![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
Abstract
This study evaluates NATO long run defense burdens by analyzing the time‐series properties of burden measures, namely growth of defense spending, defense share in national output, defense share in government spending, defense spending per capita, and defense share in total NATO spending for the time period 1949–2002. The study also compares the effect of using government Purchasing Power Parity conversion factors and Market Exchange Rates for defense share in total NATO expenditure conversions and the implications of NATO expansion in light of the defense burden measures of the newer NATO members.
ACKNOWLEDGEMENT
I would like to thank Robert McNab for helpful comments and discussions. I would also like to thank participants at the Western Economic Association meetings, the faculty at the Defense Resources Management Institute, and two anonymous referees for comments on earlier drafts.
Notes
1. NATO doctrine in the 1950s and early 1960s corresponds to the strategy of mutual assured destruction and deterrence; the doctrine of flexible response in the 1970s and 1980s; crises management from the 1990s to the present. Sandler and Murdoch (Citation2000) have an in‐depth discussion of each period.
2. Iceland maintains no military but its strategic location allows it to host NATO bases.
3. Congressional Budget Office (CBO), 2000.
4. Instead of the new alliance members fully integrating into NATO, proposals have been made where each NATO member would specialize and provide a niche capability. However, it appears that NATO is leaning toward integrating new members into its structure.
5. Oneal (Citation1992) computes the value of conscription as a percentage of military expenditures and identifies it as a budgetary savings or opportunity cost associated with conscription. Oneal provides a value of conscription estimate for three years, 1974, 1981 and 1987. The study provides a lower bound, an upper bound, and a best estimate of the value for each year. In addition, the study concludes that no adjustments are needed for the all‐volunteer forces of Canada, Luxemburg, the UK, and the USA.
6. Since peacekeeping activities are more in line with a pure public goods model and provide mostly non‐excludable and non‐rival benefits, free riding results as smaller nations depend on larger nations to maintain these activities. However, there is an element of private benefit as a result of peacekeeping participation that is symbolic. Namely, larger nations are motivated by their desire to maintain their status by reminding the world of their importance in times of both war and peace, while smaller nations enhance their prestige by contributing to world peace.
7. CBO 2001 suggests three burden sharing measures that are identified as standard: ME/GDP, ME/POP, and the proportion of the labor force in the military. We opted not to use the latter since it is a component of military expenditure and using it would be a duplication of the information used to generate ME. CBO also identifies other burden sharing measures that apply to the post Cold War environment. These are contributions to NATO’s rapid reaction forces, contributions to peacekeeping forces in Bosnia and Kosovo, and economic assistance to Central and Eastern Europe. We opted not to include these three additional measures for a number of reasons. The cost for the first and second measures are already reflected in each ally’s overall ME. The third measure is difficult to quantify as pointed out by the CBO report. Economic assistance is a small part of the economic interaction between the allies and the countries of Central and Eastern Europe. The major interaction is due to foreign direct investment. In addition, since NATO is a military alliance, it can be argued that non military assistance should not be a measure of military commitment.
8. We do not adjust ME for countries that rely on conscription based on Oneal (Citation1992). The study indicates that the necessary cost inflation to adjust for conscription is sufficiently small so as not to affect the ordering of allies according to defense burdens. In addition, Khanna and Sandler (Citation1997) establish that conscription‐adjusted data give the same conclusion as non‐adjusted data.
9. The conversion of military expenditures to US dollars is handled in a variety of ways. SIPRI converts by deflating to a common year and using average annual exchange rates reported by IMF. NATO currently uses the same methodology as SIPRI. NATO is considering using PPP conversions. The International Institute for Strategic Studies (IISS) uses exchange rates for some countries, lagged dollar values for countries with exchange rate fluctuations, and PPP for some former communist countries. The UN Group of Experts on the Reduction of Military Budgets recommends the use of PPP conversion factors, and in its 1977 report ‘The Reduction of Military Budgets’ recommends the creation of a military price index to be used for military expenditure conversions.
10. There are important differences between stationary and non‐stationary time‐series. Shocks to a stationary series are temporary and will dissipate over time. A stationary series will revert back to its long run mean level thus exhibiting mean reversion. A non‐stationary time‐series, one that has a unit root, has permanent components. The mean is time dependent and there is no long run mean to which a series reverts after a shock.
11. Dickey and Fuller (Citation1979, Citation1981) devised a procedure to test for a unit root. The Dickey–Fuller test is used for the special case of an autoregressive root of one, the AR(1) model. The augmented Dickey–Fuller test extends the Dickey–Fuller test to the AR(p) model in cases where a higher order regression is more appropriate. Three different regression equations are considered where the difference between the three is the presence of deterministic elements. The first equation Δyt = γy t − 1 + ϵt is a pure random walk model. The second Δyt = a 0 + γy t − 1 + ϵt adds an intercept or drift term. The third Δyt = a 0 + γy t − 1 + a 2 t + ϵt includes both a drift and a linear time trend. The parameter tested is γ. The time series contains a unit root if γ = 0.
12. Based on asymptotic power calculations, ERS show that the DF‐GLS local asymptotic power function lies close to the Gaussian local power envelope. ERS also show that the DF‐GLS has higher size adjusted power and smaller size distortions than the standard Dickey–Fuller test. Ng and Perron (Citation2001) apply GLS detrending and show that non‐negligible power and size gains can be made by detrending. They demonstrate that the attractiveness of GLS detrending is that it estimates the deterministic function with more precision and leads to reduced bias in the parameters of the model.
13. ERS also suggest that the DF‐GLS asymptotic derivations remain valid even in the presence of a constant or slowly evolving deterministic component. Perron and Rodriguez (Citation2003) conclude that a change in the intercept is asymptotically insignificant and confirm the ERS results. A data‐dependent lag selection criterion – the Schwarz Bayesian Information Criterion – is used.
14. Belgium, France, Germany, Italy, Luxemburg, Netherlands, Spain, and Portugal fixed their currency to the Euro in 1999, as did Greece in 2001. However, the conversion did not seem to affect their MER share of NATO military spending. With the entry of the Czech Republic, Hungary and Poland into NATO, Luxemburg changed in rank in 1999, dropping from 14th to 18th, with Portugal from 14th to 15th and Denmark from 13th to 14th.
15. Currently, the European members of NATO are modernizing equipment that is less sophisticated than US equipment. This is straining their defense budgets and there appears to be little domestic support for increasing the budget. As a result, the gap between the military capability of the USA and Europe is increasing. The failure of NATO’s European members to narrow the technological gap could result in a situation in which the USA is the only NATO member capable of using sophisticated military power to confront a crisis.
16. In Armed Forces Journal June 2005, Gen. James Jones the Supreme Allied Commander Europe expressed concern that small allies when forced to bear the costs of operations would be reluctant to participate.