What influences a bank's decision to go public?

• Date: 01 October 2019
• Author: Ping Zhang,Serafeim Tsoukas,Georgios Sermpinis
• Published date: 01 October 2019
• DOI: http://doi.org/10.1002/ijfe.1740

RESEARCH ARTICLE

What influences a bank's decision to go public?

Georgios Sermpinis

1

| Serafeim Tsoukas

1

| Ping Zhang

2

1

Adam Smith Business School, University

of Glasgow, Glasgow,G12 8QQ, United

Kingdom

2

School of Insurance, Central University of

Finance and Economics, China

Correspondence

Georgios Sermpinis, University of Glasgow,

Glasgow, UK.

Email: georgios.sermpinis@glasgow.ac.uk

Funding information

China Scholarship Council, Grant/Award

Number: 201508060273

Abstract

A bank's decision to go public by issuing an Initial Public Offering (IPO) trans-

forms its operations and capital structure. Much of the empirical investigation in

this area focuses on the determinants of the IPO decision, applying accounting

ratios and other publicly available information in nonlinear models. We mark a

break with this literature by offering methodological extensions and an extensive

and updated U.S. dataset to predict bank IPOs. Combining the least absolute

shrinkage and selection operator with a Cox proportional hazard, we uncover value

in several financial factors as well as market-driven and macroeconomic variables

in predicting a bank's decision to go public. Importantly, we document a significant

improvement in the model's predictive ability compared with standard frameworks

used in the literature. Finally, we show that the sensitivity of a bank's IPO to finan-

cial characteristics is higher during periods of global financial crisis than in calmer

times.

KEYWORDS

Equity financing, financial ratios, forecasting, LASSO, U.S. banks

JEL CLASSIFICATION

D40; G33; F23; C25; O16

1|INTRODUCTION

Market finance in the United States has become an important

source of funding for banks. According to the Federal

Reserve Board, over the period 1996 to 2016, the net new

issuance of U.S. financial corporate equities outstanding

more than tripled, from less than $50 billion to over $150

billion. The same body reports that the market value of total

U.S. corporate equity issues has risen from about $8 trillion

in 1996 to around $36 trillion in 2016. This means that

market participants have taken advantage of economic

conditions as interest rates fell to historic lows. But not all

banks were in a position to benefit from these unusual

conditions as some financial institutions may rely less on

equity financing as their funding largely comes from

customer deposits and interbank financing.

1

However,

during economic downsides, banks face tighter lending

conditions and increased levels of interbank borrowing costs

(see Farinha, Spaliara, & Tsoukas, 2019; Iyer, Peydró, da

Rocha-Lopes, & Schoar, 2013). Using new estimation

techniques across an extensive sample period that covers

both periods of crisis and calmer times, the present study

aims to identify the factors that influence a bank's desire to

issue an equity Initial Public Offering (IPO).

Our study considers the influence of bank-level financial

information and market-level indicators; we ask how these

explicators influence the decision at the level of the bank to

issue stocks for the first time. The focus is on the decision of

a bank to go public by issuing an IPO, which is a financially

significant step for a bank and provides new opportunities

for financial flexibility, increased liquidity, better

diversification, and attracting potential investors (Amihud &

Mendelson, 1988; Bodnaruk, Kandel, Massa, & Simonov,

2007; Kim & Weisbach, 2008; Lowry, 2003; Lowry,

DOI: 10.1002/ijfe.1740

wileyonlinelibrary.com/journal/ijfe

Int J FinEcon. 2019;24:1464–1485.

1464

© 2019 John Wiley & Sons, Ltd.

Michaely, & Volkova, 2017; and Pagano, 1993). In addition,

Houge and Loughran (1999) demonstrate that a bank's IPO

decision can help managers to satisfy regulatory capital

requirements, sell overvalued stock, and take advantage of

better growth opportunities. After going public, Harris and

Raviv (2014) indicate that the conditions of underlying

market discipline and capital markets have more consider-

able influence on a public bank's ability to take risk than on

a private bank. Samet, Boubakri, and Boubaker (2018)

further clarify that public banks are able to take less credit

risk during noncrisis periods compared with private banks.

