Iñaki Aldasoro Sebastian Doerr Haonan Zhou

inaki.aldasoro@bis.org sebastian.doerr@bis.org haonan@princeton.edu

Non-bank lenders in the syndicated loan market1

Non-bank lenders are an important source of syndicated credit to non-financial corporates in

most regions and industries. Their loan origination, however, is more concentrated by location
and sector than that of banks and it is also more volatile. Syndicated loans arranged by non-
banks carry a significantly higher spread relative to those by banks, consistent with the pattern
that firms borrowing from non-banks are more leveraged and less profitable, ie riskier. Non-
banks generally grant a smaller share of their new loans to foreign borrowers than banks do.
During domestic financial crises, they reduce this share further, exacerbating the global
transmission of shocks.
JEL classification: E51, F34, G21, G23.

The increasing footprint of non-bank financial intermediaries has put them front and
centre of policymakers’ agendas. While they can contribute to a more diversified and
efficient financial system, non-banks can also be a source of instability due to, for
instance, liquidity mismatches (Aramonte et al (2021)). With the March 2020 market
turmoil serving as a case study, substantial efforts have been made to understand
how such instability can unfold and what policy measures can mitigate it (Carstens
(2021)).
The reach of non-banks extends beyond financial market conditions. Non-bank
lenders represent an important source of funding for non-financial corporates (NFCs)
in general (Aramonte and Avalos (2021)) and through syndicated loans in particular
(Elliott et al (2019)). As non-banks extend syndicated loans in a highly procyclical
fashion (Fleckenstein et al (2021)) and access to syndicated credit affects firm
performance (Chodorow-Reich (2014)), their ubiquitous presence could drive real
economy developments.
This special feature provides the first systematic overview of global syndicated
lending by non-banks and contrasts it with that by banks.2 It further investigates the
role of non-banks in cross-border spillovers during financial crises.

1
The authors thank Sirio Aramonte, Fernando Avalos, Frédéric Boissay, Claudio Borio, Stijn Claessens,
Egemen Eren, Blaise Gadanecz, Ulf Lewrick, Hyun Song Shin, Philip Wooldridge, Youngsuk Yook and
Nikola Tarashev for valuable comments and suggestions, and Murphy Pan for excellent research
assistance. The views expressed in this article are those of the authors and do not necessarily reflect
those of the Bank for International Settlements.
2
Syndicated lending by banks has been studied extensively (see Güler et al (2021) for a summary).

BIS Quarterly Review, March 2022 15

Key takeaways
 Non-banks’ origination of syndicated loans to non-financial firms grew twentyfold from 1990, to $410
billion in 2019, and represents a sizeable share of the total in most regions and sectors.
 Non-banks’ syndicated lending is more concentrated across countries and industries than that of
banks and it is more volatile. It also carries higher spreads, partly reflecting riskier borrowers.
 In the wake of a crisis at home, non-banks curtail lending to foreign borrowers by more than they do
to domestic ones, thereby transmitting shocks across countries.

The main findings are as follows.

First, non-banks play an important role in the syndicated loan market, but their
lending patterns differ from those of banks. Non-banks have increased their yearly
syndicated lending twentyfold over the past three decades, serving borrowers in all
major regions and sectors. Their share in total new syndicated lending fluctuated
between 7% and 18% over the same period. Non-bank lending is more concentrated
by geographical location and sector than bank syndicated lending. While non-banks’
credit provision tends to ebb and flow with that of banks, it is more volatile and
contracted more during the Great Financial Crisis (GFC).
Second, loans arranged by non-bank lenders carry materially higher spreads than
those arranged by banks. The difference in spreads largely reflects a riskier pool of
borrowers and associated risk premia, rather than differences in loan terms.
Finally, when facing domestic financial crises, non-bank arrangers retrench
particularly forcefully from abroad. In general, non-banks exhibit a stronger home
bias than banks, as foreign borrowers account for a smaller share in their syndicated
lending. During a financial crisis in their home country, non-banks further reduce their
credit provision abroad, and do so more than banks. Such retrenchment, so far
established for banks only (Giannetti and Laeven (2012)), suggests that non-banks
transmit financial shocks across countries.
The rest of this article is organised as follows. The first section discusses the main
features of syndicated lending and the classification of lenders into banks and non-
banks (Box A). The second section documents patterns in non-bank syndicated
lending across countries, industrial sectors and over time. The third discusses
differences in loan terms and borrower characteristics between banks and non-bank
lenders, assessing to what extent they explain loan spreads. The fourth section
investigates how non-banks adjust their lending in foreign and domestic markets
during domestic financial crises, with details summarised in Box B. The final section
concludes with policy considerations.

The syndicated loan market and non-bank lenders

Syndicated loans are an important financing source for non-financial firms. They
represent around three quarters of total cross-border lending to NFCs in high- and
middle-income countries (Doerr and Schaz (2021)) and are particularly important for

