INTRODUCTION

The capital market plays an important role in bridging the gap between the investors in need of funds and the investors seeking to get returns on their investment. Stock return is an important indicator for investors to evaluate investment performance, which reflects the gains obtained from changes in stock prices and distributions to shareholders. When making investment decisions, investors take into account not only the market environment, but also fundamental corporate information that reflects the company’s ability to maintain its performance and create long-term value. From the perspective of Signalling Theory, the information of a firm’s financial condition can be a signal to investors to evaluate its prospects and make investment decisions (Brigham & Houston, 2021). The banking sector is a strategic position in the Indonesian economy and capital market. Special features are particularly evident in state owned banks under the Association of State-Owned Banks (Himpunan Bank Milik Negara—HIMBARA) and Regional Development Banks (Bank Pembangunan Daerah—BPD) because, in addition to carrying out commercial functions, these banks are closely involved in the implementation of development policies of the central and regional governments. HIMBARA banks and BPDs are involved in several economic programs, including the distribution of People’s Business Credit (Kredit Usaha Rakyat—KUR), economic recovery programs, micro-enterprise financing, and financing for strategic sectors. These features suggest that stock performance of government-owned banks may be driven not only by corporate fundamentals but also by investors’ reactions to government policies and the overall economic climate.

Financial performance is one of the key information sources that investors use to assess a firm. The financial performance of the banking industry is evaluated using a number of financial parameters, including the Capital Adequacy Ratio (CAR), Net Profit Margin (NPM), Return on Equity (ROE), and Price to Earnings Ratio (PER). While CAR measures the bank's capital adequacy to absorb future losses, NPM measures the bank's capacity to produce net profit from its activities. While ROE measures the company's capacity to provide a return on shareholders' equity, PER measures the market valuation of the company's earnings. A company's improved financial performance may provide investors a positive indication of its future prospects, according to the Signalling Theory. This could impact stock returns by raising demand for the company's stock (Brigham & Houston, 2021).

However, empirical trends have not always shown a consistent relationship between financial performance and stock returns. Harianti and Sardiana (2022) claim that stock returns were positively and significantly impacted by ROE and Earnings per Share (EPS). This implies that investor confidence increases with profitability. Nevertheless, their findings also showed that NPM had no discernible impact on stock returns. Similarly, Fitri et al. (2025) discovered that stock prices were positively and strongly impacted by CAR. This suggests that a bank's capacity to withstand possible losses and preserve its financial stability may be reflected in its high capital adequacy. From a market perspective, PER can also change due to both corporate fundamentals and investor sentiment; thus, a higher PER may indicate market expectations of future earnings growth (Rahman & Shamsuddin, 2019). These different results suggest that financial performance information does not lead to homogenous responses from investors.

For the period 2020-2024, a similar phenomenon occurs in the Indonesian banking sector. Based on data from the Financial Services Authority (Otoritas Jasa Keuangan—OJK), Indonesia Stock Exchange (IDX), and the respective banks’ annual reports, several HIMBARA and listed BPD banks have experienced a recovery and increase in net profits after the COVID-19 pandemic. However, the movement in stock prices and stock returns did not always move proportionally with improvements in financial performance . This divergence indicates that strong corporate fundamentals do not necessarily translate to higher returns for investors. This finding is indicative of the existence of other financial and non-financial factors that may influence market responses.

Figure 1. Stock Prices of HIMBARA Banks and Listed Regional Development Banks, 2020–2024

Source: OJK, Indonesia Stock Exchange (IDX), and Annual Reports (2020–2024).

The stock price movements among the observed banks were very different in 2020-2024, as shown in Figure 1. In general, Bank Mandiri (BMRI) and Bank Negara Indonesia (BNI) showed an upward trend, while Bank Rakyat Indonesia (BBRI) experienced more fluctuations during the observation period. Meanwhile, BTN (BBTN), Bank Syariah Indonesia (BRIS) and the listed BPDs were trading at different price levels and moving differently. Such differences indicate that government owned banks are not performing uniformly in the stock market. Variations of such nature provide an empirical basis for studying the extent to which differences in financial performance and non-financial information explain variations in stock returns.

The divergence between corporate fundamentals and stock returns suggests that investors' decisions may not be guided solely by financial information. Another non-financial issue which has been gaining increasing attention is Corporate Social Responsibility (CSR). CSR is a manifestation of the company’s commitment to carry out its business activities in consideration of social, environmental and stakeholder interests. Stakeholder Theory suggests that corporate sustainability is not only about satisfying the interests of shareholders but also about the ability of the company to build positive relationships with different stakeholder groups (Carroll & Buchholtz, 2015). Stronger CSR implementation could enhance a corporation’s reputation, public confidence and impact investors’ positive view of the company. However, the empirical evidence on the market responses to CSR has been mixed as well. Pratiwi and Yuliana (2023) found that CSR had a positive but insignificant effect on stock prices, while profitability indicators like ROE and EPS had stronger effects. The results suggest that financial information is more valued by investors than corporate social information, and the effects are more prominent for investors in the banking industry. However, from a theoretical standpoint, CSR can be a positive signal of the company’s commitment to sustainability and stakeholder relationships, which can enhance its reputation and market confidence (Provaroni & Checcacci, 2023). The divergence between theoretical predictions and empirical results suggests that the relationship between CSR and stock returns is an issue that requires further study.

