What Determines Military Expenditures?
Author:
Mr. Daniel P. Hewitt https://isni.org/isni/0000000404811396 International Monetary Fund

Search for other papers by Mr. Daniel P. Hewitt in
Current site
Google Scholar
Close

For the latest thinking about the international financial system, monetary policy, economic development, poverty reduction, and other critical issues, subscribe to Finance & Development (F&D). This lively quarterly magazine brings you in-depth analyses of these and other subjects by the IMF’s own staff as well as by prominent international experts. Articles are written for lay readers who want to enrich their understanding of the workings of the global economy and the policies and activities of the IMF.

Abstract

For the latest thinking about the international financial system, monetary policy, economic development, poverty reduction, and other critical issues, subscribe to Finance & Development (F&D). This lively quarterly magazine brings you in-depth analyses of these and other subjects by the IMF’s own staff as well as by prominent international experts. Articles are written for lay readers who want to enrich their understanding of the workings of the global economy and the policies and activities of the IMF.

Empirical evidence on the economic and political factors that influence national military spending

In recent years, some countries have spent as much as one fourth of GDP on the military, while others less than 1 percent. Wide variations in military expenditures exist not only between different regions of the world and different economic groups but also within different categories of countries (see “Military Expenditures in the Developing World” by the author in the September 1991 issue of Finance & Development). Can this pattern of military expenditure be explained by the economic characteristics and political circumstances of nations? Are differences in military policies due primarily to country specific circumstances, or are they attributable to inherently unobservable political phenomena—such as excessively complicated political interactions between different countries?

A recent study of military expenditures in 125 countries over a 17-year span provides useful insights into the apparent determinants of world military expenditures (see box). Contrary to popular belief, the study indicates that military spending is highly reactive to observable financial constraints. The ratio of military expenditure to GDP is found to rise with GDP in low- and middle-income countries and remain constant for high income countries. Heavily indebted countries and small low-income economies are found to spend less on the military than other nations. The level of public and publicly guaranteed foreign financing is found to have a positive association with military spending.

For details of this study, see “Military Expenditure: Econometric Testing of Economic and Political Influences” by the author, available as IMF Working Paper WP/91/53.

Among the political variables investigated, as expected, countries engaged in international war or civil war are found to spend more on the military. Additionally, countries characterized as multiparty democracies spend less on the military than others—such as countries characterized as monarchies (but without multiparty democracy), military governments, and socialist governments. Finally, land area and border length are found to have a positive influence on the level of military expenditure.

The analytical framework of the study is able to explain 55 percent of the variation and, therefore, a large proportion of the variation can be accounted for by political and economic characteristics of countries. Since a significant proportion is still left unexplained, country specific historical and political circumstances are also undoubtedly quite important influences.

Empirical results

The empirical framework envisions the leadership of a nation making two types of decisions. First, the leadership decides how much to spend on the military, and thus, the first empirical equation focuses on the determinants of military expenditures as a proportion of GDP (see table). Among the explanatory variables, the economic indicators incorporated in the equation represent an indication of the financial constraints confronting governments. The political variables are indicators either of the political circumstances (such as a state of war) or the political ideology of the leadership. The geographical variables reflect the cost of defending countries.

The second decision the leadership makes is on the level of overall government spending. Clearly, the two decisions are interrelated and should reinforce each other. A country that spends more on the military is likely to have a larger public sector. Likewise, a country that devotes a larger share of its resources to government is also likely to spend more on the military. The framework, which is specifically set up to test this relationship, confirms that the two reinforce each other.

The second equation, therefore, focuses on the determinants of central government expenditure as a proportion of GDP. The purpose of this equation is to test for the indirect influences on military expenditure transmitted through changes in central government expenditure. Since the results show a positive relationship between military and central government spending, factors that induce an increase in overall government spending indirectly induce an increase in military spending. (The results from the second equation should not be interpreted, however, as offering a full explanation of the determinants of central government expenditures.) In this equation, the level of military expenditures is viewed as an influence on the demand for overall government expenditures. The other financial variables are indicators of resource constraints on the economy and the government. The political variables are viewed as proxies for the political preferences of the leadership, as in the first equation.

