Indian Stock Market Efficiency: Evidence from Banking Sector
1. Introduction
In recent years, the Indian stock market has gained recognition as a rapidly growing and influential financial market on a global scale (Mallesha & Archana, 2023). Consequently, researchers have turned their attention towards examining the efficiency of the Indian stock market. This study aims to contribute to the existing body of literature by analysing whether the banking sector precisely follows a random walk pattern.
According to the Random Walk Hypothesis, stock prices reflect all available information and exhibit a random pattern, rendering it impossible to predict future price movements based on historical data (Fama, 1965, 1970). If the banking sector aligns with this hypothesis, it suggests that the market is efficient, and any attempts to outperform it through stock selection or market timing would be futile. On the other hand, if the hypothesis is refuted, it indicates the presence of predictable patterns or inefficiencies in the market, which presents opportunities for investors to achieve superior returns (Dsouza & Mallikarjunappa, 2015).
The S&P BSE BANKEX index, a sectoral index of the Bombay Stock Exchange (BSE) of India, represents the performance of the banking sector within the country. The banking sector plays a critical role in the Indian economy as a key financial intermediary facilitating economic growth (Lodha & Kumawat, 2022). Understanding the behaviour of the S&P BSE BANKEX index holds paramount importance for investors, policymakers, and market participants, as it provides insights into the overall health and stability of the Indian banking sector (Hossain & Maitra, 2020).
2. Literature Review
The study builds upon prior research conducted by Kalsie (2012) on weak-form market efficiency for the National Stock Exchange (NSE), which utilised run tests with 30-day average prices from 2001 to 2007. Kushwah et al. (2013) also examined the weak form of market efficiency in the NSE using run tests and concluded that the Indian stock market is efficient in its weak form.
On the other hand, Kumar & Kumar (2017) found that the real estate sector in India was not functioning optimally in terms of weak-form efficiency. Chavarkar and Nayak (2022) investigated the efficiency of Indian pharmaceutical stocks in both pre-pandemic and pandemic periods. By building on these prior studies and focusing on whether the Indian banking index follows a random walk hypothesis, this research seeks to shed more light on the efficiency of the Indian capital market.
3. Research Methodology & Econometric Tests
The study utilised secondary data as the basis of investigation. The sample comprised daily closing prices of the S&P BSE BANKEX index. The study was conducted over a substantial period from April 2012 to July 2023, encompassing 2,799 data points obtained from BSE India. The closing prices were transformed into natural logarithmic returns ($R_t = \ln(P_t / P_{t-1})$) to facilitate time series analysis.
Several robustified statistical tests were applied using RStudio (version 2023.06.1-524):
- Wald–Wolfowitz Runs Test: Developed in 1940, this non-parametric test determines whether consecutive price changes are mutually independent. The null hypothesis ($H_0$) states that price changes are independent and move at random.
- Automatic Portmanteau Test: Formulated by Escanciano & Lobato (2009), this robustified test evaluates whether the time series is devoid of serial correlation across data-dependent lags, remaining resilient against outliers and heavy tails.
- Automatic Variance Ratio Test (AVR): Formulated by Choi (1999) to improve upon the classical Lo & MacKinlay (1989) test. It employs a data-dependent procedure to determine optimal $q$ and $p$, testing the null hypothesis of zero autocorrelation by comparing multi-period variance to single-period variance (expecting a ratio of 1 for a random walk).
- Rescaled Range (R/S) Hurst Exponent: Introduced by Hurst in 1951 to analyze long-range dependence:
- $0 \le H < 0.5$: Inefficient market; anti-persistent, mean-reverting behaviour with negatively correlated returns.
- $H = 0.5$: Efficient market; random Brownian motion, uncorrelated returns, memoryless process.
- $0.5 < H \le 1$: Low market efficiency; persistent, trend-reinforcing behaviour with positively correlated returns.
4. Empirical Results & Discussions
A. Descriptive Statistics & Normality Testing
| Descriptive Statistic | S&P BSE BANKEX Daily Return Series |
|---|---|
| Mean ($\mu$) | 0.0005 |
| Standard Deviation ($\sigma$) | 0.0150 |
| Skewness | -0.6947 |
| Excess Kurtosis | 12.9016 |
| Minimum Return | -0.1840 (-18.40%) |
| Maximum Return | 0.1017 (+10.17%) |
| Jarque-Bera Test Statistic | 19,672 |
| Jarque-Bera $p$-Value | 0.0000 |
| Total Observations ($N$) | 2,799 |
As depicted in Table 1, the return distribution is negatively skewed (-0.6947), indicating a longer left tail and higher probability of extreme negative returns. The high excess kurtosis (12.9016) reflects a leptokurtic distribution with fat tails and severe outliers. The Jarque-Bera test ($p = 0.0000$) decisively rejects the null hypothesis of normality at the 1% significance level.
