfinancial
firms was similar in both recessions, but sales recovered much more quickly
during the pandemic.
Conclusion
We have evaluated the main business aid programs deployed by the U.S. gov-
ernment during the COVID-19 pandemic. Our focus has been understanding
Figure 4.12
Sales for Nonfinancial Public Firms, Indexed to

Recession Peak
120
Global financial crisis

COVID-19 pandemic
110
Share of peak sales
100
90
80
1 2 3 4 5 6 7 8 9
Quarters since peak

Note: The global financial crisis is indexed to 2008Q1 and the
COVID-19 pandemic to 2019Q1.
the potential for such programs to help speed recoveries from future non-pan-
demic-related downturns. The main conclusion is that policymakers should not
automatically interpret the rapid recovery from the pandemic as evidence that
business aid programs have strong economic benefits. Many careful studies
found that these programs had relatively small effects, suggesting that other
factors including the nature of recovery from a temporary lockdown and gen-
eral support for households likely played a more important role. There may be
circumstances in which small-business lending programs like the EIDL or bond
market stabilization programs like the CCFs could prove useful—for instance,
in cases in which other support for households is less generous—but they should
be judiciously deployed. The speed at which support programs were deployed
during the COVID-19 pandemic was admirable. However, given the rapid
rollout, it is not surprising that some of the programs were not well-designed
to achieve maximum impact.
Four concrete lessons emerge from our analysis of business support pro-
grams in the COVID-19 pandemic. First, policymakers should not blindly
redeploy the 2020 tool kit. Second, support for small businesses, like the PPP,
could have been restricted to significantly smaller firms. For instance, the
employment cap for program eligibility could have been set at 50 or 100 employ-
ees, instead of 500, without adversely affecting the program’s overall impact.
Third, support for large firms, such as publicly traded airlines, should be treated
skeptically because these firms have access to many forms of financing and can
be efficiently processed by the bankruptcy system. Finally, while the Federal
Reserve clearly can support banks and corporate credit markets, whether it
should do so involves careful consideration of the reason for a decline in credit.
References
Acharya, Viral V., Robert F. Engle III, and Sascha Steffen. 2021. “Why Did Bank
Stocks Crash During COVID-19?” Working Paper 28559, National Bureau of
Economic Research, Cambridge, MA.
Altig, Dave, Scott Baker, Jose Maria Barrero, Nicholas Bloom, Philip Bunn, Scarlet
Chen, Steven J. Davis, et al. 2020. “Economic Uncertainty before and during the
COVID-19 Pandemic.” Journal of Public Economics 191: 104274.
Anbil, Sriya, Mark Carlson, and Mary-Frances Styczynski. 2021. “The Effect of the
PPPLF on PPP Lending by Commercial Banks.” Working paper 2021-030. Board
of Governors of the Federal Reserve System, Washington, D.C.
Argyle, Bronson S., Taylor D. Nadauld, and Christopher J. Palmer. 2020. “Monthly
payment targeting and the demand for maturity.” The Review of Financial Stud-
ies 33(11): 5416–62.
Ashcraft, Adam B. 2005. “Are banks really special? New evidence from the FDIC-in-
duced failure of healthy banks.” American Economic Review 95(5): 1712–30.
Autor, David, David Cho, Leland D. Crane, Mita Goldar, Byron Lutz, Joshua
Montes, William B. Peterman, et al. 2020. “An Evaluation of the Paycheck Pro-
tection Program Using Administrative Payroll Microdata.” MIT Department
of Economics, Cambridge, MA.
Autor, David, David Cho, Leland D. Crane, Mita Goldar, Byron Lutz, Joshua
Montes, William Peterman, et al. 2022. “The $800 Billion Paycheck Protection
Program: Where Did the Money Go and Why Did it Go There?” Working Paper
29669, National Bureau of Economic Research, Cambridge, MA.
Bartik, Alexander W., Zoe B. Cullen, Edward L. Glaeser, Michael Luca, Christopher
