The Social Security Sector and Its Financing in Developing Countries
Author: Mr. Franco Reviglio1
  • 1 0000000404811396https://isni.org/isni/0000000404811396International Monetary Fund
Publication Date:
01 Jan 1967
eISBN:
9781451969139
Language:
English
Keywords:
SP; tax ratio; per capita income; Egypt; social security; government payment; sickness insurance; government employee; Securities; Tax allowances; Wages
Download PDF (2.2 MB)

THE SOCIAL SECURITY SECTOR plays an important economic and social role in many developing and developed countries. Although most developing countries are still at the beginning of the process aimed at increasing the size of the social security sector, there are indications that they are moving fast on the same road followed by developed countries. Social security programs have been introduced and expanded because of the inadequacies of older systems,1 the development of a new ethic, and the tendency of social security expenditures to grow with per capita national income. Growth of funded programs can also be explained by their role in mobilizing resources for investment.2

Abstract

THE SOCIAL SECURITY SECTOR plays an important economic and social role in many developing and developed countries. Although most developing countries are still at the beginning of the process aimed at increasing the size of the social security sector, there are indications that they are moving fast on the same road followed by developed countries. Social security programs have been introduced and expanded because of the inadequacies of older systems,1 the development of a new ethic, and the tendency of social security expenditures to grow with per capita national income. Growth of funded programs can also be explained by their role in mobilizing resources for investment.2

THE SOCIAL SECURITY SECTOR plays an important economic and social role in many developing and developed countries. Although most developing countries are still at the beginning of the process aimed at increasing the size of the social security sector, there are indications that they are moving fast on the same road followed by developed countries. Social security programs have been introduced and expanded because of the inadequacies of older systems,1 the development of a new ethic, and the tendency of social security expenditures to grow with per capita national income. Growth of funded programs can also be explained by their role in mobilizing resources for investment.2

With special reference to the developing countries,3 this paper intends (1) to measure the importance of the social security sector4 and (2) to review alternative methods of financing social security in developing countries as to their impact on income redistribution and economic growth. In this way, it will convey a better understanding of the actual and the potential role of social security as a means of savings mobilization for economic development.

I. Importance of Social Security Receipts and Expenditures

Relation of social security receipts and expenditures to gross national product

The importance of the social security sector in the national economy can be measured by the relation of social security receipts and expenditures to the gross national product (GNP). Table 1 compares total social security cash receipts and receipts net of government payments5 with GNP in 59 countries. Data on total social security cash receipts and cash receipts excluding government payments are available for 40 developing countries for a recent year.6 They show that total social security cash receipts averaged 2.5 per cent of GNP, with a median of 1.7 per cent. Exclusion of government transfers (necessary to consolidate the social security and public sectors) reduces the average ratio of the receipts for the 34 countries for which data are available from 2.65 per cent of GNP to 1.9 per cent, and the median from 1.7 per cent to 1.1 per cent. Total cash receipts exceeded 1.9 per cent in 16 countries and ranged from 0.1 per cent (Burma) to 12.6 per cent (Uruguay). In Chile and Uruguay the percentages approached the highest attained in the developed countries.

Table 1.

Social Security Cash Receipts: 1 Relation to Gross National Product, Selected Developed and Developing Countries, Recent Year

(Amount in millions of national currency)

article image
article image
Source: See Reviglio, op. cit., Appendix.

Public health services are generally excluded from social security cash receipts when programs of sickness insurance have been established. They have been included in the data for some countries: (1) Canada, Finland, Ireland, Norway, Sweden, and the United Kingdom, because only cash benefits for sickness are ordinarily provided through social insurance, whereas medical services are furnished by the Government under a program open to all residents; (2) Denmark, because public health expenditures involve a large subsidy to hold the hospital bills low for all residents; (3) Ceylon, Ghana, Malaysia, and Tanzania, because the Government provides substantial free medical care to all residents.

That is, excludes government payments of contributions as employer, for subsidies, and, when available separately, for interest on public debt held by the social security institutions.

Estimated.

Excludes government pensions for which information is not available.

GDP at factor cost.

National income.

Government payments have been estimated on the basis of previous years.

In 19 developed countries, total social security cash receipts represented higher percentages of GNP than in most of the developing countries. They ranged between 4 per cent of GNP (Japan) and 13.7 per cent (Germany); receipts net of government payments ranged between 1.3 per cent (Ireland) and 9.9 per cent (the Netherlands). Gross receipts averaged 9.1 per cent of GNP, and net receipts 5.3 per cent.

