Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
David Arseneau is an assistant director and Skander Van den Heuvel an associate
director in the Federal Reserve Board’s Division of Financial Stability. Molly Mahar
is a senior associate director in the Board’s Division of Supervision and Regulation.
José Fillat is a senior economist and policy advisor in the Federal Reserve Bank
of Boston’s Research Department. Donald Morgan is a nancial research advisor
in the Federal Reserve Bank of Ne w York’s Research and Statistics Group. Emails:
david.m.arseneau@frb.gov, jose.[email protected].org, molly.e.maha[email protected]v,
don.morgan@ny.frb.org, skander.j.vandenheuve[email protected]v.
e views expressed in this article are those of the authors and do not necessarily reect
the position of the Federal Reserve Bank of New Yor k, the Federal Reserve Bank of Boston,
or the Federal Reserve System. To view the authors’ disclosure statements, visit https://
www.newyorkfed.org/research/epr/2022/epr_2022_MSLP_arseneau.html.
I
n March 2020, it became clear that the COVID-19
pandemic would cause widespread economic disruptions
that would harm many U.S. businesses and households.
Moreover, there was acute uncertainty about the duration
and ultimate severity of the economic and nancial harm.
Many businesses with the ability to draw down on their
existing credit lines did so—either to cover revenue shortfalls
or to boost cash holdings as a precautionary measure. At the
same time, banks appeared to be tightening the supply of
new credit in response to the resulting uncertainty.
ese conditions motivated the Federal Reserve and the
Department of the Treasury to create the Main Street
Lending Program (Main Street), rst announced at the end
of March 2020. As one of several credit facilities set up in
response to the pandemic, Main Street was intended in par-
ticular to help those businesses that were too small to benet
from the Federal Reserve’s corporate credit programs but too
large to qualify for the loans and grants available through the
Paycheck Protection Program (PPP). Filling that support gap
was uniquely challenging because the targeted rms depend
• The Main Street Lending
Program was created in 2020
to support credit to small and
medium-sized businesses
and nonprots harmed by the
COVID-19 pandemic, partic-
ularly those in the “missing
middle” in terms of size, which
were unsupported by other
pandemic-response programs.
• The facility marked the
Federal Reserve’s most direct
involvement in the business
loan market since the 1930s
and 1940s and operated by
buying 95 percent participa-
tions in loans from lenders
and sharing the credit risk
with them.
• Ultimately, Main Street
supported more than 2,400
borrowers and co-borrowers
across the U.S. with loans
totaling $17.5 billion, the most
of any Federal Reserve credit
purchase facility.
• This article describes the
facility’s goals and design, the
challenges and constraints it
faced, and the characteristics
of its borrowers and lenders. It
also offers lessons learned for
future policymakers and facility
designers.
OVERVIEW
T M S L
P
David Arseneau, José Fillat, Molly Mahar, Donald Morgan, and Skander
Van den Heuvel
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
primarily on bank loans (versus bonds) that are highly dierentiated (“bespoke”) and largely
untraded. Reaching that corner of credit markets required an entirely new type of credit facil-
ity built from the ground up. It was also, incidentally, the Federal Reserve’s most direct
intervention in the bank loan market since it lent directly to businesses briey in the 1930s and
1940s (Sablik 2013). Despite the challenges, Main Street wound up supporting more than 2,400
borrowers and co-borrowers across the United States with loans totaling $17.5 billion, the most
of any Federal Reserve credit purchase facility.
1
is article tells the story of Main Street so far. We rst revisit the credit conditions in
spring 2020 that motivated the decision by the Federal Reserve and the Treasury to embark on
such a program. Second, we describe how Main Street was designed to support credit supply by
purchasing loan participations from banks and other lenders and sharing credit risk with
them. ird, we analyze the reach of Main Street, including take-up, characteristics of borrow-
ers and lenders, and factors that likely limited its take-up, such as certain program features and
much weaker loan demand aer a surge in the spring. We conclude with some lessons learned
for future policy makers and facility designers. We caution that some of these lessons are pre-
liminary, since most Main Street loans are still outstanding.
. B C C S
A crucial goal of Main Street was to reach the “missing middle” of rms—those too large for
PPP support but too small to benet from the Federal Reserve’s support of the corporate bond
market.
2
Tens of thousands of U.S. rms have more than 500 employees (the PPP cuto) but
are not rated to issue bonds; these rms instead depend on banks (or other intermediaries) for
credit.
3
So, the story of Main Street begins with bank credit conditions in the spring of 2020. By
most indications, bank credit was tight, with rms demanding additional credit at the same
time that banks were contracting supply. And since the missing middle depends on banks, the
apparent crunch would likely aect them most.
e need for credit was suggested by the remarkable, if temporary, surge in bank business
lending in the spring (Chart 1). Commercial and industrial loans on banks’ books rose by over
a half a trillion dollars in the rst few months of the pandemic. e Federal Reserve’s Senior
Loan Ocers Survey (SLOOS) also indicated increasing demand for loans at the time. e
surge in demand was important in motivating Main Street, but the eventual reversal gures
later in how Main Street played out.
Much of this borrowing reected rms drawing against their credit lines with banks.
4
Most
large, corporate rms have committed lines from a bank for working capital and to back their
commercial paper. ose rms switch betweeen bank and public debt according to which is
cheaper; they are not very bank-dependent because they have alternatives. In contrast, more
detailed, firm-level data suggested at the time that some of the credit needs of smaller rms
might be going unmet, despite the surge in total credit. As shown in Chart 2, commitment bor-
rowing by rms with less than $5 billion in annual sales (the eventual revenue cuto at Main
Street) grew notably more slowly than for larger rms above that cutoff.
5
At the same time loan demand was increasing, banks appeared to be contracting supply. e
SLOOS revealed that banks raised risk premiums (Chart 3, le panel) and tightened standards for
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
3,500
C & I loans, all commercial banks (billions of dollars)
3,000
2,500
2,000
1,500
1,000
500
Sept. 8
2006 2009 2012 2015 2018 2021
C
Business Loans at Banks Surged in the Spring of 2020
Source: Federal Reserve H.8 Statistical Release, “Assets and Liabilities of Commercial Banks in the United
States.”
Notes: C&I is commercial and industrial. Shaded areas indicate periods designated as recessions by the
National Bureau of Economic Research.
Year-over-year quarterly change (percent)
80
60
20
40
0
–20
–40
–60
–80
2015 2017
More than $5 billion $1 billion to $5 billion
2019 2021
Q2
$500 million to $1 billion Less than $500 million
C
Lower Commitment Borrowing by Firms with Less Than $5 Billion in Revenues
Source: Federal Reserve Report FR-Y14Q, “Capital Assessments and Stress Testing.”
Notes: The chart shows growth in utilized commitments of commercial and industrial (C&I) loans to U.S.
nonnancial rms that report revenue (FR Y-9C category 4).
new loans (right panel) in the rst half of 2020. “Standards” includes the sorts of loan terms, such as
covenants and collateral requirements, that distinguish loans from less bespoke (“vanilla”) bonds.
While the net fraction of banks that reported tightening credit was about equal for rms of all
sizes, it is important to note that bank-dependent rms would be more aected than larger rms
with access to public debt markets, supported by the Federal Reserve’s corporate facilities.
6
e
SLOOS in the spring of 2020 also revealed that banks were tightening primarily because of the “less
favorable or more uncertain economic outlook” and “reduced tolerance for risk.” While not
surprising, that risk aversion and uncertainty informed the design of Main Street.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
2006 202120182015201220092006 20212018201520122009
Q2
Q2
Large and middle-market firms Small firms
Net percentage of banks raising risk premiums Net percentage of banks tightening standards
100
80
60
20
0
40
–20
–40
–60
–80
–100
80
100
60
20
0
40
–20
–40
–60
–80
–100
C
Banks Tightened Credit Supply in the Spring of 2020
Source: Federal Reserve, Senior Loan Ocer Opinion Survey on Bank Lending Practices.
Notes: Large and middle-market rms are those with annual sales above $50 million. Shaded areas indicate
periods designated as recessions by the National Bureau of Economic Research.
It was this picture of surging demand and contracting supply in the spring of 2020 that led
the Federal Reserve to declare its intention to create a program to support credit to small and
medium-sized rms.
7
e actual program that emerged in the second half of 2020 is the topic
of the next section.
. T D M S
Designing Main Street was a complex undertaking. is section describes the overall objectives
of Main Street, the structure of the program, including key considerations that shaped its
design, and its implementation. As policymakers set out to design the program, they focused
on creating facilities that would make credit available to a suciently wide scope of rms
aected by the pandemic but, at the same time, limit risk to taxpayers. While a number of
policy, legal, and operational considerations shaped the program, the need to strike this careful
balance underpinned all key design decisions.
2.1 Program Objectives and Key Considerations
With Main Street, the Federal Reserve and the Treasury sought to provide credit support to
small and medium-sized businesses and nonprots impacted by the pandemic. e goal of
Main Street was to assist businesses and nonprots that faced credit constraints but were in
sound nancial condition prior to the pandemic and had good post-pandemic prospects, so
that they were in a position to benet from—and be able to repay—a loan. As already noted,
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
Main Street was intended to complement the Federal Reserve’s corporate credit and municipal
lending facilities that were launched to support larger businesses, states, and municipalities.
Several characteristics of the market for loans to small and medium-sized businesses high-
lighted above created challenges. ese loans are not traded like bonds or securitized like
mortgages; such markets (which tend to bring infrastructure, ratings, real-time prices, and a
degree of standardization) could otherwise have provided convenient on-ramps for program
design. Moreover, loans to small and medium-sized rms are some of the more individually
tailored (bespoke) nancial contracts—more bespoke than traded bonds or residential mort-
gages. Owing to the importance of relationship lending for these businesses, policymakers
were le without a readily available, standardized set of loan terms or credit metrics that could
easily be converted into a program term sheet and quickly scaled for thousands of businesses.
Additionally, while it was dicult to predict the scale and scope of demand for the program
from the outset, conditions in the spring of 2020 pointed to large potential demand. Although
the Federal Reserve is very experienced in credit analysis for its supervision and monetary
policy functions, it would have needed to hire a large number of loan ocers to directly origi-
nate and process loans to the thousands of companies that could potentially have qualied.