Economic variables. The results indicate that the ratio of military expenditures to GDP rises with GDP and per capita GDP (see table). The level of military spending rises more quickly than GDP at low levels of GDP and flattens out to a constant ratio at higher levels of GDP. Thus, low-income nations are characterized by relatively low levels of military spending, middle-income nations spend a higher proportion of their GDP on the military, and military expenditures tend to level out to a constant proportion of GDP for higher-income nations.

As expected, military expenditures are positively correlated with central government expenditures, that is, they rise and fall together. A 1 percent rise in central government expenditures is found to cause a 0.75 percent rise in military expenditures, which implies that military expenditures tend to increase at a slightly lower rate than the budget. This result, therefore, establishes the proposition that increases in central government expenditures tend to increase military expenditures and factors that influence government spending indirectly influence military spending. Military expenditures are also found to increase government expenditures.

The availability of external funding is an important element of the financing constraint on government. The variables associated with the 15 most heavily indebted middle-income nations (including ten in Latin America), one covering 1972–79 and the other covering 1980–88, illustrate this. Surprisingly, during the 1970s (prior to the world debt crisis), both central government expenditures and military expenditures in the heavily indebted countries were below the world average. During the 1980s, the level of military expenditures in these countries relative to GDP fell further, while central government expenditures rose somewhat relative to the world average. This can be interpreted as a reaction to the increased interest costs that these countries had to bear—the higher interest payments increased government expenditures and simultaneously depressed the share available for military expenditures (see “Military Expenditures in the Developing World” by the author in the September 1991 issue of Finance & Development).

Another financial proxy associated with small low-income economies shows that while their average central government expenditures were equal to the world average, their military expenditures were significantly below the world average. Apparently the low purchasing power of these nations and their emphasis on economic and social development (as indicated by the high share of such expenditures in central government budgets) led them to allocate a lower proportion of both government resources and GDP to the military. This result lends support to the notion that military expenditures are a “superior good” among developing nations, that is, middle- and high-income nations spend relatively more on the military than low-income nations.

The net flow of public and publicly guaranteed external financing, which indicates the level of foreign financing that the government contracts, also has a positive effect on government expenditures, as expected. However, it also has a positive association with military expenditures. Therefore, higher levels of government financing relative to GDP are found to induce higher military spending both indirectly, by inducing higher overall government spending, and directly, by increasing the share of the budget allocated to the military, on average. This result implies that external assistance to a nation will induce higher military spending. Further, in general, these results indicate that national levels of military expenditures do react to financial constraints and that governments take account of economic conditions and consequences, as well as political circumstances, in forming the military budget.

Political and geographic variables. While recognizing the likely importance of political circumstances in formulation of military expenditure policy, their inclusion in empirical analysis presents a challenge, since only objective and observable features of a nation can be used. Accordingly, each country has been assigned to mutually exclusive political categories (which can also vary over time). The list of categories was chosen for their observability as well as their potential relevance.

The benchmark category is a multiparty democracy, not recently engaged in either an international war or a civil war (including all nations whose political process is dominated by a multiparty system, regardless of the political ideology of the leadership or the nominal presence of a monarch). The other categories are (1) countries at war or recently engaged in an international conflict (regardless of their political leadership), (2) countries engaged in a major internal civil war (regardless of their political leadership), (3) countries ruled by monarchies, (4) countries ruled by military governments, (5) countries ruled by socialist governments, and (6) others (which consist of one-party states that do not fit into any of the other categories or highly unstable democracies).

These categories capture a number of factors that are likely to affect the priority the leadership assigns to military expenditures, while the number of options was kept to a minimum for the sake of simplicity. A state of war obviously affects the demand for military expenditures. Additionally, the historical basis for the government is likely to influence the leadership’s view of military expenditures and the priority assigned to the military. Finally, a government’s ideological view of the role of the state in an economy is a potentially important factor.

Countries engaged in international war spend the most and those engaged in civil war are second highest, as expected. Among the other variables, the ranking from highest to lowest expenditures on the military was monarchy, military government, socialist government, and others. By inference, multiparty democracies are found to spend the least on the military (see box).