It is of interest to compare country differences in the ratio of social security cash receipts to GNP and the ratio of social security outlays to GNP (Table 2). According to the definitions used in this paper, social security generally excludes public health expenditures when a program of sickness insurance has been established. However, they have been included for some countries (Canada, Denmark, Finland, Ireland, Norway, Sweden, and the United Kingdom) where only cash benefits are provided through social insurance for sickness, while medical services are furnished by the Government under a program open to all residents. Public health expenditures have been included also for another group of countries (Ceylon, Ghana, Malaysia, and Tanzania), where the Government provides substantial free medical care to all residents. Because of this treatment, ratios of social security expenditures to GNP must be used with caution in judging the size of social security expenditures.7 Several countries have developed systems of public health expenditures which are excluded from the scope of the social security sector used in this paper.

Table 2.

Social Security Expenditures: Relation to Gross National Product, Selected Developed and Developing Countries, Recent Year

(Amount in millions of national currency)

article image
article image
Source: See Reviglio, op. cit., Appendix.

Excluding investments. For the coverage, see Table 1, fns. 1 and 4.

Excluding pensions for government employees.

GDP at factor cost.

Estimated.

National income.

Another reason for care in drawing conclusions from comparisons of social security outlays is that social security outlays include direct cash payments such as family allowances and retirement pensions, and do not take into account allowances and relief from income taxes which provide similar benefits for similar social purposes.8 Moreover, some countries have programs (e.g., in collective wage agreements) that are not included in the statistics on the social security sector. The degree of protection against some risks varies considerably. In some countries, there are private insurance, voluntary schemes, and agreements under collective bargaining covering risks that are met through centrally administered programs in other countries, whereas social security expenditures shown in Table 2 cover only public programs. For example, in the United States and the United Kingdom, contributions to private social security schemes appear to be larger than in most other countries. In the United States, these contributions amounted to about $12.3 billion in I960,9 or 2.4 per cent of GNP; in the United Kingdom, total expenditures made by enterprises for social services partly financed by employees’ contributions were estimated at about £.1.5 billion, over 5 per cent of GNP.10

The size of social security expenditures—it could be maintained—should not be taken as a measure of the value of the offered services. The expenditures include cash benefits (transfers to beneficiaries), transfers of goods (such as drugs) purchased by the private sector or produced directly by the institutions, and administrative costs (mostly salaries and wages). These components are valued at their cost to the institutions and not according to the value at market price to the beneficiaries. In social security expenditures there is no market mechanism through which the supply of social security benefits is controlled by consumers’ preferences.11 However, through democratic processes, citizens can, in principle, influence the decisions made by the institutions, on grounds of both efficiency and consistency with their choices, although they may have some difficulty in exerting their influence.12

In the developing countries studied, ratios of social security expenditures to GNP were generally lower than ratios of social security receipts, but they varied widely because of differences in the relative liberality of benefits provided in the national programs. In the United Arab Republic the ratio of receipts to GNP was 4.5 per cent, against 0.4 per cent for expenditures; in Turkey the corresponding ratios were 2.3 and 1.3 per cent, in Panama 4.4 and 3.0 per cent, in Israel 4.7 and 3.8 per cent, in Ecuador 4.6 and 3.0 per cent, in Malaysia 4.8 and 2.9 per cent, and in Chile 10.4 and 8.5 per cent. These are the countries in which the long-term risk programs provide large surpluses available for investment. The extremes were represented by Burma (0.1 per cent for both ratios) and Uruguay (11.1 per cent for receipts and 12.1 per cent for expenditures).

In the developed countries studied, ratios of social security expenditures to GNP were generally close to the ratios of social security receipts to GNP, owing to the relative insignificance of fund accumulations.

Relation of social security expenditures to public expenditures

Social security expenditures are a major item of public expenditures, particularly of social public expenditures having important economic effects on welfare and growth.13 They constitute the most important class of transfer expenditures in developed countries, as well as in a large group of developing countries, where they occupy the place that once was held by interest payments on the public debt. Social security expenditures represented over one fifth of general government current expenditures in 18 of the 19 developed countries considered and over one third in 11 of these countries (Table 3). In 22 developing countries they ranged between 0.4 per cent of government current expenditures (Nigeria) and 42.0 per cent (Uruguay), and were over 20 per cent in 5 countries, all in Latin America.

Table 3.

Social Security Expenditures: Relation to General Government Current Expenditures, Selected Countries, Recent Year

(Amount in millions of national currency)

article image
Sources: For government current expenditures, United Nations, Yearbook of National Accounts and Statistics, 1965, and International Monetary Fund, International Financial Statistics; for social security expenditures, Table 2.

Factors affecting country differences

Differences in the relative importance of social security sectors among countries reflect not only the degree of economic development but also the political decisions of the governments. This can explain why countries of similar economic development have different degrees of protection, either for the number and type of various social risks covered or for coverage of the programs.