It is possible to find a research gap in the research objects and the combination of variables studied in previous research. However, prior studies have not examined HIMBARA banks extensively, and BPDs are described as specific types of banks owned by central and regional governments. These banks have specific features, since they perform commercial functions, and at the same time, the role of development agents in the implementation of government economic programs. Moreover, the simultaneous examination of CAR, NPM, ROE, PER and CSR in explaining stock returns is relatively limited. Particularly, empirical evidence on the relationships between PER, CSR and stock returns on HIMBARA banks and listed BPDs is still limited. This study attempts to fill this gap by combining Signalling Theory and Stakeholder Theory. Signalling Theory offers a theoretical basis for explaining how the financial information represented by CAR, NPM, ROE, and PER can signal to investors the corporate conditions and future prospects. Meanwhile, Stakeholder Theory offers a framework for explaining CSR as a demonstration of corporate responsibility to stakeholders that may affect corporate reputation and market confidence. The integration of these two theoretical perspectives provides a wider framework for understanding how financial and non-financial information may affect investors’ responses in the capital market.

The purpose of this study is to investigate how the stock returns of HIMBARA banks and Regional Development Banks listed on the Indonesia Stock Exchange between 2020 and 2024 are affected by corporate social responsibility (CSR) and financial performance as determined by CAR, NPM, ROE, and PER.Panel data regression is used in this study to allow for cross-sectional and temporal variance across the observation period. The results are anticipated to improve knowledge of the significance of both financial and non-financial information in investment decision making and to add to the empirical evidence on the factors influencing stock returns of Indonesia's government-owned banks.

METHOD

This study uses a quantitative associative research approach to investigate how financial performance and corporate social responsibility (CSR) affect stock returns. The financial performance was evaluated using four metrics: Price to Earnings Ratio (PER), Return on Equity (ROE), Net Profit Margin (NPM), and Capital Adequacy Ratio (CAR). The dependent variable was stock return, whereas the non-financial explanatory variable was CSR. Secondary data from Yahoo Finance, RTI Business, official business websites, annual financial reports, sustainability reports, and stock price information of companies from the Indonesia Stock Exchange (IDX) for the years 2020–2024 are used in this study. The study's sample consists of the HIMBARA and Regional Development Bank (BPD) banks that are listed on the Indonesia Stock Exchange. Bank Mandiri (BMRI), Bank Rakyat Indonesia (BBRI), Bank Negara Indonesia (BBNI), Bank Tabungan Negara (BBTN), Bank Syariah Indonesia (BRIS), Bank Pembangunan Daerah Jawa Timur (BJTM), Bank Pembangunan Daerah Jawa Barat dan Banten (BJBR), and Bank Pembangunan Daerah Banten (BEKS) are some of these. The sampling approach used was saturated sampling, which incorporates all available eligible observations for analysis throughout the study period. The data was analysed using panel data regression in the Eviews programme. The data consists of a cross-section of banks' time series from 2020 to 2024. The Random Effect Model (REM), Fixed Effect Model (FEM), and Common Effect Model (CEM) panel regression parameters are all taken into account. The Chow, Hausman, and Lagrange Multiplier (LM) tests were used to determine the optimal estimating model at the 5% level of significance. The chosen model was analysed for autocorrelation, multicollinearity, heteroscedasticity, and normality. The hypothesis about the partial impact of each independent variable (CAR, NPM, ROE, PER, and CSR) on the stock return at α = 0.05 was statistically evaluated using the t-test.

RESULT AND DISCUSSION

The empirical findings regarding the impact of financial performance and Corporate Social Responsibility (CSR) on stock returns in the Indonesian banking industry are presented and discussed in this section. The study was carried out on 17 banks that were listed on the Indonesia Stock Exchange (IDX) between 2020 and 2024 in order to raise the number of firm-year observations to 85. These banks included HIMBARA banks, Regional Development Banks (BPDs), and suitable private banks categorized into KBMI 3 and KBMI 4. Capital Adequacy Ratio (CAR), Net Profit Margin (NPM), Return on Equity (ROE), Price to Earnings Ratio (PER), and CSR indicate the financial performance; stock return is the dependent variable. Descriptive statistics, model selection and panel data estimates, hypothesis testing, and the relationship between the empirical findings and the theoretical framework and earlier research are all included in the presentation and discussion of the results.

