Beyond Market Co-movement: A Seemingly Unrelated Regression (SUR) Based Analysis of Hidden Sectoral Connectedness in Indian Equity Indices
Sohom Majumder *
Department of Management, ICFAI University, Sikkim, India and Department of Business Administration, JIS College of Engineering, Kalyani, India.
Vivek Pathak
Department of Management, ICFAI University, Sikkim, India.
*Author to whom correspondence should be addressed.
Abstract
Understanding sectoral interdependence in equity markets requires distinguishing between co-movement driven by common macro-financial factors and residual cross-sectoral linkages. This study examines conditional sectoral connectedness in the Indian equity market using a Seemingly Unrelated Regression (SUR) framework. Daily return data for eight major NIFTY sectoral indices over the period 2020–2026 are analyzed. Each sector’s return is modelled as a function of lagged own returns and lagged macro-financial variables, including the NIFTY 50 index, India VIX, USD/INR exchange rate, Brent crude oil prices, S&P 500 index, and gold prices, to mitigate simultaneity effects. The SUR–Feasible Generalized Least Squares (FGLS) approach enables efficient joint estimation while capturing contemporaneous correlations across sectoral residuals. The results indicate that global market movements, particularly the S&P 500, exert a strong and consistent influence across all sectors, highlighting the importance of international spillovers. Commodity-linked sectors such as Energy and Metals exhibit strong positive associations with oil price movements, while exchange-rate sensitivities vary significantly across sectors. Despite controlling for observed drivers, the residual covariance structure reveals a dense network of positive contemporaneous associations, suggesting the presence of latent shared influences across sectors. The strongest residual linkage is observed between Energy and Metals, indicating persistent commodity-cycle effects. Sensitivity analysis confirms that the connectedness network is robust to alternative correlation thresholds. Residual diagnostics indicate limited first-order serial dependence, although some higher-order dependence remains. Overall, the findings suggest that sectoral connectedness in Indian equity markets is conditional in nature, with both global and domestic factors playing a significant role, alongside latent cross-sectoral linkages not captured by observable variables.
Keywords: Seemingly unrelated regression, sectoral equity indices, systemic risk, market beta, volatility spillovers, financial connectedness