T. Stanton, and Adi Sunderam. 2021. “The Targeting and Impact of Paycheck
Protection Program Loans to Small Businesses.” Working Paper 27623, National
Bureau of Economic Research, Cambridge, MA.
Bartlett, Robert P., and Adair Morse. 2020. “Small Business Survival Capabili-
ties and Policy Effectiveness: Evidence from Oakland.” Working Paper 27629,
National Bureau of Economic Research, Cambridge, MA.
Bernanke, Ben S. 1983. “Nonmonetary Effects of the Financial Crisis in the Prop-
agation of the Great Depression.” American Economic Review 73(3): 257–76.
Bernanke, Ben S., and Cara S. Lown. 1991. “The Credit Crunch.” Brookings Papers
on Economic Activity (2), Brookings Institution, Washington, D.C.
Blank, Michael, and Omeed Maghzian. 2021. “Job Separations and Optimal Labor
Market Stabilization Policy.” Unpublished working paper.
Blank, Michael, Samuel G. Hanson, Jeremy C. Stein, and Adi Sunderam. 2020.
“How Should U.S. Bank Regulators Respond to the COVID-19 Crisis.” Working
Paper 63, Hutchins Center, Cambridge, MA.
Boyarchenko, Nina, Anna Kovner, and Or Shachar. 2021. “It’s What You Say and
What You Buy: A Holistic Evaluation of the Corporate Credit Facilities.” Work-
ing Paper 8679, CESifo, Cambridge, MA.
Bureau of Economic Analysis. 2021. “Gross Domestic Product, 2nd Quarter
2021.” Washington, D.C. https://www.bea.gov/news/2021/gross-domestic-prod-
uct-third-estimate-gdp-industry-and-corporate-profits-revised-2nd.
Bureau of Labor Statistics. 2021. “The Employment Situation—June 2021.” Wash-
ington, D.C. https://www.bls.gov/news.release/archives/empsit_07022021.htm.
Chetty, Raj, John Friedman, Nathaniel Hendren, Michael Stepner, and the Oppor-
tunity Insights Team. 2020. “The Economic Impacts of COVID-19: Evidence
from a New Public Database Built Using Private Sector Data.” Working Paper
27431, National Bureau of Economic Research, Cambridge, MA.
Chodorow-Reich, Gabriel. 2014. “The Employment Effects of Credit Market Dis-
ruptions: Firm-Level Evidence from the 2008–09 Financial Crisis,” Quarterly
Journal of Economics 129(1): 1–59.
Chodorow-Reich, Gabriel. 2019. “Geographic Cross-Sectional Fiscal Spending
Multipliers: What Have We Learned?” American Economic Journal: Economic
Policy 11(2): 1–34.
Chodorow-Reich, Gabriel, Olivier Darmouni, Stephan Luck, and Matthew Plosser.
Forthcoming. “Bank Liquidity Provision Across the Firm Size Distribution.”
Journal of Financial Economics.
Chodorow-Reich, Gabriel, and Loukas Karabarbounis. 2016. “The Cyclicality of
the Opportunity Cost of Employment.” Journal of Political Economy 124(6):
1563–618.
Committee for a Responsible Federal Budget. n.d. “COVID Money Tracker.”
Accessed January 13, 2022, at https://www.covidmoneytracker.org/.
Congressional Budget Office. 2020. “An Update to the Economic Outlook: 2020 to
2030, July 2020.” Congressional Budget Office, Washington, D.C.
Crane, Leland D., Ryan A. Decker, Aaron Flaaen, Adrian Hamins-Puertolas,
and Christopher Kurz. 2021. “Business Exit During the COVID-19 Pandemic:
Non-Traditional Measures in Historical Context.” Board of Governors of the
Federal Reserve System, Washington, D.C.
Dalton, Michael. 2021. “Putting the Paycheck Protection Program into Perspective:
An Analysis Using Administrative and Survey Data.” Working Paper 542, U.S.
Bureau of Labor Statistics, Washington, D.C.
Darmouni, Olivier. 2020. “Information Frictions and the Credit Crunch.” Journal
of Finance 75(4): 2055–94.
Diamond, Douglas W., and Philip H. Dybvig. 1983. “Bank Runs, Deposit Insurance,
and Liquidity.” Journal of Political Economy 91(3): 401–19.
Diamond, Douglas W., and Raghuram G. Rajan. 2011. “Fear of Fire Sales, Illiquid-
ity Seeking, and Credit Freezes.” The Quarterly Journal of Economics 126(2):
557–91.
Diamond, William. 2020. “Safety Transformation and the Structure of the Financial
System.” The Journal of Finance 75(6): 2973–3012.
Doniger, Cynthia L., and Benjamin Kay. 2021. “Ten Days Late and Billions of Dollars