Results

Table 1. Descriptive Statistics

The data set includes 85 observations of 17 banking companies over 5 years. Stock Return (Y) has a mean value of 4.843±30.874, minimum value of -51.11 and maximum value of 148.22. The Capital Adequacy Ratio (CAR) has a mean of 25.335 ± 6.022 with a variation from 16.80 to 44.72. The mean of Net Profit Margin (NPM) is 13.765 ± 68.609, ranging from −518.00 to 51.00. Meanwhile, the Return on Equity (ROE) has an average of 11.736 ± 10.755 with a minimum of −47.96 and a maximum of 27.31. Price to Earnings Ratio (PER): It has an average of 15.375 ± 17.614 with minimum of −10.84 and maximum of 120.769. Finally, the mean value of Corporate Social Responsibility (CSR) is 0.198 ± 0.198 and its values range between 0.01 and 0.97.

To determine whether the regression model's residuals had a normal distribution, the normality test was run. The Jarque-Bera test was performed to determine whether the residuals were normal using the probability value. When the probability value exceeds the 5% significance level (p > 0.05), residuals are regularly distributed. The results of the normalcy test are displayed in Figure 2.

Figure 2. Normality Test Results

According to Figure 2, the probability value was 0.545 and the Jarque-Bera statistic was 1.215. The null hypothesis (H₀), which states that the residuals were normally distributed, was accepted since the probability value was higher than 0.05. The normalcy assumption was thus met by the regression model.

A multicollinearity test was performed to see if the independent variables in the regression model exhibited significant linear relationships. Multicollinearity was tested using the Variance Inflation Factor (VIF); a VIF value of less than 10 indicated the lack of multicollinearity. Table 2 displays the findings.

Table 2. Multicollinearity Test Results

It is shown that the VIFs of CAR, NPM, ROE, PER, and CSR are 1.787, 3.251, 3.927, 1.226, and 1.063, respectively (Table 2). All values were less than 10, suggesting that there was no multicollinearity problem among the independent variables. Thus, the multicollinearity assumption of the regression model was satisfied.

The regression model's residuals' constant variance was examined using the heteroskedasticity test. The Glejser test was applied. The criteria for making decisions are that homoscedasticity is present if the probability value of ObsR-squared is larger than 0.05, and heteroscedasticity is present if the probability value of ObsR-squared is less than 0.05. Table 3 displays the outcomes.

Table 3. Heteroskedasticity Test Results

Table 3 shows that the Prob. Chi-Square was 0.1569 and the ObsR-squared was 7.987. The homoskedasticity null hypothesis (H₀) was accepted since the probability value exceeded the significance level of 5% (0.1569 > 0.05). As a result, the residual variance was homogeneous, indicating that the regression model had no heteroskedasticity issues. To determine whether the residuals in the regression model were correlated between periods, the autocorrelation test was run. The null hypothesis (H₀) for the Breusch–Godfrey Serial Correlation LM Test was that there was no serial correlation. The Prob. Chi-Square value of ObsR-squared served as the judgement criterion; a probability value more than 0.05 denotes the absence of autocorrelation. Table 4 presents the findings.

Table 4. Autocorrelation Test Results

The ObsR-squared is 2.128 and the Prob. Chi-Square is 0.3451 as shown in Table 4. The probability value was more than the 5% significance level value (0.3451 > 0.05) and hence, the null hypothesis (H0) was accepted. The result implied that there is no autocorrelation between residuals and therefore the regression model fulfilled the autocorrelation assumption.

Panel Data Model Selection

Table 5. Panel Data Model Selection Results

The Common Effect Model (CEM) is favoured over the Fixed Effect Model because the cross-section F probability of the Chow test is 0.5105, which is greater than the 5% significant level. With a probability value of 0.1298, which is also higher than 0.05, the Hausman test shows that the Random Effect Model (REM) is better than the Fixed Effect Model. The models are specified in several ways using the Chow and Hausman tests. The suitability of CEM or REM was then determined using the Lagrange Multiplier test. Since the cross-section probability of Breusch-Pagan is 0.4116, which is higher than 0.05, the CEM is recommended over REM. Consequently, the Common Effect Model (CEM) was ultimately chosen as the panel data regression model. Hypothesis testing was then performed based on the whole model selection procedure.

Hypothesis Testing Using the Common Effect Model (CEM)

Table 6. Common Effect Model (CEM) Regression Results

Based on the estimated coefficients, the panel regression equation can be expressed as follows:

Stock Return = −9.285 − 0.361CAR − 0.123NPM + 0.846ROE + 1.006PER − 2.111CSR + ε

Together, CAR, NPM, ROE, PER, and CSR explained around 30.11% of the variation in stock returns, with a coefficient of determination (R2) of 0.3011. The remaining 69.89% of the variation was explained by factors not included in the model. Additionally, the explanatory variables were significant when taken together to explain stock returns, as indicated by the F-statistic of 6.807 with a probability of 0.000025 (< 0.05).

