The independent resource on global security

The myth of Sisyphus and the security burden of the Middle East

Sisyphus was condemned to infinitely repeat the meaningless chore of pushing a boulder up a mountain, only to roll it down again and then start anew his uphill journey. Like Sisyphus, many Middle Eastern countries purchase weapons every year at exorbitant opportunity costs. The Arab Spring points to the utter discontent of the people in the Middle East with a continually evolving security web that has dried up prosperity, choked human aspirations and pushed people towards the margins.

Modern nations seldom engage in costly wars and bloody conflicts, yet they devote a sizable proportion of their economic resources to defence spending, a phenomenon known as ‘security burden’. It was widely held that security burdens had acted as deterrence against open wars. In other words, the capacity to hurt one’s rival can be used as a motivating factor for rivals to avoid war and influence other states’ rational behaviour.

Despite the fact that security burden is neither necessary nor sufficient to prevent wars, as new research confirms, the average world security burden has been stable at roughly three per cent of global gross domestic product (GDP) since 1990 with a per capita global burden of $249 each year.

Thus, security burden has a mammoth opportunity cost. Of the top 10 nations in terms of security burden, six are in the Middle East. It has long been established that emulation within a regional group—popularly known as a ‘security web’—plays an important role in triggering and perpetuating regional arms race and the Middle East is no exception.

The three main determinants of the security burden of a nation

From traditional research on security issues we know that socioeconomic, strategic and political variables are the three principal determinants of the security burden of a nation. First and foremost, socioeconomic variables include population, GDP, trade and external aid. Population has been shown to have a significant and negative effect. The impact of GDP is inconclusive. The impact of trade is shown to be positive and significant in dynamic panels, while in static and fixed effect models the impact of trade is negative. Collier and Hoeffler have shown that the effect of developmental aid is positive.

Second, past studies have highlighted external wars, civil wars, security webs, external threats, internal threats, potential enemies, emulation incentives and regional dummies as strategic variables. Third, the political variable highlights the possible role of democracy in reducing the need for security burden.

Despite being a major issue in any regional context, there is no theoretical model to explain the precise determinants of security burden. In a forthcoming paper, I have made a first attempt to offer an economic model for explaining the determination of defence burden in a bargaining framework. The paper develops a simple theory and uses direct evidence from the Middle East to highlight the possible determinants of security burden.

The hawks versus the doves

The paper posits that there are two types of agents in a country—hawks and doves—who engage in bilateral bargaining to decide security burden. Moreover, it assumes that some citizens seek to increase their nation’s security burden (the hawks), while others want to reduce security burden (the doves). The payoff/return functions of doves and hawks are simplistic: hawks like security burden while doves do not.

A Nash bargaining model, in which hawks and doves negotiate over the split of per capita GDP between defence spending and all other expenses, was developed in order to determine security burden. The construction of the model allows one to examine security burden of a nation as the per capita financial burden of maintaining and arming each member of the military in a country. The Nash bargaining solution prescribes that the optimal security burden depends on per capita GDP and the relative bargaining power of hawks and doves.

Past conflict, past per capita income, past military spending and past inequality were chosen as factors that can influence relative bargaining power of hawks. The model makes five predictions. First, the current defence burden will bear a positive relationship with the current per capita GDP. Second, past conflicts will have a positive influence on the current defence burden. Third, past inequality will have a positive impact on the current defence burden. Fourth, past per capita GDP will have an inverse relationship with the current security burden. Finally, the current security burden will be positively influenced by the size of the past military personnel.

The determinants of security burden can be estimated by using a set of panel data including observations for ten Middle Eastern countries covering the period 1963–99. Seven Arab countries—Algeria, Egypt, Jordan, Kuwait, Morocco, Syria and Tunisia—were chosen, along with three non-Arab countries—Iran, Israel, and Turkey—that play important roles in regional security webs.


As predicted by the theoretical model, the current per capita GDP has a positive and statistically significant impact on the security burden. This result is in consonance with Collier and Hoeffler’s findings. Past conflicts (with a lag of three years) tilt the bargaining power in favour of the hawks, which therefore has a positive and statistically significant impact on the current security burden in the Middle East. While all past conflicts have a positive impact on the current defence burden, the lag of three years is statistically significant and other lags are not.

This finding confirms the role of strategic variables in the determination of security burden. In addition, the per capita GDP in the previous year has a negative and statistically significant effect on defence burden. The negative sign implies that hawks have a tendency to ‘target’ defence burden.

Furthermore, the size of a country’s army (with a lag of one year) has a positive and significant impact on the current security burden, since the size of an army is an important determinant of the size and power of the pressure group represented by the hawks. Finally, past inequality, with a lag of one year, has a positive effect on security burden. This simply means that current inequality of income increases the possibility of future conflict and seems to increase the bargaining power of hawks so that hawks can increase future arms spending to prevent future conflicts.

The time profile of security burden in the Middle East turns on the pivot of the evolving balance of power between hawks and doves. If the relative bargaining position of the doves continually worsens one can then expect the myth of Sisyphus to hold its sway and the arms race will continue unabated in the Middle East.


This blog post is published as part of a collaborative partnership between SIPRI and Economists for Peace and Security (EPS).