Will you still need me, will you still feed me, when I’m 64?

Will you still need me, will you still feed me, when I’m 64?

—John Lennon and Paul McCartney

I. Introduction

Social security taxes are a major revenue source and a critical element in fiscal policy.1 Most, but not all,2 states have social security taxes. Some states with mandatory contributions to the funding of social security schemes do not call those contributions “taxes.” Nor do states always pay close attention to the interaction between their social security taxes or contributions and other taxes. Yet, most states incur social security expenditure. In some, such as those with economies in transition, social security expenditure and related social welfare costs can be the largest part of the fiscal obligations of the state. It is also accepted that the burden of social security expenditure is growing and will continue to grow during the foreseeable future.3

In Europe in particular, the law of social security has evolved together with labor relations law. Although the same enthusiasm for this approach is not present in the Americas and Asia, international agreements have been formulated to set minimum standards of entitlement to social security and to coordinate social security systems among states. These have been devised under the auspices of the International Labor Organization (ILO), an intergovernmental body, and the International Social Security Association (ISSA).4 As a result, the terminology and concepts used in social security law have tended to be drawn by analogy with labor relations law rather than with tax law in Europe. Some effects of this Eurocentric approach are noted in this chapter.

The comparative history of social security taxes and contributions in different countries shows widely varying approaches. The developed countries that do not have a compulsory social security contribution or tax are in a small minority.5 In contrast, in other countries, social security taxation is the most important single source of public revenues.6 Countries with economies in transition generally have social security taxes. In some they are a major burden and, in those countries, social security is one of the most important fiscal problems of the state.7 Many developing countries are also faced with problems of social security funding as their economies change and develop.8

Social security taxes should therefore, it is suggested, be part of the agenda for a review of the tax law of any state. The adoption and operation of any form of social security taxation should be undertaken as part of or parallel with the total tax structure of the state. It is for these reasons that this chapter analyzes the legal issues inherent in imposing social security taxes within a state’s general tax system.

Whether a state has social security taxes or not, all but the least developed or least interventionist of states have social protection expenditure. The choice between direct and indirect funding of this expenditure must affect the fiscal pattern. This is a policy issue, but one with significant practical implications at both internal and international levels. A state that funds all or most of its social expenditure by levying taxes on employees, their employers, and the self-employed is committed to significant levels of income taxation. If the state intends also to collect a general income tax, it must consider both the interaction between the two taxes and their combined effect.9 It must also consider that, while it can reach international agreements to offset double income taxation, agreements rarely do this for social security taxes. A result may be that a state that chooses to collect social security costs through high social security taxes rather than through income taxes may tax some export transactions more heavily than states that use higher income taxes offset by reliefs.

Conversely, a state that chooses to bear all or most of its social security costs from general revenue must also determine whether direct or indirect taxes will bear that burden. If the cost is transferred primarily to income tax,10 the levels of tax become significantly higher than what would otherwise be required. Otherwise, general levels of state expenditure are forced down. If a state chooses indirect taxes, then it is likely to face problems in levying them.

For example, a high level of value-added tax (VAT)11 will impose a burden on all those whom the social expenditure seeks to benefit. The level of benefits, and consequently of contributions, may have to be raised to offset the burden of the indirect tax if the base of that tax is not to be affected.12 If, instead, the form of the indirect tax is affected, then other problems about the efficacy of that tax arise.13 Again, there may be international aspects to this if the resulting tax affects internal costs and export costs without affecting import costs.

Much of the above is a matter of general state and fiscal policy and concerns macroeconomics and public finance rather than law, and so is beyond the scope of this book. However, unless all tax, benefit, and contribution patterns are looked at together, many points made elsewhere in this book about forms of taxes may be modified in unintended ways. Specific examples of this issue are raised at the end of this chapter.

A. What Is Social Security?

To clarify the scope of this topic and chapter, the terms to be used in the discussion must be defined.14 In addition, the limits imposed on the analysis in this chapter must also be set forth.

