The Role of Public Enterprises: An International Statistical Comparison
  • 1 0000000404811396 Monetary Fund
  • | 2 0000000404811396 Monetary Fund
  • | 3 0000000404811396 Monetary Fund


There is a growing awareness that public enterprises can be a major source of macroeconomic problems. One consequence is that programs of the International Monetary Fund often contain undertakings relating to the public enterprise sector or to particular units within that sector.1 There is, however, a shortage of analytical work on the macroeconomic role and impact of public enterprises. A principal reason for this is the scarcity and inadequacy of data on public enterprises.

There is a growing awareness that public enterprises can be a major source of macroeconomic problems. One consequence is that programs of the International Monetary Fund often contain undertakings relating to the public enterprise sector or to particular units within that sector.1 There is, however, a shortage of analytical work on the macroeconomic role and impact of public enterprises. A principal reason for this is the scarcity and inadequacy of data on public enterprises.

As a preliminary to such work, this paper provides statistical information on important macroeconomic dimensions of public enterprise operations. Data are presented for almost 90 countries, including both industrial and developing countries and countries with a range of political philosophies regarding public ownership. The major exclusions are many Middle Eastern countries and most centrally planned economies in Europe and Asia.

In Section I, some of the problems that arise in compiling consistent data for public enterprises are discussed. In Section II, data on important characteristics of public enterprise sectors are presented. Finally, in Section III, statistics are provided on major indicators of the macroeconomic impact and performance of public enterprises—the level and composition of their overall financial balances and the financing they obtain from, in particular, the government and the banking system. A summary is provided in Section IV.


Public enterprises have two defining characteristics: they are government owned and controlled; and they are engaged in business activities. Ideally, when making statistical comparisons, these characteristics should be given a precise and consistent meaning, but, in practice, this is difficult to do.

This is partly because it is difficult to draw a clear-cut and economically useful boundary around the public enterprise sector, especially as, in most countries, virtually all enterprises are subject to some degree of government control.2 However, more important, it is because of data problems. None of the international organizations publishes comprehensive information on public enterprises and statistical sources for individual countries generally do not separately identify public enterprise operations. Moreover, there are differences between countries—and sometimes even between different sources for the same country—in the exact definition of the public enterprise sector used for statistical purposes.

In preparing the data for this paper, an attempt was made to adjust the statistics to a common definition that appeared suitable for statistical analysis. This met with only partial success. However, as much of the basic information has not previously been brought together, it was decided to err on the side of comprehensiveness rather than comparability in choosing which statistics to include.

The problems in obtaining consistent statistics are discussed in the remainder of this section. Details of the coverage and sources of the data are given in the Appendix.

As far as possible, ownership/control was taken to mean that the government is able to control management decisions by virtue of its ownership stake alone. This will be true if the enterprise is directly operated by a government department or, in the case of a separate enterprise, it the government holds a majority of its shares, either directly, or indirectly through another public enterprise, A smaller shareholding may be sufficient to give the government effective control, but this will depend on the distribution of the ownership of the remaining shares.

For many countries, little information is provided on the criteria used in classifying enterprises under the public enterprise sector. However, where an explicit definition is given, this is most often based on majority ownership as, for example, for Finland and Botswana. For a few countries, for example, Tunisia, some enterprises with minority government shareholdings are also included in public enterprise statistics. It is possible that these holdings are sufficient to enable the government to exercise effective control, but this cannot easily be determined.

More of a problem is that, for a large number of countries, a narrower definition than majority ownership is used for statistical purposes. There are three main reasons for this. First, for several countries, public enterprise statistics cover only enterprises with particular legal forms.3 Most countries include public corporations in their statistics. However, some, for example, Malawi, exclude departmental undertakings, and some, for example, Sudan, exclude publicly owned companies. Second, for many countries, for example, Turkey, public enterprise statistics exclude enterprises owned by regional and local governments. Third, for several countries, statistics are only readily available for a number of the most important enterprises. These may be, for example, those subject to budgetary control, as in Mexico, or the largest, as in Mali, or those in receipt of government transfers.