Short: The Employment Effects of Delays in Paycheck Protection Program Financ-
ing.” Board of Governors of the Federal Reserve System, Washington, D.C.
Fairlie, Robert, and Frank M. Fossen. Forthcoming. “Did the Paycheck Protec-
tion Program and Economic Injury Disaster Loan Program Get Disbursed
to Minority Communities in the Early Stages of COVID-19?” Small Business
Economics.
Falato, Antonio, Itay Goldstein, and Ali Hortaçsu. 2021. “Financial Fragility in
the COVID-19 Crisis: The Case of Investment Funds in Corporate Bond Mar-
kets.” Journal of Monetary Economics 123: 35–52.
Farhi, Emmanuel, and Ivan Werning. 2016. “A Theory of Macroprudential Policies
in the Presence of Nominal Rigidities.” Econometrica 84(5): 1645–704.
Faulkender, Michael W., Robert Jackman, and Stephen Miran. 2020. “The Job
Preservation Effects of Paycheck Protection Program Loans.” Working Paper
2020-01, Office of Economic Policy, U.S. Department of the Treasury, Wash-
ington, D.C.
Fazzari, Steven, R. Glenn Hubbard, and Bruce Petersen. 1988. “Financing Con-
straints and Corporate Investment.” Brookings Papers on Economic Activity
(1). Brookings Institution, Washington, D.C.
Federal Reserve Board. 2020. “FR Y-9C: Consolidated Financial Statements for
Holding Companies.” Washington, D.C. https://www.federalreserve.gov/apps/
reportforms/reportdetail.aspx?sOoYJ+5BzDal8cbqnRxZRg==.
Feldman, Rob J., and Jason Schmidt. 2021. “Government Fiscal Support Protected
Banks from Huge Losses during the COVID-19 Crisis.” Federal Reserve Bank
of Minneapolis, Minneapolis, MN.
Gilchrist, Simon, Bin Wei, Vivian Z. Yue, and Egon Zakrajšek. 2020. “The Fed Takes
on Corporate Credit Risk: An Analysis of the Efficacy of the SMCCF.” Working
Paper 27809, National Bureau of Economic Research, Cambridge, MA.
Government Accountability Office. 2021. “Economic Injury Disaster Loan Pro-
gram: Additional Actions Needed to Improve Communication with Applicants
and Address Fraud Risks,” GAO Publication No. 21-589, Washington, D.C.
Granja, João, Christos Makridis, Constantine Yannelis, and Eric Zwick. 2020.
“Did the Paycheck Protection Program Hit the Target?” Working Paper 27095,
National Bureau of Economic Research, Cambridge, MA.
Greenwald, Daniel L., John Krainer, and Pascal Paul. 2021. “The Credit Line
Channel.” Working Paper 2020-26, Federal Reserve Bank of San Francisco,
San Francisco, CA.
Greenwood, Robin, Samuel G. Hanson, Jeremy C. Stein, and Adi Sunderam. 2017.
“Strengthening and Streamlining Bank Capital Regulation.” Brookings Papers
on Economic Activity (Fall). Brookings Institution, Washington, D.C.
Greenwood, Robin, Benjamin Iverson, and David Thesmar. 2020. “Sizing up Cor-
porate Restructuring in the COVID Crisis.” Brookings Papers on Economic
Activity (Fall). Brookings Insitution, Washington, D.C.
Griffin, John, Samuel Kruger, and Prateek Mahajan. 2021. “Did FinTech Lend-
ers Facilitate PPP Fraud?” Working Paper, Social Science Research Network,
Rochester, NY.
Haddad, Valentin, Alan Moreira, and Tyler Muir. 2021. “When Selling Becomes
Viral: Disruptions in Debt Markets in the COVID-19 Crisis and the Fed’s
Response.” Review of Financial Studies 34(11): 5309–51.
Hall, Robert E., and Marianna Kudlyak. 2021. “The Inexorable Recoveries of U.S.
Unemployment” Working Paper 2021-20, Federal Reserve Bank of San Fran-
cisco, San Francisco, CA.
Halling, Michael, Jin Yu, and Josef Zechner. 2020. “How did COVID-19 Affect
Firms’ Access to Public Capital Markets?” The Review of Corporate Finance
Studies 9(3): 501–33.
Hanson, Samuel G., David S. Scharfstein, and Adi Sunderam. 2019. “Social Risk,
Fiscal Risk, and the Portfolio of Government Programs.” The Review of Financial
Studies 32(6): 2341–82.
Hanson, Samuel G., Jeremy C. Stein, Adi Sunderam, and Eric Zwick. 2020. “Busi-
ness Credit Programs in the Pandemic Era.” Brookings Papers on Economic
Activity (Fall), Brookings Insitution, Washington, D.C.
Hanson, Samuel G., Adi Sunderam, and Eric Zwick. 2021. “Business Continuity
Insurance in the Next Disaster.” In Rebuilding the Post-Pandemic Economy,