“Social security” is the commonly accepted global term15 for public schemes (provided or regulated by the state)16 for the social and economic protection of individuals and families. Social security is normally classified under five headings:

  • old-age, invalidity, and survivors’ benefits;

  • benefits for sickness and maternity;

  • occupational or work-related risks;

  • unemployment protection; and

  • family assistance.17

States with social security schemes may not provide all these benefits. For example, there may be no unemployment benefit scheme because the state provides no support for the unemployed.18 The state may instead provide a means-tested benefit funded from general taxes for those without income or savings. There may also be separate organizations and separate contributions to different benefits. For instance, the state may provide family assistance for all children, financed from general taxes, while having separate contributory pension and work-related benefits.

If funding is provided on a form modeled on commercial insurance, it is often termed “social insurance.”19 It is distinguished from “social assistance,” whose chief activity is the provision of minimum incomes or material help to the poor, often on a means-tested basis.20 In some countries, health care and support for families or children are provided through “universal benefits” given to all regardless of contributions or ability to pay.21 Some are financed by contributions, and others by general tax revenues. If such schemes do not involve contributions, they are omitted from further discussion in this chapter. This chapter is concerned only with forms of social security met by specific funding from any source, but not those funded through general tax revenues. Most states with social security schemes have separate social security funds to meet this form of expenditure.22 Specific social security taxes or contributions are widely used to finance such funds.

This chapter is not about social security systems as such. There is no discussion of general policy issues, such as the desirability of public provision of social support, or of the kinds of social security benefits that might be funded through specific social security taxes. There is no discussion of the detail of social security benefits.23 Other topics excluded from this account are personally funded benefits, such as retirement benefits, the general issue of pensions, the peculiar legal problems involved with compulsory but largely privatized schemes that are found in countries such as Chile or Peru, and the voluntary provision of pensions.

The focus, therefore, is only on compulsory or state-mandated systems, which, in some countries, are also state run.24 In other countries, the administration of the system is in the hands of separate organizations or agencies.25 The administrative structure of social security is not addressed in this chapter, except insofar as it relates to the use of the state tax authorities for the collection of contributions.

B. Are Social Security Payments Taxes?

When is a payment to a social security fund a tax? The answer to this question in any particular state depends on its constitution and laws. As a general conceptual matter, it also depends on how government and others view the payments. The assumption in this chapter is that a contribution to a social security fund is a tax if there is a requirement to make payments either to state funds or to state-regulated funds from which there is an obligation to pay social security benefits.26 In short, the payment must be mandatory and must be state regulated. If the potential payer can choose whether to make the payment, then it is not a tax.27 In some countries, the system of social security payments is mixed, with some payments being compulsory and others voluntary. A complete account of social security must deal with all forms of payment. However, the system is, for present purposes, a tax system only if a substantial part of those payments are compulsory.28

The working definition does not resolve a question of characterization: are contributions, although mandatory, really taxes or are they actually insurance premiums? Some argue that the payments constitute, in essence, an insurance premium or a contribution. The payments will, or may, be returned to the payer as benefits. They are therefore, viewed from that standpoint, not taxes.29 Unless the payments have the individualized market character of premiums related to the risk presented by each person insuring (adjusted, for example, to take account of the illnesses an individual suffers or whether the individual smokes or of the risks of a particular employer’s activities), this argument is not conclusive.30 It may affect the presentation of the system to those required to pay, but it does not affect the underlying requirements of the law in imposing the contributions. If contributions are risk related and are determined by actuaries rather than by law, then at least some features of the scheme are not truly tax features.31 It must be recognized that contributions in some states imposing national standard social contributions are nonetheless regarded by all concerned as not part of taxation. It must also be recognized that the characterization is a political and cultural matter to which, in the abstract, there is no “correct” answer.

For some purposes, a compulsory payment may not be considered a tax if the payment creates an entitlement to a benefit.32 It may be relevant in this regard that social security systems generally do not involve an entitlement to benefits, because the legislature has the power to change the formulas under which benefits are determined, even with respect to benefits that relate to contributions that have already been made. Some argue that the feature that benefits can be set independently of contributions distinguishes tax-style systems.

