Do Business Borrowers Benefit from Bank Bailouts?: The Effects of TARP on Loan Contract Terms

• DOI: http://doi.org/10.1111/fima.12222
• Author: Tanakorn Makaew,Allen N. Berger,Raluca A. Roman
• Published date: 01 June 2019
• Date: 01 June 2019

Do Business Borrowers Benefit from

Bank Bailouts?: The Effects of TARP on

Loan Contract Terms

Allen N. Berger, Tanakorn Makaew, and Raluca A. Roman∗

We investigate benefits to business borrowers from bank bailouts, specifically the Troubled Asset

Relief Program(TARP). Applying difference-in-differencemethodology to loan-level data, we find

more favorableborrower contract terms in five dimensions, spread, amount, maturity, collateral,

and covenants, suggesting increased credit supply at the intensive margin by bailed-out banks.

Findings are robustto dealing with endogeneity and other issues. Riskier borrowers benefit more,

consistent with moral hazard exploitation. Small and unlisted borrowers benefit less, suggesting

fewer benefits for financially constrained firms. Benefits accrue to both relationship and nonre-

lationship borrowers. Results contribute to the research and policy debates on bank bailouts.

Do bank bailouts result in net benefits or costs for the borrowers of bailed-out banks? From a

policy perspective,whether or not bailouts are worthwhile depends upon their many consequences,

one of which is whether the borrowers from the recipient banks benefit. Many of the other

consequences, such as changes in real economic conditions, competitive advantages conferred,

and systemic risk, are coveredelsewhere in the literature, but there is very little evidence regarding

how bailouts affect borrowers. This is an important research question as treatment of loan

customers directly affects their financial conditions, which drive economic recovery and growth.

The Securities and Exchange Commission, as a matter of policy, disclaims responsibilityfor any private publications or

statements by any of its employees. The views expressedherein are those of the authors and do not necessarily reflect the

views of the Commission or of the authors’ colleagues on the staff of the Commission. Also, the views expressedherein

are those of the authors and do not necessarily reflect the views of the Federal Reserve Banks of Philadelphia, Kansas

City, or the Federal Reserve System. The authors thank the editor Rajkamal Iyer (Editor), an anonymous referee, Dean

Amel, Mitchell Berlin,Tara Bhandari, Natasha Burns, Indraneel Chakraborty,Nick Coleman, TroyDavig, Taeyoung Doh,

Andrew Foerster, Scott Frame, Bernhard Ganglmair, Todd Gormley, Bjorn Imbierowicz, Michael King, Kris Gerardi,

Rachita Gullapalli, Vladimir Ivanov,Anzhela Knyazeva, Diana Knyazeva, Mattias Nilsson, Chuck Morris, Ned Prescott,

Jordan Rappaport, Rich Rosen, Natalya Schenck,Rajdeep Sengupta, Ioannis Spyridopoulos, Anjan Thakor, Larry Wall,

Jim Wilkinson,Krzysztof Wozniak, Helen Zhang, and participants at the presentations at the ASSA Annual Meetings, FIRS

Annual Meetings, FederalReserve Bank of Kansas City Research Seminar, Federal Reserve System Committee Meetings

on Financial Structure and Regulation, Northern Financial Association Meetings, Financial Management Association

Meetings, Southern Finance Association Meetings, and the SEC’s Empirical Corporate Finance Research Group for

useful comments and suggestions. The authors also thank Lamont Black, Christa Bouwman, and Jennifer Dlugosz for

data on Discount Window (DW) and Term Auction Facility (TAF) programs. The authors also express gratitude to the

Darla Moore School of Business forthe grant that supported this research. Part of the workin this paper was completed

while Raluca was working at Federal Reserve Bank of Kansas City.

∗Allen N. Berger is the H. Montague Osteen, Jr., Professor in Banking and Finance and the Carolina Distinguished

Professor in the Darla MooreSchool of Business at the University of South Carolina in Columbia, SC and is affiliated

with the Wharton Financial Institutions Center and the European Banking Center. Tanakorn Makaew is a Financial

Economist in the Division of Economic and Risk at the Securities and Exchange Commission (SEC) in Washington,DC.

Raluca A. Roman is a Senior Financial Economist in the Risk Assessment, Data Analysis and Research(RADAR) Group

at the FederalReserve Bank of Philadelphia in Philadelphia, PA.

Financial Management •Summer 2019 •pages 575 – 639

576 Financial Management rSummer 2019

The event study evidence that does exist on this question is contradictory and only covers

borrowers with prior relationships with these banks.

We address this question by examining the effects of the US Troubled Asset Relief Program

(TARP) bailout during the recent financial crisis on loan contract terms to business borrowers of

bailed-out banks. Using difference-in-difference methodology, we find that business borrowers

get more favorable loan contract terms in all five dimensions studied. Conditional upon bor-

rower and bank characteristics, loan type, industry, and time, recipient banks granted loans with

lower spreads, larger amounts, longer maturities, less frequency of collateral, and less restrictive

covenants. This is consistent with an increase in credit supply at the intensivemargin, suggesting

that the recipient banks’ borrowers benefited from the TARP program. Our findings are statis-

tically and economically significant and are robust to dealing with potential endogeneity issues

and other checks.

Our approach departs from the existing literature in a number of important respects. First, while

most prior bailout research is at the bank level (Black and Hazelwood, 2013; Li, 2013; Berger and

Roman, 2015; Akin et al., 2016; Ng, Vasvari, and Wittenberg-Moerman, 2016) or market level

(Puddu and W¨

alchli, 2015; Berger and Roman, 2017), we use loan-level data and examine the

effects of TARPfrom the perspective of the borrowers. The loan-level data allow us to control for

borrower risk and other characteristics to determine if the supply of credit is affected by TARP

versus whether there is a change in demand that is manifested in these borrower characteristics.