edited by Melissa S. Kearney and Amy Ganz, 52–77. Aspen Institute Press,
Washington, D.C.
Hubbard, Glenn, and Michael R. Strain. 2020. “Has the Paycheck Protection Pro-
gram Succeeded?” Brookings Papers on Economic Activity (Fall). Brookings
Insitution, Washington, D.C.
ICE Data Indices. 2020. “ICE/Bank of America U.S. Corporate Index Option-Ad-
justed Spreads.” Retrieved from FRED, Federal Reserve Bank of St. Louis. https://
fred.stlouisfed.org/series/BAMLC0A4CBBB.
Ivashina, Victoria, and David S. Scharfstein. 2010. “Loan Syndication and Credit
Cycles.” American Economic Review: Papers & Proceedings 100(2): 57–61.
Iverson, Benjamin. 2018. “Get in Line: Chapter 11 Restructuring in Crowded Bank-
ruptcy Courts.” Management Science 64(11): 5370–94.
The JPMorgan Chase Institute. 2022. “Household Pulse: The State of Cash
Balances at Year End.” Last updated with December 2021 data. Accessed
March 3, 2022, at https://www.jpmorganchase.com/institute/research/
household-income-spending/household-pulse-cash-balances-at-year-end.
Kapan, Tumer, and Camelia Minoiu. 2021. “Liquidity Insurance vs. Credit Pro-
vision: Evidence from the COVID-19 Crisis.” Working Paper, Social Science
Research Network, Rochester, NY.
Kargar, Mahyar, Benjamin Lester, David Lindsay, Shuo Liu, Pierre-Olivier Weill,
and Diego Zúñiga. 2021. “Corporate Bond Liquidity During the COVID-19
Crisis.” The Review of Financial Studies 34(11): 5352–401.
Khwaja, Asim Ijaz, and Atif Mian. 2008. “Tracing the impact of bank liquidity shocks:
Evidence from an emerging market.” American Economic Review 98(4): 1413–42.
Li, Meng. 2021. “Did the Small Business Administration’s COVID-19 Assistance
Go to the Hard Hit Firms and Bring the Desired Relief?” Journal of Economics
and Business 115: 105969.
Ma, Yiming, Kairong Xiao, and Yao Zeng. 2021. “Mutual Fund Liquidity Trans-
formation and Reverse Flight to Liquidity.” Jacobs Levy Equity Management
Center for Quantitative Financial Research, Philadelphia, PA.
Minoiu, Camelia, Rebecca Zarutskie, and Andrei Zlate. 2021. “Motivating Banks
to Lend? Credit Spillover Effects of the Main Street Lending Program.” Working
paper 2021-078. Board of Governors of the Federal Reserve System, Washing-
ton, D.C.
Nasdaq. 2021. “KBW Nasdaq Bank Index.” https://indexes.nasdaqomx.com/Index/
Overview/BKX.
O’Hara, Maureen, and Xing Alex Zhou. 2021. “Anatomy of a Liquidity Crisis:
Corporate Bonds in the COVID-19 Crisis.” Journal of Financial Economics.
142(1): 46–68.
Peek, Joe, and Eric S. Rosengren. 1997. “The International Transmission of Financial
Shocks: The Case of Japan.” The American Economic Review 87(4): 495–505.
Rajan, Raghuram G. 1992. “Insiders and Outsiders: The Choice between Informed
and Arm’s-Length Debt.” Journal of Finance 47(4): 1367–1400.
Romer, Christina, and David Romer. Forthcoming. “A Social Insurance Perspec-
tive on Pandemic Fiscal Policy: Implications for Unemployment Insurance and
Hazard Pay.” Journal of Economic Perspectives.
Shleifer, Andrei, and Robert Vishny. 1992. “Liquidation Values and Debt Capacity:
A Market Equilibrium Approach.” Journal of Finance 47(4): 1343–66.
Small Business Administration. 2020. “Inspection of Small Business Adminis-
tration’s Initial Disaster Assistance Response to the Coronavirus Pandemic.”
Office of the Inspector General, U.S. Small Business Administration, Report
Number 21-02, October 28, 2020.
Small Business Administration. n.d.a. Paycheck Protection Program (PPP). PPP
Data. U.S. Department of the Treasury, Washington, D.C.
Small Business Administration. n.d.b. COVID-19 Economic Injury Disaster Loan
(EIDL). EIDL Data. U.S. Department of the Treasury, Washington, D.C.
Smialek, Jeanna. 2020. “Powell Says Federal Reserve Crossed Red Lines to Help
Economy.” The New York Times, May 29, 2020.
Stein, Jeremy. 2012. “Monetary Policy as Financial-Stability Regulation.” Quarterly
Journal of Economics 127(1): 57–95.
Survey of Professional Forecasters. 2020. “Second Quarter 2020 Survey of Profes-