Hiring such personnel quickly and in sucient numbers from the banking sector, which was
itself facing unprecedented demand for loans, was impractical—thus necessitating a role for
private lenders. e swi onset of the pandemic and the fact that the Federal Reserve lacked
previous experience setting up a small and medium-sized business credit support program
also created design complications.
e program was authorized under Section 13(3), as amended, of the Federal Reserve Act
and was capitalized, in part, by funds appropriated under the CARES Act; each act inuenced
the specic design of the Main Street facilities. Section 13(3) provides lending authority but
prohibits loans to “insolvent” borrowers and requires that the lending Reserve Bank be
“indorsed or otherwise secured” to its satisfaction.
8
(See Box 1 for a brief history of Federal
Reserve credit policy directed specically to businesses under Section 13(3).) e application
of the CARES Act set forth eligible borrower criteria and placed limits on borrowers’ ability to
distribute capital or set compensation above given thresholds.
9
2.2 Program Design
With these economic, operational, and legal considerations as a backdrop, policymakers at the
Federal Reserve and the Treasury settled on a loan participation program to support the supply of
credit. Banks would be able to sell 95 percent stakes in eligible loans at par to the Main Street special
purpose vehicle (SPV), with the credit risk shared between the SPV and lenders pro rata.
e loan participation model was chosen for three reasons. First, it leveraged lenders’ exist-
ing infrastructure for originating, monitoring, and servicing loans as well as their expertise in
assessing and controlling risk—expertise that is oen local and specialized.
Second, because the participation model transferred the bulk of the loans and associated risks
from the lenders to the Main Street program, it helped mitigate the acute economic uncertainty
and risk aversion that was driving the tightening credit supply in the spring. As an added benet,
removing 95 percent of the loan amounts from banks’ balance sheets would also free up bank
capital to recognize losses and maintain lending outside the Main Street program.
10
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
B
The Federal Reserve’s Historical Experience with Direct Lending to Businesses
e Main Street program represented the rst time since World War II that the Federal Reserve ac-
tively pursued policies to direct bank lending to the nonnancial business sector. e origins of the
Fed’s previous experience with direct lending to businesses traces back to the addition of Section
13(3) to the Federal Reserve Act, which occurred during the Great Depression.
a
In January 1932, legislation was passed to create the Reconstruction Finance Corporation
(RFC), which was designed to make short-term loans to banks and other nancial institutions,
collateralized by real bills (short-term debt from businesses). e creation of the RFC was a means
of injecting capital into the weakened banking system; however, the RFC’s ability to extend loans
outside of the banking system was limited. Recognizing this, Congress passed a bill in the summer
of 1932 that added Section 13 paragraph 3 to the Federal Reserve Act.
Congress further expanded the lending authority of the Federal Reserve by adding Section
13(b) to the Federal Reserve Act in June 1934.
b
Section 13(b) allowed Reserve Banks to directly
extend loans to businesses within their districts for periods of up to ve years. It also gave the
Reserve Banks the ability to participate in loans with lending institutions, provided those lending
institutions retained 20 percent of the risk of the loan. In contrast with Main Street, no limitations
were placed on the size of an individual loan. is Great Depression–era facility was funded in
equal part by the surplus of the Reserve Banks as of mid-1934 and the Treasury. All told, nearly
$280 million ($5.4 billion in 2020 dollars) was made available for Reserve Bank lending, with each
of the twelve Districts being apportioned a partial amount of the total. Relative to the overall size
of the economy, this quantity of funding was about 0.5 percent of GDP in 1934. In comparison,
Main Street’s capacity as a share of 2020 GDP was about six times as large.
c
By May 1935, roughly a year aer the passage of Section 13(b), the Federal Reserve System
had approved 961 loans issued directly to businesses totaling $43.9 million ($847.9 million in
2020 dollars). Interestingly, as a share of contemporaneous GDP, this uptake is nearly identical to
Main Street’s. Because each Reserve Bank had access to funds, lending was, by design, dispersed
geographically across all twelve Districts. In addition, the loans went to a broad range of industries,
including construction, lodging, manufacturing, mining, transportation, and wholesale and retail
trade—many of the same industries that took Main Street loans.
d
All told, loan volume peaked at
about $60 million by the end of 1935 ($1.2 billion in 2020 dollars). With peak volume amounting
to more than 15 percent of the total funds available, utilization was much higher relative to Main
Street.
e
Main Street’s lower utilization likely reects that it operated for only about six months, and
also that the program designs diered notably.
e Federal Reserve’s lending activity to nonnancial businesses gradually declined aer 1935 as
expanded lending through the RFC made direct loans from the Federal Reserve less attractive.
(C )
a
For an extensive treatment of this history see Hackley (1973).
b
For useful summaries of the history of Section 13(b), see Fettig (2002) and Sablik (2013).
c
With Treasury’s equity commitment and the SPV’s leverage cap, up to $600 billion was potentially available
through Main Street, about 3 percent of the size of the $20.9 trillion U.S. economy in 2020.
d
See Sablik (2013) for more details on the industry composition of 13(b) loans as of mid-1935.
e
Main Street loan volume totaled $17.4 billion at the end of 2020, about 3 percent of the $600 billion in
total available funding.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
B (C)
Section 13(b) remained in place and, in fact, activity peaked again in 1942 when the Federal
Reserve was called upon to make industrial loans during World War II. e role of the Federal
Reserve in allocating credit to businesses remained a hotly debated issue throughout the 1950s, but
ultimately Section 13(b) was repealed in 1958. e 13(3) powers, however, remained part of the
Federal Reserve Act and played an important role in implementing Main Street in response to the
COVID-19 pandemic.
Third, the participation model allowed for an appropriate balance between reach and
risk. The substantial risk-bearing by the Federal Reserve promoted reach, while the resid-
ual bank risk-bearing maintained some economic incentives for lenders to control risk. To
complement these incentives and further minimize the risk of adverse selection—the pos-
sibility that banks would offload their worst new loans to Main Street—the Main Street
program also limited borrower leverage and imposed requirements for priority and
collateral.
e program was executed through an SPV set up by the Federal Reserve Bank of Boston
that was funded with a loss-absorbing tranche of Treasury equity (that is, CARES Act funds),
as well as loans from the Reserve Bank. Given the widespread uncertainty at launch, Main
Street was created with a sizable maximum capacity of up to $600 billion in participations, in
case that much support would be needed.
Main Street ocially began purchasing loan participations on July 6, 2020. It oered to pur-
chase participations in three distinct types of loans—new loans, priority loans, and expanded
loans—through three separate facilities: Main Street New Loan Facility (MSNLF), Main Street
Priority Loan Facility (MSPLF), and Main Street Expanded Loan Facility (MSELF), respec-
tively. While certain terms were common across all three loan types, there were also important
dierences, including loan size, permissible leverage levels, and collateralization requirements
to accommodate a range of borrower and lender circumstances. e term sheets were posted
for public feedback and were adjusted in response to such feedback several times, both before
and aer the start of operations, as discussed below. e nal loan terms for the for-profit facil-
ities are shown in Table 1.
Loan terms
While the terms for small and medium-sized business loans are generally tailored to the facts and
circumstances of the borrower, some Main Street loan terms were standardized to allow the
program to function while balancing reach and risk. For example, standardized interest rates and
loan maturities enabled Main Street to purchase participations at par without the need to develop
a complex loan pricing model. An interest rate of LIBOR plus 300 basis points with zero prepay-
ment penalty implemented the Regulation A requirement that Federal Reserve emergency
lending be extended at a suciently high rate of interest relative to non-stressed conditions to
provide an incentive for rapid repayment when conditions normalize. In keeping with the
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
objective of helping borrowers bridge the pandemic, Main Street loans were given an amortiza-
tion schedule that back-loaded loan repayment, deferral of interest and principal payments for
a year (principal payments were later deferred for two years), and a five-year loan term. e
deferral was intended to alleviate short-term nancial strain on Main Street borrowers.
Lenders had discretion over loan size up to a limit, either a nominal dollar limit or a lever-
age limit, whichever was smaller. e leverage limit, which turned out to be more binding, was
a primary mechanism for limiting risk to the program. When added to the borrower’s existing
debt, the Main Street loan could not exceed four (MSNLF) or six (MSPLF, MSELF) times the
borrower’s 2019 adjusted earnings before interest, taxes, depreciation, and amortization
(EBITDA). In addition to limiting the size of Main Street loans for participants, these leverage
limits also had the eect of excluding some highly levered or unprotable rms altogether.
e choice to use 2019 EBITDA was motivated by the program’s goal to help borrowers that
were temporarily suering from the pandemic but that had been fundamentally solvent prior
to the onset of the pandemic.
In addition to the leverage limits and the lender’s risk retention, the tradeo between risk
and reach was also managed through security and priority requirements. All Main Street loans
were prohibited from being contractually subordinated to any existing borrower debt in terms
of priority in bankruptcy. While priority and expanded loans allowed higher leverage than new
loans, they were required to be senior to, or pari passu with, all existing borrower debt in terms
of collateral securing the loans, except for mortgage debt (as dened by the program). Lenders
were ultimately responsible for determining that borrowers were in sound condition prior to
the crisis and had strong post-pandemic prospects that would enable repayment of the Main
Street loan.
T
Key Main Street Loan Terms of Facilities for For-Prot Borrowers (Final Terms)
New Loan Facility Priority Loan Facility Expanded Loan Facility
Loan term 5 years
Principal payments
Principal deferred for two years.
Years 3-5: 15 percent, 15 percent, 70 percent, respectively
Interest payments Deferred for one year
Interest rate 1- or 3-month LIBOR + 3 percent
Loan size
$100,000 to
$35 million
$100,000 to
$50 million
$10 million to
$300 million
Maximum combined debt to adjusted
2019 EBITDA (including principal
amount of Main Street loan)
4 times 6 times 6 times
Lender participation rate 5 percent
Federal Reserve participation rate 95 percent
Prepayment allowed Yes, without penalty
Business size limits 15,000 employees or fewer, or 2019 revenues of $5 billion or less
Source: Board of Governors of the Federal Reserve System. https://www.federalreserve.gov/monetarypolicy/
mainstreetlending.htm.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
Finally, the program allowed borrowers to renance existing debt, but only in a single facil-
ity, the MSPLF, and only debt owed to a dierent lender, to avoid the risk that lenders would
shi poorly performing debt on their own books to the program.