The impact of these omissions varies widely. Excluded departmental undertakings are generally not significant in quantitative terms, although they include postal services, ports, and/or airports in a few countries and limited industrial or agricultural activities, such as printing or forestry, in several more. Also, in a number of cases, it was possible to obtain figures for these enterprises from government accounts. The omission of public enterprises owned by regional or local governments may be significant in some cases, especially in the larger developing countries of South America and Europe. However, in most countries where such enterprises are excluded from the statistics, their operations are mainly confined to limited public utility, transport, and marketing activities. These are probably small compared with those of enterprises owned by the national government. The most serious omissions are publicly owned companies, especially indirectly held ones, and cases where only major enterprises are included in the statistics. Impressionistic evidence suggests that such exclusions are significant in quantitative terms in several countries. In a few cases, for example, Mexico, it was possible to obtain an estimate of the importance of these enterprises using other statistical sources, but in most cases it was not feasible to adjust the statistics in this way.

These omissions limit the comparability of the data in the tables. However, the particular definitions used in individual countries often correspond to ways of classifying public enterprises that are useful for analyzing their role.

First, an important distinction between public enterprises arises from the extent and nature of the control actually exercised by the government. It is this that mainly determines the economic impact of public ownership and, particularly between countries but also often within countries, there are substantial differences between public enterprises in this respect. Actual control is extremely difficult to measure and is not a suitable basis for a statistical definition for use in international comparisons. Nevertheless, at least implicitly, it is often a factor in classifying enterprises as public enterprises for statistical purposes. This is likely to be true when the enterprises included in statistics are only those subject to budgetary control. It may also be true when the legal form is a factor in defining the public enterprise sector; for example, publicly owned companies, especially indirectly held ones, are often subject to less control than other forms of public enterprise.

Second, in several developing countries, the number of public enterprises is large, sometimes exceeding 100. It is difficult, if not impossible, for the government to exercise effective control over such a large number of enterprises, especially in countries with weak administrative capacities. The controllable public enterprise sector may, therefore, be better represented by a few of the larger enterprises.

Third, the distinction between enterprises owned by different levels of government is also important. Because of the dispersion of control and the large number of enterprises, it will be more difficult to use the public enterprise sector as a macroeconomic policy tool when the local or regional element is substantial.

As far as possible, the business characteristic of public enterprises was taken to mean that the organization's output is sold and is of a type for which, in most countries, revenue is expected to cover a substantial proportion of costs. This condition is intended to restrict coverage to activities for which a market test is a major consideration in judging performance and to exclude public goods and merit goods. Attention was also confined to nonfinancial enterprises because financial enterprises have a special role in macroeconomic policy that requires separate analysis.

Most countries use some kind of business criterion in defining the public enterprise sector, although this is generally implicit and follows from only including organizations with accounts of a commercial nature. A few restrict the coverage of their statistics to only certain types of economic activity; for example, France only includes monopoly enterprises in basic industries. However, most differences in coverage between countries reflect institutional differences in the extent to which certain activities are organized in enterprise form or operated on a commercial basis. In particular, a number of countries include activities with public or merit good characteristics in their public enterprise statistics; for example, public hospitals in Belgium, toll roads in Austria, and broadcasting services in many countries. On the whole, these differences are probably not of major quantitative importance. Also, in a few cases, disaggregated data were available to enable such activities to be excluded from the figures.

Nonfinancial public enterprises are identified separately in the statistics of most countries. However, in a few cases, for example, the United Kingdom, the only figures available include financial as well as nonfinancial enterprises. As the quantitative importance of financial public enterprises for many of the variables used in this paper, most notably capital formation, is likely to be small, this is probably not a serious problem. There are also some hybrid cases of enterprises engaged in both business and financial activities. Public agricultural and industrial development corporations, which are common in developing countries, are a prime example. Individual country practice was followed in deciding whether to classify these as public enterprises.

The differences in coverage noted above should be borne in mind when considering the figures.4 Overall, it seems unlikely that they seriously distort the general picture given by the tables, although they probably mean that the size and impact of the public enterprise sector is understated, at least on average. These differences are most relevant when comparisons are made between individual countries. They are of less significance when considering averages for broad areas or trends over time.


