Does Profitability Cause the Changes in Stock Prices? Evidence from the Turkish Banking Stocks

Stocks are the riskiest investment instruments. This situation makes it necessary for investors to concentrate on investment decisions that would maximize their returns according to the risk levels they can accept while determining the stocks they would invest. This requirement has led investors to accurately and significantly identify stock returns and the firm-specific factors affecting these returns. Assessment of company-specific factors such as companies’ liquidity, financial structure, effective use of assets, profitability, and stock market performance give investors information about the real value of the stocks of the relevant companies. The view that stock prices can be predicted to a great extent by using financial ratios that reflect the real financial conditions of firms tends to lead investors and academics to identify the relationships between financial ratios and stock returns (T​e​s​f​a​t​s​i​o​n​,​ ​2​0​0​4).

The reason for the existence of various studies examining the factors affecting stock prices in the financial literature is to increase the predictability of prices. Investors should be able to accurately and significantly determine the factors affecting stock prices and returns in making the right decisions. In this context, both macroeconomic factors such as exchange rates, interest rates, inflation, investment, capital and microeconomic factors such as the financial ratios of companies have been the subject of research studies. The banking sector has an important share of stocks (33%) in Istanbul Stock Exchange (BIST). The positive or negative movements in the banking sector are powerful enough to account for the fluctuation of the BIST Banks Index. Therefore, having information regarding the banking sector can be beneficial for the investor. Moreover, investors should be aware of the extent to which factors are used as selection criteria in determining the names of the banks to be included in the analysis.

The development of the economy and financial system depends solely on the coordination of certain economic actors. One of the most important of these actors is banks (I​ş​ı​k​ ​e​t​ ​a​l​.​,​ ​2​0​1​7).

Profit and profitability are considered as a reflection of the past business experience for all enterprises. Investors take action by predicting the future upon making their investment decisions. Accordingly, it is anticipated that investors would not make investment decisions based on the assumed profit figures formed in the past periods. However, this may not be the case in practice. Because all statistical forecasting methods make predictions based on the past while predicting the future. Apart from profit and profitability, the subject also has similar dynamics between business performance and performance-based ratios as well as stock returns. Therefore, one might ask whether it is possible that the stocks of the enterprises that had a successful financial performance in the past would be positively affected by this, or investors invest, in the purest sense, for the future. This argument raises a dilemma within itself. This research study can be helpful in comprehending the complex structure that led to the emergence of such a dilemma. Accordingly, this study aims to investigate whether or not banking profits cause banking stock prices to change. As shown in Table 1, different analysis techniques were used in studies on this subject.

Accordingly, it is tried to determine whether there is a casual relationship between profitability indicators and average of closing stock prices of BIST Banking Index over the period 2010: Q1–2020: Q4. This research study, which examines the stocks in the BIST banking index, consists of five parts. In the first part, basic information regarding the subject is introduced. In the second part, a literature review is presented. The third part explains the data and methodology of the study. In the fourth part, the analysis of empirical findings is summarized. The fifth part covers an overall evaluation.

The knowledge of how and by which factors the stock returns are affected is quite crucial for investors when creating their portfolios. Therefore, studies examining the factors affecting stock returns occupy a large portion of the financial literature. Upon analyzing the conducted research studies, it is seen that macroeconomic variables that affect stock returns such as real interest rate, inflation, investment, exchange rate is considered, and as microeconomic factors financial ratios are taken into account.

In this study, the data of deposit banks in the BIST banking index are utilized to investigate whether or not the banking profits display a signal for the investor, increasing the investor’s demand for stocks, and thus causing an increase or decrease in the prices of the banking stocks. The causality approach proposed by G​r​a​n​g​e​r​ ​(​1​9​6​9​) is based on the dynamics that an effect cannot precede its cause. E​m​i​r​m​a​h​m​u​t​o​ğ​l​u​ ​&​ ​K​ö​s​e​ ​(​2​0​1​1​) stated that the future cannot be the cause of the past, so if the variable X is Granger cause of the variable Y, the changes in X are antecedent to the changes in Y. This study examines whether or not banking profits are the cause of the changes in stock prices. Banks covered in the study include the fiscal and financial data of Akbank, Garanti BBVA, Halkbank, QNB Finansbank, Vakıfbank and Yapı Kredi Bank.