B. Wald-Wolfowitz Runs Test for Independence
| Runs Test Parameter | S&P BSE BANKEX Result |
|---|---|
| Total Observed Runs | 1,350 |
| Positive Returns ($n_1$) | 1,399 |
| Negative Returns ($n_2$) | 1,399 |
| Total Sample Sequence ($n$) | 2,798 |
| Test Statistic ($Z$) | -1.8908 |
| Asymptotic $p$-Value | 0.0587 |
Table 2 displays a $p$-value of 0.0587, which exceeds the conventional significance threshold of 0.05. This indicates a 5.87% probability that the observed sequence occurred by pure chance under the random walk null hypothesis. Consequently, the null hypothesis of independence cannot be rejected, confirming that successive daily price movements in BSE BANKEX are independent and move at random.
C. Autocorrelation Analysis: Portmanteau & Variance Ratio Tests
| Autocorrelation Methodology | Test Statistic | $p$-Value | Inference at $\alpha = 0.05$ |
|---|---|---|---|
| Automatic Portmanteau Test | 3.0656 | 0.0800 | No serial correlation; $p > 0.05$. |
| Automatic Variance Ratio Test (AVR) | 1.1567 | 0.2560 | Random walk validated; $p > 0.05$. |
The probability values for both the automatic portmanteau test ($p = 0.0800$) and automatic variance ratio test ($p = 0.2560$) comfortably exceed 0.05. This demonstrates that S&P BSE BANKEX returns are free from linear autocorrelation across time lags, rejecting predictable serial dependence.
D. Long-Range Memory: R/S Hurst Exponent Analysis
| Fractal Dimension | S&P BSE BANKEX Estimate |
|---|---|
| R/S Hurst Exponent ($H$) | 0.5044 |
The estimated Hurst exponent of $H = 0.5044$ is remarkably close to the theoretical benchmark of 0.50. This establishes that the banking return series exhibits no long-term persistence or memory, confirming that price evolution conforms precisely to a geometric Brownian motion random walk.
5. Conclusion & Policy Implications
The comprehensive empirical findings across all four robustified tests confirm that the S&P BSE BANKEX index adheres strictly to the Random Walk Hypothesis, establishing that the Indian banking stock market is weak-form efficient. Because past price changes and historical volume patterns provide zero predictive power regarding future return trajectories, active market timing strategies, chartist heuristics, and technical analysis cannot systematically generate abnormal alpha.
These findings provide vital insights for investors, institutional asset managers, and regulators. Investors are best advised to utilize low-cost passive index funds or fundamental credit-risk valuation models rather than technical momentum trading. Nevertheless, the authors acknowledge that the study is limited by its exclusive focus on the banking sector index. Further research across diverse sectoral indices and broader equity market capitalizations is recommended to enrich the empirical understanding of Indian stock market efficiency.
References
- Chavarkar, S. S., & Nayak, K. K. M. (2022). Analysis of Randomness in the Pharmaceutical Sector of Indian Stock Market: Pre- and During Covid-19 Period. Orissa Journal of Commerce, 43(3), 160–175.
- Dsouza, J. J., & Mallikarjunappa, T. (2015). Does the Indian Stock Market Exhibit Random Walk? Paradigm, 19(1), 1–20.
- Escanciano, J. C., & Lobato, I. N. (2009). An automatic portmanteau test for serial correlation. Journal of Econometrics, 151(2), 140–149.
- Fama, E. F. (1965). The Behavior of Stock-Market Prices. The Journal of Business, 38(1), 34–105.
- Fama, E. F. (1970). Efficient Capital Markets: A Review of Theory and Empirical Work. The Journal of Finance, 25(2), 383–417.
- Hossain, T., & Maitra, B. (2020). Monetary Policy, Trade Openness and Economic Growth in India Under Monetary-targeting and Multiple-indicator Approach Regimes. Arthaniti: Journal of Economic Theory and Practice, 19(1), 108–124.
- Kalsie, A. (2012). Study of the Weak Form of Market Efficiency: An Empirical Study of Indian Stock Market. FIIB Business Review, 1(2), 37–45.
- Kumar, S., & Kumar, L. (2017). Market Efficiency in India: An Empirical Study of Random Walk Hypothesis of Indian Stock Market NSE Midcap. SSRN Electronic Journal.
- Kushwah, S. V., Negi, P., & Sharma, A. (2013). The Random Character of Stock Market Prices. Journal of Business and Management, 6(1), 10–14.
- Lo, A. W., & MacKinlay, A. C. (1989). The size and power of the variance ratio test in finite samples: A Monte Carlo investigation. Journal of Econometrics, 40(2), 203–238.
- Lodha, S., & Kumawat, E. (2022). Impact of Lockdown Announcement on Indian Banking Sector: An Event Study Approach. Orissa Journal of Commerce, 43(3), 29–40.
- Mallesha, L., & Archana, H. N. (2023). Impact of Hindenburg Research Report on the Stock Prices of Adani Group Companies: An Event Study. Asia-Pacific Journal of Management Research and Innovation, 19(1), 40–46.
- Roy, S. (2018). Testing Random Walk and Market Efficiency: A Cross-Stock Market Analysis. Foreign Trade Review, 53(4), 225–238.
- Wald, A., & Wolfowitz, J. (1940). On a Test Whether Two Samples are from the Same Population. The Annals of Mathematical Statistics, 11(2), 147–162.