Borrowers
To target small and medium-sized businesses, the program limited eligibility to rms with
fewer than 15,000 employees or less than $5 billion in annual revenues (including aliates).
11
To help those businesses that lacked access to an alternative support program, these caps were
deliberately set above those used for the PPP (500 employees) or other Small Business Admin-
istration (SBA) lending (with size thresholds that vary by industry) but lower than the level at
which a company might generally have access to nancing in capital markets and thus be sup-
ported by the Federal Reserve’s corporate credit facilities. e aforementioned nominal loan
size limits, all well above the $10 million maximum for the PPP, played a similar role. In other
words, Main Street was intended to ll a gap in credit support for the “missing middle.”
In dening eligibility criteria, the Board also referenced the SBA’s exclusion of “ineligible
businesses”—a list of categories formulated especially to place reasonable limits on the types of
companies that could receive government-backed business lending.
12
is framework, particu-
larly the denition of ineligible business, was designed to mitigate fraud risk and limit evasion
of facility restrictions.
13
Further, Main Street program borrowers were subject to the require-
ments for participants in direct loan programs set forth in the CARES Act. In particular, a
borrower needed to commit to follow compensation, stock repurchase, and capital distribution
restrictions under Section 4003(c)(3)(A)(ii) of the Act. ese requirements would remain in
place until a year aer the Main Street loan was fully repaid.
Lenders
All Main Street facilities relied on private lenders and their existing underwriting infrastruc-
ture to apply appropriate expertise and enable the program to scale rapidly. In contrast to the
PPP, which allowed a broad set of eligible lenders to supply its forgivable loans, the Main
Street program limited eligible lenders to federally regulated and supervised organizations,
including banks and credit unions, to ensure that Main Street lenders’ underwriting stan-
dards and “know your customer”/anti-money laundering practices were subject to strong
and ongoing supervisory oversight.
14
While a wider set of eligible lenders might have
extended the reach of the program, the use of established and well-regulated banking orga-
nizations and credit unions was viewed as an important way to control potential taxpayer
risks in the program. As it turns out, virtually all of the participating lenders were commer-
cial banks (as we discuss later), so for brevity we will oen refer to eligible lenders simply
as “banks.”
Under the program terms, lenders were expected to underwrite Main Street loans using
their existing underwriting practices. Subsequent program guidance provided through FAQs
also claried supervisory expectations. Lenders were directed to underwrite Main Street loans
by looking at borrowers’ pre-pandemic nancial condition and post-pandemic prospects.
15
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
Lender incentives and the participation agreement
Several incentives for banks to participate were built into the program, since, to be successful,
Main Street required the active participation of lenders. First, as discussed above, the risk-sharing
with Main Street allowed banks to help existing and new customers without taking on much new
credit risk or needing to signicantly expand their own balance sheets. Second, to cover lenders’
loan origination and servicing costs and further boost incentives, lenders were able to benet
from fees: an origination fee of up to 1 percent (on the full principal) and an annual servicing fee
of 0.25 percent of the Main Street SPV’s loan share.
16
Given the banks’ limited initial investment,
these fees, together with banks’ 5 percent share in interest and principal repayments, in principle,
enabled a lender to receive reasonable returns even under the most adverse credit scenarios con-
sidered (discussed further below). at said, for loans with signicant origination or servicing
costs, the lender’s return would be lower. While data on origination and servicing costs are scant,
commercial and industrial (C&I) loan fees can be signicant, possibly suggesting that such costs
are also signicant. For example, in the market for syndicated term loans to businesses, upfront
fees (where observed) average about 80 basis points, with considerable variation around that
average (Berg, Saunders, and Steen 2016). In addition, lender incentives in the MSELF were
complicated due to interactions with the loan that was being expanded, including the possibility
that the collateral on the existing loan was diluted.
17
To operationalize the loan participation model, the Federal Reserve created a loan partici-
pation agreement based on market-standard models, with adjustments for certain features of
the program. e market-standard provisions were generally familiar to lenders that use par-
ticipations or engage in syndicated lending; this was intended to help smooth the on-ramp for
many potential lenders. While these documents were less familiar to the program’s smaller bor-
rowers, they played an important function in the program because their provisions were
generally viewed as facilitating a “true sale,” which (among other things) enabled lenders to
move 95 percent of the loan amounts o their balance sheets for purposes of bank capital
rules, thus promoting lender participation by freeing up regulatory capital.
In comment letters and outreach, lenders expressed concerns that the Federal Reserve
would “put back” nonperforming loans to the lenders by arguing that the loans were originated
imprudently. To alleviate such concerns and promote participation, the Federal Reserve added
a clause to the agreement preventing put-backs. e Federal Reserve also waived and dis-
claimed its rights to special priority in bankruptcy among unsecured lenders to enhance the
ecacy of the program and provide certainty to lenders and borrowers.
Income and loss projections during the design phase
Section 13(3) of the Federal Reserve Act and the CARES Act required that the Federal
Reserve’s investment be appropriately secured and that taxpayers be protected. Accordingly,
when deciding on loan terms, risk-sharing arrangements, and fees, the Federal Reserve and
Treasury had to gauge the eect of these choices on the potential gains or losses from Main
Street’s operations. To do so, sta projected bounds for the SPV’s net income under various
credit risk scenarios and design choices, akin to a stress test. Multiple scenarios, with varying
degrees of adversity, were used, both to ensure that the statutory taxpayer-protection
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
requirement would be satised under a range of adverse conditions and because at the time
that the program was being designed the economic outlook was extremely uncertain. e
appendix describes the scenarios and projections in more detail.
e results of these projections also guided the decision to cap SPV “leverage” at 8-to-1.
Given Treasury’s planned $75 billion equity investment, the net leverage cap dictated a
maximum program size of $600 billion. With that cap, even under adverse scenarios, the
Federal Reserve was projected to incur zero losses.
Infrastructure
Once the design was generally decided on, the next step was to build, from the ground up,
the technological infrastructure and risk control mechanisms needed to operate the
program. e loans in which Main Street would be participating could not simply be pur-
chased “in the market” as with the corporate credit programs, so the Federal Reserve Bank of
Boston (which operates the program) had to create an electronic portal through which
banks could register and submit loans for participation. To address the risk of fraud or pro-
cessing mistakes, multi-step processes that would verify lender registrations and loan
documents had to be developed. All told, building this infrastructure from scratch was a
complicated task given the lack of an existing blueprint, and this complexity slowed the
launch relative to other credit facilities implemented by the Federal Reserve or loan pro-
grams in other countries that were built on existing infrastructure. (See Box 2 for more
details on how other central banks and governments facilitated the ow of credit to small
and medium-sized businesses).
When submitting a loan, lenders uploaded the loan agreements and other relevant loan
documents to the portal. Automated eligibility checks were augmented by a manual review for
adherence to certain core program requirements; the review was done by Federal Reserve
Bank of Boston sta and hired vendors, including the Main Street credit administrator and
external counsel. Importantly, the SPV did not re-underwrite Main Street loans.
Additional program adjustments
In an eort to respond to the credit needs of nonprot organizations and smaller borrowers, a
need that became increasingly apparent in summer and early fall 2020, Main Street was
amended to introduce two facilities for small and medium-sized nonprot organizations—the
Nonprot Organization New Loan Facility (NONLF) and the Nonprot Organization
Expanded Loan Facility (NOELF) —and to enable the facilities’ participation in smaller loans.
e nonprot sector was hit particularly hard by the social-distancing requirements put in
place to slow the pace of the pandemic. Demands for their services (for example, care for
COVID-19 patients, online learning, and social services) spiked at the same time key sources
of income (such as elective surgical procedures, tuition, and donations) declined or were at risk
of declining. Designing a program for this sector presented additional challenges, given that
many nonprots were designed to minimize rather than maximize earnings, making it dicult
to meet the program’s pre-pandemic leverage thresholds, and many had limited experience
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
B
Lending Programs to Support Nonfinancial Businesses During the Pandemic:
The International Experience
The COVID-19 pandemic had a significant effect on small- to medium-sized businesses not
only in the United States but in countries throughout the world. Accordingly, an important
aspect of the policy response in many countries involved creating lending programs, some of
which were similar to Main Street, to support the flow of credit to households and nonfinan-
cial businesses.
e most similar international programs were the Bounce Back Loan Scheme (BBLS) and
the Coronavirus Business Interruption Loan Scheme (CBILS), both implemented in the United
Kingdom, and the Prêt Garanti par l’État (PGE), implemented in France. In the broadest sense,
the intent of these programs was to facilitate lending to nonnancial businesses that were hit
hard by the pandemic and that, absent support, could potentially be forced to reduce employ-
ment and economic activity.
a
One common feature of all three of these programs is that the
loans were either fully or partially backed by government guarantees of repayment in the event
that the borrower defaults. is feature signicantly reduces the amount of exposure a bank fac-
es and, as a result, makes participation more attractive. In contrast, the strong desire to protect
taxpayers by not guaranteeing loans made the Main Street program dierent from the BBLS, the
CBILS, and the PGE.
Beyond these lending programs, many central banks acted unilaterally (that is, not in con-
junction with the country’s Treasury or the Ministry of Finance) to promote credit to certain
segments of the credit market. In this regard, the most common policy response was to establish
a funding-for-lending scheme, whereby the central bank provides low-cost funding to banks
that then use those funds to extend loans to a targeted set of borrowers (small and medi-
um-sized enterprises, or SMEs).
b
Examples of targeted funding-for-lending programs intro-
duced by foreign central banks include those implemented by the Bank of England (the Term
Funding Scheme with Additional Incentives for SMEs), the Bank of Japan (the New Fund-Provi-
sioning Measure to Support Financing Mainly of Small and Medium-Sized Firms), the European
Central Bank (the modied Targeted Longer-Term Renancing Operations III), the Reserve
Bank of Australia (the Term Funding Facility), and the Sveriges Riksbank (Loans to Banks for
Onward Lending to Companies).
e Main Street Lending Program is very dierent from a funding-for-lending scheme. In the
simplest terms, the dierence boils down to what creates the incentive for a participating bank to
increase lending to a targeted set of borrowers. Under a funding-for-lending scheme this incentive
comes from low-cost funding provided by the central bank, while under Main Street it comes from
the opportunity to originate a loan and sell a large portion of the risk to the Federal Reserve while
still retaining the servicing rights.
a
See Briggs and Walker (2020) for a fuller discussion.
b
See Cantu et. al., (2021) and Cavallino and DeFiore (2021).
