The relation between bank credit growth and the expected returns of bank stocks

• Author: Priyank Gandhi
• DOI: http://doi.org/10.1111/eufm.12179
• Published date: 01 September 2018
• Date: 01 September 2018

DOI: 10.1111/eufm.12179

ORIGINAL ARTICLE

The relation between bank credit growth and the

expected returns of bank stocks

Priyank Gandhi

Rutgers Business School, Rutgers

University, New Brunswick

Email: pgandhi@rbsmail.rutgers.edu

Abstract

Higher bank credit growth implies that excess returns of

bank stocks over the next one year are lower by nearly 3%.

Credit growth tracks bank stock returns over the business

cycle and explains nearly 14% of the variation in bank stock

returns over a 1-year horizon. I argue that the predictive

variation in returns reflects investors' rational response to a

small time-varying probability of a tail event that impacts

banks and bank-dependent firms. Consistent with this

hypothesis, the predictive power, as measured by the

absolute magnitude of the coefficient on credit growth and

the adjusted-R

2

at the 1-year horizon, depend systematically

on variables that regulate exposure to tail risk.

KEYWORDS

bank credit growth, bank equity returns, tail risk

JEL CLASSIFICATION

G01, G02, G15, G21

1

|

INTRODUCTION

Understanding the relation between bank credit and equity markets has important implications for the

link between the financial sector and the real economy, and for the design of widespread government

This paper is based on my dissertation written at the University of California, Los Angeles. I would like to thank my

committee chairs, Francis Longstaff and Hanno Lustig, for their invaluable guidance. I would also like to thank Robert

Battalio, Antonio Bernardo, Shane Corwin, John Doukas, Andrea Eisfeldt, Mark Garmaise, Paul Schultz, Eduardo

Schwartz, Ivo Welch, Lu Zhang (the Editor), an anonymous referee and seminar participants at the University of California

at Los Angeles, the Ohio State University, Emory University, University of Notre Dame, Casewestern Reserve University,

and the University of Western Ontario for their invaluable comments and suggestions. All errors are my responsibility.

610

|

© 2018 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/eufm Eur Financ Manag. 2018;24:610–649.

policies regarding bank regulation and supervision. While a large literature establishes that bank credit

growth is highly pro-cyclical, and that bank credit availability affects borrower firms' valuation, studies

of the relation between bank credit growth and bank valuations are more limited.

1

In this paper, I use a monthly dataset of bank credit in the US from 1960–2017 to study its

relation with the returns on an index of bank stocks. This monthly dataset allows me to unmask

even short or mild variations in bank credit, and correlate bank credit more precisely with

variations in bank stock returns. Since the US is also one of the few markets with a long history of

equity returns for a broad cross-section of firms, this allows for additional cross-sectional analysis

that the international setting of other studies does not permit. To my knowledge, this is the first

paper that utilizes monthly data on bank credit and examines how this relation varies in the cross-

section.

I run basic forecasting regressions of future bank stock returns on bank credit growth. An increase

in bank credit growth is associated with lower excess returns on bank stocks. That is, bank credit

growth and the excess returns of banks stocks are negatively correlated. A 1% increase in bank credit

growth predicts that excess returns of bank stocks over the next one year are lower by 2.61%. Unlike

most other forecasting relations, bank credit growth tracks bank stock returns over the business cycle.

The predictive power of bank credit growth for bank stock returns is economically large. Over a

horizon of one year, bank credit growth explains 13.52% of the variation in bank stock returns. The

annual R2, when bank credit is used to forecast bank stock returns, peaks at 17.78% at a horizon of

43 months.

The negative relation between bank credit growth and bank stock returns fits within the framework

of a standard neoclassical model (such as the rare disasters framework of Barro, 2006; Gabaix, 2008;

Rietz, 1988; and Wachter, 2013) and classical models of investment behavior (such as the Q-theory of

investments of Liu, Whited, and Zhang (2009). In these models, discount rates vary over time, are

counter-cyclical, and reflect the representative agent's rational response to a small time-varying

probability of a tail event. Banks use these time-varying counter-cyclical discount rates to

determine the cost of capital and to evaluate projects they can fund. An improvement in

macroeconomic conditions (i.e. reduction in the probability of tail risk) reduces discount rates,

increases the expected present value of profits, and the optimal rate of investments by banks

increases. That is, bank credit grows. Banks are more sensitive to changes in the probability of a tail

event due to their higher leverage, higher proportion of short-debt that is payable on demand, and

lower fraction of tangible assets. In concert, these characteristics increase the likelihood that a bank

(as compared to a typical non-financial firm) will enter distress upon the realization of a tail event.