16 BIS Quarterly Review, March 2022

larger firms (Cerutti et al (2015)). Consequently, changes in the supply of syndicated
credit influence firm investment and employment (Acharya et al (2018)).3
A syndicated loan is granted by a group of financial institutions (the “syndicate”)
to a single borrower. The lending syndicate includes at least one lead institution (the
“lead arranger(s)”, sometimes called “underwriter(s)”) as well as one or more
participants. Lead arrangers negotiate preliminary loan terms and conditions and,
upon agreement with the borrower, put together the group of participants to fund
parts of the total loan amount. In return, lead arrangers receive an arrangement fee
(Gadanecz (2004)).
In the loan’s life cycle, lead arrangers typically act as the agent bank and perform
due diligence. They ensure screening and monitoring, govern the terms of the loans
and enforce covenants. In line with their responsibility as monitors, lead arrangers
retain a larger portion of the loan on their balance sheets, while participants often sell
their tranches on the secondary market (Sufi (2007)).
This feature analyses loan originations using detailed deal-level data on global
syndicated lending from Thomson Reuters’ DealScan. The sample period is from 1990
to 2019. Information on loans at origination includes the loan amount, maturity,
currency of denomination and interest rate, as well as the identity, type, location and
industry code of borrowers and lenders (banks as well as non-banks). All loan
amounts are converted to US dollars. To attribute loan portions among members, we
use the allocation provided by DealScan or, if this is not available, allocate pro rata.
Henceforth, we will refer to the syndicated loan amounts attributed at origination to
banks or non-banks as “bank loans” or “non-bank loans”, respectively. There is no
public information to track loan sales on the secondary market.4
We identify non-bank lenders based on information provided by DealScan, in
combination with Standard Industrial Classification codes, string search and manual
checks. In our sample, non-bank lenders include investment banks (or broker-
dealers), finance companies and mutual funds – in line with the categorisation used
by the Financial Stability Board (2021).5 Investment banks and finance companies
account for over four fifths of aggregate non-bank syndicated lending.
Non-bank lenders have gained a large footprint in the syndicated loan market
and their participation is now comparable with that of banks along important
dimensions. The share of all loan facilities with non-bank involvement exceeds one
third. Around two fifths of all non-bank lenders active in the syndicated loan market
act as lead arrangers, a share similar to that of banks. Likewise, individual members
of both types contribute comparable shares to the total syndicated amount on
average and extend loans for similar purposes (eg working capital, debt repayment

3
For the real effects of bank syndicated lending, see also Doerr et al (2018), Hale et al (2020) and
Chodorow-Reich and Falato (2022). These papers build on work showing that lender-borrower
relationships are sticky and that the costs of switching between lenders are high, especially during
downturns (Sette and Gobbi (2015), Bolton et al (2016)).
4
While such information is important to accurately assess lenders’ risk exposures, it is less relevant for
understanding patterns in the origination of new credit, which is key from the borrower’s perspective
and is the focus of this special feature. That said, non-banks purchase a significant amount of loans
in the secondary market (Aramonte et al (2019)). Accordingly, loan volumes at origination probably
represent a lower bound of the ultimate financing provided by non-banks to NFCs.
5
Broadly speaking, non-bank lenders differ from banks in their funding structure and their lack of
access to public backstops. For example, their funding tends to be largely wholesale (Jiang et al
(2020), Xiao (2020)) and they are not eligible for deposit insurance. There is, however, no harmonised
or universally accepted definition of non-bank lenders.

BIS Quarterly Review, March 2022 17

or capital expenditure). That said, credit lines are more common among banks at
origination and term loans are more common among non-banks.
Box A presents further details on the data construction, classification of banks
and non-bank lenders, and their role in syndicates.

Box A

Identifying non-bank lenders in the syndicated loan market

Iñaki Aldasoro, Sebastian Doerr and Haonan Zhou

For the analysis of the syndicated loan market, we categorise lenders into banks and non-banks. This box outlines the
two underlying steps.

First, we build on lender type information provided by DealScan. Consistent with the notion of banks representing
deposit-taking institutions, we define bank lenders as commercial banks, classified by DealScan as “African bank”,
“Asia-Pacific bank”, “Eastern European/Russian bank”, “foreign bank”, “Middle Eastern bank”, “mortgage bank”,
“thrift/S&L”, “US bank” or “Western European bank”. Non-banks comprise the remaining lenders. Around 20% of
lenders, however, are initially unclassified.

In a second step, we improve on this classification by using lenders’ Standard Industrial Classification (SIC) sector
codes as well as a string search procedure. In particular, we look for matches with names of major commercial banks,
descriptions of activities (such as “bank“, “life insurance”, “trust” and “securities”), or variants of spelling in different
languages (eg “banca”, “banco”, “banque”). These procedures allow us to categorise around three quarters of the
originally unclassified lenders, splitting them roughly equally between the bank and non-bank groups. We allocate
the still unclassified lenders, mostly middle-market debt management companies, to the non-bank category.
Ultimately, we classify almost 7,500 lenders as non-banks, out of a total of 19,000 lenders.

Different types of non-bank lender in the syndicated loan market Graph A

Investment banks dominate non- …serving more borrowers than other Other key loan terms do not differ
bank lending volumes… non-banks, and with larger amounts much across non-bank lenders2
USD bn USD mn Number Basis points Months

600 100 225 300 100

480 80 180 240 80

360 60 135 180 60

240 40 90 120 40

120 20 45 60 20

0 0 0 0 0
1995 2000 2005 2010 2015 Investment Finance Other Investment Finance Other
banks companies banks companies
Investment banks Other
Finance companies Amount (lhs) Borrowers (rhs)1 Spread (lhs) Maturity (rhs)

The centre and right-hand panels show average results for the full sample (1990–2019).
1
Number of distinct borrowers in a given year. 2 Spread and maturity are winsorised at the 1st and 99th percentiles. The spread refers to
the all-in drawn spread, weighted by loan volume.

Sources: Thomson Reuters DealScan; authors’ calculations.