For every explanatory variable, the partial hypothesis tests yield distinct findings. CAR (X1) has a probability value of 0.5769, a negative coefficient of -0.361, and a t-statistic of -0.560. A negative but statistically insignificant correlation between CAR and stock returns was indicated by the probability value greater than 0.05. This shows that a 1 unit rise in CAR was shown to cause a 0.361 unit decrease in stock returns when all other parameters were maintained constant.

The computed connection and zero do not differ statistically significantly at the 5% significance level. Furthermore, NPM (X2) had a negative coefficient of -0.123, a t-statistic of -1.618, and a probability value of 0.1097. Consequently, there was a negative but statistically insignificant correlation between NPM and stock returns. A one-unit increase in NPM was associated, ceteris paribus, with a 0.123-unit decrease in stock returns; however, the correlation was not statistically significant at α = 0.05.

In contrast, ROE (X3) has a positive coefficient of 0.846, a t-statistic of 1.580 and a probability value of 0.1180. The coefficient was positive which means that the higher ROE was associated with higher stock returns, however the relationship was not statistically significant. Thus, the empirical results do not provide enough statistical evidence at the 5% level to claim that changes in stock returns were systematically related to changes in ROE over the observation period.

PER (X4) is the only one of the five explanatory factors with a statistically significant coefficient. According to the findings, PER had a probability value of 0.0000 and a positive coefficient of 1.006 with t-statistics of 5.512, both of which were below the significance level of 0.05. It was therefore established that PER and stock returns had a positive and statistically significant association. A one-unit rise in PER was linked to a roughly 1.006-unit increase in stock returns, all other things being equal. The findings demonstrate a regular relationship between the change of PER and the stock returns of the tested banks between 2020 and 2024.

Lastly, CSR(X5) had a coefficient of -2.111 and t-statistic of -0.140 and probability of 0.8893. The probability value was significantly higher than 0.05, which indicates that CSR was negatively and statistically insignificantly related to stock returns. However, the coefficient indicates that a one-unit increase in the CSR measure was associated with a 2.111 decrease in stock returns. But the statistical insignificance means that this estimated negative relationship should not be interpreted as reliable evidence that greater CSR disclosure reduces stock returns.

According to the CEM calculation, there is no statistically significant partial link between stock returns and the CAR, NPM, ROE, and CSR at the 5% significance level. In the meanwhile, there is a statistically significant and positive correlation with PER. Additionally, the significant F-test shows a significant relationship between stock returns and the five explanatory factors. These findings provide the empirical basis for the discussion of each hypothesis in relation to Signalling Theory, Stakeholder Theory and previous empirical studies in the subsequent sections.

Independent-Samples t-Test: Comparison of Stock Returns by Bank Type

An independent-samples t-test was conducted to examine whether stock returns differed significantly between private banks and HIMBARA banks. Descriptive statistics and the results of the mean difference test are presented in Table 7.

Table 7. Independent-Samples t-Test Results for Stock Returns by Bank Type

The descriptive results indicate that the mean stock return of private banks was 7.039 ± 32.676, while HIMBARA banks were 2.373 ± 28.924. Compared to HIMBARA banks, private banks had a higher mean stock return. However, a statistically significant difference is not always indicated by a descriptive difference in group averages. Before the independent-samples t-test was interpreted, the assumption of homogeneity of variance was confirmed using Levene's test. The test yielded a significance value of 0.948 and an F-statistic of 0.004. The assumption of equal variances was satisfied because the significance value was greater than 0.05. Consequently, the results from the row assuming equal variances were used for the hypothesis test. The independent-samples t-test produced a t-value of −0.693 and a two-tailed p-value of 0.490 with 83 degrees of freedom (df = 83). The absolute t-statistic was below the critical t-value (|−0.693| < 1.989), but the p-value exceeded the 5% level of significance (0.490 > 0.05). Consequently, the null hypothesis cannot be rejected, suggesting that there is no statistically significant difference between the mean stock returns of commercial banks and HIMBARA institutions. The mean difference between the two groups was −4.666 with a 95% confidence range of −18.051 to 8.720. This result offers more proof that there isn't a statistically significant difference between the two groups because the confidence interval contains zero. Therefore, even although the average stock return of private banks was higher than that of the bank of HIMBARA, the difference was not statistically significant.

Discussion

The Effect of Capital Adequacy Ratio (CAR) on Stock Returns

The outcome demonstrates that the Capital Adequacy Ratio (CAR) has a slight but negative impact on stock performance. Above the 5% threshold of significance, the CAR coefficient was −0.361, the t-statistic was −0.560, and the probability value was 0.5769. Higher CAR was therefore associated with lower stock returns but this association was not significant. The results indicate that variations in stock returns over the sample period could not be explained by differences in the banks’ capital adequacy.

This finding does not fully support Signalling Theory from a theoretical perspective. Brigham and Houston (2021) state that strong financial conditions may send positive signals to investors about the stability of the firm and the ability to manage risk. In the banking world, a higher CAR is a sign of more capital adequacy and a stronger ability to absorb potential losses. Hence, theoretically, a high CAR is expected to enhance investor confidence, market valuation, and eventually, higher stock returns. However, the weak relationship found in this study suggests that CAR may not have given a sufficiently strong signal to affect investors’ market responses.

