Measuring arms production

The SIPRI programme on arms production was initiated in 1989 to study developments in the arms industry. The programme collects information on the major arms-producing companies and describes and analyses their adjustment to the changed economic and political context and the resulting changes in industrial structures.

The SIPRI arms industry database, also created in 1989, contains financial, employment and other data for arms-producing companies in most significant arms-producing countries (except China). This includes information on total sales, arms sales, profits, employment, exports and (where available) research & development spending and arms exports. It also includes qualitative data on the sectors of the arms industry (e.g. aircraft, shipbuilding, missiles, services, etc.) in which the company is engaged, and information on changes in ownership and control. At present, around 700 companies are included in the SIPRI arms industry database, with up-to-date financial data available for around 250. SIPRI also collects comprehensive annual information on significant mergers and acquisitions within the arms industries in OECD countries.

Data on the arms sales, total sales, profits, employment and industry sectors of the 100 largest arms-producing companies in these countries are presented annually in the SIPRI Yearbook and online. A list of significant mergers and acquisitions within OECD arms industries is also published.

Defining arms production

There is no generally accepted international definition of arms production, or alternatively ‘defence production’, the arms/defence industry, etc. In particular, there is no Standard Industrial Classification (SIC) code or collection of SIC codes that corresponds to any reasonable definition of the arms industry. In reality, the ‘arms industry’ is not a single industry but a segment that cuts across a number of industries, such as aerospace, shipbuilding, IT, etc. – as well as the very specific industrial sector for weapons and ammunition. This lack of clear definitions and boundaries for the arms industry is a significant problem in collecting internationally-comparable data on arms production.

SIPRI defines ‘arms sales’ fairly broadly as including all military-specific goods and services sold to military customers (by private or state-owned corporations, but not by production arms of the armed forces itself, where they exist). This does not include all goods and services supplied to the military – general purpose supplies and services such as food, office supplies, oil, electricity, cooking and cleaning services, etc. are not included. It does however include items such as computer software where this is written for military-specific purposes, ‘ruggedized’ laptops adapted to operate in battlefield conditions, management of military bases, maintenance and repair of equipment, logistics services etc., that might not immediately be thought of as ‘arms’.

Estimating company arms production

In practice, data corresponding precisely to the SIPRI definition is rarely if ever available. Most arms-producing companies are also engaged in at least some civil production, and have no particular interest in reporting their revenues in accordance with a definition devised by a non-governmental organisation. Some companies do report the proportion of their revenues related to ‘defence’; while the exact definition of this may not be clear, it will usually correspond fairly closely to the SIPRI definition. Others (in the US) report the proportion of their sales to the US Department of Defense (DoD) – as this includes Foreign Military Sales, which go through the DoD, it also includes most exports. In some cases, figures for a company’s operating divisions may correspond fairly closely to the split between military and civil production. A few companies respond to specific SIPRI requests for information on their arms sales if suitable figures are not otherwise publicly available. In other cases however, SIPRI must make estimates based on such information as is available in the public domain.

The main sources of data for arms-producing companies

Sources of data for arms-producing companies varies according to both country and type of company. There are four principle types of company, as far as data availability is concerned:

Publicly-traded companies – that is private-sector companies (although they may have a state-owned share) whose shares are traded on a stock market. These must publish annual reports with comprehensive financial data. Nowadays such annual reports are almost always available online. In the US, such companies also submit detailed “10-K” annual reports to the US Securities and Exchange Commission (SEC), which are frequently valuable sources of data.
Privately-held companies – companies owned by an individual or small group of shareholders, and which are not traded on a stock exchange. This category also includes employee-owned companies and in at least one case communally-owned companies (the Chugach Alaska Corporation, whose non-tradable shares are distributed to members of the Native Alaskan Chugach tribe.) These companies do not have the same public reporting requirements, so that information on them is often harder to obtain. Most such companies usually publish some information, but other sources such as government contract information or occasionally secondary sources such as press reports may be required.
State-owned companies – the level of information provided will depend on the policy of the government in question, although information may be more easily accessible by citizens of the country in question in some cases. In many cases, such companies do provide relevant financial information in annual reports or online.
Subsidiaries of larger companies. The degree of information provided on subsidiaries is the choice of the parent company, and is as such highly variable. As the SIPRI Top 100 is based on the arms sales of parent companies, a systematic coverage of subsidiaries is not attempted. However SIPRI does attempt to gather information, where available, on larger subsidiaries of major companies, especially subsidiaries in countries other than that of the parent company, or where the arms sales of a large, mostly civil company are concentrated in one or more subsidiaries. (E.g. Pratt & Whitney and Sikorsky as subsidiaries of United Technologies, or the IVECO subsidiary of Fiat.)

