The independent resource on global security

Frequently asked questions

Find answers to your frequently asked questions below.

  1. What is the SIPRI definition of military expenditure?
  2. Is military expenditure the same as spending on arms?
  3. What is the difference between the SIPRI data on military expenditure, arms transfers, and arms production?
  4. Does SIPRI have military expenditure data before 1988?
  5. Where does SIPRI military expenditure data come from?
  6. How reliable is SIPRI military expenditure data?
  7. How well does military spending measure military capability?
  8. What proportion of military spending goes on arms?
  9. How does world military expenditure compare with government expenditure on health and education, or with overseas aid to developing countries?
  10. Do you have disaggregated military expenditure data (e.g. by categories, such as personnel, equipment, operations & maintenance, or by armed service)?
  11. What is meant by military expenditure in current and constant dollars, and how are they calculated?
  12. What exchange rates do you use to convert military expenditure to US dollars? Do you provide figures based on Purchasing Power Parity (PPP) exchange rates?
  13. Do you produce forecasts of military expenditure?
  14. Where can I find data on military aid?
  15. Where can I find data on countries' military manpower and forces?
  16. Can I use your data in my blog/paper?


1. What is the SIPRI definition of military expenditure?

The SIPRI definition of military expenditure aims to include all spending on current military forces and activities. Specifically, the SIPRI definition of military expenditure includes current and capital spending on:

  • the armed forces, including peace keeping forces;
  • defence ministries and other government agencies engaged in defence projects;
  • paramilitary forces when judged to be trained, equipped and available for military operations; and
  • military space activities.

Such expenditures should include:

  • personnel, including:
    • all expenditures on current personnel, both military and civil,
    • retirement pensions of military personnel; and
    • social services for personnel and their families;
  • operations and maintenance;
  • procurement;
  • military research and development;
  • military construction; and
  • military aid (in the military expenditures of the donor country).

The following military-related expenditures are excluded:

  • civil defence;
  • current expenditure for previous military activities, including:
    • veterans' benefits;
    • demobilization;
    • conversion of arms production facilities; and
    • destruction of weapons.

In practice, it is not always possible to obtain data corresponding precisely to the SIPRI definition. In some cases it may not be possible to determine what is or is not included in a given data source. SIPRI's priority is first to obtain data series that are consistent over time for each country, and then to obtain data corresponding as closely as possible to the SIPRI definition. Where data sources for a particular country are known to diverge from the SIPRI definition in particular respects (e.g. excluding military pensions), this information is provided in a footnote to the data.


2. Is military expenditure the same as spending on arms?

No. Military expenditure means spending on the military in general, including spending on personnel (i.e. the salaries and benefits of troops and civilian staff), operations and maintenance (i.e. spending on general supplies, services and transport), equipment (e.g. arms, other military equipment and non-military equipment), construction (e.g. of military bases) and research and development. In general, spending on weapons, weapons systems and platforms, and other specifically military equipment (including the research and development for such equipment) forms no more than a third of military spending, and much less in non-arms producing countries. For more information see the SIPRI definition of military spending given above.


3. What is the difference between SIPRI data on military expenditure, arms transfers and arms production?

Military spending refers to all expenditure on countries' military forces. The largest part of this is usually the salaries and benefits of soldiers and civilian staff, with actual spending on arms making up a small part of the total. It is a financial measure, measuring inputs, and does not necessarily measure military capability.

Arms transfers refer to arms that are traded or donated as aid across international borders. Therefore, arms transfers data do not include arms that are procured for the armed forces of the country in which they are produced, which in fact account for the majority of spending on military equipment. The main SIPRI measure of arms transfers, the Trend Indicator Value (TIV), is not a financial measure, but a measure of the volume of arms transferred between each pair of countries. For more information, see the SIPRI Arms Transfers Database. (There is also some SIPRI data on the financial value of the international arms trade but this only provides total export figures for certain major exporter countries.)

SIPRI's Arms Industry project measures sales by companies of military equipment and services. The data (chiefly the annual table of the Top 100 Arms Producing Companies) is at the level of corporations, rather than countries. The figures for the 'arms sales' of each company refer to the revenue received for sales of military equipment and services both to the company's own government (very often the majority, especially for US companies) and overseas sales.


4. Does SIPRI have military expenditure data before 1988?

From 21 November 2016, SIPRI has made available an extended military expenditure dataset, with data going back in some cases to 1949, and to at least 1957 for a majority of countries that were independent at the time. This data can be found in an excel file on the main SIPRI Military Expenditure Database webpage.