In a number of countries the social security system does not include pension schemes or provident funds but is limited to work injuries, maternity benefits, family allowances, and special programs for public employees.14 In other countries the system is limited to work-injury programs and special systems for public employees,15 or is limited (as in Ethiopia) to public employee programs. The countries with a high ratio of social security expenditures usually cover a greater part of the population under their programs.

Only in 8 countries of a sample of 32 developing countries in a recent year did 10 per cent or more of the population contribute to social security plans; in 17 countries less than 5 per cent contributed (Table 4).

Table 4.

Social Security Coverage: Selected Developing Countries,1 Recent Year

article image
Sources: Population figures from United Nations, Statistical Yearbook, 1965. Other information mainly from Daniel S. Gerig, “Social Security in the New African Countries,” Social Security Bulletin, U.S. Department of Health, Education, and Welfare, Social Security Administration, Vol. 29, January 1966, pp. 29-41; Pan American Union, Department of Social Affairs, Estudio Social de America Latina, 1963/64 (Washington, 1964); and Reviglio, op. cit., p. 325.

For bibliography on the existing programs, see Reviglio, op. cit., pp. 362-65.

Pension fund only.

Excludes petroleum workers, oilmen, and military employees.

Main social security institution only.

Excludes government employees, covered by special systems for pensions.

But all residents are covered by the National Health Program.

But about three fourths of the population are covered by the sickness fund.

In this sample with a total population of about 886 million, only about 34 million, less than 4 per cent, were contributing to some social security program. If India is excluded, the ratio is 6.4 per cent. Since many developing countries excluded from the sample have no social security program or have a program limited to government employees, this figure overstates total coverage in the developing world. In 1960 only 2.7 per cent of the population of less developed countries was estimated to contribute to social security, against 23 per cent for the developed world. The population protected, including families, is estimated to be over twice that of the number of contributors; therefore about one half of the total population in the developed world is protected, in contrast to only 6.7 per cent in the developing world. Among developing countries, the highest protection was given in South and Central America (contributors represented about 7.3 per cent of total population and the protected population about 13.1 per cent) and the lowest in Asia (contributors were 1.6 per cent of the total population and the covered population 1.8 per cent).16 In Africa both contributors and the protected population were 4.8 per cent of the total population.

This low coverage may be objectionable on equity grounds, but it can be explained on economic, administrative, and financial grounds. Coverage is restricted by the prevalence of a large agricultural and small monetary sector, dispersed settlements, internal migrations, difficulties of communications, inadequate medical personnel and facilities outside the large cities, and limited financial resources.

II. Social Security Financing and Tax Structure

Basic methods of financing social security programs

We shall now consider the methods of financing social security programs in developing countries and the conditions for the choice of the method.17 Conceptually, there are two basic ways of financing a social security program: (1) payroll taxes, aimed to support provident funds and insurance programs, and (2) transfers from government budgets according to the social approach.

Provident fund method

Under the provident fund method, employees pay premiums, which, with additional contributions by their employers, accumulate as savings for future contingencies, such as retirement. There is no pooling of risks: everyone receives what he accumulates. National provident funds are in force in Malaysia (1951), India (1952), Singapore (1955), Ceylon (1958), Nigeria (1961), Tanzania (1964), and Ghana (1965).

Insurance method

Under the pure insurance method, participants pay premiums on their salaries and wages to cover their own risks; there is no redistribution of wealth in the participating group. However, often social insurance programs involve standardization of contributions and risks so that everybody pays the same contribution.18 The premium paid by each participant is not rigidly bound to the amount of the risk, as in private insurance, so that the social insurance program averages the risks of all individuals. On the other hand, many insurance programs provide for redistribution of wealth in the participant group as in benefiting married persons more than bachelors and low-paid workers more than higher-paid workers. Some redistribution results when contributions are made by temporary workers or others who do not qualify for the benefits. Other redistribution results from the payment of benefits larger than the premiums paid, for example, benefits to persons who join the scheme late in life.

Social method

The social method aims at providing benefits based on need and financed by those who are best able to pay. This system considers the financing of social security programs as a general government function supported by general taxation, with benefits which favor the needy. Benefits, therefore, are not bound to the payment of premiums. General programs financed through progressive tax systems may have merits, not only on grounds of equity but also, when fully financed from the general budget, on grounds of lower cost in eliminating the cost of assessment and collection of social security taxes. Costs of assessment and collection of social security taxes appear high in some countries.