18 BIS Quarterly Review, March 2022

 Finance companies and investment banks account for the bulk of non-bank syndicated lending. While only 3%
of non-bank lenders are investment banks, these account for around half of the total amount of non-bank lending to
NFCs from 1990 to 2019 (Graph A, left-hand panel). This group of non-bank lenders can include, among others,
brokerage and securities underwriting firms or an international bank‘s broker-dealer subsidiary. Finance companies,
in turn, comprise 44% of the non-bank lenders and account for 34% of total non-bank lending. This group can include
financial institutions that specialise in industrial loans or the financial arms of large companies. Other lenders, such as
insurance companies, pension funds and mutual funds, have a small presence in the primary market for syndicated
loans in the data. Over the sample period (1990–2019), investment banks served more clients and extended larger
loans than did finance companies and other non-bank lenders (centre panel). Loan spreads and maturities were of
comparable magnitude across non-bank lender types (right-hand panel).

Non-banks participate in a significant number of syndicated loans, contributing – at the level of individual lenders
– similar amounts to syndications as banks, and often acting as lead arrangers. Overall, the share of loan facilities with
non-bank involvement is 37%, vs 88% with bank involvement. Individual non-bank syndicate members account, on
average, for 19% of the total syndication amount in a given syndicate, vs 17% for banks. On average, 36% of all non-
banks active in the syndicated loan market have acted as lead arrangers at least once, compared with 42% for banks.
While 72% of all syndicates have only bank lead arrangers, just 7% of all loans are arranged by non-banks exclusively.
These patterns are similar across deals involving borrowers in advanced vs emerging market economies (EMEs).

When it comes to the type of loan and loan purpose, patterns are broadly comparable between non-banks and
banks. Credit lines comprise 41% and term loans 48% of all non-bank loans, compared with 50% and 40% for banks.
For both lender types, term loans are more common among EME borrowers. As regards loan purposes, there are no
significant differences, with “corporate purposes” and “debt repayment” dominating for both banks and non-banks.

 Prior to the classification, we drop incomplete deals and those involving government-linked lenders. In particular, we drop deals with the
status “cancelled”, “suspended” or “rumour”, and deals with missing information on amounts and loan packages that amend previous deals.
We further drop deals involving supranational organisations, governments and development banks based on string matches and SIC codes.
In cases where only the aggregate size of the loan is known or for loan facilities with lending shares totalling more than 110%, we impute the
lending share of each participant using pro rata splits.  For instance, we assign lenders with SIC code 6211 (“Security brokers, dealers,
and flotation companies”), classified as commercial banks by DealScan, to the non-bank group. This is in line with policy practice (Financial
Stability Board (2021)) and academic literature (Elliott et al (2021)).  Other types include bridge loans or leases.

Patterns in non-bank syndicated lending

Non-bank syndicated lending has grown with the overall market, both domestically
and internationally. Total new syndicated lending to NFCs increased substantially,
from under $300 billion originated in 1990 to around $4,750 billion in 2019. Non-
bank lending increased from $20 billion to $410 billion per year over the same period
(Graph 1, left-hand panel). Its share of total loan originations steadily increased to
around 14% in 2007 but contracted sharply during the GFC. This share trended up
again from 2010 to 2019. Foreign loans, ie loans to borrowers located in a jurisdiction
other than the lender’s home country, follow a similar pattern, albeit at a lower level.
Growth in non-bank syndicated lending is more volatile than that of banks. On
average across borrower countries, the standard deviation of the growth in loan
origination is almost 60% higher for non-bank than bank loans (Graph 1, centre
panel). While the volatility is in general higher for emerging market economy (EME)
borrowers than for those from advanced economies (AEs), non-bank lending is more
volatile than bank lending in both regions. Higher volatility meant greater
procyclicality during the GFC, when non-banks cut their syndicated loan origination
by twice as much as banks, in a pattern typical for crisis times (Fleckenstein et al
(2021), Aldasoro et al (2022b)).

BIS Quarterly Review, March 2022 19

Non-bank syndicated lending to non-financial corporates Graph 1

Volumes and shares of total Lending by non-banks is more …but growth in non-bank and bank
syndicated lending volatile than that of banks…2, 3 lending is strongly correlated2, 4
Per cent USD bn Standard deviation in growth rates

20 600 1.0 0.6

Growth in bank lending (%)

16 480 0.8 0.4

12 360 0.6 0.2

8 240 0.4 EMEs 0.0

4 120 0.2 –0.2

AEs
0 0 0.0 –0.4
1990 1995 2000 2005 2010 2015 AEs EMEs –1 0 1 2
Share (lhs) Non-bank lending (rhs): Bank Growth in non-bank lending (%)
Total Domestic Non–bank
1 1
Foreign Foreign

AEs= advanced economies; EMEs= emerging market economies.

1
A loan is classified as foreign if the location of the lender’s headquarters differs from that of the borrower. 2 Data are aggregated to the
borrower-country-year level and split by the region of the borrower. The underlying sample covers 1,863 borrower-country-year observations
from 119 countries from 1990 to 2019. Loan growth is computed as the difference in the logs of loan origination volumes and is winsorised
at the 1st and 99th percentiles. 3 Standard deviation of loan growth from 1990 to 2019, averages across borrower countries within a group,
for a given lender type. 4 Binned scatter plot, which enhances readability by grouping observations into equally spaced bins and reporting
the average value within each bin; region-specific fitted lines.

Sources: Thomson Reuters DealScan; authors’ calculations.