Since the deposit banks sample is used in the study, Albaraka Turk Participation Bank and the Industrial Development Bank of Turkey included in the BIST banking index are excluded. Besides, due to the presence of different stocks of İşbank such as A, B, and C; it is possible to analyze the casual relationship between banking profits and stock prices based on type of the stocks. Since this situation would not provide an integrated perspective for the rest of the sample, the relevant banks could not be included in the sample. Because ICBS Turkey Bank and Şekerbank’s annual balance sheet of 2020 were not published at the time of the study, they are excluded from the sample. Within all these limitations, the sample period is from 2010:Q1 to 2020:Q4. The financial data of the banks included in the sample are obtained from the Banks Association of Turkey (BAT), whereas the data relating to the closing prices of the stocks are obtained from a licensed distributor institution. As a basis for the analysis, the average closing prices for each quarter are calculated using the closing figures of the banking stocks in the sample. Moreover, the following calculations are made in accordance with the financial literature using such items as Total Assets, Total Equities, Outstanding Shares and, Net Earnings included in the financial statements of the relevant banks:

Return on assets and return on equity are the two basic profitability criteria used in the financial literature. Earnings per share is a measure of profit evaluated in terms of investors and expressed as the “bird in the hand” in financial literature. Since the nominal value of the banking stocks included in the sample is 1 TL, the outstanding shares are considered equal to the number of stocks of the banks. In this context, earnings per share can be calculated by dividing the net earnings by the outstanding shares. In order to provide basic information about the financial structure of the research sample, the values of average closing, return on assets, return on equity, and earnings per share calculated for the banks included in the sample are presented in Table 2.

Upon examining the table, it is seen that Halkbank has the highest closing price whereas Yapı Kredi Bank has the lowest closing price in terms of quarterly average closing prices. Besides, it can be claimed that the highest average of return on assets belongs to Garanti BBVA, whereas the lowest average belongs to Vakıfbank. In terms of return on equity, it can be considered that Halkbank has higher profitability compared to other banks, while the lowest figure belongs to Vakıfbank. Upon evaluating the issue in terms of earnings per share, it can be stated that Halkbank differs from other banks in the sample with a profit per share amount of 1.22 TL. The lowest profit per share in the table (0.35 TL) belongs to Yapı Kredi Bank.

In this study, by considering ROA, ROE and Earnings per Share as performance indicators, the following hypotheses can be established in order to determine whether any causal effect exists between bank performance and bank stock prices:

H₀: There is no causal relationship between bank performance and bank stock returns.

H₁: There is a causal relationship between bank performance and bank stock returns.

Firstly, the unit root test should be carried out to detect whether or not the return on assets, return on equity, and earnings per share estimated in this study cause the average closing prices is the unit root test. In this regard, the unit root test, which is widely used in practice and developed by I​m​ ​e​t​ ​a​l​.​ ​(​2​0​0​3​), can investigate the existence of unit root separately for each bank in the series. Heterogeneous panel data analysis is developed by combining individual unit root tests in the I​m​ ​e​t​ ​a​l​.​ ​(​2​0​0​3​) test, which would be expressed as the IPS test. In the IPS unit root test, individual unit root tests are combined to obtain results.

