Imports are vital as one of the determinants of gross domestic product (GDP), and they are principally recognized among the main items in the balance of payments (BOP) of all countries. In this way, all the changes in the volume and value of imports can exert a significant influence on domestic production, and eventually a country's growth and development. Over the last couple of decades, Iran's economy has been highly dependent on agricultural imports, especially the basic goods. The restrictions facing farming and crop production, attributable to diverse reasons, have further led to the compensation of some demands by imports, in such a way that up to 90% of domestic demands have sometimes been met by imports. In reference to the empirical model developed by Bahmani-Oskooee and Goswami, (2004) a significant unilateral or bilateral relationship can be often assumed between imports and economic growth. The given model explicitly presents imports as the function of the national income, the relative import price (i.e., the ratio of prices of imports to domestic goods), and the exchange rate (ER). Following the rises in the national income accompanied by the growth in per capita income (PCI), the demand for imports elevates. Another determinant of imports is their prices. Therefore, an inverse relationship exists between the prices and volume of imports. If there are comparable domestic products, then their prices can change the volume of imports. In view of that, the price ratio for imports to domestic goods is typically one of the determinants of imports. Against this background, the present study attempted to investigate the asymmetry in the relationship between Iran's imports and agricultural growth for the period of 1987-2020.
The self-explanatory non-linear autoregressive distributed lag (NARDL) framework, proposed by Shin, Yu, and Greenwood-Nimmo (2014), as the asymmetric mode of the self-explanatory autoregressive distributed lag (ARDL), is often employed to investigate the non-linear and asymmetric relationships between economic variables in the short and long term. Like ARDL, the NARDL framework has some advantages over other models for testing long- and short-term relationships between variables. This model does not incorporate short-term dynamicity into the error correction, so it can be utilized with not many observations. Besides, it is applicable when there are endogenous explanatory variables. In this line, the agricultural value added (AVA) at constant prices in 2011 in billions of Iranian rials (IRR), the value of agricultural imports in millions of dollars (AIM), the agricultural labor force per thousand (AL), the annual precipitation in millimeters (PRE), the agricultural capital stock at constant prices in 2012 in billions of IRR (AC), and the free market exchange rate in IRR per US dollar (ER) were accordingly selected as the study variables, in accordance with the theoretical underpinnings of growth models.
Investigating asymmetry in the relationship between Iran's imports and agricultural growth for the period of 1987-2020 demonstrated that the positive shock of agricultural imports in the long term could have a significant positive effect on agricultural growth, whereas the same effect for the negative shock was inversely estimated in a non-significant manner. In addition, both effects were significantly different. Indeed, asymmetry in the effect of imports on agricultural growth in the long term was proven. A similar analysis could be further inferred about the short-term impact of agricultural imports. Reflecting on AL, no long- and short-term effects were found to be significant at the 5% level, which implied the saturation of agricultural sector in this respect. The positive shock of AC correspondingly revealed a significant effect, in such a way that it influenced agricultural growth with the coefficient of 0.23, and this represented the capacity of the agricultural sector to absorb more capitals, because the relative lack of capital compared to AL was obvious under the current conditions. The other significant coefficient was associated with PER, which seemed logical, and then justified considering the high share of rain-fed agriculture. The error correction coefficient (λ) was subsequently estimated to be 0.54 and significant, which illustrated that the shock effect had been amortized within less than two time periods.
The present study aimed to empirically investigate a non-linear relationship between Iran's imports and agricultural growth for the period of 1987-2020. In addition to the imports, other growth stimuli, such as AL, AC, PER, and ER (owing to their effects on the supply of inputs imported into the agricultural sector) were recruited. The main findings indicated an asymmetry in the effect of all stimuli (except for the AL) on agricultural growth. In other words, the effect of increasing and decreasing shocks in the stimuli on the target variable (i.e., agricultural growth) was different, which was statistically significant in the long term. Another important point was that the effect of the increasing shocks was estimated to be stronger than the decreasing ones in respect of the absolute value. Thus, it was argued that the shocks of increasing agricultural imports, AC, PER, and ER could more strongly influence the target variable. Considering the decreasing effect of the negative shocks, it is thus essential to counteract the negative impact of the decreasing shocks on agricultural growth by implementing the shock-absorbing policies. In this regard, the precautionary reserves policy is highly recommended, mainly on the subject of the inputs demanded in the agricultural sector.