Differences in volatilities notwithstanding, bank and non-bank lending move in

tandem, especially for loans granted to borrowers from AEs. The overall correlation
between bank and non-bank loan growth stands at 0.50 on average across borrower
countries. For loans granted to NFCs in EMEs, the correlation is 0.36, whereas it is 0.54
in AEs (Graph 1, right-hand panel). Together with the greater volatility in loan
originations, this positive correlation casts doubt on the notion that non-banks
provide a “spare tyre” for borrowers when banks retrench.
Non-banks serve firms in all regions and sectors. In terms of geographical reach
over the full sample period, the share of non-bank loans in total loans is largest in
South Asia and that in foreign loans is largest in East Asia and the Pacific (Graph 2,
left-hand panel). Among the major economies, the share of syndicated borrowing
from non-banks averages 14% for the United States and 3% for China. In terms of
sectoral presence, the average share of non-bank loans in total new syndicated
lending ranges from 6.4% in the construction sector to 14.2% in services (centre
panel).
Despite their global presence, non-banks’ origination of loans is more
concentrated than that of banks. The average non-bank serves fewer distinct
borrowers than the average bank (around 330 vs 740 per year) and does so in a
narrower set of regions and sectors. Accordingly, loan origination by non-banks is
more concentrated across borrower countries and industries (Graph 2, right-hand
panel).

20 BIS Quarterly Review, March 2022

Non-bank lending is pervasive but less diversified than bank lending Graph 2

On aggregate, non-banks serve non- …and all major sectors… …but individual lenders’ loan
financial corporates in all regions… origination is more concentrated
Per cent Per cent HHI
Middle East & Agriculture,
North Africa forestry and fishing Across Across Across
Europe &
countries industries borrowers 0.8
Mining
Central Asia
Sub-Saharan Construction
Africa 0.6
Manufacturing
Latin America &
Caribbean Transportation &
public utilities 0.4
East Asia &
Pacific Wholesale trade
North America 0.2
Retail trade
South Asia Services
0.0
0 5 10 15 0 5 10 15 nk nk nk nk nk nk
1 Ba n-ba Ba n-ba Ba n-ba
Share of non–bank lending Share of non–bank lending1
No No No
Total lending Foreign lending2 Herfindahl-Hirschman index3
Average results for the full sample (1990–2019).
1
Total amount of syndicated loans originated by non-bank lenders over the total origination amount of syndicated loans in the borrower
region/industry. 2 A loan is classified as foreign if the locations of the lender parent and the borrower differ. 3 Averaged Herfindahl-
Hirschman index (HHI) calculated for each lender across all borrower countries, two-digit borrower industries, and individual borrowers,
respectively.

Sources: Thomson Reuters DealScan; World Bank; authors’ calculations.

Loan terms and borrower characteristics

We first provide a comparison of banks’ and non-banks’ borrower characteristics and

syndicated loan terms. Then, motivated by the patterns thus revealed, we zoom in on
potential drivers of the notable difference in loan spreads.
Loans with non-bank participation carry materially higher spreads than those
with bank participation only. The average all-in spread, defined as the annual spread
over Libor, including fees and interests and weighted by loan size, equals 253 basis
points on syndicated loans with non-bank involvement, vastly exceeding that on
bank-only loans (170 basis points). While average spreads trended upwards until
2010 and receded somewhat since then for both lender types, the difference in loan
spreads has remained relatively constant (Graph 3, left-hand panel). These patterns
are robust to zooming in on lead arrangers only.
Banks and non-banks originate loans of similar amounts and maturity. The
average loan amount granted by banks stands at $44 million, roughly the same as
the $42 million granted by non-bank lenders (Graph 3, right-hand panel). The
corresponding maturities are also similar, at 60 and 55 months respectively. Again,
the patterns persist when considering lead arrangers only.

BIS Quarterly Review, March 2022 21

Non-bank loans carry higher spreads, but are of similar amount and maturity Graph 3

Non-bank loans carry higher spreads…1 …but banks and non-banks extend loans of similar
amount and maturity
Basis points USD mn Months

350 60 60

300 50 50

250 40 40

200 30 30

150 20 20

100 10 10

50 0 0
1990 1995 2000 2005 2010 2015 Banks Non-banks
2
Banks Non–banks Amount (lhs) Maturity (rhs)
Average results for the full sample (1990–2019) are shown in the right-hand panel.
1
“All-in drawn spread”, defined as the spread over Libor, including fees and interests. 2
Averages weighted by loan size.

Sources: Thomson Reuters DealScan; authors’ calculations.

By contrast, firms borrowing from syndicates with non-bank lenders appear to

be riskier.6 The average firm receiving syndicated credit from non-banks has a lower
return on assets, higher leverage and a lower interest coverage ratio than firms
obtaining syndicated loans from banks only (Table 1). Unreported tests show that
these differences remain material even among firms of similar size and operating in
the same location and industry. They are also present when we classify firms by their
connection to bank or non-bank lead arrangers.

Firms borrowing from non-banks have lower returns and higher leverage Table 1

Assets Employees Interest Leverage Return on

(USD mn) (’000) coverage ratio (ratio) assets (%)

Borrows from banks only 5,180 13.65 22.81 1.92 6.42

Borrows from non-banks 10,703 24.99 12.48 2.61 6.01
The table compares non-financial corporates (NFCs) that borrow only from banks with NFCs that borrow from at least one non-bank lender
on the syndicated loan market. Assets are in USD millions. Employees are in thousands. The interest rate coverage ratio is computed as
earnings over interest expenses. Leverage is defined as the sum of long-term debt and current liabilities over total equity. Return on assets
is defined as operating income net of depreciation over total assets, in per cent. The sample covers 14,443 unique NFCs with 56,295 borrower-
year observations in total. The sample period is from 1990 to 2019.