One possible explanation is that a relatively high CAR does not necessarily mean more efficient utilisation of bank capital. A strong capital position improves a bank’s ability to absorb risk, but investors may also look at whether the capital on hand is being used effectively to support lending and other productive activities. In the absence of profitability or business expansion, the additional capital may not be associated with an increase in expected returns when capital adequacy increases. Therefore, investors may be more inclined to look at indicators that are more closely related to earnings generation and market performance rather than capital adequacy alone. Furthermore, CAR is closely related to banks’ prudential and regulatory requirements. Hence, the market may see an adequate CAR as a basic prerequisite for banking operations instead of an additional piece of information distinguishing one bank from another. With capital adequacy at acceptable levels in most banks, CAR variations may offer little incremental information to investors. This condition may explain why the changes in CAR did not elicit a statistically significant response in stock returns during the 2020–2024 observation period.

The finding is in line with Purba (2023) who found that CAR did not significantly affect stock returns, implying that investors do not necessarily regard capital adequacy as a key determinant in making investment decisions. The present result thus strengthens the case that CAR has little explanatory power for stock-return movements when considered apart from other aspects of bank performance. Likewise, Enjela and Wahyudi (2022) found that CAR did not significantly affect stock returns. One reason is that banks generally maintain capital adequacy in keeping with regulatory requirements and this lowers the ability of CAR to differentiate the attractiveness of investment in individual banks. In such cases, investors may be more responsive to information regarding profitability, growth prospects, market valuation, and other indicators that are more directly related to expected investment returns.

The Effect of Net Profit Margin (NPM) on Stock Returns

The outcome demonstrates that the impact of Net Profit Margin (NPM) on stock performance is negative but statistically negligible. The probability value of 0.1097 was higher than 5% of the significance level, and the NPM coefficient was −0.123 with a t-statistic of −1.618. A one-unit increase in NPM was linked to a 0.123-unit decrease in stock returns, despite the correlation not being statistically significant. Therefore, throughout the period under consideration, an increase in net profit margin did not always translate into an increase in stock returns. Although the negative coefficient also shows an inverse link, the claim that more NPM consistently lowers stock returns should not be supported by this relationship because it is statistically insignificant.

According to the Signaling Theory, a greater NPM is typically anticipated to give investors a good signal since it demonstrates the business's ability to efficiently produce net income from operating revenue. A strong profit figure is often taken as a sign of good financial performance and can raise investors’ expectations of future returns. However, the study's findings indicate that NPM's signal was insufficiently powerful to generate a significant market response. However, this may be the case because investors consider other financial metrics, such as Return on Equity (ROE) and Return on Assets (ROA), which more accurately reflect return on capital and assets, in addition to net profit margins when evaluating banking stocks.

The unique features of the banking industry can account for NPM's minimal influence. Banks’ revenue and cost structures are more complex than those of non-financial companies, with interest income, interest expenses, fee-based income, credit risk, funding costs and other financial activities. Hence, a higher net profit margin does not necessarily imply an enhancement in the bank’s overall capacity to generate sustainable returns for shareholders. For instance, the rise in NPM may originate from short-term cost efficiency rather than sustainable improvements in core operating performance. In such a situation, investors may put less weight on changes in NPM when assessing the investment prospects of a bank.

This result is consistent with that of Harianti and Sardiana (2022), who discovered a negative and negligible correlation between NPM and stock returns. According to their findings, investors might not give NPM much thought when making investing decisions. NPM is a measure of the percentage of net profit generated in relation to the revenue and does not reflect the returns generated with the shareholders’ equity directly. So even if a company reports an increase in NPM, investors may not necessarily view this information as sufficient evidence that the stock will perform better in the future. This result is also in line with Purba's (2023) discovery that NPM has a negligible and negative impact on stock returns in the banking sector. The finding suggests that net profit margin may not be very useful in providing a strong market signal in banking companies. Investors may also pay more attention to the banks’ asset and capital management efficiency, and external factors affecting the banking performance and stock market conditions. Therefore, changes in NPM alone may not be sufficient to induce significant changes in investors’ valuation of bank stocks.

The Effect of Return on Equity (ROE) on Stock Returns

The result shows that Return on Equity (ROE) has a minor but positive effect on stock return. The coefficient of ROE was 0.846, above the significance level of 5%, with a probability value of.1180 and a t-statistic of 1.580. It shows that, when all other factors remain constant, stock returns increased by 0.846 units for every unit increase in ROE. But the association was statistically insignificant. However, although a positive relation is found, there is not enough empirical evidence to claim ROE is a significant determinant of stock returns in the period of observation. The positive sign of the ROE coefficient is consistent with theoretical predictions according to Signalling Theory. ROE is a measure of profitability, showing investors how well a company is using shareholders' equity to generate earnings. A company's return on equity (ROE), which can be a good predictor of its financial success and prospects, increases with how well it uses its equity to generate profits. In theory, this type of signal should boost investor confidence and demand for the company's shares, leading to stronger returns and higher stock prices. However, the insignificant finding in this study suggests that the positive signal sent by ROE was not strong enough to lead to a systematic market reaction.

