The Market Perception of Banking Industry Risk: A Multifactor Analysis

Article excerpt

Introduction

One of the most important issues in the ongoing debate on the risk of the commercial banking industry and regulatory reform is whether movements of bank stock prices can provide accurate assessments of bank risk and reliable signals of the risk-taking behavior of individual banks.(1) Therefore, insights into the main factors that underlie movements in bank stock prices, which reflect the stock market's perception of the risk of banking firms based on available information, are useful to owners and managers of banks as well as regulators.

Numerous studies indicate that several events in the 1970s and 1980s which affected the banking industry (e.g., less developed countries' loan problems, deregulation and other regulatory events, monetary policy changes, and domestic or regional economic problems) had significant differential impact on the stock prices of banks, depending on the nature of the event and the size or type of bank [Madura, 1990]. Also, several studies provide evidence that, in addition to market and economic factors, subgroup and bank-specific characteristics such as size, the maturity composition or duration of assets and liabilities, and accounting variables are reflected in bank stock prices [Akella and Greenbaum, 1992; Neuberger, 1991; Saunders and Yourougou, 1990; Bae, 1990; Unal and Kane, 1988; Brewer and Lee, 1986; Flannery and James, 1984].

The purpose of this paper is to complement prior research through the employment of a different methodology to analyze the underlying factors that influence bank stock returns. Specifically, this study uses both a multifactor model and principal component (PC) analysis to provide evidence on the stock market perception of banking industry risk over the 1969-89 period and on the relative influences of market, economic, industry, and bank-specific sources of risk as reflected in bank holding company (BHC) stock returns.

The multifactor model is an extension of Stone's model [1974] to include economy-wide and banking industry-specific default risk variables in addition to proxies for changes in both short-term and long-term interest rates. The correlations between factors extracted using principal component analysis from bank stock returns and the variables of the multifactor model are then analyzed.

The results provide useful additional insights into the relative influences of the factors underlying the stock returns of commercial banking firms beyond what is usually evident from studies that use the Stone model and event studies. First, the two default risk variables provide insights on the perceived higher risk of the banking industry from the mid-1970s through the 1980s that may not be evident from the market and interest rate variables of the Stone model. Second, the correlations between the factors of the multifactor model and those extracted using PC analysis depict the significant relationship between industry and intra-industry PC factors and the interest rate and default risk factors of the multifactor model. Finally, the PC analysis indicates only one clearly discernible subgroup factor beyond the common market or industry factor and it is apparently related to bank size or characteristics for which size is a proxy.

This paper is presented as follows. In the second section, the multifactor model and results of regressions of the model are presented. The results of the principal component analysis are discussed in the third section in relation to the factors of the multifactor model. The fourth section summarizes and concludes the study.

The Multifactor Model: Framework and Tests

A Five-Factor Model

A multifactor model is used to analyze the influence of market, economic, and industry factors in BHC stock returns at the industry level. The model consists of interest rate variables in addition to the market factor. The use of interest rate variables is predicated on the notion that interest rates contain information about expected economic performance and, therefore, changes in interest rates reflect changes in expectations about economic variables and expected stock returns [Merton, 1973; Chen et al. …