Not all social security programs can be properly organized according to the social method. For some programs (e.g., work-injury and unemployment protection), it seems more appropriate to use the insurance basis because of the high degree of risk dispersion. In fact, if these programs were financed by general taxation, the industries with a high risk coefficient would be favored against other industries, a distortion that would not be approvable on grounds of efficiency.19

The desirability of financing a family allowance program through general taxation can also be debated. General programs of family allowances financed out of consolidated revenues are in force in the United Kingdom, Canada, and the Scandinavian countries, where programs cover the entire population. Contributory systems are in effect in most of the other European countries, in 20 French-speaking African countries, and in some Asian and Latin American countries, such as Cambodia, Iran, Israel, Lebanon, Argentina, Bolivia, Brazil, Chile, Colombia, and Uruguay.

A general family allowance program is aimed at two main goals: (1) to encourage the full development of human resources and (2) to redistribute income from high-income families to lower-income families. It does not seem reasonable, therefore, to put the burden of financing directly on employers and workers to achieve these aims. On the other hand, payroll taxes may be suitable when the program is limited to dependents of employed persons; in this case, the income redistribution is effected only within the group of those paying the premiums, and it may be unfair to ask other groups to finance the program. Unless the program of family allowances is general, the pension and sickness programs remain the principal fields of application of the social approach.20

Apart from family allowances with general coverage, an evolution in favor of the social method can be clearly seen in some of the more developed countries for sickness programs, especially for medical and hospital assistance. The concept of social insurance, established for the benefit of a particular group of the population, with medical and hospital services proportional to the payments made by the contributor, sometimes integrated with government transfers, has yielded to the concept of protecting most or all of the population with benefits that are not bound to the payments made.

Some Nordic countries and the United Kingdom have established national health programs open to all residents and supported by the general budget. In these countries the tax systems do not present serious loopholes, administrative inefficiencies, and differences of treatment, so that the change from contributory to general budget financing has been made without psychological and administrative difficulties.21 The same evolution has not been followed for pension programs, at least for those which provide benefits beyond minimum subsistence. In fact, the Scandinavian countries have established, in addition to general pension programs supported by the general budget, supplementary programs financed by contributions. It is still commonly held in these countries that contributory financing should not be altogether abandoned for pensions which provide more than minimum subsistence.

Four arguments are advanced for maintaining the insurance principle for pensions beyond minimum subsistence: (1) The total elimination of old-age risk may reduce willingness to work and to assume risk, weakening also an important motive to save. (2) Pensions can be restricted to those who contribute, because unlike other types of government benefits, such as public health, defense, roads, and education, pension programs do not have significant external economies.22(3) Covered individuals receive benefits in relation to their payments, thereby reducing demands on the state to raise benefits arbitrarily.23 (4) Payment of the pensions does not place an unreasonable burden on the general budget in any one year because, with the building up of reserves, the costs are spread more evenly over time.24

Developing countries are generally unable to support a social security program with a general coverage, which is considered a prerequisite to financing through general taxation. In fact, except for sickness protection, no developing country has established a social security program that is wholly financed by the government budget. Sickness protection entirely financed by the government budget is provided in Ceylon, Malaysia, Ghana, and Tanzania, mainly for historical reasons—the British tradition—but it appears quite limited, one reason being the heavy burden that would be placed on the public budget. All other developing countries that have established social security programs for one or more contingencies have adopted the provident fund or insurance method, although many have provided for government transfers to subsidize the program. The coverage is in fact restricted to a part of the population. In the beginning, it extended only to employees in the government sector and in the largest industrial enterprises; later, more people were incorporated, often with longer protection and for more contingencies, in the industrial and commercial sectors. But developing countries generally exclude, in addition to independent workers, persons engaged in agriculture, particularly those who are not wage earners. Under these conditions, contributory financing seems generally suitable. In fact, in most developing countries, because of the regressivity of the existing tax systems, financing through the budget would not make for a more equitable distribution of the tax burden than a program based on payroll taxes. It would be particularly unfair to support with general taxation a part of the population which already enjoys better economic conditions in relation to the unprotected portion (generally those not earning wages). Moreover, payroll taxes can be justified if the proceeds are used for public investments in infrastructures or industrial investments to increase future income and consumption.25

Contributory financing in developing countries

Both provident funds and social insurance programs are generally financed by premiums paid jointly by employers and employees at a fixed percentage of the payroll. Frequently, for purposes of convenience, the tax is not applied to the actual earnings of every participant but is based on several wage classes, or the rate of the premium is calculated within the limits of certain ceilings.

In 46 developing countries, rates of premiums for employers and employees (1964) ranged from 4.0 per cent of payrolls in the Republic of China and Burma to 42.5–48.8 per cent in Chile (Table 5). They averaged 15.1 per cent of payrolls, with a median of 11.8 per cent. This level is considerably lower than the premium rates in several developed countries (e.g., Germany, the Netherlands, Belgium, France, and Italy) where the social security sector is still organized on the insurance basis (Table 6). In 4 countries (United Arab Republic, Bolivia, Uruguay, and Chile), social security taxes exceeded 30 per cent of payrolls, a very high level reached by few European countries (e.g., the Netherlands, Belgium, France, and Italy).

