The stock market typically generates a substantial amount of data, and intricate networks can be employed to represent stock market dynamics and the correlation of stock prices. To assess this, we utilize stock data from the Tehran Stock Exchange (TSE), covering the years 2013 to 2023, to develop a stock correlation network using the threshold approach. Given that highly central stocks are more strongly interconnected with other firms, they are expected to be more exposed to aggregate shocks and market-wide spillovers, in addition to firm-specific risk. This raises the question of whether centrality can serve as a risk factor within a multifactor asset pricing framework. To address this issue, we augment the Arbitrage Pricing Theory (APT) model with a centrality factor and evaluate its pricing implications using both cross-sectional and time-series methodologies. The Fama–MacBeth (1973) two-step procedure is employed to estimate the risk premia associated with centrality, excess market return, firm size, inflation, and exchange rate fluctuations. In parallel, the Gibbons–Ross–Shanken (GRS) test is applied to assess the joint validity of the proposed asset pricing specification. The empirical findings indicate that centrality carries a statistically significant risk premium, suggesting that a firm’s structural position within the return network represents an independent dimension of systematic risk. Moreover, the GRS test supports the overall validity of the augmented APT model. • Constructs a stock return correlation network for Tehran Stock Exchange (2013–2023) using a threshold-based method. • Introduces network centrality as a novel risk factor, capturing firms’ exposure to market-wide shocks and spillovers. • Estimates an augmented APT model via cross-sectional and time-series approaches to evaluate centrality’s pricing power. • Reveals that firms’ network positions constitute an independent source of systematic risk, with implications for asset pricing and portfolio diversification.
Moghadam et al. (Wed,) studied this question.