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
managing longer-term debt. e terms of the nonprot facilities sought to balance these chal-
lenges by setting out dierent and additional eligibility requirements to capture those for
which a loan product would be most beneficial.
18
Similarly, policymakers received repeated feedback during the life of the program that some
small businesses and nonprots would benet from a loan smaller than the minimum size per-
mitted originally. In response, the program was adjusted to allow for loans as low as $100,000
in the MSNLF, MSPLF, and NONLF. e program fees were also adjusted upward for the
smallest loans, in order to compensate lenders for the proportionally larger potential cost asso-
ciated with originating small loans.
. M S A
Over its six-month run, Main Street purchased 1,830 loans with a combined principal amount
of $17.5 billion, more than any of the Federal Reserve’s other debt-purchase programs. Its
volume, although small relative to capacity, was a meaningful addition to the ow of credit—
roughly comparable, for example, to the amount of lending by the largest banks (those with
consolidated assets greater than $100 billion) over the second half of 2020 to borrowers with
similar characteristics, that is, within the eligibility parameters but outside the Main Street
program. is section describes in detail Main’s Street activity and its limits, including loan,
lender, and borrower characteristics.
A look at the portfolio yields the following high-level observations. e average loan was
$9.5 million, substantially larger than the average PPP loan, suggesting the program supported
rms too large for PPP loans. Loan size was oen dictated by the program’s leverage limits
dened above (of four and six times EBITDA). e lenders were nearly all commercial banks.
Most active lenders were in the $250 million to $10 billion asset-size range, although the
largest banks (those with assets of more than $1 trillion) also participated to some extent. e
program’s reach was wide, with borrowers from nearly all states, and state-level activity tended
to correlate positively with COVID-19 cases and increases in a state’s unemployment rate. Bor-
rowers were, on average, somewhat riskier than the typical borrower found in the portfolios of
the largest banks, possibly reecting dierences between the types of borrowers that seek loans
from the largest banks and those that seek loans from other banks (that is, those with assets of
less than $1 trillion).
4.1 Overall Activity
e program began accepting participations on July 6, 2020, and ended on January 8, 2021.
Activity grew modestly but steadily until early December, when it surged in advance of the
December 14 deadline for submitting new participations (see Chart 4, le panel). Roughly
half of the overall volume of the program occurred in the nal month of the program.
19
All
told, the late surge in loan purchases pushed Main Street’s volume above that of any debt
purchase (versus liquidity) facility created by the Federal Reserve during the pandemic
(right panel).
20
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
Main Street loans also constituted a meaningful addition to the overall ow of credit during
the program’s active phase. As shown by Bräuning and Paligorova (2021), the cumulative
volume of Main Street lending was about 60 percent of the volume of term loans originated
during the same time span by large banks (FR Y-14Q lers) to borrowers of similar size and
leverage (that is, borrowers with less than $5 billion in annual revenues and leverage below six
times EBITDA). Moreover, when considering smaller rms (those with less than $50 million in
EBITDA), Main Street lending substantially exceeded the supply of credit by the largest banks
to borrowers of comparable size. When also imposing the six times EBITDA leverage limit in
the Y-14Q data, Main Street lending was about twice as large as lending by the largest banks to
comparable borrowers.
21
At the same time, Main Street volumes were low when compared with the surge in C&I
lending from credit line drawdowns in March 2020, or when compared with the maximum
capacity of the program, as noted. In part, this reected much weaker loan demand aer the
launch of the program in July 2020, as discussed in Section 2. Reach was likely also constrained
by certain program features, a theme we return to below.
Table 2 summarizes the Main Street purchases by loan type and size. e bottom line shows that
priority loans and new loans turned out to be more in demand than expanded loans. e 1,173 pri-
ority loans accounted for nearly three-quarters (74 percent) of total volume while the 616 new loans
made up 15.5 percent. e 26 expanded loans accounted for the balance. As stated above, expanded
loans entailed modifying existing credit agreements, which may have reduced demand for these
loans. e Nonprot New Loan Facility (NONLF) was very small both in the number and volume
of loans, and the Nonprot Extended Loan Facility (NOELF) was not used at all.
Table 3 summarizes the size distribution of loans made across the dierent facilities. Most
loans were in the range of $1 million to $50 million, with an average of $9.5 million and
median of about $4 million. In comparison, the average PPP loan was just $101,000, suggest-
ing that Main Street succeeded in targeting rms that were too large for the PPP but too
small to access the bond market. At the program’s inception, the minimum loan size was
0
5
10
15
20
MSELF NONLFMSPLF MSNLF
–5
0
5
10
15
20
25
Main Street
Corporate bonds
Municipal bonds
Commercial paper
TALF II
2020 2021
Sep 14
Aug 17
Jul 20
Billions of dollars
Total Volume of Accepted MSLP
Submissions by Facility
Asset Purchases at Federal Reserve
Credit Facilities
Billions of dollars
Nov 9 Dec 7
Oct 12
C
Loan Purchases at Main Street and Other Credit Facilities
Sources: MSLP data; Federal Reserve, H.4.1 Statistical Release, “Factors Aecting Reserve Balances.”
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
$250,000, but this threshold was lowered to $100,000 for certain facilities on October 30 to
better target support for small businesses. ere were, however, only 22 loans smaller than or
equal to $250,000 at the end of the program. On the other end of the size distribution, there
were a small number of loans made through the MSELF that were larger than $50 million,
together totaling $1.5 billion—almost 10 percent of the overall Main Street volume. e
largest loan made through this facility was $300 million, the maximum loan size for
expanded loans.
4.2 Borrower Characteristics
Altogether, 2,453 borrowers and co-borrowers took out a total of 1,830 loans.
22
Table 4 proles
borrowers in terms of revenue, leverage, and assets as of 2019. e average revenue was
T
Loan Volume (in Millions) and Count, by Loan Type and Size
Expanded Loans New Loans Priority Loans Nonprot Loans Total
Loan Size
(Dollars)
Volume Count Volume Count Volume Count Volume Count Volume Count
≤250K 4 19 0.3 2 0.2 1 5 22
250-500K 26 68 12 28 0.9 2 39 98
500K-1M 95 118 65 82 1 2 161 202
1-10M 20 2 1,221 350 3,034 671 40 10 4,314 1,033
10-35M 238 10 1,349 61 5,809 304 7,396 375
35-50M 81 2 3,997 86 4,078 88
>50M 1,466 12 0 0 1,466 12
All loans 1,805 26 2,695 616 12,917 1,173 42 15 17,459 1,830
Source: Authors’ calculation using MSLP and Call Reports data.
Note: Entries may not sum to total due to rounding.
T
Main Street Loan Size Distribution, by Type
Loan Size (in Millions)
Mean Min p10 p25 p50 p75 p90 Max
Expanded loans 69.4 10.0 11.0 22.0 40.5 90.0 148.0 300.0
New loans 4.4 0.1 0.4 0.8 2.0 4.5 10.0 35.0
Priority loans 11.0 0.1 1.1 2.4 6.0 14.8 30.0 50.0
Nonprot loans 2.8 0.2 0.4 0.6 2.5 5.0 5.0 8.5
All facilities 9.5 0.1 0.7 1.5 4.0 10.6 25.0 300.0
Source: Authors’ calculation using MSLP and Call Reports data.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
T
Main Street Borrower Financial Characteristics
Metric Mean p25 p50 p75 Count
2019 Revenue (millions of dollars) 33.9 3.9 11.5 31.8 1,830
2019 Leverage (multiple of EBITDA) 1.1 0.0 0.6 1.8 1,830
Assets (millions of dollars) 26.2 1.5 6.3 21.6 1,830
Source: Authors’ calculation using MSLP and Call Reports data.
$33.9 million. e pre-pandemic levels of leverage were relatively low, with the average being
just above one multiple of EBITDA. Borrowers’ average asset size was $26.2 million, consistent
with the program’s target of reaching medium-sized rms. Moreover, Main Street borrowers
saw an average revenue decline of about $7 million during the rst two quarters of the pan-
demic, relative to their most recent pre-pandemic reporting in 2019. is illustrates that Main
Street helped many borrowers that were hit hard by the pandemic but were solvent and viable
businesses before the crisis started.
Main Street supported borrowers across a diverse range of industries (see Table 5). e top
industries by loan volume were accommodation and food services; manufacturing; real estate
and rental and leasing; mining, quarrying, and oil and gas extraction; and transportation and
warehousing. By loan count, professional services was second to accommodation and food ser-
vices, with manufacturing third and construction rms fourth. e least active industries in
terms of both loan volume and counts were utilities, agriculture and forestry, and public
administration.
The geographic reach of Main Street borrowers was also wide, with borrowers in nearly
every state. The most active states by volume were Texas ($3.1 billion), Florida
($2.1 billion), California ($2.1 billion), Ne w Yor k ($700 million), and Missouri
($700 million).
23
It is also useful to look at loan volumes relative to state GDP, as shown in
Exhibit 1. Using this normalization, the top five states were Oklahoma, Arkansas, Missouri,
Florida, and Texas.
Chart 5 provides further evidence suggesting that Main Street reached borrowers in indus-
tries and regions that were hit hard by the pandemic. e le panel shows that 72 percent of
total Main Street lending went to COVID-affected industries.
24
e right panel shows that
state-level Main Street borrowing in any month was positively correlated with the previ-
ous month’s COVID-19 positivity rate in the borrower’s state, controlling for state GDP per
capita and time xed effects.
25
4.3 Lender Characteristics
A total of 643 lenders successfully registered to participate in the Main Street Program, all but
27 of which were commercial banks. at represents about 1 in 7 of all FDIC-insured banks, a
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
meaningful share for a six-month program. About half of these banks (316) sold loans to Main
Street, while 327 did not actively participate despite being registered.
Chart 6 shows that Main Street lending activity was dominated by banks that were small to
medium-sized in terms of total assets. Most active banks were in the $250 million to
$750 million range or the $1 billion to $50 billion size range (le panel). e share of registered
lenders increases with each size group (right panel, blue and gold portions of the bars). Very
small banks (less than $250 million in assets) were underrepresented.