Table 5.

Social Security Tax Rates on Employees and Employers, Selected Developing Countries, 1964

(In per cent)

article image
article image
Sources: U.S. Department of Health, Education, and Welfare, Social Security Administration, Social Security Programs Throughout the World, 1964 (Washington, 1964) and Social Security Bulletin, various issues; Gerig, op. cit.; and Jørgen R. Lotz, “Taxation in the United Arab Republic,” Staff Papers, Vol. XIII (1966), pp. 121-53.

Excludes premiums paid by the Government as employer to finance the special pension system for public employees.

This percentage is generally on a maximum basis.

Plus whole cost of work-injury program.

5.2 per cent of self-employed, plus about 1 per cent of sickness fund.

Plus about one half of sickness fund.

Plus 80 per cent of sickness fund.

Pensions for government employees are financed mostly by the employees; pensions for government workmen by the employees (40 per cent) and by the Government (60 per cent).

The Government finances a pension fund for government employees with a contribution of 14 per cent of wages and salaries. This rate has not been included in this table.

12.0 per cent for government employees.

21.0 per cent for government employees.

9.0 per cent for salaried employees.

Table 6.

Social Security Tax Rates on Employees and Employers, Selected Developed Countries, 1964

(In per cent)

article image
Sources: See Table 5.

Plus contributions of about one half the cost of sickness and maternity benefits and one third of unemployment insurance.

Plus whole cost of work-injury insurance.

Maximum income for old age, SKr 30,000 a year.

Maximum earnings for contribution, f. 10,400 a year per person under work-injury and family-allowance programs.

Maximum monthly contribution for old age, invalidity, and death DM 1,550; for sickness and maternity DM 900, and unemployment DM 750.

Maximum monthly contribution for pension BF 9,000 and for sickness and maternity, unemployment, and family allowances BF 11,825.

Maximum basis, F 970 a month.

Family allowance contributions (17.5 per cent) are related to a maximum basis (Lit 17,500 a month).

Premium rates for employers are generally more than twice as high as those for employees. In 46 developing countries the employees’ contributions averaged 4.2 per cent of wages and salaries, against 10.5 per cent of payrolls for employers’ contributions. Moreover, employers generally have to finance the whole cost of work-injury programs. In 30 of the 46 countries, maximum earnings for the calculation of contributions are established.

Regressivity and efficiency distortions of social security taxes

Before considering the implications for developing countries, of the argument of regressivity (which is generally raised against payroll taxes), we shall discuss some distortions that can be produced by contributory financing.

Under certain circumstances, high premiums can produce undesirable economic and social effects. Under inflationary conditions, when an automatic escalator clause exists for wages and salaries, determination of contributions on the basis of a fixed percentage of payrolls can have a cost-push effect.26 An inflationary impact on the demand side may result when the flow of the benefits is not offset by the deflationary impact of the payment of the contributions if, as seems reasonable, a part of the premiums is paid out of savings, whereas all benefits represent an increase in consumption expenditures. On the other hand, an increase in social security receipts owing to inflation can have a stabilizing effect when it reduces the level of consumption.

In some Latin American countries, when money has been scarce during inflation (or deflation), high premium rates, together with a low interest rate on overdue payments, have encouraged delinquency in the employers’ payments. Employers in many developing countries lag in the payment of social security premiums and use the premiums as a source of financing for their own company operations; some of them not only fail to pay their own contributions but also retain the premiums withheld on employees’ wages and salaries. Employer delinquency reaches unusual proportions in some countries, such as Argentina, Bolivia, Brazil, Paraguay, and Uruguay. In Argentina (1964), accumulated overdue premiums were estimated at M$N 45 billion (compared with M$N 92 billion for total receipts); in Paraguay (1962) the principal institution had accumulated claims on employers for

article image
46.8 million (compared with
article image
783 million for total receipts). In Uruguay, Bolivia, and Brazil the ratio of the increase in employers’ delinquency to total cash contributions in a ranged from at least 2.4 per cent to 13.0 per cent (Table 7).

Table 7.

Increase in Employers’ Delinquency in Social Security Payments: Relation to Total Cash Social Security Receipts, Three Developing Countries, Recent Year

(Amount in millions of national currency)

article image

Excludes large claims of the Industries and Trade Fund, for which information is not available.

Sometimes the government implicitly permits delinquency as a means of overcoming the financial difficulties of the enterprises, but, in doing so, it relinquishes the opportunity to direct the flow of credit in the public interest to selected industries or enterprises.