A second argument against categorizing compulsory contributions as taxes is the identity of the fund-holding or administering body. In some states, the fund is held by the state itself as part of the general tax and budget exercise.33 In others, it is held by the state, but in separate funds.34 In yet others, it is held by extrabudgetary bodies that are state entities.35 Finally, in other states, the funds are run by separate individual funds under general state control.36 For present purposes, the identity of these bodies is not relevant if payments to them are compulsory, provided only that they are state, or state-regulated, bodies. Nonetheless, the primary focus of the chapter is on state social security funds rather than on state-regulated private bodies.

Regardless of whether social security contributions are treated as taxes under the constitution and laws of a particular country, they are justifiably considered as taxes for purposes of this book because the contributions are imposed by legislation that involves the same issues as other tax legislation and that interacts with other tax laws, particularly the individual income tax law. Therefore, the approach to drafting such laws and other tax laws should, to the extent the laws fulfill parallel functions, be a common or parallel one.

The definition of when a contribution is, for present purposes, a tax, also does not resolve the question of terminology. Two contrasting pressures apply in practice. In some states, social security taxes are neither called “taxes” nor treated as such in the legislation. Terms such as “contributions,” “insurance contributions,” or “premiums” are used instead. The standard general term in English is “contribution,” the term used in this chapter.37 In contrast, other states avoid the use of “social security” and use tax terms, such as “payroll tax,” that do not refer to the purpose of the tax.38

C. Is the Legislation for Contributions Tax Legislation?

Responsibility for legislation on social security contributions varies between states, and so does its form. Sometimes, the legislation is presented as part of the general social security legislation, so that the contributions and the benefits are presented as an entity.39 Alternatively, the legislation may be treated as social security legislation, but kept apart from the details of benefits payable.40 In contrast, the legislation imposing contributions may be treated as tax legislation and kept entirely separate from benefit legislation.41 Is that important? The practical answer is that the matter is unimportant unless there are either special constitutional or legislative procedural requirements for taxation. If such requirements exist, it may be possible to avoid them by presenting contributions law as part of general social security law. In other states, social security contributions may be given the form of taxes so as to take advantage of broad taxation powers under the constitution or general tax legislation.42

A final constitutional issue is that of the level of imposition within federal states. In some states, elements of social provision take place at the local rather than at the national level, thus preventing a state scheme from operating.43 This may also mean that social security taxation is operated at a different level of government from the income tax.44

D. What Are the Forms of Social Security Tax?

The inconsistencies of national practices about social security funding are reflected in the varying forms of contribution adopted. States have a series of choices in deciding the form that contributions take.45 These choices are influenced by two conflicting pressures, the fiscal context of the contributions and the linkage between contributions and benefits.

A key decision about the contribution structure of a scheme is whether there is to be one global contribution to all forms of funded social security46 or separate contributions for each separate form of social protection.47 This issue is usually decided by the institutional structure of the social security system. If different funds or institutions are responsible for the different elements of social security, there is strong pressure to provide different contributions to each institution.48 There are also policy arguments for separating, at least in name, the contributions to different funds so that different policies may be followed about whether and how much of the contribution is paid by the employer and how much by the employee.49

This raises a further issue. Is the contribution treated as a series of separate contributions50 or as a single payment to be made by the contributor but consisting of separate amounts for different funds?51 Having separate amounts of contribution for each separate form of benefit reflects the view that there should be separate contributions to each fund and also separate accounting and actuarial analysis.52 The rationale is that the independence of the separate funds is recognized by separation of the contribution payments. This is, it is suggested, neither necessary nor an efficient use of contributors’ money.53 The collection of a composite contribution is usually more efficient, and therefore less expensive, than multiple collections. The authority acting as agent in collecting the contributions can, of course, arrange subsequent distribution of contributions to those organizing the individual funds. In this discussion, only one contribution is assumed to be collected from any contributor for any contribution period, regardless of how the contribution is calculated or shared out after collection.