Chart 7 shows lending intensity by bank asset size. Banks in the $1 billion to $10 billion
asset-size category account for 34 percent of the total number of loans and 34 percent of the
total volume of loans; banks in the $10 billion to $50 billion asset-size group account for
29 percent of loans and 21 percent of volume; and banks in the $250 million to $500 million
and $500 million to $750 million size groups together account for 17 percent of loans and
volume. While the volume of Main Street loans issued by banks with assets of $1 billion or
T
Main Street Borrowers by Industry
Industry
Volume
(Millions of Dollars)
Percentage of
Volume
Loan
Count
Percentage of
Loan Count
Accommodation and food services 2,182 12.5 268 14.6
Manufacturing 1,711 9.8 169 9.2
Real estate 1,659 9.5 141 7.7
Mining, oil and gas extraction 1,468 8.4 90 4.9
Transportation 1,397 8.0 107 5.8
Arts and recreation 1,242 7.1 117 6.4
Professional services 1,159 6.6 171 9.3
Construction 1,132 6.5 166 9.1
Wholesale trade 961 5.5 112 6.1
Information 907 5.2 92 5.0
Health and social care 837 4.8 71 3.9
Administrative support services 796 4.6 60 3.3
Retail trade 618 3.5 92 5.0
Other services 352 2.0 64 3.5
Educational services 307 1.8 26 1.4
Finance and insurance 237 1.4 49 2.7
Management 230 1.3 18 1.0
Utilities 186 1.1 8 0.4
Agriculture and forestry 76 0.4 8 0.4
Public administration 1 0.01 1 0.1
Total 17,459 100 1,830 100
Source: Authors’ calculation using MSLP and Call Reports data.
Note: Entries may not sum to total due to rounding.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
more account for 55 percent of the total Main Street lending, these banks’ total assets represent
95 percent of the U.S. banking system’s assets.
Most banks active in the program sold just one or two loan participations (Chart 8). e
banks that sold multiple participations tended to sell fewer than 10, though several lenders
sold more than 20, and there were a few extremely active participants that sold more than 30
loans, suggesting once a lender had experience with the loan process, scale was possible.
Registered banks tended to have a higher concentration in C&I lending than nonregistered
banks, regardless of their size. e le panel of Chart 9 shows that dierences in C&I concen-
tration between registered and nonregistered lenders are signicant for any size bin. Moreover,
the panel on the right shows that the intensive margin is positively correlated with the concen-
tration. Banks that were more active in the Main Street program tended to have a higher
concentration in C&I lending measured before the pandemic.
4.4 Program Features and Take-up
Many of Main Street’s features were chosen to balance the tradeo between the reach of the
program and the riskiness of the loans made to borrowers. is section takes a very prelimi-
nary look at how the program performed in terms of striking that balance—preliminary since
the ultimate credit performance of the Main Street loans is not yet known.
Regarding determinants of reach, Table 6 shows that loan size was more oen limited by
the leverage cap than by the nominal maximum loan size. About 30 percent of borrowers were
within 5 percent of the relevant leverage limit. In addition, on the extensive margin, the
1
2
3
GDP−normalized
volume
Exhibit 1
Main Street Loan Volume Divided by State GDP
Source: MSLP data.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
0
20
40
60
80
0
2
4
6
8
10
12
100
Billions of dollars
Share Monthly (log) total MSLP uptake
Share to highly affected industries (right scale)
Total biweekly flows (left scale)
6
5
4
2
3
0
02468101214161820
Relationship Between State-Level COVID-19
Positivity Rate and Main Street Uptake
Lagged (one-month) COVID-19
testing positivity
Jul 27
Aug 10
Jul 13
Aug 24
Sep
7
Sep 21
Oct
5
Oct 19
Nov 2
Nov 16
Nov 30
Dec 14
C
Main Street Credit Reached Highly Affected Industries and States
Sources: Authors’ calculation using MSLP data (left panel). Authors' calculation using MSLP data, Bureau of
Economic Analysis, and Opportunity Insights data (right panel).
Notes: COVID-aected industries include entertainment and recreation, oil and gas, real estate, retail, and
transportation services. Data in right panel are orthogonalized with respect to calendar month and state GDP
per capita. Each dot represents a U.S. state.
0
0.1
0.2
0.3
0.4
0.5
Fraction of banks in each status group
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Fraction of banks in each status group
Registered active Registered inactive
Not registered
<100M
100-250M
250-500M
500-750M
750M-1B
1-10B
10-50B
50B-1T
>1T
<100M
100-250M
250-500M
500-750M
750M-1B
1-10B
10-50B
50B-1T
>1T
Banks by asset size (dollars) Banks by asset size (dollars)
C
Lenders Size Distribution by Registration Status
Source: Authors’ calculation using MSLP and Call Reports data.
Notes: Registered active banks are those with accepted special-purpose vehicle (SPV) loans. Registered
inactive banks have either zero or rejected submissions to the SPV.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
leverage limits also completely excluded some potential borrowers with high leverage. Con-
versely, less than 4 percent of borrowers were within 5 percent of the loan size upper limit, also
across all three facilities.
Although it is still too early to fully assess the riskiness of loans made through the Main
Street program, it is nonetheless informative to compare the characteristics of Main Street
0
0.1
0.2
0.3
0.4
0
0.1
0.2
0.3
0.4
Fraction of total loans Fraction of total loans
Loan Count Loan Value
<100M
100-250M
250-500M
500-750M
750M-1B
1-10B
10-50B
50B-1T
>1T
<100M
100-250M
250-500M
500-750M
750M-1B
1-10B
10-50B
50B-1T
>1T
Banks by asset size (dollars) Banks by asset size (dollars)
C
Main Street Lending Activity by Lender Size
Source: Authors’ calculation using MSLP and Call Reports data.
0
25
50
75
100
125
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
Number of loans
Number of banks
C
Number of Loans Sold per Bank
Source: MSLP data.
Note: Selected banks were excluded for legibility.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
0
0.05
0.10
0.15
0.20
0.25
0.30
0.35
5
10
15
20
25
<100M
100-250M
250-500M
500-750M
750M-1B
1-10B
10-50B
50B-1T
>1T
0 50 100 150 200 250 300 350
Banks by asset size (dollars)
MSLP volume (millions of dollars)
Weighted average ratio Percent
C&I/RWA among Active Banks
Nonregistered lenders
Registered lenders
C&I/RWA by Bank Size and
Registration Status
C
C&I Loan Concentration by Lender Registration and Activity
Source: Authors’ calculation using MSLP and Call Reports data.
Notes: Left panel: Ratios as of 2019:Q4 are weighted by risk-weighted assets (RWA). * denotes statistically
signicant dierence in mean of registered versus unregistered lenders, by asset size, at the 95 percent
condence level. Right panel: Selected banks were excluded for legibility.
loans with those of a set of similar loans made outside the program. For loans made outside
the program, we use loan-level data from the Federal Reserve’s Y-14Q (Y-14) data covering the
largest banks that were subject to stress tests over the same time period as the Main Street
program.
26
Chart 10 (le panel) shows that Main Street borrowers tended to be smaller and more lev-
eraged than a large bank’s typical C&I borrower.
27
About half of Main Street loans went to
rms with total assets in the range of $5 million to $100 million, very comparable to the frac-
tion of large bank C&I lending to rms of that size. However, 25 percent of large bank C&I
loan volume went to borrowers with total assets exceeding $100 million, whereas Main Street
borrowers of that size represent only 4.3 percent of the loan volume. e right panel shows that
Main Street borrowers also tended to be more leveraged. Almost half of Y-14 borrowers had
leverage between zero times and two times EBITDA. In contrast, almost 90 percent of Main
Street loans went to borrowers with leverage between two times and six times EBITDA. While
most large-bank borrowers tended to have leverage within program limits, the fraction with
leverage exceeding those limits (that is, exceeding six times EBITDA, or with zero or negative
EBITDA) was still signicant (16.9 percent).
For a deeper analysis, we name-matched Main Street borrowers to those also present in the
Y-14 to come up with a set of 149 borrowers that have a loan both in the Y-14 (as of the fourth
quarter of 2019) and through the Main Street program. is matched dataset, though small,
provides a more detailed understanding of the borrower risk prole and terms for loans made
through Main Street compared with loans made outside of Main Street but to the
same borrower.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
Chart 11 compares internal bank ratings for our matched sample of Main Street borrowers
that are also found in the Y-14 (green bars) and borrowers from the Y-14 more generally (blue
bars).
28
e le panel shows that the distribution of ratings for the two groups was roughly
similar before the pandemic. e panel on the right shows that during the pandemic, the distri-
bution of ratings for Main Street borrowers was considerably skewed toward worse credit quality
relative to Y-14 borrowers more generally. Moreover, Chart 12 shows the evolution of ratings aer
origination for Main Street-matched borrowers compared with the rest of Y-14 borrowers. Main
Street borrowers show a signicantly faster deterioration of credit quality according to the banks’
own internal rating systems. As a caveat, note that the internal rating given by the Y-14 bank may
not coincide with the Main Street lender’s rating of that same borrower.
Because the Y-14 has data on loan spreads, we can compare the pricing of loans made
outside Main Street to the uniform pricing (LIBOR + 300) on all Main Street loans.
Table 7 shows (unsurprisingly) that smaller, more leveraged Y-14 rms paid higher spreads
on average prior to the onset of the pandemic, with an interquartile range of 150 to
255 basis points over LIBOR. Spreads were slightly higher in the second quarter of 2020,
when restrictive health policy measures were in eect. Before the pandemic, 13.5 percent of
the bank loans paid a spread over LIBOR higher than 300 basis points, rising to 16.5 percent
in the second quarter of 2020. is rise occurred despite tighter (non-price) lending stan-
dards and the shi to safer borrowers by banks during the spring and summer, as noted
previously. Most Y-14 borrowers were able to secure lending below 300 basis points even
during the crisis, which may explain the initial slow pace of uptake in the Main Street facili-
ties by companies that already had banking relationships with large nancial institutions
(Y-14 lenders). However, the lack of comparable data from smaller lenders that do not le
Y-14 data and the lack of data indicating the number of loan requests denied by lenders
make it dicult to draw conclusions about the impact of Main Street pricing on
program demand.
e prole so far suggests that Main Street borrowers were, on average, riskier than compa-
rable Y-14 borrowers. is is not entirely surprising, as higher-quality borrowers were probably
able to secure credit at a lower rate through their already established relationship with a Y-14
lender. ese conclusions also need to be taken with caution, as the matched sample represents
a small fraction of all Main Street borrowers and a tiny fraction of Y-14 borrowers overall. e
T
Share of Loans by Distance from the Leverage and (Nominal) Loan Size Limits
Leverage Limit Loan Size Limit
Facility At Limit Within 1% Within 5% At Limit Within 1% Within 5%
Expanded loans
3.8 11.5 26.9 3.8 3.8 3.8
New loans 5.7 21.6 31.7 2.1 2.1 2.1
Priority loans 5.9 18.6 29.2 3.8 4.1 4.4
Nonprot loans 0.0 0.0 0.0 0.0 0.0 0.0
All facilities 5.8 19.5 30.0 3.2 3.4 3.6
Source: Authors’ calculation using MSLP and Call Reports data.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
Percentage of loans
Borrower asset size (dollars) Borrower leverage (debt-to-EBITDA ratio)
4–62–40–2 <0 or >65–100M2.5–5M1–2.5M<1M >100M
0
10
20
30
40
50
60
0
10
20
30
40
50
60
Percentage of loans
MSLP Y-14Q
C
Main Street and 14Q Borrower Size and Leverage
Source: Authors’ calculation using MSLP and FR Y-14Q (H.1) data.