This is important, as certain types of loan customers may self-select to borrow from TARP banks

or non-TARP banks, which would, in itself, cause loan contract terms to change. Numerous

studies of the effects of TARP on bank lending supply at the bank level are unable to account

for borrower characteristics and credit demand. As discussed below, this literature reports mixed

results.

In addition, to our knowledge, we are the first to use multidimensional information about bank

loans by examining the effects of TARP on five key loan contract terms. Previous literature

examines loan supply at the extensive margin, the quantity of loans. This is insufficient to

determine the benefits to borrowers due to the possibility that the banks extract all of the gains

from the borrowers in the form of higher loan spreads or other harsher loan terms. In additional

tests, we find that banks increased the probability of issuing loans during the post-TARP period

to their pre-TARPperiod bor rowers. This complements our main analysis by showingan increase

in credit supply at the extensive margin, as well as at the intensive margin.

Our loan-level analysis also allows us to identify which types of borrowers benefit most from

the bailout program. Our results suggest that improvements in loan contract terms are greater

for riskier borrowers than for safer borrowers. This is consistent with TARP banks attempting

to attract riskier borrowers that yield higher expected returns for the banks. This preferential

treatment of riskier borrowers is consistent with an increase in the exploitation of moral hazard

incentives. We also find that improvements in loan contract terms are greater for less financially

constrained borrowers than for more constrained borrowers, suggesting fewer benefits for those

borrowers most in need of funding, limiting the potential positiveeffects of TARP on the economy.

In addition, we find improved credit terms to both relationship and nonrelationship borrowers,

consistent with the notion that TARP banks used bailout funds to reach out to both new and

existing borrowers. This finding suggests that other studies focusing only on borrowers with prior

relationships with TARP banks may overlook some benefits of the program.

Our paper contributes to several strands of literature. We add to the literature regarding the

effects of bailouts on bank borrowers.Two eventstudies consider the valuation effects of TARP on

relationship customers and document opposing results. Norden, Roosenboom, and Wang (2013)

find that TARP led to a significantly positive impact on relationship firms’ stock returns around

Berger, Makaew, & Roman rDo Business Borrowers Benefit from Bank Bailouts? 577

the time of TARP capital injections. In contrast, Lin, Liu, and Srinivasan (2014) determine that

relationship borrowers suffered significant valuation losses around the time of the TARPapproval

announcements.

Our work adds to this research in three main ways. First,the valuation changes in these studies

may be due to expectations of better or worse direct treatment of the borrowers by TARP banks,

but it is unknown from these studies alone whether these expectations were met in practice. In

contrast, we examine actual changes in borrowertreatment. In effect, the event studies may reveal

a noisy signal about borrower treatment, while we measure it more directly.

In addition, stock returns around TARP dates may be partially driven by other indirect factors

that are not specifically related to the treatment of the loan customers, such as expectations

of changes in local economic conditions, but are correlated with bailouts of their banks. As

discussed below, the TARP selection criteria targeted “healthy, viable institutions,” which may

mean that TARP was more often given to banks in markets with improving local conditions

that, in turn, may be related to positive stock market returns for their relationship borrowers.

Unlike event studies in which all of the control variables must be measured before or on the

TARP announcement dates, we are able to control for borrower characteristics at the time the

loans are issued and examine the actual effects of TARP on the borrowers’ loan contract terms.

Controlling for borrower characteristics at the time of loan issuance is crucial for alleviating

the identification problem. This is because changes in local economic conditions and borrower

characteristics between TARP initiation and the time when the loan is issued may be correlated

with TARP acceptance, but not caused by TARP itself.

Finally, event studies are, by construction, limited to borrowers with existingr elationships with

banks and cannot measure the effects of TARP on nonrelationship borrowers. In contrast, we are

able to measure the latter effects and find that nonrelationship borrowers benefited slightly more

than relationship borrowers from the bailout program.

We also add to the studies that investigate the effects of bank bailouts on credit supply. As

briefly noted above, a number of studies examine the effects of bailouts on the quantity of credit

(i.e., the credit supply at the extensive margin) and reach conflicting results. Berrospide and

Edge (2010), Li (2013), Puddu and W¨

alchli (2015), Chavaz and Rose (2017), Berger and Roman

(2017), and Chu, Zhang, and Zhao (2017) find that TARP banks expanded their quantities of

credit. Black and Hazelwood (2013) find mixed results. Lin, Liu, and Srinivasan (2014) confirm

a decline in credit supply, while Bassett, Demiralp, and Lloyd (2016) and Duchin and Sosyura

(2014) do not find any evidence of a change in credit supply.We are able to extend the research to

cover the intensive margin orhow the borrowers that receive credit are treated using information

on five loan contract terms.

Weargue that a full understanding of credit supply effects requires examination at the intensive

margin. Prior studies of changes in loan quantities that disregard loan contract terms may fail

to capture important components of the TARP effects. For example, if TARP banks grant loans

with longer maturities, their borrowers may be ableto fund larger projects with longer investment

horizons. Additionally,lowercovenant intensity may relax borrower credit constraints and increase

investment flexibility, while less restrictive collateral requirements may allow borrowers to use

assets to secure financing from other sources. Our analysis of multiple loan contract terms

quantifies TARP’s economic impacts beyond changes in credit supplymeasured in a more limited

sense by the volume of new loan credit.

Our paper also supplements the bank bailout and moral hazard literature. Bailouts could raise

expectations of future bailouts increasing moral hazard incentives for banks to take on greater

risk (Acharya and Yorulmazer, 2007; Kashyap, Rajan, and Stein, 2008). Alternatively, bailouts,

such as TARP, could reduce moral hazard incentives due to additional bank capital or because