The Effects of Ownership Structure on Bank Performance During Recessions

“The effects of ownership structure on bank performance during recessions”

By Katie Lindsay

Abstract

In this study, I investigate the relationship between ownership structure of banks and performance during recessions. I focus on S-corporation and non-S-Corporations to explore whether the tax advantage of S-Corporations outweighs the disadvantages of restrictions on ownership structure during recessions. While previous literature discusses the superiority in performance of S-Corporations, it is possible that ownership restrictions may make it more difficult to raise equity capital during recessions thus diminishing some of the performance advantages of S-Corporations. Specifically, I test whether S-Corporations have higher performance during recessions than their non-S-Corporation counterparts compared to pre-recessionary periods. I find that not only do S-Corporation banks have better performance during recessions than non S-Corporation banks, but the margin of performance increases from pre-recessionary periods.

Introduction

In this study, I investigate the relationship between ownership structure of banks and performance during recessions. I focus on S-corporation and non-S-Corporations to explore whether the tax advantage of S-Corporations outweighs the disadvantages of restrictions on ownership structure during recessions. Specifically, I test whether S-Corporations have higher performance during recessions than their non-S-Corporation counterparts. This study is relevant in two ways. First, the effect of S-Corporations on bank value and performance is still a topic of conversation as financial institutions are often ignored in financial literature due to their different characteristics. However, banks are a major contributor to the United States economy and should not be overlooked; failing to understand such an important industry can only exacerbate future economic problems.

Second, this study is important because academics are still exploring the characteristics, causes, and effects of the recent recession on different industries and types of institutions. Fresh off the 2008 recession, there are a number of still unanswered questions that may help us to better understand imminent economic downturns. This study relates to the recession literature because the failure of the banking system is often discussed as the cause of the recent economic crisis. Thus, an understanding of how banks operate during recessions is imperative for avoiding similar future crises.

Why are S-Corporations unique? S-corporations are not subject to the normal double taxation laws of regular corporations; instead, they are only taxed at the shareholder level allowing these firms to still enjoy the benefit of the limited liability of corporations. However, to qualify as an S-Corporation, the institution must meet a number of restrictions. Though S-Corporations have been around for some time, banks were not allowed to organize in this manner and thus could not capitalize on their tax advantages. The passing of a law in 1996 allowed banks to begin organizing as S-Corporations in 1997 if the proper requirements were met. In order to achieve the optional tax advantages, the institution must be a domestic corporation, have no more than 100 shareholders, have only one class of stock, and restricted ownership types. Shareholders must be individuals or certain trusts and estates; partnerships, corporations, and non-resident aliens are not allowed to hold shares of S-corporations. So, while an S-corporation has tax benefits for banks, the qualification restrictions on ownership structure as well as the cost to convert could be disadvantageous. According to subsbanks.org, nearly a third of all banks in the US are now classified as S-corporations.

The rest of the paper is organized as follows. A summary of related literature is covered in Section 2. Section 3 discusses the motivation and summary statistics inspiring this study. Section 4 discusses the data while Section 5 explains the methodology and results. Section 6 summarizes the main findings and Section 7 discusses paths of future development.

Literature Review

Much of the early S-Corporation literature investigates the advantage of an S-Corporation over a C-corporation. The main questions are: Why would a bank convert to an S-Corporation? And how does this affect bank features? Harvey and Padget (2000) explore the effect Subchapter-S Corporations have had on the banking industry by analyzing characteristics of S-Corporation banks and changes in behavior and performance of converting banks. Harvey and Padget find that corporations which convert to Subchapter S banks tend to be slightly smaller and slower-growing, have lower level of capital adequacy, and slightly higher ROAs prior to conversion. They also look at post-conversion performance and find that firms which convert to S-Corporation tend to increase their dividends which decreases their capital levels post-conversion, and have higher bank performance ratings. Hodder, McAnally and Weaver (2003) take it one step further and find that banks tend to convert when conversion minimizes taxes. Though S-Corporation conversions erase the double taxation standard of regular corporations, it is not always the most beneficial. For example, it might be tax-minimizing to not convert to an S-corporation if a bank has tax loss carryforwards, any penalties due to conversion, or significant deferred tax assets. Furthermore, if it restricts access to equity capital, the bank will forgo conversion. The authors add to the literature by also uncovering that “converting banks alter their capital structures, deliberately sell appreciated assets, and strategically set dividends to augment net conversion benefits.”