5. Where does SIPRI military expenditure data come from?

The sources for military expenditure data are, in order of priority: (a) primary sources, that is, official data provided by national governments, either in their official publications or in response to questionnaires; (b) secondary sources which quote primary data; and (c) other secondary sources.

The first category consists of national budget documents, defence white papers and public finance statistics as well as responses to a SIPRI questionnaire that is sent out annually to the finance and defence ministries, central banks, and national statistical offices of the countries in the SIPRI Military Expenditure Database. It also includes government responses to questionnaires about military expenditure sent out by the UN and, if made available by the countries themselves, the Organization for Security and Co-operation in Europe (OSCE).

The second category includes international statistics, such as those of the North Atlantic Treaty Organization (NATO) and the International Monetary Fund (IMF). The data for the 16 pre-1999 NATO member states has traditionally been taken from military expenditure statistics published in a number of NATO sources. The introduction by NATO of a new definition of military expenditure in 2005 has made it necessary to rely on other sources for some NATO countries for the most recent years.

The data for some developing countries is taken from the IMF's Government Finance Statistics Yearbook, which provides a defence heading for most IMF member countries, and from country reports by IMF staff. Other commonly used secondary sources are data from the Asia Development Bank, the UN Statistical Yearbook (UNSY), the Europa Yearbook, the German Central Statistical Office's 'Statistisches Jahrbuch für das Ausland', and Country Reports of the Economist Intelligence Unit. UNSY, the Europa Yearbook and the 'Statistisches Jahrbuch für das Ausland' are only used for data up to the early 1990s, as later editions did not include military spending data.

The third category of sources consists of specialist journals and newspapers.


6. How reliable is SIPRI military expenditure data?

The SIPRI military expenditure project makes every effort to gather as many good sources of data as possible for each country, carefully assess their reliability and consistency, and produces series that are consistent over time for each country and, as far as possible, provides figures that are as close as possible to the SIPRI definition of military spending.

That said, SIPRI data cannot be more reliable than the sources on which they are based. There are two key questions to be asked. First, are the sources accurately reporting what they claim to measure? Second, what are they actually measuring, and how does this compare with SIPRI or other understandings of military expenditure?

The first question relates to the honesty and accuracy of data collection and reporting. So, if the figure is given for the expenditure of the Ministry of Defence (MOD) in a given year, how close is this likely to be to what the MOD actually spent? The second question relates to coverage and definition: how closely does the MOD budget match the SIPRI definition of military expenditure? Are some important items of military spending left out? If so, can additional data for these items be obtained?

In most developed and many developing countries, the accuracy of data reporting is generally high. However, in many countries with weak state capacity or very low levels of transparency, this may not be the case. Monitoring and recording of expenditure may be weak, due to poor systems and management. Secret transfers between ministries may not be recorded. Corruption may lead to the disappearance of significant sums and, as a result, serious discrepancies in reporting.

The accuracy of SIPRI's reporting may also be reduced where we have had to make estimates of military expenditure for certain countries and years. These estimates are displayed in square brackets in data tables. Estimates are usually made to combine overlapping sources of data that do not agree with each other. In this case, one series is generally raised or lowered by a fixed percentage so as to make it consistent with the other in the year at which they are joined. In a few cases, estimates of the whole series are made, usually based on expert analyses, to obtain a series more consistent with SIPRI's definition. The extent of estimation varies: in the cases of Russia and Israel, most of the total is from actual data. For China and the United Arab Emirates (UAE), significant portions are based on estimates. In both instances, the estimation process introduces further uncertainty into the data.

The second question, as to what is being measured, is also very variable. In many countries, the spending of the Ministry/Department of Defence is a good approximation of the SIPRI definition of military spending, with two major exceptions in certain cases: military pensions (or contributions to fully-funded pensions schemes) are not always included, and paramilitary forces counted as military spending by SIPRI may come under other ministries, such as the Interior Ministry. In many such cases, SIPRI is able to find additional sources of data for such items. Where additional data is not available, this is noted in a footnote. However, for countries with a very limited breakdown of military spending, it may not be possible to determine if such items are included.

Furthermore, in many countries (both developed and developing), expenditure by the MOD may omit numerous significant items of military spending (other than pensions and paramilitaries). Such expenditures may be extra-budgetary (i.e. coming from other parts of the state budget), or off-budget (i.e. coming from outside the state budget altogether). An example of the latter could be dedicated natural resource funds for the armed forces (often used for arms purchases), or from the commercial activities of the military. In some countries, these may constitute a large proportion of total military spending.