Arguments have been raised against the distribution effects of contributory financing and the way the amount of contributions is calculated. Social security contributions of employers and employees are probably regressive as to personal incomes.27 A flat-rate contributory system pays no regard to individual circumstances and excludes from the tax base important components of total income, such as rents, interest, and dividends; and the ceilings on contributions exclude wages and salaries in the higher brackets. Therefore, social security taxation, unlike personal income taxation, is contrary to equity and ability-to-pay principles.28 Moreover, social security benefits are generally tax free;29 they are worth much more than their face value to persons in high-income brackets, and this can worsen the regressivity of social security taxes.

The degree of regressivity of payroll taxes is determined by their incidence. The incidence of the tax on employees depends on a country’s stage of development and on its economic and social conditions, as well as on the extent of the coverage of the programs. In general, employees can shift their portion of the tax only if there are alternative employment opportunities in a sector not subject to the tax. In developing countries, payroll taxes are not general but are limited mostly to the government sector and the more advanced industrial sector. Because of limited alternatives of employment not subject to the tax, a general payroll tax on employees in a developed country is likely to be borne by the wage earner.

Against this conclusion it could be objected that if we assume the existence of unlimited supply of unskilled workers and of disguised unemployment in agriculture,30 we can expect that payroll taxes in the long run will be shifted to employers or to consumers because the basic wage level is determined by the average earnings in the subsistence sector plus a difference (a premium) to meet the cost of the transfer to the market sector. This premium cannot be reduced if the existing or the required increased supply of labor is to be maintained. Therefore, although in the short run the tax on wages may not be shifted because the supply of workers is inelastic, in the long run, accepting the above assumptions, it has to be shifted.

However, the realism of the assumptions of an unlimited supply of labor and disguised agricultural unemployment in some developing countries has recently stirred up a vigorous and continuing debate. It has been said that in some Latin American countries, transfer of agricultural labor has resulted in a decline in agricultural output,31 and that, in any case, there seems to be little empirical evidence of the existence of much disguised unemployment (hardly more than 5 per cent) in developing countries.32 Therefore, wages in the capitalist sector of these countries can rise far above the subsistence level; actually a wide degree of inequality in the wage structure in different industries, firms, and areas has been observed in many developing countries.33

In these conditions, it seems that the incidence of payroll taxes on workers cannot be excluded. In general, we may conclude that workers who are paid above the subsistence level plus the premium are likely to bear at least a portion of the tax. This conclusion applies particularly to government employees. The weakness of most trade unions in developing countries (compared with their relative strength in developed countries) and the infrequency of collective bargaining reinforce this conclusion. To the extent that taxes are not borne by workers, they are more likely to be shifted to profits, at least in export industries and some import-substitute industries where prices are determined by the world market.

For payroll taxes on employers also, it is difficult to assess the incidence. Generally, it depends on the extent of the coverage, the market structure, the demand for products, the supply of labor and capital, the impact of the use of social security contributions on savings-investment and consumption patterns, and monetary conditions. In developing countries the forward shifting of the tax is limited since prices in export industries and some import-substitute industries are determined by the world market, the labor supply is elastic and labor unions are generally weak, and funded programs are relatively important. The weakness of labor unions enhances the possibility of shifting the tax to the employees.34 However, this shift can occur only if salaries necessary to attract the workers from the subsistence sector are above the subsistence level plus the premium. Under certain conditions, incidence on profits appears likely because of the difficulties of raising prices in export industries and import-substitute industries and the contraction in general consumption induced by the funded programs. On the other hand, inflationary forces in the economy make it easier for employers to raise the prices of their products to cover higher social security costs, although it is often difficult to ascribe price increases to only higher contributions rather than to monetary and credit conditions.

As long as the incidence of payroll taxes is on employees, the taxes are mildly progressive in the lower ranges and regressive thereafter. The shift to consumers involves regressivity throughout most of the coverage. Effective ceilings on the contributions bring about different incidence in sectors with lower wages and salaries.35 When incidence is shifted to consumers, the balance of payments may be worsened because of the effects on the terms of trade for imports and goods produced for export.

When the incidence of payroll taxes falls on profits, premiums distort the choice between capital and labor, stimulate employers to prefer combinations of factors of production with a smaller quantity of labor, and encourage capital-intensive (or labor-saving) development. Such distortion can occur also when the incidence falls on consumers. When prices increase, enterprises that save labor may enjoy a sort of fiscal rent. Naturally this discrimination is effective only when the entrepreneurs’ choices are affected by the relative costs of labor and of capital and by the relative supply of labor. It is reasonable to expect that the higher the social security tax on employers, the more likely the rearrangement of the combination of factors of production in favor of untaxed goods. In a developed country, part of this efficiency distortion can be eliminated if the forward shifting of the higher labor costs also raises the prices of capital goods.36 But the similar compensatory effect would not readily occur in a developing country which must import most of its capital goods. When employers’ taxes are shifted, at least partially, to wages and salaries, the distortion can be avoided. To the extent that the distortion does occur, it appears undesirable when labor is underemployed.37

Other distortions may be produced by ceilings for contributions; they can distort the choices between unskilled and skilled workers, stimulate employers to employ more skilled labor, and push employers to resort to overtime work in order to avoid the payment of the tax. While the former distortion can be acceptable because it gives incentive to the development of human resources, the latter can reduce employment. Generally, when premiums are shifted to profits, investment and employment may decline and the development of the country can be impeded.