sional Forecasters.” Federal Reserve Bank of Philadelphia, Philadelphia, PA.
Timiraos, Nick, and Kate Davidson. 2020. “Fed, Treasury Disagreements Slowed
Start of Main Street Lending Program.” Wall Street Journal, July 12, 2020.
U.S. Census Bureau. n.d. “Business Dynamic Statistics (BDS).” Last updated Octo-
ber 26, 2021. U.S. Census Bureau, Washington, D.C.
U.S. Courts Bankruptcy Filing Statistics. 2021. “Business and Nonbusiness Cases
Filed, by Chapter of the Bankruptcy Code.” Washington, D.C. https://www.
uscourts.gov/report-name/bankruptcy-filings.
Wang, Jialan, Jeyul Yang, Benjamin Iverson, and Renhao Jiang. 2021. “Bankruptcy
and the COVID-19 Crisis.” Working Paper 21-041, Harvard Business School,
Cambridge, MA.
Wheat, Chris, and Chi Mac. 2021. “Did the Paycheck Protection Program Support
Small Business Activity?” JPMorgan Chase Institute, New York City, NY.
Zwick, Eric, and James Mahon. 2016. “Tax Policy and Heterogeneous Investment
Behavior.” Working Paper 21876, National Bureau of Economic Research, Cam-
bridge, MA.
The COVID-19 pandemic posed an extraordinary threat to lives
and livelihoods. In the United States, the pandemic triggered a
sharp downturn. Yet, the ensuing economic recovery was faster
and stronger than nearly any forecaster anticipated due in part
to the swift, aggressive, sustained, and creative response of
U.S. fiscal and monetary policy. But when the next recession
arrives, it most likely won’t be triggered by a pandemic.
Recession Remedies examines and evaluates the breadth of

