What Were the Economic Costs of the Islamic Revolution and the Iran-Iraq War?

This research note presents the main findings of a paper published online in the journal Defence and Peace Economics in October 2020. The original article can be seen at this link.

Four decades ago, one of the major revolutions of the 20th century occurred in Iran, ending its system of monarchy. Mass protests and strikes intensified over the course of 1978, and in response, the Shah enforced martial law. He fled amid increasing unrest in January 1979. The monarchy collapsed on 11 February 1979. In September 1980, Iraq invaded Iran, and the world observed one of the longest ever interstate wars between two major oil producers (1980–1988).

What would Iran’s economy look like in the absence of revolution, war, and violence of that period? Gaining evidence-based insights into the economic costs of conflict and regime change is becoming more important, given the increasing tension between Iran and the United States, especially under Donald Trump’s administration. 

A limited number of studies have examined the economic costs of the Iran-Iraq War. However, none has estimated the economic costs of war and revolution for the average Iranian citizen, in terms of lost income. I aim to do so by using a synthetic control method (‘SCM’), an approach which will help to identify the effect of conflict on economic development and quantify its economic size by comparing Iran with a set of similar countries that did not experience the shocks of war and revolution over the period of analysis (1978–1988).

Data and Model

I use SCM to construct a synthetic control unit for Iran, representing expected GDP figures under a scenario with no revolution, and no war after 1978. I refer to this control unit as “Synthetic Iran.” An outcome variable (in this case, GDP per capita in constant prices) should be comparable between Iran and its synthetic twin before the revolution and war with Iraq. In the case that the trends of outcome show a significant diversion between Iran and synthetic Iran after the shock, it becomes possible to make suggestions about the economic effects of the revolution and war. I will then be able to quantify this diversion, as well as the economic costs of conflict for an average Iranian.

To generate Synthetic Iran, I use country-level panel data for Iran and a sample of MENA and OPEC countries from 1970 to 1988. Restricting my set of potential control countries to the MENA region helps control for cultural, religious and geographical similarities. Considering OPEC members in generating Synthetic Iran makes sense, due to their common natural resource rent dependency.

For the outcome variable, I use GDP per capita in constant 2010 U.S. dollars. In order to have unbiased estimates of the post-revolution-war trajectory of Iran, the control countries for generating Synthetic Iran should not have experienced a main exogenous shock (e.g. war or revolution) from 1978 to 1988. To avoid such bias, I exclude countries affected by Iran’s revolution and war with Iraq. This eliminates Iraq itself. Israel and Lebanon also experienced a series of significant conflicts after Iran’s revolution. With these and some other adjustments (for example for missing data), 11 countries out of the initial 20 remain as possible candidates to generate Synthetic Iran.

The generated Synthetic Iran should have a comparable economic and demographic structure to Iran for the average period of 1970–1977 (1978 was selected as the treatment year, since revolutionary protests and large-scale strikes intensified during this year, leading to the collapse of the monarchy in February 1979). In particular, I use different predictors of real GDP per capita to generate Synthetic Iran before the joint treatment of revolution and war. These predictors are available for all included countries for the period of analysis (1970–1988) and are helpful in producing a counterfactual Iran with similarities to the real Iran before the shock.

Findings

Table 1 shows that Synthetic Iran is best generated by a weighted average of five countries, with Tunisia (56 percent), Venezuela (16 percent), Saudi Arabia (13 percent), Oman (12 percent), and Algeria (1.6 percent) having the highest weights. Table 2 shows the average pre-1978 values of the covariates for Iran and Synthetic Iran. The latter reflects the pre-1978 performance of the GDP per capita covariates for Iran relatively closely. Synthetic Iran is similar to actual Iran in terms of pre-1978 GDP per capita as well as the associated shares of imports, gross capital formation, final consumption (private and public) in total GDP, life expectancy, and population growth rate. Some similarities between the selected countries and Iran in the 1970s are shown in Table 2.