In addition to the generally available sources discussed above, SIPRI has a network of experts in various countries who provide information on arms-producing companies in their countries. This network currently covers arms companies in France, Italy, Spain, Greece, Turkey, Canada and South Korea.

Data on national arms production

SIPRI has collected a limited amount of data on national arms production. The quality and availability of the data is highly variable between countries. In general, data on total arms production in a given country are surprisingly hard to find, even in major developed arms-producing countries such as the USA. This stems from the lack of a clear definition of arms production, or of any set of SIC codes covering arms production. In some countries, either the government or for example a defence industry association may publish information on the total volume of arms sales (or defence production) in that country, but in many cases there are no regularly-produced figures corresponding to any standardized or explicit definition of arms production.

Measuring the volume of arms production nationally may be attempted either from the supply side or the demand side. A supply side approach is based on company data of arms sales, where national arms production is the sum total for all companies in the national arms industry. There are numerous problems with this approach:

Identifying all (or at least all sufficiently large) companies within the national arms industry.
Separating military from civil sales for each company (as discussed above, this is not always possible).
Avoiding double-counting – which happens if the sales of components and subsystems of one company are added to the sales of the final systems in which they are included of another company.
Cross-border ownership of arms companies: in estimating arms production within a country, it would be necessary to include the arms sales of foreign-owned companies within the country, and exclude the arms sales of foreign subsidiaries of companies from the country in question. The division of a company’s sales between operations in different countries is not always available.

The demand side approach often offers more potential, as governments in developed economies generally publish reasonably detailed expenditure data, for example for equipment expenditure by the ministry/department of defence, or volume of contracts with industry. Many countries also publish annual figures on the value of arms exports. There are nonetheless often problems with this approach too:

The private arms industry provides not only equipment to armed forces, but a range of outsourced services, which would come from the operations and maintenance budget, rather than equipment. Therefore figures for expenditure or contracts with industry, rather than just equipment spending, is needed. (This is available in some cases).
Figures for government spending with industry must also exclude arms imports, or some other source of data on arms imports must be available.
Different countries employ different definitions of arms exports, so figures may not be comparable.
Most problematic is distinguishing between a defence ministry’s spending on military equipment and services, in accordance with the SIPRI definition, and general-purpose supplies and services. This is rarely possible to find precise data for. For example, the US government provides very detailed data on the contracts awarded by the DoD to private industry, but assessing how much of this is for military-specific goods and services would require a contract-by-contract analysis, an enormous task. The UK provides information on Ministry of Defence spending with UK industry, broken down by SIC codes – but while some SIC codes clearly relate to purchases of military equipment and services, and some clearly relate to civilian supplies (e.g. food), others are general services categories whose split between military and civil services cannot be determined.

The uses of arms production data

For these reasons discussed above, SIPRI data on arms production is primarily focused on data at a company level, which is more readily available than data on a national level. Given that such a database cannot possibly cover all companies engaged in arms production worldwide, or even in major producer countries, and given the problems discussed above, such data cannot be used to provide a reliable estimate of total arms production either globally or by country. In this respect it differs from SIPRI data on military expenditure which, despite the limitations of the data, does provide such estimates. Primarily, the information contained in the SIPRI arms industry database, and published in the Top 100 should be seen as industrial-level information, giving a guide to trends in the industry and the corporations engaged in it, rather than macro-level economic data. However there are many potential uses of SIPRI arms industry data:

Despite the above qualification, the total arms sales of the SIPRI Top 100 almost certainly gives a reasonable guide to the overall trend in the volume of global arms production (at least within the countries covered, excluding in particular China).
The SIPRI Top 100 may also give a rough guide to national trends in some countries, and to the relative importance of the arms industries of different countries. This must be treated with extreme caution, however; some countries’ arms industries, e.g. Canada and Australia’s, are largely foreign-owned (and thus do not fully show up in the Top 100), since this dataset covers primarily parent companies, while UK companies own very substantial overseas operations, especially in the US; the volume of arms sales by “UK” companies in the Top 100 therefore overstates the size of the UK arms industry.
The Top 100 also gives an indication of which sectors of the arms industry are growing or shrinking at a given time.
SIPRI data on the arms industry can give a picture of the changing structure of the arms industry – the degree of concentration and competition, patterns of cross-country ownership, vertical and horizontal consolidation, the movement of companies into new sectors, etc.
Information on the arms industry is useful for those analyzing the concept of a “military-industrial complex” – the interaction between government, the military, and the arms industry, and its relevance for the formation of foreign, defence and industrial policy.
Information on arms-producing companies may be important to ethical investors, in terms of which companies are highly dependant on arms sales for their revenues. NGOs concerned with arms production and trade may likewise be interested in this data.