Second, contributions are collected in a variety of ways. They may be collected from the groups covered by the schemes only54 or from taxpayers generally.55 They may take any form adopted in general taxation or in insurance and pension practice. The cost of social protection for employees tends to be imposed on employers generally and sometimes on their employees. However, there is no consistent pattern of the share of the contribution burden between the employer and the individual. The payment is usually a payroll tax on the employer or a form of income tax imposed on the individual or the employer (or both). By their nature, these forms of tax are income related. Sometimes contributions are risk related at a generalized level of risk. The ultimate form of tax on this basis is a poll tax. All those within the scheme pay the same contribution in money terms and so share the risks.56 The self-employed may also be asked to make an income-related contribution or a flat-rate contribution (or both); this may be set to imitate the contributions of employers and employees.

If the employee or employer’s contribution is income related rather than risk related, it may take the form of a second income tax. This could be—and sometimes is—collected with the main income tax.57 It can be set perhaps as an extra rate of income tax.58

It is frequently assumed that contribution liability arises in respect of the earnings of the contributor. Although the definition of earnings is relatively straightforward for most employees, it is less obvious for the self-employed (or for those who are both employed and self-employed).59

Whether individuals should pay contributions on unearned income is a complex question of policy. If the social security system is designed essentially as an income replacement system for earners, then principle suggests that contribution liability should be based only on earnings. The extreme case is that of the individual who has no earned income, but has a significant source of investment income. Such an individual will in most states fail to qualify for social security benefits on the ground that he or she is not a member of the social security scheme. There is no risk of loss of income through unemployment, sickness, disability, or retirement, and, therefore, no need to ensure against these risks. Use of health facilities can be made subject to payment. Social security is therefore largely irrelevant to the limited number of individuals in this group, and principle suggests that they should be excluded from both benefits and contribution liability. Similarly, the unearned income of those within the system can be ignored for both benefit and contribution purposes, which simplifies administration. However, it demands a clear distinction between forms of income inside and outside the scope of contribution liability. There is also a contrary argument based on social solidarity and the use of income taxes to redistribute wealth between citizens. This argument leads some states to use nonearned income that is liable to income tax as a base for charge to social security contributions as well.60

In the discussion that follows, it is assumed that only earned income is used for contribution liability. Some problems of definition of earned income are discussed in connection with the contribution liability of the self-employed. The discussion of the liability of employees ignores any income other than employment income.

While other forms of social tax are possible, such as the social equivalent of a VAT or a sales tax61 or a levy on company profits,62 these are not usual. The link between such a tax and the benefits it funds is far from transparent, except to the extent that the name of the tax clarifies the link. With these forms of tax, there is no clear relationship between those paying the tax and those benefiting from it, except that both are within the tax and social security jurisdictions of the state. There is little reason, save political expediency, to identify a general tax with social security expenditure. In addition, it is doubtful that the link could be maintained in the longer term. For example, a link between a sales tax and the cost of state health services may lead to inappropriate changes of tax rate or inappropriate levels of funding of the health service. While budget pressures may suggest use of these forms of social taxes in the short term, no special design or drafting issues arise because of the “social” labeling of the tax. Therefore, they are not discussed further in this chapter.

E. Links Between Contributions and Benefits

There is an inevitable linkage in any social security scheme between contributions to the scheme and benefits paid from the scheme. If contributions to the scheme and the income that the scheme itself generates do not match the total level of benefits that must be paid from the scheme, then the scheme will fail. Either further forms of funding must be found or benefit levels must be reduced. It is impossible, except in the very short term, to have a viable social security scheme if the benefits and contributions are decided independently of each other. In consequence, the level of funding of a scheme must be decided by primary reference either to the intended levels of contribution or to the intended levels of benefit. The method selected will influence the structure of the fund. The existence of the fund does not of itself determine how the contributions of any one contributor relate to the benefits to be received by that contributor unless the fund has only that contributor as a member.