0
10
20
30
40
50
60
70
0
10
20
30
40
50
60
AAA AA A BBB BB B CCC CC C D NR AAA AA A BBB BB B CCC CC C D NR
Internal bank rating
Internal bank rating
Percentage of total firms
MSLP borrowers in Y-14
All Y-14 borrowers
Rating as of 2019:Q4 Rating as of 2020:Q3
Percentage of total firms
C
Ratings of MLSP Borrowers in the Y-14 and All Y-14 Borrowers
Source: Authors’ calculation using MSLP and FR Y-14Q (H.1) data .
dierences noted may also reect dierences between the types of borrowers at small and
medium-sized banks (that were most active in Main Street) relative to the types of borrowers
at the large banks covered in the Y-14.
In sum, Main Street borrowers historically paid higher spreads for bank loans and experi-
enced more severe rating downgrades than a comparable reference group (Y-14). Additionally,
the fact that riskier borrowers were able to obtain credit from Main Street facilities can be inter-
preted as consistent with program objectives, since the goal of Main Street was to share risk with
banks during the severe economic downturn caused by the pandemic. In the initial months, bor-
rowing was driven by more highly levered rms, but the scope of lending increased over time to
reach less levered rms. However, leverage ended up being the binding constraint for most of the
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
0
0.1
–0.1
–0.2
–0.3
–0.4
–0.5
0.2
0 1 2 3 4 5 6 7 8
Quarters since new origination
Average net change in rating
All Y-14 borrowers
MSLP borrowers in Y-14
C
Changes in Credit Risk Ratings for Main Street Borrowers in the Y-14
and All Y-14 Borrowers
Source: Authors’ calculation using MSLP and FR Y-14Q (H.1) data.
borrowers, and this was true across all industries. Finally, the program reached industries and
geographies that were most aected by the economic eects of the pandemic.
4.5 Capital Channel
Main Street loans allowed banks to preserve capital buers, since banks are required to
maintain capital against only their retained (5 percent) share. An implication is that, apart from
risk-sharing, Main Street might have also supported lending through a capital channel whereby
banks benet from originating loans but do not pay the full capital cost of carrying those loans
on their balance sheets.
Chart 13 shows that registered banks tended to have lower capital ratios than nonregistered
banks across all but the smallest size category. ese dierences are statistically signicant for
all but the smallest size groups. Moreover, there is a signicant dierence in capital ratios
between banks that actively participated and those that did not register or registered but were
not active. To investigate the capital channel further, we calculated the aggregate reduction in
required capital facilitated by the Main Street program for all active banks and found it to be a
modest 0.24 percent.
29
e median capital savings across banks is 1.1 percent, the average is
10.2 percent (reecting outliers), and the interquartile range is 0.23 to 6.8 percent. Looking
across bank size groups, the largest percentage reductions in required capital were at smaller
banks. For example, the 43 active banks in the $100 million to $250 million size group save
53 percent on average, with a median saving of 12.8 percent. For the largest banks (more than
$50 billion), the reductions are insignificant.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
T
Loan Spreads Relative to LIBOR on Newly Originated Y-14 Term Loans, by Date and Size-Eligible Borrower Characteristics
2019:Q4 2020:Q2
Percent of
Total Loans
>300 BPS
Percent of
Total Volume
>300 BPS
Percent of
Total Loans
>300 BPS
Percent of
Total Volume
>300 BPSMean Median p25 p75 Mean Median p25 p75
Panel A: Total Assets
Less than $1M 2.47 2.25 1.75 2.78
17.4% 10.5%
2.09 2.14 1.27 2.63
4.3% 3.4%
Between $1M and $2.5M 2.38 2.20 1.75 2.75
11.3% 5.1%
2.75 2.30 2.01 2.93
8.3% 1.3%
Between $2.5M and $5M 2.43 2.27 1.75 2.86
14.1% 15.0%
2.62 2.45 2.00 3.00
23.1% 17.5%
Between $5M and $100M 2.26 2.00 1.64 2.75
15.2% 18.8%
2.44 2.20 1.50 2.75 18.2%
26.6%
Greater than $100M 1.96 1.63 1.36 2.25
9.5% 11.2%
2.22 1.85 1.50 2.50
15.8% 14.3%
Total (Size of Assets) 2.21 2.00 1.50 2.55
13.5% 13.8%
2.35 2.00 1.58 2.75
16.5% 16.4%
Panel B: Leverage
Between 0 and 2 2.16 2.00 1.50 2.60
12.4% 11.3%
2.15 2.00 1.52 2.50
10.8% 13.0%
Between 2 and 4 2.11 1.97 1.50 2.50
10.5% 15.1%
2.48 2.25 1.83 2.98
13.5% 19.0%
Between 4 and 6 2.29 2.25 1.59 2.50
13.6% 10.2%
2.20 1.75 1.50 2.50
17.9% 5.6%
Less than 0 or greater than 6 2.38 2.00 1.60 2.75
19.7% 17.7%
2.57 1.88 1.25 3.50
31.6% 31.2%
Total (Leverage) 2.20 2.00 1.50 2.54
13.4% 13.4%
2.32 2.00 1.50 2.75
16.5% 16.9%
Total (Aggregate) 2.21 2.00 1.50 2.57
13.8% 14.2%
2.30 2.00 1.56 2.61
15.6% 15.9%
Source: Authors’ calculation using MSLP and Call Reports data.
Note: Size-eligible borrowers are rms with annual revenue up to $5 billion.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
All told, the evidence presented here supports the capital channel. For the largest banks, the
capital channel may have provided an incentive to actively participate, but, in practice, capital
savings were likely modest. In contrast, the capital savings for smaller banks that used the program
more intensely were estimated to be more substantial. Minoiu et al. (2021) also nd evidence in
favor of the capital channel using a more sophisticated multivariate regression framework.
. L L C
With most Main Street loans still outstanding, it is too early to discuss denitive lessons. In
particular, the credit performance of the loans is not yet known. However, now that Main
Street has stopped purchasing loan participations, we attempt to outline a few conclusions and
preliminary lessons learned.
e program helped many borrowers hit hard by the pandemic.
Main Street facilitated more than 1,800 loans to businesses across the nation, representing a
wide range of industries. Volume, at about $17 billion in total, was modest relative to the
maximum size of the program, but it represented a meaningful addition to the ow of bank
credit while the program was in operation, leading Main Street to become the largest credit
0.25
0.20
0.15
0.10
0.05
0
0.30
CET1 capital ratio
Banks by asset size (dollars)
Registered lendersNonregistered lenders
*
*
*
*
*
<100M
100-250M
250-500M
500-750M
750M-1B
1-10B
10-50B
50B-1T
>1T
C
Capital Ratios by Registration Status and Lender Size
Source: Authors’ calculation using MSLP and Call Reports data.
Notes: CET1 is common equity tier 1. Ratios are weighted by riskweighted assets. * denotes statistically
signicant dierence in mean of registered versus unregistered lenders, by asset size, at the 95 percent
condence level.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
purchase facility operated by the Federal Reserve.
30
Moreover, many Main Street borrowers
were hit hard by the pandemic, and lenders indicated that they made loans they would not oth-
erwise have made, in line with the goals of the program.
Speed is essential, but setting up a novel loan purchase program takes months.
Loan demand was most pronounced in the spring of 2020, before Main Street was opera-
tional. Looking at the experience across PPP and similar programs abroad, about half to
three-fourths of the uptake occurred by the end of the second quarter of 2020.
31
is pattern
suggests that, in a crisis, speed of execution may need to be prioritized to ensure that support is
available when needed. With Main Street, about four months passed between its announce-
ment and the rst loan purchase, longer than other emergency lending programs of the Federal
Reserve (Morgan and Clampitt 2021).
e length of the rollout time reected the unprecedented nature of the program: e
Federal Reserve had not operated a credit program for small and medium-sized businesses
since the 1940s, and it had never deployed a program to purchase loan participations. So there
was no blueprint, as there was for most other emergency programs rolled out by the Federal
Reserve in response to the pandemic. In addition, the program was operationally complex,
reecting the bespoke nature of the C&I loan market for small and medium-sized businesses,
and necessitated development of many legal agreements and roughly 100 pages of FAQs in
coordination with the Treasury. e program also required the development of
information-technology, credit-risk, and accounting systems to execute the purchase of loan
participations, all of which took time to build. Even with this experience, any future loan par-
ticipation program (or direct lending program) would likely require more time to
operationalize than other market-based emergency lending programs. Finally, policymakers
made several adjustments along the way to rene the program in response to feedback and
evolving conditions. ese changes meant lenders had to incorporate new aspects of the
program in their origination process, which created some delays in underwriting. e changes
also introduced new operational elements that required time to incorporate.
e program’s structure and complexity limited its attractiveness to lenders and borrowers.
e program’s participation structure, which was designed to be consistent with Federal
Reserve authorities and to give banks an incentive to undertake a degree of risk-screening
through banks’ risk retention, likely limited lender appetite to underwrite loans to riskier borrow-
ers, compared with, for example, a full loan guarantee program. Most lenders entered the
pandemic with stronger balance sheets and more lending capacity than in past economic down-
turns. is cushion prevented a more severe reduction in loan supply than might otherwise have
occurred and reduced demand for programs without loan forgiveness.
