The Effect of Oil Price Increase on Economic Growth and Income Convergence (Case Study: Evidence from GCC and Iran)

3. The Empirical Model

In order to estimate the model related to convergence that was mentioned in the previous section, the specification of the empirical framework follows Barro and Martin’s (1995) Model. This framework is in conjunction with the real per capita income growth rate to the initial level of state variables (such as physical-capital stock and human-capital stock) and some control variables such as life expectancy, literacy rate, and so on. In addition, in Solow-Swan model, control variables are determiners of steady-state level of production. According the theory of Barro and Martin (1995), it is assumed that if the initial level of per capita GDP is high, physical-capital stock will also be high. Moreover, according to Soto (2000), the human-capital stock is reflected in lagged value of per capita output. Solow-Swan model predicts that for given values of the control variables, an increase in initial levels of state variables reduces the growth rate. Therefore, per capita production growth rate in the panel data model can be written as follows (Ledyaeva & Linden, 2006):

where y_it is per capital production in  economy in t period. y_i,t-1 is the initial level of per capital production of economy i in t-1 period, a is a negative parameter and represents the speed of convergence, matrix X_it is related to the control variables in economy i  in the period t, and β is the parameters related to it. v_i is the country-specific effect of an economy, and τ­_t is the period-specific effect and ε _it­ is error term of the model.

Assuming , equation (7) will be re-written as follows:

If lny_i,t-1is transferred from left to right, dynamic model of panel data will be as follows:

As it was stated, X includes a matrix of control variables. Generally, these variables are those which affect the economic growth of countries. In neo-classical models such as Solow-Swan and Ramzi, the effect of control and environmental variables on growth rate is the same as their effect on steady-state. Barro and Martin (1995) suggested some control variables such as measures of market distortions, home investment, and rate of openness of economy, financial development and political instability.

In this study, four control variables- among those stated by Barro and Martin- were used which seemed to have effects on the growth rate of the countries in the considered period. The first variable was the logarithm of government expenditure to GDP ratio (Barro & Martin, 1995). Regarding the fact that the countries under study are among developing countries, government plays an important role in these countries, hence this ratio affects the economic growth and it is expected that when it increases, the steady-state of per capita production increases. The second control variable was the logarithm of ratio of investment to GDP that according to the theory, an exogenous increase in the rate of  leads to the increase in steady-state of effective per capita production, and growth rate will increase in proportion to specific rate of state variable (Barro & Martin, 1995). The third reason is that the oil price affects investment directly. The third variable was the logarithm of ratio of trade (sum of exports and imports) to GDP which was considered as a representative for rate of openness of an economy and it was predicted that this variable has also positive relationship with economic growth (Ledyaeva and Linden 2006). The fourth variable was also the logarithm of ratio of savings to GDP which was related to the previous matters and affected the steady-state of economic growth.

3.1. Empirical Results

It is assumed that in the period of oil price increase, per capita income of member countries of GCC and Iran converge to a steady-state. The following equation is used to test defined:

As it is observed in the equation, if the only explanatory variable in the model is per capita income logarithm of the previous period, the hypothesis of absolute convergence will be tested. If <β0 and is statically significant, absolute convergence hypothesis among selected countries will be confirmed; this means all of these countries converge to a steady-state. In this case, growth speed of poorer countries of the sample which have lower initial per capita income will be higher. The result of the test is presented in the following equation:

T=(2.08)        (22.06)

The results show that β is negative, therefore, the hypothesis of the absolute convergence among member countries of GCC and Iran is rejected.

As it was determined, although the economies of selected countries in this study have oil-based structures and oil income plays an important role in their economies, their per capita incomes in the period of oil price increase do not converge to a common steady-state. In other words, the hypothesis of absolute convergence is rejected. Hence, in this section, the factors which affect growth rate are added to the model.

According to Barro and Martin (1995), variables such as literacy rate, life expectancy, ratio of government consumptions to GDP and democracy index are the main determinants affecting the steady-state equilibrium. Among these variables those which have effects on the economic growth rates of the selected countries are added to the model to justify the difference in income level of various economies' steady-state. As it was stated in previous sections, these variables are: logarithm of the ratio of government consumptions (G) to GDP, logarithm of the ratio of investment (I) to GDP, logarithm of the ratio of trade (volumes of both export and import) to GDP and logarithm of the ratio of savings (S) to GDP. Model (8) is applied to test the absolute convergence as follows:

A framework of the panel data model is defined as follows:

where βs are related to each country in the sample and the estimated coefficient related to the oil price for each country is presented in Tables 1 and 2. Except for KSA and Qatar, according to Table 1, all estimated βs are statistically significant with the negative signs.

Table 1. Absolute Convergence Test for GCC + Iran

                                Source: Authors

Table 2. Oil Price Coefficient Estimated of Economic Growth Model for GCC + Iran

Source: Authors

In addition, Table 2 reports that estimated coefficients of the oil price are significant and positive for all the sampling countries.

According to the results, the negative sign of the β coefficient and the rejection of absolute convergence hypothesis across the selected countries represent the existence of difference in steady-state of these countries and heterogeneity of their economic conditions. Thus, the factors which have effects on the steady-state of these countries should be taken into consideration and consequently the conditional convergence is considered that is each country should move toward its steady-state.

Regarding the rejection of conditional convergence hypothesis in the case of the countries under study, it can be said that these results represent that conditional convergence hypothesis in the case of Bahrain, United Arab Emirates, Iran, Kuwait and Oman is rejected, that is, these countries do not converge to their steady-state and this represents that such economies are not in their way to long-run growth. In the case of Saudi Arabia and Qatar where β<0, while is not statistically significant, it can be said that conditional convergence hypothesis is not acceptable. Rejection of conditional convergence hypothesis is due to the absence of steady state equilibrium in the above countries' perspective of planning.

As it was stated, oil price coefficient was significant in growth model and the positive effect of oil price on economic growth was confirmed. This result was predictable beforehand, because economy in all these countries is based on oil and main part of their income is related to oil sector. Moreover, the selected countries are among developing countries where there is no enough planning for using the obtained income equivalently, so economic growth variation will have a process similar to the process of oil price.

4. Conclusion

The results of the estimation of models show that absolute convergence hypothesis in the case of member countries of GCC and Iran is rejected. That is, poor countries in the sample under study do not necessarily grow faster than rich countries.

According to the results of conditional convergence hypothesis test in the case of member countries of GCC and Iran, although oil price significantly affected the economic growth of these countries, this hypothesis in the case of the selected countries has been rejected. This is because of the lack of long-run solid planning as well as the existence of political tension in the region.

Considering the rejection of absolute and conditional convergence hypothesis and also the effect of oil price on trade between the countries of the region on the one hand and sensitive situation of the region in political geography of the world on the other hand, it is suggested that these countries which have main resources of energy of the world and have common cultural similarities attend in the international arena with more power and union and use their chances such as an increase in oil price for developing and gaining power and independence for the region instead of having enmity and conflict to each other.

Hence, considering the results, although oil price affected the economic growth of the selected countries, effectiveness coefficient of this variable was too low and these economies did not converge toward their steady-state. On the other hand, regarding the fact that the main income source of these countries is oil, so it is suggested that these countries use their chances effectively by long-run and accurate planning and consider long-run political and economic benefit instead of entering superficial political tensions.