One explanation might be that investors do not use one profitability indicator in valuing banking stocks. Decision on investment could also be taken with wider range of information such as other financial ratios, macro-economic conditions, exposure to risk, government policies and market sentiment. So, even if a bank shows a relatively high ROE, investors may not react instantly by increasing the demand for its shares if other financial or economic information does not foster favourable expectations. This result implies that a higher accounting profitability does not necessarily directly lead to a higher stock return.

The characteristics of the banks included in the study may also be the reason for the insignificant relationship. Particularly the role of HIMBARA banks and Regional Development Banks (BPDs) is strategic, apart from the fundamental purpose of commerce. 5. Government owned banks may participate in the economic development policies and financing programs. 6. Development agent Therefore, their financial performance could be partly due to policy related roles and objectives other than profit maximisation. In such cases, investors may look at ROE as part of a bigger problem, such as government policies, bank strategies, risk and possible future growth rather than a single indicator of investment attractiveness.

From the standpoint of Signaling Theory, the findings also imply that a good financial signal does not always result in a notable market response. But investors may also look beyond the magnitude of reported profits to their quality and sustainability. If investors believe that the improvement in earnings is temporary, due to non-operating factors, or unlikely to continue in future periods, then a high ROE may be a weaker signal. So, the market’s valuation of ROE may be influenced by the credibility and sustainability of the underlying earnings of the reported ratio. This result is supported by Sholihah and Susilo (2021) study which stated that the ROE is not necessarily the main consideration of investors in making investment decisions. Such similarity further corroborates the suggestion that ROE may not be a good proxy for stock-return movements under certain market and industry conditions. Since investors may use other information they deem more pertinent to determine projected returns, changes in ROE may not have a statistically significant independent effect on stock returns.

The findings of this study, however, disagree from those of Harianti and Sardiana (2022), who claimed that ROE significantly and favorably affects stock returns. The difference in the results indicates that the relationship between ROE and stock returns may not be robust after taking into account the sample characteristics, industry context, time period of observations and market conditions. Thus, the effect of profitability information on investors' reactions should not be assumed to be the same in different research settings.

The Effect of Price to Earnings Ratio (PER) on Stock Returns

The findings demonstrate that the Price to Earnings Ratio (PER) significantly and favorably affects stock return. The probability value of 0.0000 was below the 5% significance level, and the coefficient of PER was 1.006 with a t-statistic of 5.512, which was greater than the critical t-value of 1.990. This finding suggests that, ceteris paribus, a unit rise in PER was associated with a roughly 1.006 unit increase in stock returns. Thus, PER was the only explanatory variable in the model that had a statistically significant partial relationship with stock returns during the observation period, unlike CAR, NPM, ROE, and CSR. The positive relationship implies that the higher PER values were accompanied by higher stock returns in the observed banking firms. Investors frequently use PER, a measure of the link between a company's share price and earnings per share, to ascertain what the market anticipates regarding a company's future earnings potential. Because they expect better outcomes in the future, investors may be ready to pay more for earnings today if the PER is high. Investor interest and demand for a company's shares may result from such expectations, raising share prices and, consequently, stock returns.

This is consistent with Signalling Theory, under which the financial information dispensed by a firm can act as a signal to investors while they are assessing the firm’s prospects and making investment decisions. Harahap (2016) stated that the financial statement information published in the form of financial statements could be good news (favourable news) or bad news (unfavourable news) about the company's future prospects. In this study, PER appears to be an appropriate market signal, since it reflects investors’ assessment of the relationship between share prices and corporate earnings. The positive and statistically significant coefficient indicates that stock returns in the sample under observation were positively impacted by a higher market value in relation to earnings. The results are also consistent with Husnan and Pudjiastuti (2015) and Ross et al. (2015) who describe that PER reflects market expectations about future earnings and growth prospects of a company. A high PER could indicate that investors are willing to pay more for the company's stock and have high hopes for its future performance. Therefore, in addition to the current accounting performance, PER takes into account the market's evaluation of the company's potential to generate future profits. Perhaps this characteristic explains the fact that PER is found to be statistically significant with stock returns while some of the accounting-based financial ratios tested in this study are not.