An implication of this framework is that the relation between bank credit growth and equity returns

should vary in the cross-section. Since small banks, banks with more leverage, and banks with more

short-term debt are more exposed to tail risk, bank credit should correlate more with equity returns of

banks with these characteristics.

2

Indeed, this is exactly what I find in the data. The predictive power of

bank credit growth for bank stock returns, as measured by the absolute magnitude of the coefficient on

bank credit growth and the adjusted- R2at the 1-year horizon, depends systematically on size,

leverage and the proportion of short-term debt. The predictive power decreases monotonically in

1

A notable exception is Baron and Xiong (2017) who use a dataset of 46 countries over 100 years to study the relation

between bank credit and bank stock returns. I discuss below the incremental contribution of this paper compared to theirs.

2

Ivashina and Scharfstein (2009) show that banks with more short-term debt are more exposed to tail risk. I show in

Section 3 below that small banks are more exposed to tail risk.

GANDHI

|

611

size, as measured by market capitalization, and increases in leverage, and in the proportion of short-

term debt employed by the bank.

The forecasting relation between bank credit growth and bank stock returns is robust to several

empirical specifications. The results hold when I exclude the data for the financial crisis, and when I

include other forecasting variables used in the literature. Bank credit growth does not predict future

cash flows of bank stocks. It does, however, predict returns for investment banks and bank-dependent

firms. Tail events that impact banks can also affect investment banks. Like banks, investment banks

employ high leverage, rely on short-term debt, do not own substantial tangible assets and are more

sensitive to changes in the probability of a tail event. Finally, an increase in tail event risk also increases

expected returns of any direct equity claim on bank-funded projects, and this explains the negative

correlation between bank credit growth and returns of bank-dependent firms.

Bank credit growth does not strongly predict returns of any other asset class (non-financial firms,

treasury bonds or corporate bonds). This is possible if tail events result in large losses for projects

funded by banks, but do not impact other projects in the economy. Examples of such events are the Less

Developed Country Debt crisis of 1982, the Mexico crisis of 1994, the East Asian crisis of 1997, and

the Long-Term-Capital-Management crisis of 1998. Each of these events resulted in a large loss of cash

flows from projects funded by banks, and a significant drop in bank profitability and valuation.

However, there was no measurable effect on the performance of other projects in the economy. These

crises were not accompanied by a recession, and other important asset markets such as the stock and

housing markets, were relatively unaffected.

3

My paper is related to seminal work by Ba ron and Xiong (2017) and Schularic k and Taylor

(2012), who analyze the relati on between bank credit, macroe conomic conditions, and equit y

markets in more than 46 countries ov er 100 years. Schularick and Tayl or (2012), show that bank

credit expansion increases macroeconomic (crash) ris k in the near future. Baron and Xio ng (2017)

also document a negative cor relation between bank cred it and equity returns. Howev er, given the

results in Schularick and Tay lor (2012), they attribut e this negative relation to th e neglect of crash

risk by investors.

My results contrast with Baron and Xiong (2017) and Schularick and Taylor (2012) and are

consistent with a significantly different insight: bank credit simply responds to exogenous changes in

macroeconomic conditions as one would expect. As macroeconomic (tail) risk declines, bank credit

increases, and vice-versa. In other words, worsening macroeconomic conditions increase the risk of

potential borrowers (they hurt borrower's net worth or collateral values), and forward-looking

decisions by bank managers anticipate time variation in macroeconomic conditions. Thus, the

predictive ability of bank credit for equity returns is due to a simple ‘price of risk’mechanism: bank

credit increases, but expected equity returns are reduced with improving macroeconomic conditions.

My results are also consistent with the NPV rule (or the Q-theory) of investments that prescribes

banks evaluate lending opportunities by discounting projected cash flows at discount rates that

appropriately reflect investment risk. Thus, bank credit growth should be negatively correlated with

future macroeconomic (tail) risk.

Yet, Baron and Xiong (2017) and Schularick and Taylor (2012) document that bank credit

growth is sometimes positively correlated with future macroeconomic (crash) risk. How does one

reconcile the results? One possibility is that market frictions can sometimes cause deviations from

the NPV rule (Q-theory) of investments. An example of such a market friction is the agency cost of

3

One implication of this hypothesis is that banks actively fund projects with higher exposure tail risk. I return to this

question in section 3 below.

612

|

GANDHI