Size of the Public Enterprise Sector

The quantitative importance of public enterprises is clear from Table 1, which shows two measures of the size of the public enterprise sector: the percentage shares of public enterprises in gross domestic product (GDP) and in gross Fixed capital formation. In the mid-1970s, public enterprises accounted for an average5 of 13½ percent of capital formation for the countries for which investment figures are available; excluding the United States, whose share is untypically low, the average share exceeded 16½ percent. They accounted for an average of 9½ percent of GDP for the smaller group of countries, again excluding the United States, for which output statistics are available.

Table 1.

Output and Investment Shares of Public Enterprises

article image
article image
article image
article image
article image
article image
article image
Source: Appendix.

Share in GDP at market prices where indicated. (For most countries for which both measures are available, the share at market prices is lower than that at factor cost because of the importance of subsidies in the factor incomes received by public enterprises.)

Share in gross domestic capital formation, including change in stocks, where indicated.

Weighted average for 1974–77 or closest available period using GDP at market prices and gross fixed capital formation expressed in U.S. dollars for 1974–77 as weights. Figures in parentheses are averages for countries for which both GDP and capital formation figures are available. Differences between countries in the measures of GDP and capital formation that are available were ignored in calculating averages.

Figures in parentheses exclude public enterprises at the regional or local level.

Figures in parentheses are for eight large enterprises only.

Figures in parentheses exclude iron ore and petroleum enterprises nationalized in 1975.

Major enterprises only.

Excluding industrial enterprises.

Figures include financial enterprises except for those in parentheses which exclude such enterprises. Figures for investment in 1976 relate to gross fixed capital formation; figures for other years relate to total gross capital formation.

Figures relate to state economic enterprises; figure in parentheses includes an estimate of investment by other public enterprises. See Appendix for further discussion.

Figures in parentheses include 22 major public enterprises only.

Refers to Taiwan Province of China.

It is also clear that these high averages are not simply a reflection of large public enterprise sectors in a few countries: public enterprises are now of major quantitative importance in most of the countries included in Table 1. There are also broad similarities in their shares, especially in output, in countries that are diverse in terms of geographical location and economic development. African countries generally have noticeably larger public enterprise sectors than other areas. However, average output and investment shares are very similar for developing countries in Asia, South America, and Europe. Although, in the mid-1970s, the investment share for all developing countries of 27 percent was much higher than that of 11 percent for industrial countries, the output share of both groups was close to 9 percent. The majority of countries in both groups had investment shares of between 10 percent and 25 percent and output shares of between 5 percent and 15 percent.

These similarities probably result in part from differences in the coverage of public enterprise statistics. However, more important is the fact that they reflect the variety of factors that have contributed to the development of substantial public enterprise sectors in a large number of countries.6

First, socialist policies have been of some significance in several countries and the dominant factor in a few, for example, Burma and Tanzania. Also of importance is that, in this group of countries, government intervention in public enterprise operations is often particularly great, especially in pursuit of noncommercial objectives.

Second, historical and other political factors have resulted in substantial extensions in public ownership. Many public enterprises were established as a result of the political upheavals of World War II as, for example, in France and Korea. Others were set up following the financial collapse of large private companies, as, for example, in Argentina and the United Kingdom, or were established to reduce foreign ownership of industry, especially of former colonial powers, as, for example, in Egypt and Indonesia. In principle, these enterprises could eventually have been sold to the private sector. However, in practice, there have been few instances of public enterprises being returned to private ownership, at least not on a large scale.7 This partly reflects political factors, notably a reluctance to give up the additional policy instrument provided by public ownership, and partly the large size and persistent financial problems of many public enterprises, making it difficult to find private sector buyers.

Third, the pursuit of economic objectives has been a major reason for the establishment of public enterprises.8 Probably the most fundamental, and certainly the most widespread, objective in this respect has been the promotion of allocative efficiency. As is shown below, certain industries where economies of scale or externalities are important are almost universally operated by public enterprises because of the failure of private markets to satisfy the conditions for an optimum allocation of resources. Another notable factor in developing countries, especially with respect to enterprises in the manufacturing sector, has been the encouragement of economic growth. Public enterprises have been established to boost investment and saving because of the perceived failure of the private sector to achieve socially optimal levels as a result of such factors as excessive risk aversion, inadequate capital markets, or high rates of time preference. Improving the distribution of income and economic stabilization, especially of prices or employment, have also played a role, although, generally, these goals have been tied to other objectives and have not been the prime motive for the establishment of public enterprises.