In addition, the result is in line with Hartono’s (2022) view that fundamental analysis considers financial information and company prospects as important considerations in assessing securities. Investors can assess how the market values a company's earnings and forecast its future course by using valuation measures such as PER. Higher PER is perceived by investors as good growth prospects which could increase demand for the shares and thus could result in share-price appreciation and higher realised stock returns. But the positive relation should be interpreted with caution. A high PER does not necessarily mean better fundamental performance, as it can also mean a relatively high market valuation relative to current earnings. Thus, the result should be interpreted specifically in the empirical context of the 17 banking companies observed during the period 2020–2024: higher PER was significantly associated with higher stock returns after controlling for CAR, NPM, ROE, and CSR. The finding does not imply that higher PER will necessarily result in higher stock returns in all banking companies and market conditions.

The Effect of Corporate Social Responsibility (CSR) on Stock Returns

According to the findings, corporate social responsibility (CSR) has a negative but statistically insignificant impact on stock returns. The CSR coefficient was -2.111 with t-statistics -0.140 and probability value was 0.8893. This was greater than 5% significance level. Thus, a 1-unit increase in CSR is associated with a 2.111-unit decrease in stock returns, ceteris paribus, but the relationship was not statistically significant. Hence, the negative coefficient should be read as the sign of the estimated relationship and not as proof that higher CSR systematically reduces stock returns. The results suggest that further CSR disclosures and activities implemented by the observed banking firms were not linked to a significant increase in the returns received by the investors. However, the estimated coefficient was negative. The absence of statistical significance implies that variations in CSR were insufficient to account for variations in stock returns over the study period. This result implies that, in comparison to information that more closely represents financial performance and market expectations, investors' value of banking stocks may not be much influenced by CSR information.

Stakeholder theory, the results do not fully support the theoretically expected relationship. Carroll and Buchholtz (2015) contend that by engaging in socially responsible activities in response to stakeholder interests, firms can improve their reputation, legitimacy, and stakeholder support, which in turn may contribute to corporate value. From this perspective, CSR activities are supposed to improve the relationship between stakeholders, corporate reputation and long-term business sustainability. However, the insignificance of the relation found in this study implies that these potential benefits were not sufficiently priced in stock returns during the observation period. Signalling Theory offers an alternative interpretation of the results. According to Brigham and Houston (2021), when a corporation discloses pertinent information about its performance and future prospects, it may have an impact on investors. Investors may not have viewed CSR information as a sufficiently strong economic signal to predict future financial performance given the circumstances surrounding this study. CSR disclosure may be viewed by investors as being partly influenced by regulatory, social or institutional pressures rather than as information directly signalling better profitability or higher future returns for shareholders. Thus, the variations in CSR disclosure may not result in variations in market valuation and stock returns in an immediate period.

One other possible explanation is that investors in the banking sector may look more closely at financial fundamentals when considering investment opportunities. Profitability, earnings growth, operational efficiency, valuation metrics, dividend prospects, and other financial considerations may be viewed as more directly related to expected investment returns. Therefore, CSR information may be weighted relatively less when making short-term investment decisions. This interpretation is also consistent with the present study, in which PER was statistically significant, and CSR was not, suggesting that the market valuation information had a more visible relationship with stock returns in the estimated model. This outcome is consistent with the findings of Pratiwi and Yuliana (2023), who discovered a statistically insignificant but negative correlation between CSR and stock market performance. The findings imply that when making investment selections, investors in Indonesia's banking industry might prioritize financial metrics over corporate social responsibility initiatives. The similarity in these findings adds credibility to the argument that the potential economic benefits of CSR may not necessarily be reflected immediately in stock-market performance, especially when such benefits are expected to materialise over a longer time frame.Based on Stakeholder Theory, the results do not fully support the relationship that is expected based on theory. Carroll and Buchholtz (2015) argue that firms responding to stakeholder interests by engaging in socially responsible activities can increase their reputation, legitimacy, and stakeholder support, which ultimately may add to corporate value. From this point of view, CSR activities are expected to enhance the relationship between stakeholders, corporate reputation and long-term business sustainability. However, the insignificance of the relation found in this study implies that these potential benefits were not sufficiently priced in stock returns during the observation period. The findings also can be interpreted through Signalling Theory. Brigham and Houston (2021) state that the information disclosed by a company may influence investors when it is relevant information about the company’s performance and prospects in the future. In this study’s context, investors may not have viewed CSR information as a sufficiently strong economic signal to predict future financial performance. Investors may perceive CSR disclosure as a response to regulatory, social, or institutional expectations rather than as information that directly signals an improvement in profitability or future shareholder returns. Thus, changes in CSR disclosure may not immediately impact market valuation and stock returns.