Sources: Chava and Roberts (2008); S&P Compustat; Thomson Reuters DealScan; authors’ calculations.

6
We link the DealScan data set to balance sheet data on listed firms from Compustat Global and North
America following the updated matching table of Chava and Roberts (2008). Around 40% of all
observations are matched, as not all borrowing firms are covered in Compustat. Firms borrowing
from at least one non-bank in a given year account for 19,500 firm-year observations. The small
sample of firms borrowing exclusively from non-banks (ca 2,500 observations) exhibits similar
patterns in the measures of risk. Among firms with credit ratings, the share rated high-yield is
significantly higher for NFCs borrowing from non-banks.

22 BIS Quarterly Review, March 2022

 We now examine to what extent loan and borrower characteristics drive the
spread differential in a regression analysis. The analysis focuses on lead arrangers as
they negotiate loan terms and usually contribute material shares to the syndication
amount.
Loan characteristics explain only a small part of the spread differential. We start
by recording that the average loan spread on loans with non-bank arrangers is 84
basis points higher than on those without (Table 2, column 1). Accounting for loan
amounts, maturity and loan currency narrows the gap only marginally, to 77 basis
points (column 2). This is consistent with the earlier finding that loan terms are
broadly comparable across lender types.
Borrower characteristics, on the other hand, explain a significant fraction of the
higher spread on loans arranged by non-banks. Accounting for borrowers’ country
and industry as well as the year of origination reduces the difference by one third, to
50 basis points (Table 2, column 3). Further considering differences in borrower
characteristics such as size, profitability and leverage halves the remaining difference

Borrower characteristics largely account for higher spreads on non-bank loans Table 2

(1) (2) (3) (4)

Dependent variable: Spread Spread Spread Spread

Non-bank 83.981*** 76.832*** 49.536** 26.883**
(20.911) (19.240) (19.966) (11.749)
Loan maturity 37.339*** 35.748*** 16.066***
(3.915) (3.101) (1.599)
Loan amount –14.311*** –20.571*** –7.674***
(2.190) (1.833) (1.456)
Size –23.015***
(1.068)
Return on assets –345.909***
(19.293)
Leverage 5.282***
(0.200)
Interest coverage ratio –0.108***
(0.028)
Currency fixed effects (FEs) No Yes Yes Yes
Country-industry-year FEs No No Yes Yes
Number of observations 379,516 379,516 379,516 130,559
R-squared 0.039 0.100 0.483 0.608
Observations are at the lender-borrower-year level and the sample is restricted to lead arrangers. The dependent variable is the all-in spread,
defined as the spread over Libor, including fees and interests. It is winsorised at the 1st and 99th percentiles. Column (1) includes the dummy
non-bank, which takes on a value of one if the lender is a non-bank. Column (2) adds the logarithm of loan maturity in months (loan maturity)
and the logarithm of loan amount in million US dollars (loan amount), as well as currency fixed effects. Column (3) adds borrower country-
industry-year fixed effects. Finally, column (4) adds borrower-specific balance sheet controls obtained from Compustat (see the notes of Table
1 for variable definitions). The reduction in sample size in this last column is due to the unavailability of balance sheet data for some of the
firms borrowing in the syndicated loan market. Standard errors are clustered at the lender level. ***/**/* indicates statistical significance at
the 1/5/10% level. The sample period is from 1990 to 2019.

Sources: S&P Compustat; Thomson Reuters DealScan; authors’ calculations.

BIS Quarterly Review, March 2022 23

to 27 basis points (column 4). Thus, borrower riskiness is a key driver of the spread
differential.7
The remaining difference in loan spreads between bank and non-bank arrangers
could be explained by lender-specific factors. For example, a closer lender-borrower
relationship has been shown to lead to higher spreads to compensate for better
access to credit during shocks (Bolton et al (2016)). Lower geographical diversification
has also been associated with higher spreads (Keil and Müller (2020)). Further, a
higher market share in a sector could allow lenders to charge relatively higher interest
rates (De Jonghe et al (2020)). Exploring these dimensions in the context of non-banks
would be an interesting avenue for future work.

Non-banks and the global transmission of shocks

Non-banks’ large global footprint could have implications for shock transmission
across borders. As the volume of non-banks’ syndicated lending expanded, so did the
attendant share of credit to foreign borrowers, from 3% for the average lender in the
early 1990s to 11% in 2019 (Graph 4, left-hand panel). This share is even higher for
larger lenders. A retrenchment of such international presence could have material
implications in the borrower countries.
To investigate how non-banks’ global lending responds to negative shocks, we
analyse whether non-bank lenders exhibit a “flight home” effect. So far established
for banks only, this effect refers to banks cutting their lending in foreign markets by
more than they do in their domestic market following a financial crisis in their home
country (Giannetti and Laeven (2012)). The relative retrenchment has been shown to
be stronger for banks with higher exposure to risky clients and less stable funding
sources.
Non-banks are likely to exhibit a more pronounced flight home effect relative to
banks. Not only do non-banks serve riskier borrowers, but they also rely more on
wholesale funding (Jiang et al (2020), Fleckenstein et al (2021)). Such funding is more
sensitive to price changes than retail deposits are, and makes lenders – non-banks
and banks alike (Aldasoro et al (2022a)) – more vulnerable to negative liquidity shocks
(Demirgüç-Kunt and Huizinga (2010), Ivashina and Scharfstein (2010)).
We find that, during a financial crisis in their home country, non-bank lead
arrangers reduce the share of loans to foreign borrowers by more than bank lead
arrangers.8 Both banks and non-banks curtail credit to foreign borrowers by more