Finally, the similarity in shares can partly be explained by structural factors. As is shown in Section II under “Sectoral Distribution of Public Enterprises,” there are substantial differences between sectors in the extent of penetration by public enterprises that are common to most countries. Also, countries with a relatively large public enterprise presence in most sectors often have relatively large shares of their total output or investment in the sectors where public enterprise penetration is generally low. These two factors operate in opposite directions and may therefore offset one another. This effect is especially important in explaining the similarity of shares between industrial and developing countries. In developing countries, public enterprises are generally more important in individual sectors, especially manufacturing, than they are in industrial countries. However, developing countries also have relatively large shares of output in agriculture, where penetration by public enterprises is low in all countries, and correspondingly small shares in manufacturing.

A further notable feature is the change over time in the shares of output and investment. In industrial countries there appears to have been no dramatic change in recent years in the relative size of the public enterprise sector, but in developing countries there has been a large increase. For industrial countries, the shares of public enterprises in output and capital formation actually fell between the early 1960s and the end of the 1970s, although by only half a percentage point in each case.9 10 For these countries, the major increases in the size of the sector occurred in earlier periods than those covered by available statistics. For example, there were substantial increases in the 1930s or 1940s in France, Italy, and the United Kingdom.

For developing countries, the share of public enterprises in GDP rose by 4½ percentage points between the late 1960s and the end of the 1970s and the share in investment increased by 10½ percentage points. While this partly reflected spectacular growth in a few countries, such as Venezuela, shares rose in most of the developing countries for which data are available. Moreover, there also appear to have been very large increases in the size of the public enterprise sector in many developing countries, especially African ones such as Zaïre, for which adequate runs of data could not be obtained.

Sectoral Distribution of Public Enterprises

Although public enterprises have traditionally been viewed as operating only in a limited number of natural monopoly industries, they can now be found operating in virtually all types of economic activity. This is evident from Tables 2 and 3. Table 2 shows the percentage shares of public enterprises in output or investment in major sectors; Table 3 provides qualitative information on the industries in which public enterprises are most frequently found.

Table 2.

Percentage Shares of Public Enterprises in the Output or Investment of Major Sectors

article image
article image
article image
article image
Source: Appendix.

GDPF = gross domestic product at factor cost; GFCF = gross domestic fixed capital formation; GDPM = gross domestic product at market prices; GO = gross value of output.

Share in electricity, gas, and water, plus transport, storage, and communications.

Share in electricity, gas, and water, plus mining and quarrying.

Share in agriculture only.

Share in GDPM at 1970 prices.

See note to Table 1, page 122.

Table 3.

Industry Coverage of Public Enterprises1

article image
article image
article image
article image
article image
article image
article image
article image
article image
article image
article image
Sources: International Monetary Fund, Government Finance Statistics Yearbook (Washington, 1982); Appendix.

Industries in which at least one public enterprise is operating. Coverage of sector-wide and subsidiary enterprises and of enterprises owned by local governments is incomplete.

A = agricultural production; FO = forestry; FI = fishing; AD = agricultural development (enterprises mainly engaged in the promotion of private sector production, for example, through the provision of inputs, rather than directly in production).

C = coal; P = petroleum and natural gas; M = other and general mining and quarrying.

S = sugar; F = other food and agricultural processing; D = drink; T = tobacco.

T = textiles; C = clothing.

L = leather; F = footwear.

W = wood products; F = furniture.

PA = paper; PR = printing.

P = petroleum refining and/or distribution.

PH = pharmaceuticals; FE = fertilizer; C = other and general chemicals.

B = bricks; C = cement; P = pottery and glass.

I = iron and steel; B = other basic metals.

M = motor vehicles; A = aircraft; S = shipbuilding.

E = electricity, gas, and/or water.

A = agricultural marketing; T = other wholesale or retail trade.

R = railways; B = bus and other land transport.

A = airlines; AP = airports; W = water transport; P = ports.

P = posts and/or telecommunications.