Against the charge of regressivity, it has been said that, although social security taxes are a much larger percentage of the income of the poor than of the rich, under the insurance principle the poor and the rich are paying for future benefits. This argument does not appear to be defensible when the coverage of the programs is general. It is subject to the same criticism made against financing public expenditure on the basis of the benefit principle. However, the argument appears to be better grounded when the coverage of the social security programs is limited mostly to the industrial workers and the civil service, excluding most of the lower income groups. As we have seen before, in most developing countries, social security financing from the general budget would not appear more equitable so long as the tax systems rely mostly on regressive customs and excise duties, and when progressive taxes are generally only a minor source of revenue? In general, the regressivity of social security taxes should not be judged alone but in relation to the available alternatives; it is likely that these do not have better features on equity grounds and may not have the advantages that payroll taxes have from the standpoint of feasibility, administrative convenience, and income elasticity.

Payroll taxation, therefore, finds justification from the fact that the group which contributes to the cost of the program receives the benefits. Financing benefits with payroll taxes limits vertical redistribution; in this way, the effects of the social security taxes and benefits on income redistribution are mostly confined to participants within the same income group.38 Individuals within the group may receive benefits that are larger or smaller than their contributions. The redistribution can be particularly important for family allowances, as has been shown for some developed countries, such as Italy, France, and Canada.39

The same argument may be used also in judging the opportunity for the government to subsidize social security programs. While a vertical redistribution of income from the higher to the lower income group has been achieved in some industrial countries through government subsidies to social security programs, these subsidies in most developing countries would benefit people who are better off relative to the large uncovered portion of the population.

In any case, when the coverage, though not general, becomes wider and includes also the poorest portions of the population, the regressivity of the contributions can be redressed, at least partially, by devising a distribution of benefits which is not bound in a rigid accounting way to the paid contributions.40

Social security financing through the general budget in developing countries

Besides payroll taxes, government transfers form an important source of financing of social security in many developing countries. In most social security plans, the government is responsible for a certain proportion of the costs, a percentage of the combined employer-employee contributions, or for any residual cost of meeting the deficits (Table 8). Many government payments are financed by earmarked taxes.

Table 8.

Government Transfers to the Social Security Sector, Selected Countries

(Excludes payments made by government as employer and interest on government debt held by social security institutions)

article image
article image
Sources: U.S. Department of Health, Education, and Welfare, Social Security Administration, Social Security Programs Throughout the World, 1964, and supplemental reports; International Labour Review; and Bulletin of the International Social Security Association.

Provision not applied.

Provision only partly applied.

Table 9 shows, for 31 developing and 19 developed countries, payments actually made by the Government to the social security sector. These include not only payments made to support some programs, but also those made as employer of civil and military personnel.41 Employer payments by a public corporation outside the budget are generally considered in this paper as nongovernment social security tax payments.

Table 9.

Government Cash Payments to Social Security Sector: Per Cent of Total Receipts, Selected Developed and Developing Countries, Recent Year

(Amount in millions of national currency)

article image
article image
Source: See Reviglio, op. cit., Appendix.

Central and local government payments as transfers cover special taxes allocated to social security, state participation, and participation of other public authorities.

Estimated.

Budgeted.

Payment made to meet the deficit of the social security institutions. Allocation of this payment as subsidy is uncertain because the Government in the same year did not pay a large amount of contributions due.

In the developed countries, government payments to the social security sector ranged between 12 per cent of total revenues of the sector (Spain) and 83.9 per cent (Ireland) and exceeded 31 per cent in 10 of the 19 countries.

In the 31 developing countries, the government contribution averaged about 30 per cent of the total social security receipts, ranging from 7.7 per cent (Mali) to 89.2 per cent (Tanzania). In 6 countries (Costa Rica, Bolivia, El Salvador, Malaysia, Ceylon, and Tanzania), government payments rise to over 40 per cent of total receipts. This rate is comparable to government participation in developed countries (e.g., the United Kingdom, Ireland, Canada, and the Nordic countries) where social security programs such as medical care have been established primarily on the basis of the social principle. Malaysia, Ceylon, and Tanzania have health programs open to all residents, wholly financed by the Government through general revenues. In the other three countries, participation of the Central Government (and of the state governments in India) as employer is particularly high in relation to the low coverage of the programs for nongovernment employees. Government transfers in Costa Rica are also high.