7
Results are qualitatively similar when all syndicate participants are included (the coefficient estimates
on the dummy variable non-bank are 83 basis points, 72 basis points, 50 basis points and 30 basis
points in columns 1, 2, 3 and 4, respectively). They are similarly robust in regressions weighted by
each lender’s contribution to the total syndicated amount (105 basis points, 95 basis points, 66 basis
points and 50 basis points in columns 1, 2, 3 and 4, respectively). Re-classifying investment banks as
banks results in coefficient estimates of 80 basis points, 69 basis points, 50 basis points and 29 basis
points. Finally, comparing loans exclusively arranged by non-banks (7% of all loans) with those
arranged by banks only (72%) yields coefficient estimates of 163 basis points, 151 basis points, 115
basis points and 76 basis points. The robust effect of borrower characteristics on spread differentials
is in line with evidence for mid-sized US companies (Chernenko et al (2021)).
8
In terms of total lending, banks and non-bank lenders reduce their loan origination by a respective
3.3% and 5.4% during domestic crises. For the identification of financial crises, we rely on Laeven and
Valencia (2020). The two conditions defining a banking crisis are significant signs of financial distress

24 BIS Quarterly Review, March 2022

 than to domestic ones, even after controlling for crises in foreign countries. However,
while the share of new loans to foreign borrowers declines by 7.3 percentage points
for banks, it falls by 18.2 percentage points for non-banks (Graph 4, right-hand panel).
Box B provides details of the estimation.
The relative retrenchment is more pronounced for non-bank lending to risky
borrowers and among non-banks with less stable funding. Non-banks contract their
share of foreign loans by 22 percentage points when considering only the subset of
risky borrowers, defined as those obtaining loans with a spread above the yearly
industry median. Excluding investment banks, in turn, leads to a decline of 26.5
percentage points. This may reflect the fact that some investment banks tend to have
more stable funding than other types of non-bank lender (eg when they are part of
a bank holding company).
These findings suggest that non-banks transmit financial shocks across
countries, and by more than banks do. As a large literature has found that
contractions in syndicated lending reduce firm investment and employment,9 non-
banks’ retrenchment is likely to have consequences for the real economy.

Non-banks have a stronger home bias that strengthens during a domestic crisis Graph 4

Non-banks increased lending to foreign firms, but less so During domestic crises, non-banks cut credit to foreign
than banks1 borrowers by more than banks2
Per cent Percentage points

60 –0.1

40 –0.2

20 –0.3

0 –0.4
1995 2000 2005 2010 2015 Banks Non-banks NB risky NB excl
(NB) borrowers investment
Share of loans to foreign borrowers: banks
Banks Banks (weighted)
Non-banks Non-banks (weighted) Flight home coefficient estimate
90% confidence interval
1
Average share of new loans extended to foreign borrowers. “Weighted” refers to the weighted average, with weights given by the total
syndicated loan amount originated by a lender in a given year. 2 Based on lender parent-borrower country-year level panel regressions for
lead arrangers only. The dependent variable captures the share of new loans by lender l in country c in quarter t out of total new syndication
by that lender in that quarter. It is regressed against the dummy foreign loan that equals one if the location of the arranger’s headquarters
differs from the location of the borrower, as well as the interaction of foreign loan with crisis lender country that captures financial crises in
the arranger’s home country. The displayed coefficients are those on the explanatory variable foreign loan X crisis lender country. Regressions
control for the effect of borrower country crises on foreign loans as well as the (logarithm of) the lender country-specific loan spread and
include time-varying borrower country fixed effects. “Banks” and “non-banks” refer to coefficients obtained from separate regressions for
bank and non-bank lead arrangers. “Risky borrowers” focuses on lending by non-bank arrangers to borrowers obtaining loans with a spread
above the yearly industry median. “Excl investment banks” excludes investment banks from the sample of non-bank arrangers. The sample
period is from 1990 to 2019.

Sources: Laeven and Valencia (2020); Thomson Reuters DealScan; authors’ calculations.

in the banking system and significant banking policy intervention measures in response to losses in
the banking system.
9
See above for the real effects of bank syndicated lending. Elliott et al (2019, 2021) and Aldasoro et al
(2022b) establish the real effects of non-bank syndicated lending.

BIS Quarterly Review, March 2022 25

 Box B

The flight home effect among non-bank lenders

Iñaki Aldasoro, Sebastian Doerr and Haonan Zhou

In this box, we investigate the flight home effect (FHE) among non-banks. The FHE refers to the finding in Giannetti
and Laeven (2012) that, during financial crises in their home country, lead arranger banks cut lending to foreign
borrowers by more than they did to domestic borrowers. Giannetti and Laeven also find that the effect is driven by
banks that are more exposed to riskier borrowers and with more volatile funding.

Following the methodology in Giannetti and Laeven, we estimate regressions at the lender parent-borrower
country-quarter level. We do so separately for bank and non-bank lead arrangers:

𝑙𝑜𝑎𝑛 𝑠ℎ𝑎𝑟𝑒 𝛼 𝑓𝑜𝑟𝑒𝑖𝑔𝑛 𝑙𝑜𝑎𝑛 𝛼 𝑓𝑜𝑟𝑒𝑖𝑔𝑛 𝑙𝑜𝑎𝑛 𝑐𝑟𝑖𝑠𝑖𝑠 𝑙𝑒𝑛𝑑𝑒𝑟 𝑐𝑜𝑢𝑛𝑡𝑟𝑦 𝜃 𝜏 𝑐𝑜𝑛𝑡𝑟𝑜𝑙𝑠 𝜖 . 1

The dependent variable captures the share of new loans by lender l in country c in quarter t out of total new
syndication by that lender in that quarter. Using the share as a dependent variable has the benefit that it is unaffected
by shocks changing the lender’s overall supply of loans. Instead, it captures the allocation of a lender’s new loans to
domestic and foreign markets in response to changes in economic conditions. Foreign loan is a dummy variable that
equals one if the location of the arranger’s headquarters differs from the borrower’s, and zero otherwise. Crisis lender
country captures financial crises in the arranger’s home country, obtained from Laeven and Valencia (2020).

