پژوهشهای اقتصادی (Sep 2024)
Applying Markov-Switching Vector Autoregressive Method in Examining the Hypotheses of Import-output and Export-output Growth in Iran
Abstract
Introduction Empirical analysis of export-led growth (ELG), growth-Led export (GLE), import-led growth (ILG) and growth-Led imports (GLI) hypotheses, are supported by a review of the trade literature and economic growth, which creates verifiable evidence using scientific methods for interpretation. To start with the first hypothesis, ELG is also expressed as the role of exports in economic growth in most empirical researches. The ELG hypothesis is described as a development strategy that focuses on foreign exports while simultaneously aiming to strengthen productive capacity that is consistent with economic growth. This hypothesis includes the promotion of exports and the acquisition of foreign exchange reserves by adopting certain policies supported by advanced technology can potentially benefit economic growth. Exporting is considered a tool for long-term economies of scale. Exports promote economic growth in the domestic market through the use of more technology and skilled labor. This process leads to improved efficiency and productivity in the economy. In line with the above, it can be argued that there may be a non-linear causal relationship between output, export and import, and awareness of this issue and its extent is of great importance for planners and policy makers. Therefore, how to investigate the relationship between non-linear causality and mutual effects of output, export and import needs to be experimentally investigated in Iran. For this purpose, the present study examines the analysis of the non-linear causality relationship between output, export and import and confirms the hypotheses of import-output growth and export-output growth in Iran using quarterly data during the period 1988-2022. In this regard, the theoretical foundations related to the subject will be examined first, and then some related studies will be reviewed. In the following, the introduced model will be estimated and analyzed and the conclusion will be presented. Methodology In this study, the non-linear causality relationship between output, export, and import is investigated and the hypotheses of import-output growth and export-output growth in Iran is examined using a MS-VAR model. This paper employs a MS-VAR model to determine the asymmetric relationship between the variables. In this model, the parameters are time-dependent and the variables in the VAR model behave based on the types of regimes (states) and the transition probabilities between them. This model is used to explore the regime-dependent responses of the output to export and import under different regimes. In the MS model, regimes are expected to pursue a latent random process. One of the most prominent peculiarities of the MS model is its ability to specify the shock performances differently in diverse manners. They are a subset of time series models that are able to analyze the dynamic behavior of variables under different circumstances. In addition, these models are generally suitable for capturing unobserved asymmetries in time series. Findings Since the Iranian economy is export-dependent, it seems that in case of structural breaks, the linear correlation method of the model is insufficient to estimate the total unit effect. Therefore, the Markov regime switching vector autoregression model (MSVAR) is used to analyze the nonlinear causality relationship between economic growth, export and import and to confirm the hypotheses of export- output growth and import-output growth. Three main data sets including real GDP, real exports and real imports are considered in logarithmic and differential form. The results of the unit root test show that all variables are at a stationary level. According to the results obtained in table (2), lag 5, which has the lowest value of Akaike and Schwartz, is determined as the optimal lag order. As can be seen in table (3), in the first stage, the value of the probability value of the χ2 test, which is less than one percent, indicates the non-linearity of the relationship between the variables. Hamilton states that the regime with intercept negative origin represents the bust regime and the regime with intercept positive origin indicates the boom regime. Here, the effect of intercept on economic growth in the first regime is positive and significant, but in the second regime, its effect on economic growth is negative and insignificant. Therefore, here the first regime represents the boom regime and the second regime represents the bust one. According to the results of the probability matrix, it can be said that the boom regime is more stable than the bust. Also, the results obtained from the causality relationship indicate a two-way non-linear causality relationship and confirm the feedback hypotheses, i.e. the hypotheses of export-output growth and import- output growth in Iran. In addition, the results show that in the boom regime, there is a one-way non-linear causal relationship between imports and exports from the export to import side. There is a two-way causality relationship between imports and exports in the recession regime. Discussion and Conclusion In the present study, the non-linear causality relationship and the confirmation of export-output growth and import-output growth hypotheses in Iran have been investigated using quarterly data during the period from 1988 to 2022. For this purpose, the non-linear approach Markov regime switching vector autoregression model (MSVAR) was used to investigate the non-linear causality relationship. The results show that the first regime (boom) is more stable and attractive than the second regime (bust). The results obtained from the causality relationship also indicate a two-way non-linear causality relationship and confirm the feedback hypotheses, i.e., export-output growth, import-output growth in Iran.