This paper also attempts to measure the burden of social security on the government budget by showing general government payments to the social security sector (contributions as employer and subsidy), as a per cent of government current expenditure (Table 10). In 23 developing countries, government social security payments ranged between 0.2 per cent of current expenditures (Nigeria) and 16.6 per cent in Malaysia, and were 5 per cent or more in 10 countries. Compared with developed countries, these proportions are relatively low. In developed countries, government social security payments ranged between 2.7 per cent of current expenditures in the United States and 33.7 per cent in Denmark; with the exception of the United States, these ratios are relatively higher than those in the first half of the developing countries.

Table 10.

Government Cash Payments to Social Security Sector: Per Cent of Government Current Expenditures, Selected Countries, Recent Year

article image
article image
Sources: For government current expenditures, United Nations, Yearbook of National Accounts and Statistics, 1965; for government payments to social security, Reviglio, op. cit., Appendix.

Employers’ contributions and subsidies.

Government arrears

Governments, particularly in Latin America, often fail to meet their obligations to social security institutions because of budget deficits and inflationary or deflationary conditions and sometimes because of government failure to consider subsidies to social security institutions as necessary expenditures.42 Government arrears involve transfers (subsidies) to the social security sector, payment of premiums as employer, and in a few countries (e.g., Argentina), payment of interest on the public debt held by the social security institutions. Some governments do not meet their obligations with cash payments and issue government securities.

Failure of the government to pay a substantial part of its social security obligations in cash was notable in some countries. Unpaid government obligations in a recent year represented over one fourth of the total cash receipts of the social security sector in some Latin American countries (Nicaragua, Bolivia, Brazil, and Argentina) and over 10 per cent in Costa Rica and Uruguay (Table 11).

Table 11.

Increase in Government Nonconsolidated Debt to the Social Security Sector, Selected Developing Countries, Recent Year

(Amount in millions of national currency)

article image
Source: See Reviglio, op. cit., Appendix.

Only accounts due by the Government as employer.

On equity grounds, failure of the government to meet financial obligations to the social security sector has a different meaning, depending on whether it is due to unpaid transfers as subsidy, as employer, or as interest. When social security benefits are provided only to a small part of the population (generally those who have the highest wage levels), the government subsidy actually benefits those who are relatively better off than those in the uncovered population; moreover, the subsidy is financed by tax systems which are generally regressive. Therefore, the subsidy would appear to be unfair, especially when the country’s resources are inadequate to satisfy the more essential needs of the population. The same criticism cannot be raised against payments by the government as employer and for interest, both of which are rather necessary expenditures, due by the government for services received.43

Do payroll taxes substitute for other taxes or add to them?

What are the economic, social, and political determinants of the absolute and relative size of social security taxes in the tax structure of various countries? This question can be posed for every tax. The answer is important for defining the tax policy of a country in a certain stage of development, when the objective is to maximize tax revenues consistent with given economic, social, and political constraints. Knowledge of the factors which influence the existence and the importance of the different taxes (including the existing tax structure) allows the policymaker to exploit the tax system in the most feasible way.

Among these factors, we have considered the number and type of social risks covered, the coverage of the programs (generally limited because of the large agricultural sector and small monetary sector), and the lack of adequate medical and administrative facilities outside the large cities. Another determinant is the size of financial resources which can be channeled to social security. These resources depend, above all, on the amount and distribution of income and on the amount and kind of other taxes collected by the government.

Social security tax revenues are generally a sizable part of total taxes in developing and developed countries. Table 12 presents for 55 countries, ranked according to per capita income (in U.S. dollars), the importance of social security taxes relative to (1) central government tax revenue excluding social security taxes and (2) total tax revenue for those countries for which the data are available. Such taxes ranged from as low as 0.5–0.6 per cent of all other central government tax revenue (Tanzania and Ghana) to 73.0 per cent (Germany) and to 92.8 per cent (Uruguay). They represented as little as 0.5–0.6 per cent of total government tax revenues (Tanzania and Iraq), and as much as 35.8 per cent (France) and 44.6 per cent (Uruguay).

Table 12.

Social Security Taxes: Relation to Total Tax Revenues, Selected Developed and Developing Countries, Grouped According to Per Capita Income, Recent Year

article image
article image
Sources: Mainly from Reviglio, op. cit., Appendix, and Organization for Economic Cooperation and Development, National Accounts Statistics, 1955/1964, 1966.

Excludes social security.

Excludes municipal taxes which are not available.

Eleven agencies only (about 80 per cent of the total).

Includes some returns on investment