In the case of extra-budgetary funding, it is frequently possible to find data for the extra-budgetary sources, especially in Europe and the Americas. However, it is rarely possible to find data for off-budget sources of spending, as these are by their nature not the subject of public accountability and reporting. The problem of untraceable extra- or off-budgetary spending is particularly severe in many countries in Africa and the Middle East, as well as some countries in Asia.

In the case of China, SIPRI estimates that various additional funding sources (of which extra-budgetary military research and development is the largest,) account for about 15 per cent of the total SIPRI estimate for China in 2011. Again, if extra- or off-budget sources of military spending exist but it is not possible to obtain figures or reasonable estimates for these, this information is provided in footnotes to the data.


7. How well does military spending measure military capability?

While there may be other political or economic motivations, the primary reason countries carry out military expenditure is to acquire military capability of one sort or other. However, extreme caution should be exercised in drawing a link between a country's level of military expenditure and its degree of military power or military capability, as many factors contribute to military capability. Further, other intervening factors may affect the degree to which military expenditure succeeds in buying military capability.

First, it should be recognized that military expenditure is a flow measure, that is a measure of current resources devoted to renewing, replacing and expanding military capability. It therefore does not measure the acquired stock of capabilities represented both by previous stocks of equipment, and by accumulated knowledge, experience, infrastructure, organization and doctrines (or lack thereof) within the military establishment. For example, while Russian military expenditure plummeted following the collapse of the Soviet Union in 1991, it also retained the Soviet Union's nuclear arsenal, and much of its stocks of conventional weapons. Of course, as such equipment has aged this may well have led to a commensurate drop in Russia's military capability.

Second, the value of military expenditure in creating military capability will vary enormously from country to country. An important factor will be whether the mix of military expenditure between, for example, personnel and equipment expenditure, is appropriate for the types of military task desired by the country. Many countries have sought to reduce the size of their armed forces since the end of the cold war, seeking to create a smaller but better trained and equipped force suitable for engaging in modern armed conflicts. Conversely, if most of a country's high levels of military spending is going towards maintaining an excessively large army, this spending may not translate into much meaningful military capability. Another factor will be the efficiency of military expenditure, which may be adversely affected by corruption, poor management and organization of forces, or poor planning and execution of equipment projects. A further factor is a country's technological absorption capability. Large sums of money spent on major high-tech equipment may be of little value if a country lacks the trained personnel, military organization and doctrine to effectively use the equipment.

Third, a country's military capability will depend also on factors which may be hard to measure in financial terms, including morale, military preparedness, combat experience, doctrine and organization. Even more so, a country's military power-its ability to use its military capability to achieve desired political goals-will depend on a whole range of other political, geographical and economic factors, such as the overall economic and industrial strength of a country, the size of a country in terms of borders and coastlines to be defended, the terrain in which armed forces may be expected to operate, the quality of communications between different areas of operation, the strength of their potential adversaries and alliances with other countries, a country's position within the international community, and many more.

Therefore, while one could be fairly confident that a country with annual military expenditure of $50 billion would almost certainly have considerably more military capability than one with annual expenditure of $5 billion, attempts to draw conclusions about a country's level of military capability from its level of military expenditure should be regarded with considerable skepticism.


8. What proportion of military spending goes on arms?

This is extremely variable, and the data is not available for all countries, as some countries only provide a single total figure for military spending, or a very basic breakdown that does not identify spending on military equipment.

Relevant information can be found for many countries in online budget documents. Other sources for certain countries include:

Note that these various sources frequently use varying definitions of military spending and of categories of military spending. In particular, the returns to the UN instrument are from countries' own self-reporting, and frequently diverge considerably from SIPRI's figures.

Some examples:

  • In the USA, procurement and research and development have usually accounted for about 30 per cent of total “National Defense” outlays since 2005.
  • By contrast, Burkina Faso's equipment spending has varied between 1.4 per cent and 6.8 per cent of total military spending between 2007 and 2010.
  • According to NATO data, Germany has allocated roughly 16-17 per cent of its military spending to equipment in recent years. 
  • According to Chinese defence white papers, about one third of China's official defence budget goes on equipment. Figures for this broad budget category 'mainly cover research on, experimentation with, and procurement, maintenance, transportation and storage of weaponry and equipment.' [1] This only relates to the official defence budget, and not to numerous sources of extra-budgetary military spending.
  • A little over 25 per cent of India's military spending in 2011 went on equipment procurement and research and development.


9. How does world military expenditure compare with government expenditure on health and education or with overseas aid to  developing countries?