Links between contributions and benefits can be of two kinds, reflecting the differences between funded and unfunded schemes. A funded scheme is a scheme in which the contributions paid in are used to create a fund from which benefits will, in due course, be drawn. If the fund has numerous contributors, the level of contributions required by the fund is based on actuarial advice about the probable pattern of contributions, fund income, and benefits in the predictable future.

In funded schemes, the contributions paid in by an individual are saved to fund that individual’s pension, on either an actual or an actuarial basis. In either case, the relationship between the contributor’s contributions and benefit entitlement is provided in the structure of the scheme. Benefit entitlement, therefore, is based on the total actual contributions of or for the contributor, or the assumed total of contributions. The assumption of total contributions is often decided by reference to the length of service of the contributor in employments covered by the scheme. The actuarial basis of a fund will assume full pension entitlement only for a contributor who has contributed on all earnings throughout a set maximum period working for the employer, such as 30 or 40 years.63 Those who contribute for less than that period will not receive a full pension.

An unfunded scheme does not retain contributions to meet future obligations to pay benefits to those making the contributions. Instead, it is run on the basis that current contributions meet current benefit expenditure.64 This is subject perhaps to the maintenance of a buffer or reserve to ensure a smooth flow of both contributions and benefit payments. In an unfunded scheme, there is no economic link between an individual’s contributions and the benefit entitlement of that individual. A person’s pension contributions this year fund this year’s pensioners. This person’s own pension will be funded, it is to be hoped, by those contributing during the person’s retirement.

Funded schemes therefore need links between the contributions and benefit entitlements of individuals, while unfunded schemes do not necessarily need such links. There are, however, good reasons to establish links within unfunded schemes as well. A formal link between benefit entitlement and contribution ensures that only those who have contributed can benefit from the fund. It is both a justification to contributors of why they are expected to contribute and an inducement to voluntary compliance with the obligations of contribution. It also reduces the effective cost to the fund of those who avoid or evade contributions.

The link between contributions and benefit entitlements may be either direct or indirect. A direct link requires a contributor to have made a defined level of contributions before receiving benefits.65 That requirement may be small in the case, for example, of entitlement to benefit for industrial injury.66 For long-term benefits, such as retirement pensions or survivors’ benefits, the requirement may be for payment of contributions at a set level in each year of the contributor’s working life (or until death before retirement). Assuming that contributions are related to earnings, the total of contributions may affect not only entitlement to any benefit, but also the level of benefit. A direct link imposes administrative requirements for an accurate record of the contributions of every individual who may have benefit entitlement. It also requires that the contribution record be readily available for the determination of any claim to benefit. This requirement may create problems if a complex contribution condition is imposed in respect of short-term benefits, such as benefits during sickness or unemployment.

A scheme may also use an indirect link, of which there are two effective forms. The first is the period during which the contributor is employed in employment subject to contributions.67 In effect, this link measures the total time during which the contributor has contributed, but uses the contributor’s employment record to decide this rather than the contribution record. The other form of link is by reference to the contributor’s residency status.68 To comply with entitlement under this form of link, the contributor is required to show that he or she resides in the territory covered by the scheme for a required period or periods. This basis can work if the territory is covered by only one scheme for the contributor. It also assumes that residents pay their contributions or taxes. A direct link between contributions and benefit entitlements may be more effective than the indirect alternatives in ensuring that contributions are paid.

II. Issues in Social Security Taxation for Employees

A. General Terms

For the reasons discussed previously, the terminology used in this chapter follows that normally used by social security lawyers.69 The terms set out in this section are those used in the chapter. Most require definition, and these definitions are discussed in the following paragraphs.

Contributions: payments of social security tax, whether made directly by potential beneficiaries or by others. Contributions are assumed throughout to be compulsory. Any contributions that are not compulsory are termed “voluntary contributions.” In this part of the chapter, “tax” therefore refers only to general taxation.

Contributor. A physical or legal person making or required to make a contribution is a “contributor,” rather than a “taxpayer” or “insured person.”

