Document Type : Original Article
Author
Assistant Professor of Economics, Nahavand Higher Education Complex, Bu-Ali Sina University, Hamedan, Iran,
Abstract
Introduction: Understanding the dynamic interdependence among industries within stock market indices is crucial for investment decision-making and the design of economic policies. The Pearson correlation coefficient only captures linear dependence. However, the return distribution of financial variables is not elliptical. Moreover, extreme events occurring in the tails of the distribution during crisis periods can rapidly propagate across financial markets, leading to stronger interdependence, which necessitates the use of more advanced and sophisticated models. It is noteworthy that copula-based models enable the modeling of nonlinear dependencies using flexible choices of marginal distributions. Among the various families of copulas, vine copulas allow for flexible modeling of complex dependence structures by utilizing a wide class of bivariate copulas. The interdependence of individual stock returns is depends on the dependence between different sectors of the stock market. Therefore, it is essential to understand tail dependence across sectors, as each sector often responds differently to economic circumstances. The aim of this study is to reveal the dependence structure of 70 stocks across 10 industries using C-vine, R-vine, and D-vine models.
Methods: In this paper, a ARMA-EGARCH (1,1) model with Student-t innovations is employed for the marginal distributions. To define copula data, the cumulative distribution function (CDF) corresponding to the Student-t distribution is employed as a probability integral transform. Next, copula functions are selected using the marginal data, and vine structures are developed. In this paper, parameters are estimated using the sequential estimation (SE) method and the joint maximum likelihood estimation (MLE).
Results and discussion: The root nodes or industry representatives in the R-vine trees are as follows: Seshahed represents the real estate development industry; Vaomid serves as the representative of the industrial conglomerates sector, while Desobha represents the chemicals and pharmaceutical products sector; Kegol represents the metal ore extraction industry, while Femeli is considered the representative of the basic metals industry; Automobile and auto parts companies are represented by Khazin, while Beterans serves as the representative of companies active in the machinery and electrical equipment industry. The investment industry, the chemical products industry, and the cement, lime, and plaster industry are represented by Pardis, Sharak, and Ceshomal, respectively. The dependence among industry representatives in the R-vine specification is extracted. Furthermore, Node 1 plays a central role among various firms in the chemicals and pharmaceutical products, basic metals, automotive and auto parts, and investment industries. The results also confirm that R-vine models are preferred over D-vine and C-vine models.
Conclusions: According to the results, the first hypothesis—“The dependency structure among industries listed on the Tehran Stock Exchange does not follow a symmetric (normal) pattern and exhibits tail dependence”—is supported. The dependency structure among the ten industry representatives is modeled using various copula families, including the BB8 and survival BB8 copulas. According to the findings, the second hypothesis—“The interconnections among industries listed on the Tehran Stock Exchange possess a multilayered and network-based nature and cannot be reduced to a purely centralized (C-vine) or chain-like (D-vine) structure”—is also supported. The empirical superiority of the R-vine model confirms the hypothesis of networked and multi-source interdependencies within the Tehran Stock Exchange, which is consistent with the structural characteristics of the Iranian economy. In emerging markets such as the Tehran Stock Exchange, dependencies often arise from multiple sources of risk (e.g., exchange rate fluctuations, commodity price shocks, and monetary policy). The economic structure is diversified yet imbalanced, and shocks may propagate simultaneously through multiple transmission channels. R-vine models exhibit superior explanatory power and goodness of fit in modeling inter-industry dependencies compared to C-vine and D-vine structures. In the context of the Tehran Stock Exchange—where dependencies are heterogeneous, shock transmissions occur with varying intensities, and ownership structures follow distinctive patterns—the R-vine framework provides a more flexible and better-fitting representation of inter-industry dependence dynamics.
From an economic perspective, the superiority of the R vine model suggests that shock transmission in the Tehran Stock Exchange does not occur through a single pathway or via a dominant industry. Rather, different industries—depending on their positions within the dependency network—play distinct roles in either absorbing or transmitting risk. This finding is consistent with the structure of the Iranian economy, which is simultaneously influenced by factors such as exchange rate movements, global commodity prices, domestic policies, and institutional constraints. Consequently, the co movement of industries in the Tehran Stock Exchange should be understood as the outcome of interactions among multiple sources of risk, rather than merely a response to a single common factor.
Keywords
