| Abstract | There is a growing body of literature on the determinants of military spending, mostly either starting from the premise of a dyadic arms race, or a welfare maximisation model
incorporating economic, political and strategic variables. This thesis takes the latter approach, to analyse a large sample of developing countries for the period 1981-1997. Two cross-section regressions are estimated, one for the Cold War period, one for the post-Cold War, and a panel data model for the whole period. The results of this analysis suggest that military expenditure is roughly proportional to GNP, depends positively on war and the hostile neighbours' milex, and negatively on population and democracy. There is little evidence that the coefficients differ between the two cross-section regressions, but there is strong evidence of a structural break at the end of the Cold War in the panel data results. There is also evidence of regional differences in the results, which suggests that there may be different regional dynamics in the demand for military spending. To investigate further, the South American continent was chosen, and case
studies conducted for Argentina, Brazil and Chile for the period 1970-2000. The main strategic influences on milex are hypothesised to be tension between Argentina and Chile
and between Peru, Bolivia and Chile, the Falklands war, and Brazil's ambitions as a great power, while national income, debt and inflation are potential economic influences. Another concern is the differing circumstances of democratic transition in the three countries, especially the level of continuing military influence. This is also hypothesised to
affect whether the transition led to reduced military spending. To test these hypotheses, ARDL regressions are run for each country. Argentine milex depends positively on GDP, Chilean milex, and post-Falklands rearmament, and negatively on debt and increasing inflation. Brazilian and Chilean milex seem to be independent of GDP, following an upward trend, but both are negatively affected by debt and inflation. There are no significant external influences on Brazilian milex, but some evidence that tension with Argentina affected Chilean milex. Democracy had no effect in Brazil, a clear negative
effect in Chile, and a negative effect in Argentina, but only during the Menem administration. Taken overall, the results of the thesis produce a strong and consistent picture that relates the demand for military spending to overall economic resources and to the level of external and internal threat; however, in countries such as Brazil and Chile with strong continuing military influence, an 'institutionalist' understanding may also be needed, with the military seen as a rent-seeking institution. |
|---|