Employee. An individual working for another person is an “employee,” and the nature of the work that the “employee” undertakes is “employment,” Employees are sometimes also termed “workers,” but that term is ambiguous because it may include workers who are not employees. The alternative phrase sometimes used for employment (particularly in civil law countries) is that of “dependent personal services,” but this has only limited use in English. It contrasts with “independent personal services” or the services supplied by an “independent worker” or “self-employed person.”

Employer. The person employing the employee is the “employer.”

Self-employed. A person working independently, or engaged in independent economic activities of any kind (other than those of managing investments), is referred to by the usual English term of “self-employed person.”

Unemployed. An individual who is neither an employee nor self-employed is said to be “unemployed.” “Unemployed” is used to describe those not engaged in employment or self-employment. If the individual, because of independent resources, dependency, marriage, or for any other reason, is not seeking employment, it may be better to term the person “nonemployed.”

Working age. Employees and the self-employed are of “working age” if they are over the age at which the state requires everyone to attend school on a compulsory basis and under the age at which an old-age pension becomes payable. States normally have general rules deciding the school-leaving age and the age at which the old-age pension becomes payable. It is normal to expect employees and the self-employed to pay contributions throughout the period when they are of working age, except while unemployed. Contributions might not be required below or above that age.

Old-age pensioner. An individual over the retirement age (or pensionable age, as it is sometimes called) is normally called an “old-age pensioner” or, simply, a “pensioner,”

Contribution period. Contribution liability arises with respect to defined periods, as does income tax liability. The standard income tax approach defines liability by reference to a fiscal year (often a calendar year), but it may also do so by reference to the calendar month or to the amount of a payment. Contribution liability may also arise with reference to a year. The analogy to labor relations law may mean that the period to which the contribution is related is the primary period for payment at work, usually a week or a month. The question of the length of this period is addressed below.70 For ease of discussion, the term “earnings period” describes the period used to decide the amount of a contribution. An earnings period can be as short as a day or as long as a year depending on the administrative policy of the scheme involved.

B. What Is Employment?

The distinction between economic activities that are characterized as employment and those that are considered self-employment is fundamental to the laws that impose liability to contribute to a social security system and to the laws providing for benefits from a system. The distinction is found in the laws of countries of all legal traditions save those where the whole economy is regulated and the roles of individual workers are controlled.71 It is widely used for social security purposes, although some states deal, instead, with categories of economic activity.72

The distinction is also important for several reasons other than social security. For example, it is often important in defining the liability of an individual to the income tax.73 It is important for VAT purposes because an employee is not a taxable person for VAT.74 The rights of an individual under employment law or labor relations law depend on the individual establishing that he or she is an employee. Civil contract and liability rights, and rights under insurance legislation, also depend on the same distinction.

Although the concept of employment or dependent personal services is most important, it has generally proved difficult to define. For some groups, the definition is relatively easy. For example, those working in government service can usually be regarded without difficulty as state employees if they are under the authority of the state. This includes those working for the armed forces of the state and the state’s diplomatic service. This is so for other large employers. The employees work on standard terms and within a clear structure that sets out the duties of each employee and the wages or other benefits received by the employee.

There are other groups of workers for whom the position is less clear. Are professional workers regarded as employees, or are they regarded as members of a profession and treated as self-employed whether actually employed or not? Should those working for several employers be regarded as engaged in part-time employment for a series of employers or as self-employed (or both)?

The definitions of employer, employee, and employment are all linked. The key to these definitions is the employment relationship between the employer and the employee and the other terms are best defined by reference to it. It is also important to ensure that the definition of these terms is the same for both contributions to the social security scheme and entitlements from it, so that the link between contributions and benefits is clear. In this way, a state may establish a more formal link, such as making benefit entitlement dependent on the payment of contributions.

The social security meaning of employment has traditionally been linked to the meaning of employment that is accepted as a matter of labor relations law (often called “employment law”).75 This means that an individual who is regarded as an employee for general labor relations purposes is also regarded as an employee for social security purposes. This definition will often also be used for other legal purposes. For example, it is used for civil liabilities of the employer to the employee (and the reverse), or of either to third parties.