Data on government health and education expenditure by country and region is available from the World Bank World Development Indicators (WDI). Useful comparisons to military expenditure can be found by analysing the figures for 'Public expenditure on education, total (per cent of gross domestic product, or GDP)', and 'Health expenditure, public (per cent of GDP)'. These indicators include both central and local government expenditure on health and education, but exclude private expenditure. 

In 2012, the most recent year for which data is available, public education expenditure amounted to 4.4 per cent of world GDP. Public health expenditure, for which the most recent year of data was 2014, reached 9.9 per cent of global GDP, a steady rise from 2008’s 5.7 per cent. Military expenditure according to SIPRI’s estimates in 2015 was 2.3 per cent of global GDP. Converting to US$, the 4.4 per cent spent on education in 2012 amounts to $3260 billion compared to the $1748 billion spent on the military. For public health, in 2014, this was $7726 billion compared to the $1747 billion for military expenditure.

The WDI also have data for overseas aid. Total net Official Development Assistance (ODA) and official aid received worldwide in 2014 was $161.1 billion, compared to world military spending of $1747.4 billion.

More data on health, education, overseas aid and other indicators in comparison with military expenditure can be found at


10. Do you have disaggregated military expenditure data?

A variety of sources of data in the military expenditure project archives are disaggregated to varying degrees, including budget and expenditure reporting documents, responses by governments to SIPRI military expenditure questionnaires, responses by governments to the UN Reporting Instrument for military expenditure, and NATO data on states' military spending. However, these vary considerably both in their definition of military expenditure as a whole, and in their definition of different categories of spending. Therefore, we have not as yet been able to carry out a systematic analysis of this data that would allow for a consistent reporting of disaggregated data between countries and over time.


11. What is meant by military expenditure in current and constant dollars, and how are they calculated?

SIPRI presents military expenditure data for each country in several forms: first, in current-price, local currency terms. Secondly, in current price US dollars; thirdly, in constant price US dollars for a given base year, and fourthly, as a percentage of Gross Domestic Product (GDP). These each have a different meaning and use, as explained below.

The raw military expenditure data that SIPRI collects, as well as the estimates we make for certain countries based on this data, is almost always expressed in the local currency of the country in question. It is also in nominal (or current) prices, which is just to say that no adjustment has been made for inflation. Such raw data is the basis of all other SIPRI data, but it is not on its own very useful; knowing that Nepal’s military expenditure in 1999 was 3486 million Nepalese rupees does not tell you much if you do not know what a Nepalese rupee is worth. Likewise, discovering that Brazil’s military spending went up from 272 million reais in 1993 to 7040 million reais in 1994 makes it sound like an enormous increase, but one cannot say this without knowing about inflation in Brazil from 1993 to 1994, affecting the purchasing power of the real.

The raw military expenditure data is based on governments’ budgets and expenditure for their financial year. A majority of countries have a financial year running from January to December, but for those that don’t, SIPRI first converts the data to a calendar year basis, by assuming an even spread of spending throughout each financial year, i.e. that 1/12 of the total is spent each month.

Current dollars

The easiest thing to do with the raw local currency military expenditure data for each country is to convert it into a common currency, such as US dollars. This conversion is done at the average exchange rate for each country and year. This gives SIPRI’s figures in current US dollars. While this is useful for cross-country comparisons of military spending in a given year, it is a poor measure of changes over time in a country’s military spending, as exchange rates can fluctuate considerably from year to year in a way that does not reflect the domestic purchasing power of a currency.

Constant prices and constant dollars

To assess trends in a country’s military spending (or any other sort of spending) over time, it is necessary to adjust the nominal spending figures for inflation in that country to get a consistent measure of the purchasing power of the spending in each year. Figures that are adjusted for inflation are said to be in constant prices, based on a given base year. Spending in the base year is left unchanged in the calculations, while figures for each other year are adjusted based on the change in the Consumer Price Index (CPI), an overall average of prices of a basket of consumer goods and services in the economy, between each given year and the base year.

For example, suppose that prices (measured by the CPI) in a particular country had risen by 20% between 2005 and 2014. Then to convert the 2005 figure to constant prices (in local currency) we would add 20% to the nominal figure to obtain a figure in constant 2014 prices. Conversely, if prices have gone up by 5% from 2014 to 2015, then we must divide the 2015 nominal figure by 1.05 to get the constant 2014 prices figure.

These constant local currency figures are then converted into US dollars at the average exchange rate for the base year, 2014. The resulting figures are said to be in constant (2014) US dollars, or in US dollars at constant 2014 prices and exchange rates.