The findings of the present study are also consistent with the study of Irawan et al. (2025) which suggested that in government-owned banks, CSR implementation may be perceived as being driven by institutional or regulatory considerations rather than solely as a business strategy to generate economic value. Under such circumstances, investors may not always interpret higher CSR activity as a sign of future higher profitability. The potential economic benefits of CSR expenditure may not easily be quantifiable immediately, which may also be considered as a short-term cost, and may help explain the negative sign of the estimated coefficient. However, as the coefficient in the present study was statistically insignificant, the negative direction should not be construed as evidence that CSR expenditure reduces shareholder returns. The findings are also relevant when considering the institutional characteristics of HIMBARA banks and the Regional Development Banks (BPDs). Being banks with a significant government ownership and development-oriented functions, their social responsibility programs may target broader objectives such as national and regional development, MSME empowerment, financial inclusion, and community welfare. Although such activities may produce substantial social benefits, their economic effects on firm valuation may take time to materialise and may not be directly reflected in immediate stock returns. Thus, the market may not immediately respond to increased CSR activities with large changes in share demand or stock returns. Further, many of the possible advantages associated with CSR such as corporate reputation, customer loyalty, public trust, stakeholder relationships and social legitimacy are intangible and are likely to be long-term in nature. The lack of a meaningful correlation between CSR and stock returns over the study period may be due to this temporal factor. A longer time horizon may be necessary to observe the benefit of CSR to company performance since it may generate value through mechanisms that are not immediately reflected in market movements.

Differences in Stock Returns between HIMBARA and Regional Development Banks (BPDs) and Private Banks

The study's findings indicate that there is no discernible difference in stock returns between HIMBARA, commercial banks listed on the Indonesia Stock Exchange, and Regional Development Banks (BPDs). For the independent-samples t-test, the p-value (0.490) was more than the significance level of 5% and the absolute value of the t-statistic (0.693) was less than the critical t-value (1.989). Thus, the empirical evidence does not suggest that the two groups differ significantly in terms of stock returns. This finding implies that the ownership status of the bank (government related or private) was not enough to differentiate the stock returns that investors received during the observation period. Investors react to information that indicates the company's financial situation and prospects, according to the Signaling Theory. In the banking sector, investors may be more interested in information regarding profitability, asset quality, capital adequacy, operational efficiency, risk management and earning generating ability than in ownership status alone. Hence, if HIMBARA and BPD banks and private banks convey similar signals about financial performance and future prospects, the ownership structures may not necessarily result in different market reactions as reflected in stock returns.

The results also suggest that the ownership status might not be a strong enough signal to differentiate market performance of major Indonesian banks. HIMBARA banks are, however, government owned and institutionalised, which usually helps investors to have confidence. However, well-known private banks with competitive financial results and solid market reputations include Bank Central Asia (BCA), Bank Mega, Bank Pan Indonesia (Panin Bank), and CIMB Niaga. So maybe investors should look at individual banks on their financial fundamentals and business prospects, not just whether they are government or privately owned. This interpretation is in line with the explanation of Indro et al. (2022) that the returns of banking stocks are more closely related to bank soundness and financial performance indicators than to ownership features. Their results show that investors respond more to information on profitability, asset quality and bank risk than to differences in ownership structure. This is consistent with the current conclusion that ownership classification alone did not significantly affect stock returns.

Likewise, Soma (2025) found that the main determinants of banking stocks’ returns were risk and profitability. The ability of the bank to generate earnings is an important consideration for investors as this is linked to the future prospects for dividends and possible increases in the value of the firm. The importance of fundamental signals in understanding investors’ responses to banking stocks was also emphasised by Iswandi (2020). Investors are inclined to react to information concerning financial performance and prospects of companies, so that differences in ownership characteristics do not necessarily lead to differences in stock returns. The absence of difference may also be associated with the relatively similar external environment in which the two groups operate. The Indonesian banking industry consists of government-related and private banks, which are all under the same regulatory and macroeconomic environment. They are subject to banking regulation and supervision and experience similar changes in interest rates, inflation, economic growth and financial-market conditions. Similar stock-market dynamics for the two groups may be driven by common external factors, thus reducing the degree to which ownership status per se differentiates their stock returns.

When considered in the context of the Efficient Market Hypothesis (EMH), the result is also comprehensible. According to this view, market prices reflect all the available information relevant to firm value and prospects. Thus, information of ownership by the government or the private sector would not necessarily imply consistently different stock returns if such information is already known to market participants. Instead, stock prices and returns may be responding to new information about financial performance, risk, growth prospects and other factors that investors consider relevant.

CONCLUSION

The findings indicate that while ROE has a positive direction and CAR, NPM, and CSR have a negative direction, none of these factors significantly affect stock returns. On the other hand, it was discovered that PER significantly and favorably affected stock returns. Thus, the market valuation reflected in PER was more strongly associated with investors’ returns than the other variables examined. The findings also indicated that stock returns were not significantly different between HIMBARA and private banks, meaning that the ownership status alone did not matter in stock-market performance. Thus, bank management needs to enhance the optimal utilisation of productive capital, improve the transparency of financial information, and strategically implement CSR to support long-term value creation. Instead of relying on a single measure, investors should look at banking stocks through a comprehensive mix of financial fundamentals, valuation measures, risks and external economic conditions. In order to provide a more thorough explanation of stock returns, further research is recommended to incorporate additional variables such as interest rates, inflation, credit risk, dividend policy, and other macroeconomic factors, as well as to extend the observation time and the scope of the study.