As a point of both practice and principle, the definition of employment should be the same for both liability to pay contributions and entitlement to receive benefits. Any differences weaken the link between contributions paid and benefits received and may also increase the administrative burden of running the system. It is particularly important that the system ensure that a contributor cannot avoid contributions and then claim benefits. For example, the contributions due from or for employees are usually higher than those for the self-employed. If an individual can claim to be self-employed while contributing, but then claim to be an employee to obtain higher benefits, then the system is open to abuse.

There is also a clear linkage between the definition of employment for social security purposes and the use of the same term to define the liability to other taxes. For example, income tax law also requires a definition of the status of an employee and of income from employment.76 A person can register for VAT, or be required to register, only if the person is, in the terminology used here, self-employed.77

There is much to be said for establishing a common approach to definition across all these laws and taxes, although the different laws will require specific provisions to deal with problems specific to each tax. A common definition ensures that those who claim rights as an employee must also be shown to have paid tax and contributions as an employee. This does much to stop abuse of the system.

If a common definition is used, clear references should be introduced between the laws ensuring that key definitions are applied uniformly. This simplifies the application of the law, but it does not by itself ensure uniform resolution of disputes. Even with a common definition, it must be decided who is to handle a disputed question about the status of an individual. In practice, different authorities must decide for their own purposes, and their conclusions may not be consistent. While there are obvious advantages in defining and applying laws consistently, it may not be easy to accomplish. For example, the labor relations courts or tribunals decide disputes for labor relations purposes. If a labor relations tribunal is not involved, the parties to a relationship decide by mutual agreement whether they treat a relationship as employment or not. For income tax and VAT purposes, the tax authorities may be called upon to decide how an individual is to be taxed, subject to appeal to tax courts or tribunals. This may result in divergence, particularly because tax tribunals and labor relations tribunals are likely to be involved at different stages of an employment. It is difficult to ensure consistency among the different courts and tribunals without affecting their jurisdictions or independence. A practical alternative may be to secure agreement between administrators. For example, the tax and social security authorities may agree that they will normally respect each other’s rulings in individual cases.78 If there is a common definition, the superior courts can also ensure consistency and, at this level, bring together the approaches of lower tribunals.

This general approach to the definition of employment raises some difficult issues for social security. Problems particular to social security contribution liability are discussed below.79 Two general issues need further attention here: problems of contribution avoidance and the status of officeholders. A third issue is that of excluded categories of employee or employment.

C. Independent Workers: Are They Employees?

Experience in many countries80 shows that employees will be strongly tempted to claim to be self-employed if the liability to tax and social security contributions is greater for an employee than for a self-employed individual. This temptation is particularly marked when a high level of contributions or payroll taxes is levied on the employer. If an employer employs an employee, the employer contribution is payable. If the employee is instead taken on under a contract as a self-employed person working independently, there is no payroll tax. The saving to the employer may be considerable, even if the “employee” receives more money. It is also harder for the social security authorities to obtain a full contribution from a self-employed individual than from an employer. One solution to the noncollection problem is to levy a flat-rate tax on all payments made to persons in these categories, whether they are employees or self-employed. However, if the rate of withholding tax applied is too high, a genuinely self-employed person may be overtaxed because this approach makes no allowance for business expenses. This may not be a problem with genuine employees.81

This problem may require a case-by-case approach, with rulings being made for, and agreements being secured with, different groups of employers or employees. These agreements can reflect the fact that contributions and benefits are linked, at the state level if not at the individual level.

Experience also shows that the problem of sham self-employment is more likely to occur in some areas of activity than others. One solution is to provide special rules under which an individual is treated as an employee whatever the actual legal status of the person. This is done, for example, with those whose work involves neither high levels of skill nor the use of tools or equipment provided by the “employee.” Examples are casual staff in a catering establishment or on a building site.

D. What of Officeholders?

The other general problem in defining employment is the status of an officeholder. An individual appointed to an office is not an employee and is not self-employed because of the appointment. This applies to public offices such as judges, government ministers, or members of parliament. It applies to noncommercial positions, such as trustees of a charity or senior members of a religious organization. In addition, it applies to commercial organizations, in particular, to the directors of a company.