A number of studies discuss the opportunities available for banks which convert to S-Corporations through enhanced bank value. Denis and Sarin (2002) develop a model that shows that the magnitude of the tax advantage of S-corporations depends on the firm’s payout policy as well as marginal personal and corporate tax rates. The authors theoretically model that, on average, S-corporations are valued at a premium over C-corporations. One problem is that the model developed is based on a number of assumptions which are likely to not be realized in practice. To empirically test this theory, the authors compare limited partnerships (MLPs) which act like “publicly traded pass-through entities” (p 5) to their publicly traded industry peers. To value the differences, the authors use market value of invested capital to sales ratio and the market value of invested capital to book value of assets ratio. For the sample conducted, the authors find that MLPs are valued at a premium over their industry peers, consistent with their theoretical findings. Cyree, Hein and Koch (2005) also find that S-corporations have a higher value over non S-Corporations during their investigation on the motivation of banks which convert to S-corporations. First off, the authors confirm that S-corporation and C-corporation banks have a number of fundamental differences. For example, S-corporations tend to be smaller, have lower capital ratios, loan-loss reserves to assets, loans to deposits and asset growth. Furthermore, S-corporation banks have higher dividends and use diverse business strategies to serve differing geographic markets. Upon comparison of Sub-S and non-Sub-S banks, the authors find that “Sub-S banks have significantly higher pre-tax return-on-assets and…return-on-equity… [The] results suggest that Sub-S banks have higher profits, but lower asset growth than their peer institutions, on average” (p. 5). Mehran et al. (2008) find similar results when comparing bank characteristics from three years before conversion to S-corporation to three years post-conversion. The authors find that when banks convert to subchapter-S, they tend to increase their performance as measured by ROA, as well as their riskiness as determined by declining capital ratios and increasing nonperforming loans. Furthermore, they find that banks which convert to subchapter-S are significantly less likely to be acquired than their C-corporation peers.

Interestingly, when exploring the paths of literature related to S-Corporation characteristics, there are conclusions which both support and contradict the result that S-corporations are better performers. For example, S-Corporations tend to be smaller firms with more condensed ownership. The literature on size of firms and market value support the S-Corporation literature (Fama and French, 1992). In a related study, it was found that capital helps small banks to increase their probability of survival and market share at all times (during banking crises, market crises, and normal times). Second, capital enhances the performance of medium and large banks primarily during banking crises (Berger and Bouwman).

However, to become an S-Corporation, a number of ownership restrictions must be met. The literature on ownership structure and performance is not as supportive. Contrary to the literature on S-Corporation performance, Cole and Mehran (1998) study thrifts which convert from mutual to stock ownership, and they show that “restrictions on stock ownership may harm a company’s performance, because restrictions prevent owners from choosing an optimal structure.” The authors find that when restrictions on ownership are removed, performance increases significantly. Furthermore, much of the literature on agency theory suggests that with ownership diffusion comes costly agency problems having an adverse effect on performance. However, some empirical evidence suggests there is not a clear relationship between ownership structure and performance (Demsetz and Villalonga, 2001; Iannotta, Nocera and Sironi, 2007). Iannotta, Nocera and Sironi (2007) find that while “ownership concentration does not significantly affect a bank’s profitability, a higher ownership concentration is associated with better loan quality and lower insolvency risk.”

Motivation and Summary Statistics

As one can see, the literature depicts S-corporations as having higher valuation. The characteristics of S-Corporations support these studies. For example, S-Corporations tend to be small firms, and small firms tend to have better performance during normal times as well as during times of crisis (Fama and French, Berger and Bouwman). However, literature on ownership restrictions explains that performance and value decrease as restrictions are placed on ownership. S-Corporations have a long list of ownership restrictions in order to qualify. Thus, these two characteristics appear to lead to inconclusive results on performance.

This article explores the relationship between S-corporations and performance a little further. Specifically, I will test whether S-Corporations or non-S-Corporations perform best during recessions. Prior literature suggests that S-corporations would perform better in recessions because they are generally small banks (Cyree, Hein and Koch (2005); Berger and Bouwman (2013)).

Data

The data for this study is quarterly banking data collected from the FDIC and compiled into a single data source by Dr. Will Hippler. For the purposes of this study, I use data over the time span of December, 2004 through June, 2009. I split the data into two periods: the pre-recession period which includes the 3 years prior to the start of the recession, and the recession period which includes the last quarter in 2007 (when market pressures began to intensify, as stated by the St. Louis Fed) through the second quarter of 2009 (when market performance began to see positive increases.) I delete any banks which do not have data in both periods. I require that each bank have at least 4 quarters of observations during the pre-recession period, and at least 2 quarters of observations during the recessionary period. I shorten the recession period observation requirement to reflect the increased likelihood of failing or closing banks.