Following the unit root test, the causality analysis is performed. E​m​i​r​m​a​h​m​u​t​o​ğ​l​u​ ​&​ ​K​ö​s​e​ ​(​2​0​1​1​) causality test, using the meta-analysis developed in the study of F​i​s​h​e​r​ ​(​1​9​9​2​) which was a panel adaptation of the T​o​d​a​ ​&​ ​Y​a​m​a​m​o​t​o​ ​(​1​9​9​5​), is a test that can be performed when the variables are not stationary at the same level. In this method, it is assumed that the heterogeneity among units is valid regardless of whether the variables are stationary or cointegrated. Critical values derived from bootstrap distributions are utilized in the method. Furthermore, the relevant test may yield separate results for both individual units and the overall panel. E​m​i​r​m​a​h​m​u​t​o​ğ​l​u​ ​&​ ​K​ö​s​e​ ​(​2​0​1​1​) bootstrap causality test can be performed without the need for a preliminary unit root test. Nevertheless, in the application of the E​m​i​r​m​a​h​m​u​t​o​ğ​l​u​ ​&​ ​K​ö​s​e​ ​(​2​0​1​1​) causality test, it is practically more appropriate to perform a unit root test, since the calculations are repeated by taking the vector numbers containing unit-roots into account.

Descriptive statistics of the estimated values of the banks in the sample can be seen in Table 3. Upon examining the descriptive statistics, it is seen that the average closing has the highest standard deviation value. This situation can be noticed from the high level of the maximum value in the relevant series. It is among the information that can be obtained from Table 3 that the descriptive values of return on assets and return on equity are similar.

It is an appropriate process in practical methodology to investigate whether or not a unit root exists in series before performing a causality analysis. Accordingly, in compliance with the causality methodology of the study, the unit root test proposed by I​m​ ​e​t​ ​a​l​.​ ​(​2​0​0​3​), which provides an advantage in examining the existence of unit root individually for each series, is performed. In Table 4, the results of the IPS unit root test performed on average closing, return on assets, return on equity, and earnings per share series can be observed.

Upon examining Table 4, it is determined that the average closing and earnings per share series are stationary at 95% confidence level, I(1); whereas the return on assets and return on equity series are stationary at level, I(0). Even if the analyzed series are not stationary at the same level, an E​m​i​r​m​a​h​m​u​t​o​ğ​l​u​ ​&​ ​K​ö​s​e​ ​(​2​0​1​1​) causality analysis can be performed. By courtesy of the E​m​i​r​m​a​h​m​u​t​o​ğ​l​u​ ​&​ ​K​ö​s​e​ ​(​2​0​1​1​) study, it has become possible to apply the Toda & Yamamoto approach to the Granger causality for the panel datasets. Accordingly, causality analysis can be performed regardless of whether the variables are stationary or integrated. The motivation underlying this study is determining whether or not banking profits cause investors’ demand for stocks, thus changing the stock prices. In compliance with the aim of the study, it is tried to examine whether or not the calculated return on assets, return on equity, and earnings per share cause the changes in the stock prices. The results of E​m​i​r​m​a​h​m​u​t​o​ğ​l​u​ ​&​ ​K​ö​s​e​ ​(​2​0​1​1​) bootstrap causality analysis performed unilaterally are presented in Table 5.

The direction of causality, the Wald statistics, lag length coefficients selected according to Akaike criteria, and Fisher test statistics where causality can be examined in the panel as a whole are presented in Table 5. The results shown in Table 5 separately report the findings of whether the return on assets, return on equity, and earnings per share cause the change in average closing prices of stocks. The causal effect of return on assets on stocks’ closing prices is determined only for Halkbank at the 5% significance level. Therefore, results suggest that Halkbank investors change their demand for stocks depending on return on assets. Similarly, upon examining the causal relationship between the return on equity and the average closing price, a causality at the 5% significance level is detected only for Akbank. Upon evaluating the results regarding the return on assets and return on equity as a whole, the Fisher test statistic reveals that no causal relationship exists at the 5% significance level. The results become more significant once the subject is analyzed in terms of earnings per share. Earnings per share is a prioritized concept for investors. Dividend earnings are based on distributable earnings per share. The return on equity and return on assets offer profitability criteria for the enterprise, whereas earnings per share can be considered the most crucial information for the investor. Accordingly, the Wald test results reveal that earnings per share of Akbank, Garanti BBVA, QNB Finansbank, and Vakıfbank stocks influence the average closing prices. Besides, upon evaluating the earnings per share results as a whole, the Fisher test statistic reveals the existence of causality at the 1% significance level. Accordingly, upon simultaneously evaluating the obtained results, it can be claimed that return on assets and return on equity do not affect closing prices of stocks, however, the earnings per share influence average closing prices.