The political categories are slightly different in the government expenditure equation. The two war-related options were dropped because theory suggests that the presence of war should not have a direct effect on the overall level of government spending; instead, war will increase military expenditures and thereby indirectly influence government expenditures. The results indicate that socialist governments have the highest level of central government spending relative to the GDP, then in descending order are multiparty democracies, other forms of government, monarchies, and military governments. Thus, relative to multiparty democracies, socialist governments tended to have higher central government expenditures and to allocate a higher proportion of government spending to the military. Military governments and monarchies tended to have smaller public sectors, but tended to allocate a higher share of the budget to the military; the same is true of countries in the “other’ category, but to a lesser extent. However, the direct effect of the form of government on military spending dominates the indirect effect in each case. Therefore, monarchies and military governments allocate a larger share of GDP to the military than socialist governments when both the direct and indirect effects are combined.

The geographic variables included in the estimation equation confirmed prior expectations. Land area and border lengths are found to have a positive influence on military spending, presumably because they increase the costs of defense. Additionally, as expected, coastal borders have a less powerful effect than land borders, since coastal borders provide natural protection in most cases, and are, therefore, less costly to defend.

Central government expenditure. The analysis of central government expenditures in this study is not intended to provide a comprehensive overview of the determinants of government spending. Instead, its purpose is to indicate how military expenditure interacts with central government expenditure (see equation 2 in the table).

As will be recalled, a 1 percent increase in central government spending was found to lead to a 0.75 percent increase in military spending on average. Thus, increases in government spending are shared between military and nonmilitary spending. The central government expenditure equation examines the opposite perspective—how changes in military expenditures affect the overall level of central government expenditure. The result is that a 1 percent increase in military expenditures induces a 0.18 percent increase in central government spending. Since this is remarkably close to the average ratio of military expenditures to GDP (0.17), it implies that autonomous increases in military expenditure are exactly accommodated by increasing overall budgetary expenditures.

An explanation of the empirical framework

In order to indicate how the results in this study are derived, a brief discussion of some of the technical details may be useful. The quality of empirical results always reflects the accuracy of the data. The military expenditure data used for this study are based primarily on estimates of the Stockholm International Peace Research Institute (SIPRI), which is widely believed to represent the best available comprehensive data source. The other major sources include IMF published data, World Bank, United Nations, and US government publications.

Another important characteristic of an empirical study is the framework in which it is cast. Every empirical study, either explicitly or implicitly, is based on a model. The particular model or framework that is used, by necessity, imposes a certain interpretation on the data and the results indicate whether or not the framework that the researcher used is plausible. However, the process rules out other points of view or other plausible interpretations of observed tendencies simply by ignoring them.

The analysis of government behavior in this study is based upon what economists call a “public choice” framework. In this setting, the leadership is assumed to act as if it were an economic entity, such as a business making a production decision or a household deciding how to spend its available income. The leadership is treated as if it maximizes its own welfare subject to its specific national economic, political, and geographic circumstances.

The leadership is envisioned as making two fundamental decisions in formulating the military budget: It chooses the overall level of government spending and, simultaneously, decides the proportion of the budget to allocate to the military. Such a process is represented as an allocation of national resources between private expenditures, military expenditures, and non-military government expenditures (consisting primarily of social and economic programs). Although there are numerous other decisions and subcategories of expenditures that are extremely important for the economy, this delineation is the focus of the study. This framework produces a two equation system. In the first, the determinants of military expenditures as a proportion of GDP are tested. The second equation tests the determinants of central government expenditures as a proportion of GDP.

The statistical technique is designed to isolate the effects of each variable from the other variables. For instance, the effect of central government spending on military expenditures is isolated from the effects of population or GDP variations. Therefore, the coefficients provide much more information than might be derived from a simple correlation analysis.

The coefficients associated with most of the major variables are elasticities that indicate the percentage change in the dependent variable associated with a percentage change in the explanatory variable. For the other variables, the coefficients indicate the direct influence.

How important are individual factors in military and central government expenditures?

article image
Source: Hewitt, 1991, See box on model.

Significant at the 95 percent level of confidence.

Significant at the 99 percent level of confidence.

The natural log of the variable.

The square of the natural log of GDP.

In the first equation, the dependent variable is the natural log of the ratio of military expenditures to GDP. Among the explanatory factors, the following have coefficients that represent elasticities (because in each case the natural log of the variable was used in the empirical test): the natural log of GDP (in US dollars, based upon 1980 purchasing power parity weights); the natural log of GDP squared; population; the ratio of central government expenditures to GDP; land area (in square kilometers); land borders (in kilometers); and coastline (in kilometers).