Methodology and Results

To explore the difference in performance between S-Corporation and Non S-Corporation banks during recessions, I use univariate analysis as well as OLS Regression Analysis. I look at three measures for performance. The first measure or bank performance is return on assets (ROA) which is defined as the bank’s net income divided by its total assets. A second profitability measure I use is net interest margin (NIMY) which is defined as the net investment returns divided by average invested assets. Lastly, I also use the return on equity (ROE) to measure bank performance; ROE is defined as the bank’s net income divided by its total equity.

Univariate Analysis

First, I split the dataset into pre-recession and recession groups to evaluate the difference in means for measures of performance for the whole sample. Consistent with intuition, bank performance decreases during the recession for all three measures. A difference in means t-test shows this difference is statistically significant at the one percent significance level. The means and t-test results can be shown in Table 1.

Second, I split the pre-recession and recession groups even further to evaluate the univariate differences in S-corporation and Non-S-corporation performance. For all three measures of performance in both time periods for this sample, S-Corporations out-perform Non S-Corporations. It also appears that performance of S-Corporations decreases by a smaller margin during the recession than the performance of Non S-Corporations. A difference in means t-test shows that the difference in means between the pre-recession period and the recession period are statistically significant at the one percent level for both S-Corporation banks and Non S-Corporations. The means and t-test results for S-Corporation and Non S-Corporation banks can be shown in Table 2.

Third, I use the difference in means test once more to explore the difference in means between S-Corporation and Non S-Corporation banks. I look at both the pre-recession period and the recession period. I find that the performance measures for S-Corporation and Non S-Corporation banks are not statistically equal at the one percent significance level during both the pre-recession and recession periods. The difference in means results are displayed in Table 3. The results of the univariate difference in means tests discussed above show there are statistical differences in how the performance measures of S-Corporation Banks and Non S-Corporation Banks are affected by the recession. The next section explores if these differences are supported by regression analysis using panel data.

Regression Analysis

In order to further explore the differing performance of S-Corporation and Non S-Corporation banks during recessions, I use regression analysis. I use return on assets (ROA) as my measure for performance. My independent variable describing the type of bank is a dummy variable, Subchaps, where Subchaps = 1 if the banks is Subchapter S Corporation. Subchaps = 0 if the bank is not an S-Corporation. I use four control variables. The first control variable I use is the log of total assets as a proxy for bank size since this is an important differing characteristic between S-Corporations and Non S-Corporations. The other three firm performance control variables are net operating income, total equity, and leverage ratio to control for other factors which influence bank performance.

I use OLS to regress ROA on Subchaps and the control variables for both the pre-recession and recession period groups. I then use a difference in means test to compare the difference in the coefficients from the two periods. The coefficients and p-values are displayed in Table 4. The results indicate that, on average, an S-Corporation bank will have statistically better performance as measured by ROA.

During recessions, the margin for performance between S-Corporations and Non S-Corporations becomes even wider as the ROA for Non S-Corporations decreases by a greater rate, on average. The difference in means test concludes the coefficients for Subchaps in the Pre-Recession Period and Recession period are statistically different at the 1% level.

Last, I use the full panel dataset to regress bank performance on two dummy variables. The coefficients and their significance are displayed in Table 5.

Again, Subchaps equals 1 if the firm is an S-corporation, 0 otherwise. The second dummy variable, After, differentiates the pre-recession and recession periods. After equals 1 if the quarter is in the recession period, 0 otherwise. Sub*After is an interaction variable of the two dummy variables. The results indicate that the type of corporation of a bank significantly impacts its performance as measured by ROA. A positive Subchaps dummy variable indicates that banks which are Subchapter-S tend to have higher ROA than Non Subchapter-S banks. This is true for both time periods, and the coefficient increases dramatically when the interaction variable is included.

One interesting observation which I have not yet been able to explain is the positive coefficient for After. One possible explanation for this result is that bank assets decreased dramatically during the recession relative to the net income causing the ROA ratio to increase. However, this conflicts with my earlier univariate results which show that, on average, ROA decreases during the recession. Furthermore, a smaller ROA during the recession makes sense intuitively since banks are being squeezed financially. Moving forward, I would like to explore this variable further. Possible explanations include omitted variable bias or low degrees of freedom due to Panel Regression. Thus, I would like to try to deepen my understanding of the variables used to explain ROA even further. Secondly, because I am using Panel regression, the low degrees of freedom may be affecting my results. In the future, I would also like to try running my equations as a pooled regression to compare any differences in results. However, the After variable is not the main focus of this paper, so I will put these idea extensions on pause
momentarily.