The interpretation of the key coefficients is as follows. The coefficient 1 indicates whether lenders systematically
issue a higher or lower share of their new loans to the average foreign country, relative to their home country. The
coefficient 2 captures the FHE. A negative estimate suggests that lenders reduce lending to foreign borrowers relative
to domestic borrowers when their home country experiences a crisis. A higher absolute value of 2 indicates a more
pronounced FHE.

The flight home effect is stronger for non-bank lenders Table B

(1) (2) (3) (4) (5)

Sample: Banks Non-banks (NB) NB risky NB excl NB excl GFC

borrowers investment
banks
Dependent variable: Loan share Loan share Loan share Loan share Loan share
Foreign loan (FL) –0.422*** –0.328*** –0.224*** –0.343*** –0.326***
(0.004) (0.015) (0.017) (0.019) (0.016)
FL X crisis lender country –0.073*** –0.182*** –0.220*** –0.265*** –0.178***
(0.006) (0.028) (0.043) (0.039) (0.033)
Number of observations 53,987 7,479 5,066 5,205 6,916
R-squared 0.513 0.670 0.643 0.684 0.669
The dependent variable in all regressions is loan share, defined as the share of new syndicated loans by lender l in country c in quarter t out
of total new syndication by that lender in that quarter. Foreign loan is a dummy variable that takes on a value of one if the location of the
lender’s headquarters differs from that of the borrower, and zero otherwise. Crisis lender country refers to a financial crisis in the home country
of the lender. All columns include borrower country-time fixed effects, an interaction term of foreign loan with the dummy crisis in borrower
country and the log of the lender-borrower country-specific spread. In column (3), risky borrowers are defined as those borrowing at prices
above the industry median in any given year. Column (4) excludes investment banks from the group of non-bank lenders. Column (5) excludes
the Great Financial Crisis (GFC) from the sample. Standard errors are robust. ***/**/* indicates statistical significance at the 1/5/10% level. The
sample period is from 1990 to 2019 and focuses on lead arrangers only.

Sources: Laeven and Valencia (2020); Thomson Reuters DealScan; authors’ calculations.

26 BIS Quarterly Review, March 2022

 The allocation of new loans could also be affected by crises in borrower countries, and more broadly by changes
in demand. We control for the differential effect of borrower country crises on foreign loans directly by interacting
foreign loan with a dummy for crisis years in borrower countries, which implies that the coefficient 2 reflects relative
changes in lending to foreign countries after accounting for the effects of crises there. To account for demand shocks
and borrower risk, regressions include time-varying borrower country fixed effects, and control for the (logarithm of)
the median lender country-specific loan spread. However, we cannot fully account for (unobservable) changes in
borrowers’ loan demand, which could also be affected by crises outside the borrower country. Estimated coefficients
hence reflect lender characteristics, but also differences in the pool of borrowers within a country. That said, a domestic
financial crisis is likely to have a stronger effect on the performance – and hence loan demand – of domestic borrowers
compared with foreign borrower countries, implying an attenuation bias in the estimates.

Results in Table B show that the FHE is stronger for non-bank arrangers. Column (1) first shows a FHE among
lead arrangers that are banks, with estimates quantitatively similar to those in Giannetti and Laeven. In column (2) we
estimate the same specification for non-bank arrangers. As indicated by the point estimate of 1, the average
difference in the shares of new loans between foreign and domestic countries is larger for banks than non-banks (42.2
percentage points vs 32.8 percentage points). More importantly, the FHE is stronger for non-bank arrangers (2 of
18.2 percentage points vs 7.3 percentage points for banks). In column (3) we show that the FHE becomes more
pronounced when considering only the subset of riskier borrowers, defined as those obtaining loans with a spread
above the yearly industry median. Column (4) shows that results strengthen further when excluding investment banks,
which could have more stable funding than other non-banks. Finally, the results are also robust to excluding the Great
Financial Crisis in column (5).

 The views expressed in this box are those of the authors and do not necessarily reflect those of the BIS.  In unreported regressions,
we find that the existence of a stronger flight home effect for non-banks is also robust to the exclusion of the euro zone crisis, as well as the
exclusion of different regions and countries that are home to a significant number of non-bank lenders (eg Japan, the United Kingdom and
the United States).

Conclusion

Policymakers have so far mostly focused on the role of non-banks during severe
market stress. Yet, non-banks are also important providers of credit to non-financial
corporates across the globe. Indeed, this feature finds that, relative to banks, non-
banks lend to a riskier pool of borrowers. To the extent that such firms have limited
access to financing (Chernenko et al (2021)), non-bank lenders contribute to a more
diversified and efficient financial system. That said, the global footprint of non-bank
lenders may be a destabilising force. Not only is their credit provision more volatile,
possibly reflecting their greater reliance on wholesale funding, but it is also more
concentrated than that of banks. The resulting procyclicality of non-bank lending
warrants continued analysis, as does the tendency of non-bank lenders to transmit
shocks across borders to a larger extent than banks do.