Findings obtained from the study reveal a causal relationship sequence in the direction illustrated in Figure 1. Accordingly, Figure 1 analytically depicts the finding asserting that the amount of earnings per share for the banking sector causes the changes in the stock prices through investor preferences, whereas the return on assets and return on equity do not have a causal effect.

This study examines whether banking profit indicators lead to a change in stock prices or not. Accordingly, E​m​i​r​m​a​h​m​u​t​o​ğ​l​u​ ​&​ ​K​ö​s​e​ ​(​2​0​1​1​) bootstrap causality test is applied to determine whether the return on assets, return on equity, and earnings per share of six Turkish banks included in the BIST banks index cause any change in the average closing prices of their stocks. Covering the period between 2010: Q1 to 2020: Q2, the results reveal that especially the earnings per share variable affects the stock price of banks. Accordingly, the alternative hypothesis (H₁) implying a causal relationship between bank performance and bank stock returns would be accepted in terms of earnings per share and for all the banks included in the sample except for Halkbank and Yapı Kredi Bank.

This finding complies with the findings of B​ü​y​ü​k​ş​a​l​v​a​r​c​ı​ ​(​2​0​1​1​) and K​u​r​t​ ​&​ ​K​ö​s​e​ ​(​2​0​1​7​). However, no such causal relationship for neither return on assets nor return on equity is detected, and both K​a​r​c​ı​o​ğ​l​u​ ​&​ ​Ö​z​e​r​ ​(​2​0​1​4​) and Ç​a​l​ı​ş​ ​&​ ​S​a​k​a​r​y​a​ ​(​2​0​2​0​) revealed similar findings.  

In the study of S​e​n​o​l​ ​e​t​ ​a​l​.​ ​(​2​0​1​8​), return on assets was found to have a significant impact on stock prices, however, the relevant study was conducted on companies in Metal Goods, Machinery and Equipment Maintenance sector, not on banks. This indicates that the causal relationship envisaged in this study may differ from sector to sector.

There may be various factors that determine the demand motivations of securities investors. The profits of the invested enterprise are also considered among such motivations. Nonetheless, the results obtained from the study reveal that investors should consider the causal relationship, especially in terms of earnings per share. In this case, issues such as the size and equity level are not considered by investors. The existence of a regulatory institution such as the Banking Regulation and Supervision Agency (BRSA) and the activities of banks being carrying out without exceeding certain risk levels can be considered as the reasons why no causal relationship can be detected between factors such as return on assets or return on equity and the average closing prices of the stocks.

Earnings per share, on the other hand, are considered as a quite determinant factor for investors. The causal relationship detected between earnings per share and average closing prices of stocks can be evaluated in this aspect.

Upon evaluating the findings obtained from the study in terms of investors, it can be claimed that it is the right strategy to invest according to the earnings per share expectation rather than the return on assets or return on equity in the periods during which the financial statements of the banks are announced. This situation reveals that the investment strategy is appropriate regardless of the size of the bank. Undoubtfully, besides the contribution of the study, there are some shortcomings of the study. Especially the number of banks included in the sample is limited. Also, the subject of the study yields merely generalized results for Turkey in terms of the related literature. Accordingly, it would be quite useful to analyze the subject with a sample consisting of different sectors in order to understand causal relationships comprehensively. The relevant investigation would constitute the subject of future studies.