32
Additionally, the com-
plexity of the program likely made origination and servicing costs large, and hence the lender’s
return may have been attractive only for larger loans, safer borrowers, or at high volumes. Indeed,
many banks indicated that they preferred to lend outside the program when possible to avoid its
administrative and operational complexities, including the program’s certications and covenants
as well as perceived uncertainty about partnering with the government in the event of future
workout situations. Further, lenders cited the reporting requirements over the life of the loan,
necessary to track credit quality, as a signicant deterrent to smaller borrowers not accustomed to
providing regular quarterly nancial statements as part of a lending arrangement. Finally, for
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
lenders who did participate, the program’s complexity necessitated an investment in new pro-
cesses that delayed underwriting. e surge at the close of the program provides some evidence
that the program pipeline among participating lenders had been building up over time.
Binding leverage limits, relatively inexible loan terms, security and priority requirements, and
limits on renancing all limited risk, but did so at the expense of the program’s reach.
Leverage limits were a binding constraint on loan size for many borrowers and likely
excluded some vulnerable borrowers with an ability to repay, such as those with higher lever-
age levels that traditionally relied on asset-based borrowing. is was particularly true for the
nonprot facilities, where potential borrowers, which operate with low earnings in normal
times, were required to meet a large number of nancial and operational thresholds to be eligi-
ble for the program. In addition, the loan terms oered little exibility, including no allowance
for revolving credit facilities. Allowing some exibility on the loan interest rate might have
created room for more risk-based pricing—that is, loan rates that reected lenders’ assessment
of borrowers’ risk.
33
Credit programs in the United Kingdom and France allowed for more ex-
ibility on rates than Main Street. At the same time, such exibility would have increased
complexity further, and high interest rates may not have been viewed as consistent with the
program’s goals. e requirement at some Main Street facilities that loans be senior to or pari
passu with the borrower’s other loans may also have discouraged lenders from expanding
credit to their existing borrowers. Finally, lenders and borrowers repeatedly asked for greater
exibility to renance existing loans through the program, particularly those that were matur-
ing in the near term. While renancing limits were important in reducing the risk that lenders
would simply shi their existing exposure to risky borrowers to Main Street, additional options
for lenders to roll over maturing debt would likely have fostered broader program reach.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
To ensure compliance with the requirements of Section 13(3) of the Federal Reserve Act, the
Federal Reserve had to assess potential gains and losses from Main Street’s operations. ese
projections were akin to a stress test, starting with the development of several credit risk sce-
narios. At the time the program was being designed, still early in the pandemic, the economic
outlook was extremely uncertain. It was impossible to know how long the economic disrup-
tions would last or how deep the economic damage would be. Against that background, sta
considered a range of loan-loss scenarios. As in a stress test, some of the scenarios were
intended to be fairly conservative—severe yet plausible.
One approach was to consider the worst cumulative gross charge-off rates on bank C&I
loans that had been historically observed over any four- or five-year period. is resulted in
elevated projected loss rates.
34
Still, in light of the unprecedentedly severe nature of the down-
turn, Main Street’s goal of helping borrowers hit hard by the pandemic, and the risk of adverse
selection in the program’s portfolio, it seemed prudent to consider more severe scenarios with
loss rates two to three times the (historically) worst case.
A second approach relied on results from severely adverse scenarios in the Federal
Reserve’s stress tests of large banks in 2018 and 2019.
35
Staff used the projected loss rates
on unsecured and non-investment-grade loans, which seemed consistent with Main
Street’s targeting of small and medium-sized business borrowers, for which an
investment-grade rating is less common than for large corporate borrowers. In addition,
staff also considered the 75th percentile of loan losses across all unsecured,
non-investment-grade business loans, which suggested substantially higher loss rates, in
the range of 10 to 20 percent.
36
A third approach employed forecasts of default rates by a major credit rating agency for
the institutional leveraged loan market. These forecasts incorporated early estimates of the
effects of COVID-19–related disruptions on credit performance. Leveraged lending is
generally riskier than the broad class of lending eligible for Main Street, so this approach
was also plausibly conservative. After some adjustments, this resulted in a scenario with a
14 percent default rate over the term of the loans. To obtain loan-loss rates from default
rates, assumptions for the loss-given-default (LGD) were needed. Given the likelihood of
stressed economic conditions, at least for the coming months or years, the projections
assumed relatively high LGDs, in the 60 to 90 percent range.
37
Again, multiples ultimately
up to two times the default rates were also considered for robustness (holding LGDs
constant).
With these credit scenarios in hand, sta was able to project gains and losses for Main Street
under alternative design choices for the loan terms, fees, and risk-sharing arrangement.
Defaults were assumed to be concentrated at the end of year two of the loan, when the rst
principal repayment becomes due. e less adverse scenarios, including the worst historically
observed C&I loan charge-off rate and the stress testing portfolio-average losses, were pro-
jected to result in net gains for the Main Street SPV, with interest income outweighing credit
losses. However, the more adverse scenarios were projected to result in net losses to the Main
Street SPV and therefore to the Treasury’s equity investment.
A: I L P
D P
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
A (C)
ese projections guided the decision to cap SPV “leverage” at 8-to-1. Given the Treasury’s
planned $75 billion equity investment, the leverage cap dictated a maximum program size of
$600 billion. With that leverage, even under the more adverse scenarios, the Federal Reserve
was projected to incur zero losses.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
Acknowledgments: e authors thank, without implicating, William Bassett, Steanie Brady, Jie Chen, Julian Di Giovanni,
Michael Kiley, Andreas Lehnert, Kelley O’Mara, Joe Peek, Mark Van Der Weide, and an anonymous referee for valuable
input. Jake Faber, Frankie Lin, and Mary Zhang provided expert research assistance.
T M S L P
N
1
See “Funding, Credit, Liquidity, and Loan Facilities,” https://www.federalreserve.gov/funding-credit-liquidity-and-
loan-facilities.htm. The comparison excludes liquidity facilities, some of which had larger peak outstanding amounts,
for example the Paycheck Protection Program Liquidity Facility, the Money Market Mutual Fund Liquidity Facility,
and the Primary Dealer Credit Facility.
2
We use “missing middle” as short-hand for medium-sized firms that depend on banks (or other intermediaries)
for credit and that are too large for PPP loans. Note, though, that there is no standard cross-industry definition of
“small,” “medium-sized,” or “mid-sized,” and the definitions in our analysis vary somewhat according to the data
we cover. The cutoffs for Main Street are discussed in the next section.
3
Based on 2018 Census data, firms with between 500 and 5,000 employees employ about 23 million people. Most of
these firms are private and cannot access public debt markets. Even among the publicly traded firms covered in
the Compustat database, the smaller rms (which are still larger than most private firms) rely more on bank nancing
(Rauh and Su 2010). Calomiris, Himmelberg, and Wachtel (1995) find that only 20 percent of manufacturing firms
in the Compustat database have a bond or a commercial paper rating.
4
The later phase of this lending surge also reflected PPP lending by banks.
5
Chodorow-Reich et al. (2020) find that this difference reflects the reality that smaller firms were less likely to
have credit lines or faced stricter (pre-COVID) terms that limited their takedowns.
6
The SLOOS defines small firms as those with annual sales of less than $50 million. Large and middle-market firms
have sales greater than $50 million.
7
See “Federal Reserve Announces Extensive New Measures to Support the Economy,” Federal Reserve Board press
release, March 23, 2020, https://www.federalreserve.gov/newsevents/pressreleases/monetary20200323b.htm. Note that
the Board announced its intention to establish Main Street before the passage of the Coronavirus Aid, Relief,
and Economic Security Act (CARES Act), and Congress, in the CARES Act, expressly gave the Board wide discretion
in designing a program to “support lending to small and mid-sized businesses on such terms and conditions as
the Board may set consistent with Section 13(3).” 15 U.S.C. § 9042(c)(3)(D)(ii).
8
12 U.S.C. § 343(3).
9
See 15 U.S.C. § 9042, 9054.
10
Lack of regulatory capital or lack of funding at banks were not considered the primary constraints on lending
at the time. Had either been, a very different type of program might have been deemed appropriate, such as
a funding-for-lending initiative. However, these factors did not appear to be as important as heightened risk aversion.
Although more bank capital, or a greater distance from regulatory capital requirements, can generally help reduce
banks’ risk aversion somewhat, it is far from clear that this could have overcome the extreme uncertainty encountered
in 2020.
11
As noted previously, there is no standard U.S. definition of “small or “medium-sized.”
12
By using the SBA’s framework, the Board was able to quickly implement definitions that had been promulgated
pursuant to notice-and-comment rulemaking, tested in bank-intermediated government lending, and elucidated
through SBA guidance. Further, these definitions were familiar to many lenders and had been recently incorporated
into provisions of the PPP established under the CARES Act.
13
Borrowers certified their eligibility for program loans through the Borrower Certifications and Covenants.
The use of certifications for purposes of borrower compliance with program requirements has a foundation in
the statutory text of both the Federal Reserve Act and the CARES Act. (12 U.S.C. § 343(3)(B)(ii); 15 U.S.C. § 9042(c)
(3)(D)(ii), 9054(c)). In general, the Borrower Certifications require the borrowers to establish their own eligibility,
although lenders had an obligation to conduct due diligence with respect to the borrower’s formation under law.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
N (C)
14
The following organizations could be an eligible lender: a U.S. federally insured depository institution (including a
bank, savings association, or credit union), a U.S. branch or agency of a foreign bank, a U.S. bank holding company,
a U.S. savings and loan holding company, a U.S. intermediate holding company of a foreign banking organization,
or a U.S. subsidiary of any of the foregoing. These entities all have existing supervisory relationships with the Federal
Reserve or other federal regulators.
15
Lenders generally had to establish their eligibility at the time of their registration through Lender Registration
Certifications and Covenants, while the Lender Transaction-Specific Certifications and Covenants primarily required
lenders to establish that a particular loan was eligible for sale to the Main Street SPV.
16
MSNLF and MSPLF loans under $250,000 were permitted to have an origination fee of up to 2 percent,
while MSELF loans (which entailed a $10 million minimum loan size) featured an origination fee of up to
75 basis points.
17
Analysis predicted that MSELF participation would generally still be attractive to the lender provided the loan
expansion reduced the borrower’s probability of default. This proviso was broadly in line with the program’s goal
of helping borrowers hit hard by the pandemic but otherwise in sound financial condition.