In addition to data on Iran and Synthetic Iran, I present an unweighted average of variables for countries with a weight > 0 during 1970–1977 in Table 2. The predicted outcome (real GDP per capita) in the pre-treatment period is very close between Iran and Synthetic Iran (with optimally selected weights, as shown in Table 1). However, there is a significant gap between the outcome of real Iran and the unweighted average of real income per capita for countries with weight >0. This shows the importance of using a Synthetic control method for this analysis, which generates our counterfactual Iran by assigning the optimum weights to relevant countries.  

Figure 1 shows the GDP per capita trajectory of Iran and its Synthetic counterpart for the 1970–1988 period. Synthetic Iran almost reproduces the per capita GDP of Iran during the entire pre-revolution period, making it possible to closely reproduce the economic characteristics of Iran before the 1978 uprisings without extrapolating outside the support of data from the donor pool. My estimate of the effect of the revolution and war on the per capita GDP of Iran is shown by the difference between actual Iran and its synthetic twin (Table 3).

We can see that the two lines diverge from each other significantly after 1978. While per capita GDP falls in Iran, in Synthetic Iran, per capita GDP keeps its earlier path during the early 1980s. The difference between the two series remains significant towards the end of the sample period. My results therefore imply mainly negative effects from the revolution and war on the economic development of Iran.

Figure 2 shows the estimated income gap between Iran and synthetic Iran with confidence sets (lower and upper bounds). The negative effect of the joint treatment of revolution and war is statistically significant in 9 out of 11 years following revolution.

Conclusion

During 1978–1988, the average annual economic costs for an Iranian were USD 3,150. The lowest average annual cost was USD 1,572 (in 1978), while the highest annual financial burden is USD 5,135 (in 1981).

If Iran had not experienced the revolution and subsequent war with Iraq, it could have allocated oil revenues, in reality devoted to military spending, to education, health, and physical infrastructure instead, creating higher productivity in the long run (see Farzanegan 2011 for a study on oil and government spending in Iran, and Farzanegan 2014 for the nexus between military spending and economic growth in Iran).

Furthermore, the Islamic revolution and war affected the economic position of social classes in Iran, with significant consequences for economic development. Economic disruptions and negative growth rates, which were partly caused by the Iran-Iraq War (1980-88) led to a significant decline in the size of the middle class, with its lowest level coming in at just above 15 percent of the total population in 1988.

Political factionalism after the revolution was another negative factor for economic growth. According to Bjorvatn, Farzanegan, and Schneider (2013), since the Islamic Revolution, reformers, conservatives, and several other factions have been involved in a competition for political dominance in Iran. Bjorvatn et al. (2013) use theoretical and empirical methods and show that in the case of Iran, an increase in oil rents is negatively associated with economic development when the degree of political fractionalization is high. They also find similar results using a sample of oil-based economies worldwide (Bjorvatn et al., 2012). In a related study, using theory and panel regression analysis, Bjorvatn and Farzanegan (2015) show that resource rents can promote political stability, but only when political power is sufficiently concentrated. There is little evidence for the concentration of political power in the post-revolutionary period in Iran.

My study, and the others mentioned, show a significant income loss for Iranians, mainly due to the political instability associated with regime change and the destructive war with Iraq.

References

Bjorvatn, K., Farzanegan, M.R., 2015. Resource rents, balance of power, and political stability. Journal of Peace Research 52, 758-773.

Bjorvatn, K., Farzanegan, M.R., Schneider, F. 2012. Resource curse and power balance: evidence from oil rich countries. World Development 40, 1308–1316.

Bjorvatn, K., Farzanegan, M.R., Schneider, F., 2013. Resource curse and power balance: evidence from Iran. Review of Middle East Economics and Finance 9, 133–158.

Farzanegan, M.R. 2011. Oil revenues shocks and government spending behavior in Iran. Energy Economics 33 (6), 1055-1069.

Farzanegan, M.R., 2014. Military spending and economic growth: The case of Iran. Defence and Peace Economics 25, 247-269.

Farzanegan, M.R., 2020. The Economic Cost of the Islamic Revolution and War for Iran: Synthetic Counterfactual Evidence. Defence and Peace Economics. (in press).

Firpo, S., Possebom, V., 2018. Synthetic control method: inference, sensitivity analysis and confidence sets. Journal of Causal Inference 6(2), 1-26.

Photo: IRNA