Moreover, if banks go public, market discipline can improve

credibility and transparency in the banking industry and

force public banks to maintain operational quality because of

regular announcements of their financial health (Delis,

Molyneux, & Pasiouras, 2011).

In this paper, we extend the literature methodologically,

by developing a series of Cox proportional hazard (CPH),

discrete hazard (DH), and logistic models combined with a

more intuitive, yet innovative model, which is based on the

variable selection technique, pioneered by Tibshirani

(1996)—the least absolute shrinkage and selection operator

(LASSO). This model, also known as L1 norm penalty, has

proved very useful in identifying the most relevant predictors

from an extensive set of candidate variables, without consid-

ering a preselection of these potential variables (van de Geer,

2008). The LASSO selection approach has a number of

appealing characteristics: It not only helps identify the most

relevant predictors from an extensive set of candidate

variables but it also improves the predictive power (Fan &

Li, 2001; Tian, Yu, & Guo, 2015). In addition, LASSO does

not require strict assumptions such as a preselection of the

variables considered, and it is consistent statistically, as the

number of observations approaches infinity (van de Geer,

2008). Importantly, LASSO can potentially sidestep the

problem of multicollinearity, which is fairly common in

reduced-form models, and it is computationally efficient

even when considering a large set of potential predictors.

An additional important contribution of the present paper

is that we test the estimator with superior predictive ability

utilizing a panel of U.S. banks over an extensive time period.

This approach not only allows us to compare our results with

previous research but also consider different time periods.

Intuitively, banks respond in a different manner to extreme

economic events as opposed to noncrisis periods, when they

time their IPOs. Our sample covers the most recent global

financial crisis and calmer (pre and post crisis) periods. We

argue that across time periods, there is a differential sensitiv-

ity to bank and market information when it comes to the

probability of banks going public.

To preview our findings, we discover value in several

bank-specific financial factors as well as market-driven and

macroeconomic variables in predicting the decision of banks

to go public. In terms of the models' predictive ability, when

we apply the LASSO estimator in a CPH model, we note a

significant improvement in predicting a bank's IPO, and the

penalized CPH model outperforms other candidates. Specifi-

cally, we note improvements compared with a CPH, DH,

and logistic models with or without LASSO. On the other

hand, we show that the CPH model underperforms DH and

logistic models, which highlights the effect of LASSO on

our algorithms. Our L1 penalized models are tuned through

the Akaike information criterion (AIC) and the Bayesian

information criterion (BIC) criteria. We observe increased

predictability on our dataset when the latter criterion is

applied. Finally, when we put the model with superior pre-

dictive ability to the data and split our sample into crisis

period and noncrisis periods, we find that the above

variables become more potent in determining banks'

IPOs, which signifies the ability of banks to time their IPOs

relative to the economic conditions.

The rest of this work is laid out as follows. In Section 2,

we present a brief overview of the relevant literature.

Sections 3 and 4 contain the data statistics and methodolo-

gies, respectively. Section 5 explains the empirical results of

the forecasting simulation, and Section 6 presents the econo-

metric results of an empirical application. Section 7 provides

conclusions.

2|LITERATURE REVIEW

Pagano, Panetta, and Zingales (1998) present the first

systematic study on the determinants of firms' IPO. They

rely on a number of accounting indicators to predict Italian

firms' probability of issuing an IPO and conclude that larger

firms, those with greater growth rates or improved future

investment opportunities, are more likely to go public.

Several other studies confirm the importance of financial

health in determining access to the public market in

Germany (Boehmer & Ljungqvist, 2004), the United

Kingdom, and India (Albornoz & Pope, 2004; Mayur &

Kumar, 2016). In a slightly different setting, Helwege and

Packer (2003) exploit the requirement of the Securities

and Exchange Commission to obtain information about

U.S. public firms. The authors show that variables

measuring size, profitability, leverage, interest coverage,

R&D investment, capital structure, growth rate, future

investment opportunities, ownership information, and

riskiness all have an important role in influencing the

decision to issue an IPO.

2

As well as evaluating the importance of financial infor-

mation, Chemmanur, He, and Nandy (2009) find that total

factor productivity is a key contributor to the probability that

a firm will issue IPOs. They find that a pr ivate firm which is

SERMPINIS ET AL.1465