18
The terms for the nonprofit facilities can be found at https://www.federalreserve.gov/newsevents/pressreleases/les/
monetary20201229a4.pdf and https://www.federalreserve.gov/newsevents/pressreleases/les/monetary20201229a5.
pdf.
19
It is unclear whether the rush was a function only of the impending closure or whether it also reflected the
time required by lenders to originate Main Street loans. Discussions with lenders active in the program indicated
that familiarizing themselves and their clients with the legal and operational elements of the program required
a considerable investment in time. Both factors likely contributed to a backloading of the loan participations, with by
far the largest volumes occurring in the program’s waning days.
20
The announcements of the corporate credit and municipal facilities had significant real-time effects on prices,
and thus yields, of existing corporate and municipal bonds. Because such bonds are actively traded in secondary
markets, announcement effects can be observed. Notably, price impact was seen even outside the range of bonds
that would later be purchased by the facilities. In contrast, there is no active secondary market for business loans of
the type targeted by Main Street. us, there is no way to gauge the announcement effect of Main Street in a
similar fashion.
21
The comparison is not perfect since loans with balances below $1 million are not required to be reported in
the FR Y-14Q schedule. In addition, very small firms are more likely to borrow from smaller banks. However,
as Chodorow-Reich et al. (2020) show, FR Y-14Q loans represent 82 percent of the total C&I bank credit.
22
The number of borrowers exceeds the number of loans because some Main Street loans had multiple borrowers that,
in most cases, consisted of subsidiaries of the same parent firm.
23
Conversely, the U.S. Virgin Islands, Maine, Montana, and Vermont all had volumes totaling less than $10 million.
24
COVID-affected industries include entertainment and recreation, oil and gas, real estate, retail, and transportation
services.
25
We find similar results when we use other measures of economic slowdown, such as unemployment claims,
and population mobility measures as shown by Bräuning, Fillat, and Wang (2021).
26
Banks with $100 billion or more in consolidated assets are required to submit these data. The Y-14 data contain
extensive supervisory information about the borrowers and about the loans, allowing us to compare the distribution
of lending to Main Street borrowers and the rest of Y-14 borrowers along several dimensions. Information on
borrower and loan characteristics is limited in the Main Street data, but it is much more comprehensive in the Y-14.
27
We consider only potentially eligible borrowers in the Y-14 data, for comparability. Hence, the largest rms,
those with revenue greater than $5 billion, are excluded from our comparison.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
N (C)
28
Regarding the risk profile, Main Street participants are (by design) too small to have access to market finance
and therefore to be rated by rating agencies. However, Y-14 banks are required to disclose borrower-level internal
ratings as well as the correspondence to a common scale for comparison purposes. In our matched sample of 149
borrowers, we find that 139 Main Street borrowers had loans outstanding with internal (bank) ratings in the third
quarter of 2020.
29
We compute the reduction in risk-weighted assets (RWA) as the volume of Main Street loans removed from the
banks’ balance sheets through the sale of participations (that is, 95 percent of their total Main Street volume).
Because risk-based capital requirements are expressed as fractions of RWA, the percentage reduction in RWA als o
equals the percentage reduction in required capital (CET1, tier 1, and total).
30
See Chart 4 and “Funding, Credit, Liquidity, and Loan Facilities,” https://www.federalreserve.gov/funding-credit-
liquidity-and-loan-facilities.htm. The comparison excludes liquidity facilities, some of which had larger peak
outstanding amounts.
31
The U.K.’s Coronavirus Business Interruption Loan Scheme (CBILS) and Bounce Back Loan Scheme (BBLS),
France’s Prêt Garanti par l’État (PGE), and the U.S.’s Paycheck Protection Program saw, respectively, 48 percent,
63 percent, 76 percent, and 65 percent of their total uptake by the end of the second quarter of 2020. See,
for CBILS, https://www.gov.uk/government/collections/hm-treasury-coronavirus-covid-19-business-loan-scheme-
statistics#Coronavirus-Business-Interruption-Loan-Scheme; for BBLS, https://www.gov.uk/government/collections/
hm-treasury-coronavirus-covid-19-business-loan-scheme-statistics#Bounce-Back-Loan-Scheme; for PGE, https://
www.data.gouv.fr/fr/datasets/donnees-relatives-aux-prets-garantis-par-letat-dans-le-cadre-de-lepidemie-de-covid-19/
and for PPP, https://www.sba.gov/funding-programs/loans/covid-19-relief-options/paycheck-protection-program/
ppp-data.
32
In a special Senior Loan Officer Opinion Survey on Main Street, a vast majority of nonregistered banks cited their
ability to address the credit needs of Main Street-sized borrowers without participating in the program as an important
or very important reason for not registering. See https://www.federalreserve.gov/data/sloos/sloos-202009.htm
33
English and Liang (2020) have argued for more flexibility in Main Street’s loan terms, including their interest rates
and banks’ risk retention share.
34
Maximum cumulative gross charge-off rates amounted to 7.4 percent for a four-year period (2007:Q2 through
2011:Q1) and 8.7 percent for a five-year period (2006:Q3 through 2011:Q2). These rates were calculated using FFIEC
Call Reports, where the relevant data are available from 1985 onward. Gross rates, which exclude recoveries, were used
for robustness.
35
From each of these stress tests, staff used the portfolio-average loss rate on unsecured, non-
investment-grade business loans, taking the weighted average of financial and nonfinancial borrowers, which resulted
in loan-loss rates of 8.3 and 5.4 percent for the 2018 and 2019 CCARs, respectively. The weights used are the shares
of each sector of the stress-tested banks’ total unsecured non-investment-grade business loans. See “Dodd-Frank Act
Stress Test 2019: Supervisory Stress Test Methodology,” Board of Governors of the Federal Reserve, March 2019,
https://www.federalreserve.gov/publications/les/2019-march-supervisory-stress-test-methodology.pdf, and
“Dodd-Frank Act Stress Test 2020: Supervisory Stress Test Methodology,” Board of Governors of the Federal Reserve,
March 2020, https://www.federalreserve.gov/publications/les/2020-march-supervisory-stress-test-methodology.pdf.
36
Specically, the loan-loss rates calculated from the 2018 and 2019 stress tests were 17.2 and 11.5 percent, respectively.
37
Reflecting the higher priority and security embedded in the terms of the PLF and ELF facilities, LGDs were set as
90 percent for NLF, 75 percent for PLF, and 60 percent for ELF. This implied loss rates ranging from 8.4 to 12.7 percent.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
T M S L P
R
Berg, T., A. Saunders, and S. Steen. 2016. “e Total Cost of Corporate Borrowing in the Loan Market:
Don’t Ignore the Fees.” J F 71: 1357-92.
Bräuning, F., and T. Paligorova. 2021. “Uptake of the Main Street Lending Program.” Federal Reserve Bank
of Boston, C P P, March 19.
Bräuning, F., J. L. Fillat, and J. C. Wang. 2021. “A Helping Hand to Main Street Where and When It Was
Needed.” Federal Reserve Bank of Boston, C P P, May 27.
Briggs, J., and B. Walker. 2020. “US Daily: Supporting Small Businesses in a Pandemic: Lessons from
Other Countries.” Goldman Sachs Economic Research, November 17.
Calomiris, C. W., C. P. Himmelberg, and P. Wachtel. 1995. “Commercial Paper, Corporate Finance, and
the Business Cycle: A Microeconomic Perspective.” C-R C S
P P42, no. 1: 203-50, June.
Cantu, C., P. Cavallino, F. De Fiore, and J. Yetman. 2021. “A Global Database on Central Banks’ Monetary
Responses to COVID-19.” BIS Working Paper no. 934.
Cavallino, P., and F. De Fiore. 2021. “Central Banks’ Response to COVID-19 in Advanced Economies.” BIS
Bulletin no. 21.
Chodorow-Reich, G., H. Cooperman, O. Darmouni, S. Luck, and M. Plosser. 2020. “Weathering the Storm:
Who Can Access Credit in a Pandemic?” Federal Reserve Bank of Ne w YorkLiberty Street Economics,
October 13. https://libertystreeteconomics.newyorkfed.org/2020/10/weathering-the-storm-who-can-
access-credit-in-a-pandemic.html
English, W. B., and J. N. Liang. 2020. “Designing the Main Street Lending Program: Challenges and
Options.” Hutchins Center Working Paper #64, Brookings Institution.
Fettig, D. 2002. “Lender of More an Last Resort.” Federal Reserve Bank of Minneapolis.
Hackley, H. H. 1973. “Lending Functions of the Federal Reserve Banks: A History.” Washington, D.C.:
Board of Governors of the Federal Reserve System.
Minoiu, C., R. Zarutskie, and A. Zlate. A. 2021. “Motivating Banks to Lend? Credit Spillover Eects of the
Main Street Lending Program.” Working paper.
Morgan, D. P., and S. Clampitt. 2021. “Up on Main Street.” Federal Reserve Bank of New York Liberty
Street Economics, February 5. https://libertystreeteconomics.newyorkfed.org/2021/02/up-on-main-
street.html.
Federal Reserve Bank of New York
Economic Policy Review 28, no. 1, June 2022
R (C)
T M S L P
Rauh, J. D., and A. Su. 2010. “Capital Structure and Debt Structure.” R F S 23,
no. 12 (December): 4242–80.
Sablik, T. 2013. “Fed Credit Policy during the Great Depression.” Federal Reserve Bank of Richmond
E B 13-03 (March).
e Economic Policy Review is published by the Research and Statistics Group of the
Federal Reserve Bank of New York. e views expressed are those of the individual authors
and do not necessarily reect the position of the Federal Reserve Bank of New York or the
Federal Reserve System.
Series Editor: Julian di Giovanni. Editorial Staff: Robert Powell, Trevor Delaney,
Maureen Egan, Anna Snider, and Peter Stevens. Design Sta: eresa Izzillo, Alyona Botogoeva,
and Laura Gharrity.
Economic Policy Review articles may be reproduced for educational or training purposes,
provided they are reprinted in full; include credit to the author(s), the publication, and the
Bank; and include the publication’s disclaimer.
©2022 e Federal Reserve Bank of New York
www.newyorkfed.org/research
Follow us on Twitter: @NYFedResearch
ECONOMIC POLICY REVIEW
FEDERAL RESERVE BANK OF NEW YORK