the economic-policy response to COVID-19. Chapters address
Unemployment Insurance, Economic Impact Payments, loans
and grants to businesses, assistance to renters and mortgage
holders, aid to state and local governments, policies that
targeted children, Federal Reserve policy, and the use of non-
traditional data to monitor the economy and guide policy.
These chapters provide evidence and lessons to apply to the
next recession.
Contributors
Anna Aizer • Robin Brooks • Tomaz Cajner
Gabriel Chodorow-Reich • Wendy Edelberg • Laura Feiveson
Jason Furman • Peter Ganong • Tim Geithner • Michael Gelman
Kristopher Gerardi • Fiona Greig • Laurie Goodman • Ben Iverson
Christopher Kurz • Lauren Lambie-Hanson • Pascal Noel
Claudia Persico • Jonathan Pingle • Louise Sheiner
Melvin Stephens, Jr. • Daniel Sullivan • Adi Sunderam
Stacey Tevlin • Joseph Vavra • Susan Wachter
David Wessel • Paul Willen

Lessons Learned From The US Economic Policy Response To Covid-10

Uploaded by

Copyright:

Available Formats

Lessons Learned From The US Economic Policy Response To Covid-10

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lessons Learned From The US Economic Policy Response To Covid-10

Uploaded by

Copyright:

Available Formats

RECESSION

Lessons Learned from Support to

policymakers may be tempted to return to these programs in the hope of

Background on Business Performance

Actual and Projected Macroeconomic

2020 2021 2020 2021

Source: Bureau of Economic Analysis n.d.; Bureau of Labor Statis-

Unemployment Rate and Job Vacancies

1948 1960 1970 1980 1990 2000 2010 2021

B. Total Vacancies, 2000–2021

Source: Bureau of Labor Statistics n.d.; authors’ calculations.

Business Bankruptcies and Unemployment,

Source: Bureau of Labor Statistics n.d.; U.S. Courts Bankruptcy

immediate onset of the pandemic, well before government support programs

Year-Over-Year Change in Business Bankruptcy

Source: U.S. Courts Bankruptcy Filings Statistics n.d.

alternative indicators of business exit (e.g., paycheck issuance and phone-track-

temporary or permanent. Given these difficulties, it will likely be necessary to

Framework for Evaluation

We instead start by highlighting conditions under which governments should

UI—can alternatively target the decline in consumption by laid-off workers,

Summary of Major Programs

Distribution of Major Business Aid Programs,

Source: Bureau of Economic Analysis n.d.; Federal Reserve n.d.;

Receipt of Paycheck Protection Program, by

Dollar amount (billions)

Source: Small Business Administration n.d.a.

A second approach uses the 500-employee threshold as a natural experi-

loan activity until Congress appropriated an additional $310 billion on April 24

SBA Forbearance and the EIDL Program

Economic Injury Disaster Loans, by Loan Size

Dollar amount (billions)

Source: Small Business Administration n.d.b.

Other Federal Subsidies

Federal Reserve Programs

The Corporate Credit Facilities

Corporate Credit Spreads and Select Policy

6/15: SMCCF purchases to track index

Source: ICE/Bank of America n.d.

Consistent with Figure 4.7, academic studies, including Haddad, Moreira,

The Main Street Lending Program

Bank Stock Prices and Regulatory Capital

B. Bank Regulatory Capital Ratios

Source: Nasdaq n.d.; Federal Reserve Board n.d.

State and Local Programs

Other Factors Impacting Businesses

Quarterly Net Income and Financing for

300 Net Change

Source: Compustat n.d.

Sales and Operating Costs for Nonfinancial

Source: Compustat n.d.

Net Financing for Nonfinancial Public Firms,

Source: Compustat n.d.

Sales for Nonfinancial Public Firms, Indexed to

Global financial crisis

Source: Compustat n.d.

Survey of Professional Forecasters. 2020. “Second Quarter 2020 Survey of Profes-

Recession Remedies examines and evaluates the breadth of

You might also like