This ensures that we have a meaningful comparison between this country's military expenditure in different years, taking account of inflation - we have consistency over time, and we also have some degree of international comparability. However, if we want to compare, say, country A’s 2005 military expenditure with that of country B, the comparison is distorted by the fact that we are converting at 2014 prices and exchange rates. Therefore, for cross-country comparisons it is better to use the figures expressed in current US dollars, in particular when comparing countries in years far from the base year.


12. What exchange rates do you use to convert military expenditure to US dollars? Do you provide figures based on Purchasing Power Parity (PPP) exchange rates?

SIPRI uses market exchange rates, or where applicable fixed official exchange rates, to convert local currency military expenditure data into US dollars (whether current or constant prices). Market exchange rates are determined by the supply and demand of currencies used in international transactions. However, the prices of many goods and services on domestic markets are determined in partial or complete isolation from the rest of the world. Therefore, the MERs do not always accurately reflect differences in price levels between countries. Fixed, official exchange rates may be even less reflective of the actual purchasing power of a currency within the country that uses it.

An alternative is to use purchasing power parity (PPP) conversion factors (or PPP exchange rates). The PPP dollar rate of a country's currency is defined by the World Bank as 'the number of units of a country's currency required to buy the same amount of goods and services in the domestic market as a U.S. dollar would buy in the United States'.  The only PPP rates available for all countries are GDP- based basket of goods and services that are major components of the gross domestic product. Such GDP-based PPP rates are designed to control for differences in price levels and thus to provide a measure of the real purchasing power of the GDP of each country.

Using GDP-based PPP rates instead of MERs for currency conversion results in much higher output and expenditure figures for many developing and transition countries since they have relatively low prices for non-traded goods and services—thus giving the currency higher purchasing power. A unit of local currency therefore has greater purchasing power within a developing country (which is better reflected by using PPP rates) than it has internationally (which is what is reflected by using MERs). For those such developing and transition countries for whom data was available for 2014, the median increase in military expenditure figures from using PPP rates instead of MERs was by a factor of 2.13. Three-quarters of these countries would see their relative figures increase by at least 79 per cent. However, for most of the 'developed' countries in Western Europe, Canada, Australia, New Zealand and Japan, the use of PPP rates on GDP and military expenditure figures leads to a fall in these totals relative to the USA, by a median amount of 9 per cent. For countries in Central Europe and East Asia that have more recently entered the ranks of rich nations, the picture is somewhere in between. However, the reliability of such PPP rates is lower than for MERs, since PPP rates are statistical estimates, calculated on the basis of collected price data for a basket of goods and services for benchmark years. Between benchmark years, the PPP rates are extrapolated forward using ratios of prices indexes, either GDP deflators or consumer price indexes. Like all statistical estimates they are subject to a margin of error.

Furthermore, GDP-based PPP rates are of limited relevance for the conversion of military expenditure data into US dollars. Such PPP rates are designed to reflect the purchasing power for goods and services that are representative of spending patterns in each country, that is, primarily for civilian goods and services. Military expenditure is used to purchase a number of goods and services that are not typical of national consumption patterns. For example, the price of conscripts can be assumed to be lower than the price of a typical basket of goods and services, while the prices of advanced weapon systems and of their maintenance and repair services can be assumed to be much higher. The extent to which this data reflects the amount of military goods and services that the military budget can buy is not known. Due to these uncertainties, SIPRI uses market exchange rates to convert military expenditure data into US dollars, despite their limitations.


13. Do you produce forecasts of military expenditure?

No. We do not consider attempts to forecast military spending to be reliable or useful, as they are subject to much uncertainty relating to economic, political and security developments.


14. Where can I find data on military aid?

SIPRI does not collect detailed data on military aid. However, country-by-country information on US foreign military assistance can be found at the US State Department's website. For years after financial year (FY) 2008, see the page for the Office for Military Assistance; for years up to FY2008 see the page of the U.S. Department of State. Some countries report total military aid as part of their returns to the UN reporting instrument on military expenditure.


15. Where can I find data on countries' military manpower and forces?

The annual Military Balance report released by the International Institute for Strategic Studies (IISS) provides information on the armed forces of most countries worldwide, including troop numbers and holdings of major weapons systems. The Military Balance can be ordered from IISS, or can be read online for free by IISS members.

There are a number of free online sources of information on military force numbers, although their reliability cannot always be easily assessed.


16. Can I use your data in my blog/paper?

Yes, once the user abides by the terms and conditions of use.


 [1] Information Office of the State Council, The People's Republic of China, China's National Defense in 2010, p94.