The other variables have ordinary coefficients instead of elasticities. These are the ratio to GDP of the net flow of public and publicly guaranteed external financing (for net debtor developing nations only); 15 heavily indebted nations in 1972–79; 15 heavily indebted nations in 1980–88; 31 small low-income economies; and the political variables described in the text. With the political variables, for technical reasons, there is no coefficient associated with the benchmark category, multiparty democracies. Instead, all the other coefficients are interpreted as indicating the effect of being in that particular category in comparison to being in the benchmark category. Finally, to adjust for changes in the overall level of military expenditure over time, variables for each year were included in the analysis. The results, not reported in the table, indicate that a considerable drop in the level of world military expenditure occurred in the 1980s.

In the second equation, the dependent variable is the ratio of central government expenditures to GDP. The explanatory variables are the natural log of the ratio to GDP of military expenditures, and a development index based on per capita GDP and life expectancy (see the working paper for full details). The other variables are the same as in the first equation, except that the two war-related variables have been dropped. In this case, the time variables indicate a steady increase in central government expenditures occurred over the period.

The strict interpretation of this result is that other types of expenditures are virtually unaffected by increases in military expenditures, instead private spending is crowded out. This is, of course, an average result. It is more likely that the effect of higher military expenditures on other types of government expenditures varies from country to country. In some circumstances, where the budget constraint is not tight, increases in military spending lead to higher spending on all items—when the government increases military spending, it simultaneously spends more on social programs to appease competing interest groups. In countries where the government is financially constrained, it accommodates higher military expenditures by decreasing other types of government expenditures.

Another result in equation 2 is that the development index, which is an indication of the level of economic development of a country (see box), has a positive coefficient. This confirms that central government spending relative to GDP tends to rise with the level of economic development.

Conclusion

The primary motivation behind the study described in this article was to determine if any discernible pattern could be identified in the military expenditure policies across countries. The military expenditure data, which were collected in connection with another study, offer a suitable basis for such a study since they are reasonably comprehensive, cover a long period, and represent the best available estimates.

Opinions differ on the proper level of military expenditure in any given country and on what motivates countries to pursue the widely different military expenditure policies that are observed. One prominent view is that no worldwide pattern exists because military policy is an inherently country specific phenomenon. For instance, a country’s preference for military expenditure might primarily reflect historical experience. A complementary view, which arrives at essentially the same conclusion, is that military policy is primarily driven by external forces and, therefore, is largely outside the control of a nation’s political leadership.

The empirical results of our study, while not necessarily refuting these theories, support an alternative explanation. Financial factors emerge as an important determinant of military expenditures. Thus, affordability appears to be a major consideration for government policymakers in the realm of military expenditures, as in other areas. Further, the observed pattern suggests that governments view the military as a luxury item instead of a necessity. If military expenditures were a necessity, low-income and small countries would allocate a higher share of their GDP to the military. Instead, the opposite pattern emerges. Military expenditures of middle-income nations are found to be higher both in absolute terms and in proportion to GDP, while high-income nations spend about the same relative to GDP. An implication of these findings is that financial assistance of any sort to low-income countries is likely to induce higher military expenditures, either directly or indirectly, through an easing of budgetary constraints.

The results also support the hypothesis that geographical features of a nation influence military policies, as do political factors such as being in a state of war and the form of government. One interpretation of these results is that these variables are proxies for the political situation and the ideology of the leadership. However, the reverse causality could exist. The geopolitical status of a country could influence both the type of government that takes power and the level of military expenditures.

International Monetary Fund announces

The IMF’s Statistical Systems in Context of Revision of the United Nations’

A System of National Accounts

This volume, sponsored by the IMF’s Statistics Department and edited by Vicente Galbis, contains a selection of papers presented at meetings sponsored by the IMF for the revision of the United Nations’ System of National Accounts (SNA). Experts from the IMF, UN, OECD, EC, and other statistical institutions examine issues pertaining to external sector transactions, public sector accounts, and financial flows and balances.

Available in English. ISBN 1-55775-159-5 (paper) 1991 US$35.00

To order, please write or call:

International Monetary Fund

Publication Services Box FD-104

700 19th Street, N.W.

Washington, D.C. 20431 U.S.A.

Telephone: (202) 623-7430

Telefax: (202)623-7201

Cable Address: INTERFUND

American Express, MasterCard

and VISA credit cards accepted.

Daniel P. Hewitt

  • Collapse
  • Expand