BIS Quarterly Review, March 2022 27

References

Acharya, V, T Eisert, C Eufinger and C Hirsch (2018): “Real effects of the sovereign debt
crisis in Europe: evidence from syndicated loans“, Review of Financial Studies, vol 31,
no 8, pp 2855–96.
Aldasoro, I, J Caparusso and Y Chen (2022a): “Global banks’ local presence: a new
lens”, BIS Quarterly Review, March, pp 47–59.
Aldasoro, I, S Doerr and H Zhou (2022b): “Non-bank lending during financial crises“,
working paper.
Aramonte, S and F Avalos (2021): “The rise of private markets”, BIS Quarterly Review,
December, pp 69–82.
Aramonte, S, S J Lee and V Stebunovs (2019): “Risk taking and low longer-term
interest rates: evidence from the US syndicated term loan market”, Journal of Banking
& Finance, forthcoming.
Aramonte, S, A Schrimpf and H S Shin (2021): “Non-bank financial intermediaries and
financial stability”, BIS Working Papers, no 972, October.
Bolton, P, X Freixas, L Gambacorta and P Mistrulli (2016): “Relationship and
transaction lending in a crisis”, Review of Financial Studies, vol 29, no 10, pp 2643–76.
Carstens, A (2021): “Non-bank financial sector: systemic regulation needed”, BIS
Quarterly Review, December, pp 1–6.
Cerutti, E, G Hale and C Minoiu (2015): “Financial crises and the composition of cross-
border lending”, Journal of International Money and Finance, vol 52, pp 60–81.
Chava, S and M Roberts (2008): “How does financing impact investment? The role of
debt covenants”, Journal of Finance, vol 63, no 5, pp 2085–121.
Chernenko, S, I Erel and R Prilmeier (2021): “Why do firms borrow directly from
nonbanks?”, working paper.
Chodorow-Reich, G (2014): “The employment effects of credit market disruptions:
firm-level evidence from the 2008–09 financial crisis”, Quarterly Journal of Economics,
vol 129, no 1, pp 1–59.
Chodorow-Reich, G and A Falato (2022): “The loan covenant channel: how bank health
transmits to the real economy”, Journal of Finance, vol 77, no 1, pp 85–128.
De Jonghe, O, H Dewachter, K Mulier, S Ongena and G Schepens (2020): “Some
borrowers are more equal than others: bank funding shocks and credit reallocation”,
Review of Finance, vol 24, no 1, pp 1–43.
Demirgüç-Kunt, A and H Huizinga (2010): “Bank activity and funding strategies:
the impact on risk and returns”, Journal of Financial Economics, vol 98, no3,
pp 626–50.
Doerr, S, M Raissi and A Weber (2018): “Credit-supply shocks and firm productivity in
Italy”, Journal of International Money and Finance, vol 87, no 10, pp 150–71.
Doerr, S and P Schaz (2021): “Geographic diversification and bank lending during
crises“, Journal of Financial Economics, vol 140, no 3, pp 768–88.
Elliott, D, R Meisenzahl, J-L Peydro and B Turner (2019): “Nonbanks, banks and
monetary policy: US loan-level evidence since the 1990s”, working paper.
Elliott, D, R Meisenzahl and J-L Peydro (2021): “Nonbank lenders as global shock
absorbers: evidence from US monetary policy spillovers”, working paper.

28 BIS Quarterly Review, March 2022

Financial Stability Board (2021): Global monitoring report on non-bank financial
intermediation, December.
Fleckenstein, Q, M Gopal, G Gutierrez Gallardo and S Hillenbrand (2021): “Nonbank
lending and credit cyclicality”, working paper.
Gadanecz, B (2004): “The syndicated loan market: structure, development and
implications”, BIS Quarterly Review, December, pp 75–89.
Giannetti, M and L Laeven (2012): “The flight home effect: evidence from the
syndicated loan market during financial crises”, Journal of Financial Economics,
vol 104, no 1, pp 23–43.
Güler, O, M Mariathasan, K Mulier and N Okatan (2021): “The real effects of banks’
corporate credit supply: a literature review”, Economic Inquiry, no 59, pp 1252–85.
Hale, G, T Kapan and C Minoiu (2020): “Shock transmission through cross-border bank
lending: credit and real effects“, Review of Financial Studies, vol 33, no 10,
pp 4839–82.
Ivashina, V and D Scharfstein (2010): “Bank lending during the financial crisis of
2008”, Journal of Financial Economics, vol 97, no 3, pp 319–38.
Jiang, E, G Matvos, T Piskorski and A Seru (2020): “Banking without deposits: evidence
from shadow bank call reports”, working paper.
Keil, J and K Müller (2020): “Bank branching deregulation and the syndicated loan
market”, Journal of Financial and Quantitative Analysis, vol 55, no 4, pp 1269–303.
Laeven, L and F Valencia (2020): “Systemic banking crises database II”, IMF Economic
Review, vol 68, pp 307–61.
Sette, E and G Gobbi (2015): “Relationship lending during a financial crisis“, Journal of
the European Economic Association, vol 13, no 3, pp 453–81.
Sufi, A (2007): “Information asymmetry and financing arrangements: evidence from
syndicated loans”, Journal of Finance, vol 62, no 2, pp 629–68.
Xiao, K (2020): “Monetary transmission through shadow banks”, Review of Financial
Studies, vol 33, no 6, pp 2379–420.

BIS Quarterly Review, March 2022 29