Many officeholders are in broadly the same position as an employee in that they earn their income by working in the positions they hold. Others are in the position of a self-employed person. For example, a lawyer who holds a part-time directorship in a company that is a client of the lawyer’s firm so that the client may be given legal advice is in reality a holder of that office only as part of the lawyer’s professional activities. Rules must be provided to deal clearly with these cases.82

The most difficult area is that of the individual who runs a small business through a company. An extreme, but not uncommon, situation is that of an individual who has a small business that she runs on her own and then turns into a small company. The individual owns all the shares in the company, is the managing director of the company, and works full time in the business of the company. The individual takes money out of the company as director’s fees. She may also be paid as an employee of the company under an employment contract in addition to her status as director. She can also receive dividends. Is the individual employed, self-employed, both, or neither? If the company also employs the individual, then clearly there is an employment. However, what is the position of the individual as company director? Again, a clear rule is needed.83

E. What Categories of Employees May Be Excluded?

Employees may be excluded from a state’s social security scheme for a variety of reasons. However, in any national scheme the reasons for exclusion should be objective. If the scheme is truly compulsory and universal, no one has a right to exclude an individual for personal or voluntary reasons. Therefore, those groups of employees that may be excluded from schemes must be noted. In each case, the rules for exclusion may apply differentially to the employee and the employee’s employer. In this section, the exclusion of employees is addressed. The effect on their employers is discussed in the next section.

1. Age-Based Exclusions

A general reason for exclusion is age. A scheme can apply an upper age limit and a lower age limit. In most schemes, those over retirement age (or pensionable age) become entitled because of age to a pension. Those receiving pensions are not required to contribute from those pensions. A series of subsidiary issues must also be decided to determine whether a person over pensionable age has any liability for contributions. For example, if a person is allowed to draw a pension but also continues to work, are the earnings from the work liable to contributions? A similar point arises when a person who could draw a pension on age grounds chooses not to, but decides instead to continue in full-time work. Should that individual pay contributions after reaching pensionable age? Again, what of the individual who, on reaching pensionable age, is, for some reason, not entitled to a pension and is therefore obliged to keep on working? In each case, the answer reflects both issues of funding and issues of fairness. It may be decided that, upon reaching pensionable age, an individual should, on age grounds, not pay any further contributions. On the other hand, it may be decided that contributions should be paid on all earnings, regardless of the identity (and therefore age) of the earner. The answer may also reflect any linkage between contributions and benefits. If the contributions of an individual over retirement age cannot earn the individual any further pension, then it may be questioned if they are truly contributions.84 They may have become an income tax to be justified on other policy grounds. Alternatively, if the employee remains covered for some risks but not for others, a different rate of contribution may be appropriate.

Similar issues arise with young earners. Schemes may have a lower age limit. This may reflect a general rule of law preventing children below a certain age from remunerative work. As any child working below that age is working illegally,85 it is administratively simpler to exclude any social security involvement in such cases. A second reason for a low age limit is an assumption of a lower limit to the working age used for the scheme. A lifetime scheme is based on all earnings of an employee of working age. The state, or the scheme, may assume that the working age starts at a certain age or when the individual is no longer required to attend school. If so, it may be appropriate on grounds of fairness and administrative efficiency to exclude an employee below that age, or still at school, from a requirement to contribute. Such contributions will not, in the normal course of events, be needed for the individual to get a full pension entitlement. Inclusion of such earnings may also encourage evasion or nondeclaration.

2. Education and Training

An issue related to a lower age limit is that of continuing education or training. This covers those continuing to pursue a full-time education or training after reaching the minimum age of the scheme and after having left school. In particular, those attending university or full-time professional training should be considered. The income of a student from a government or private scholarship is not usually regarded as earnings. A student may have a low level of casual earnings from, for example, part-time catering work, and it is for consideration whether such earnings should be within a system. A related question, beyond the scope of this chapter, is whether a student or trainee should be able to pay voluntary contributions to make up any “missing” years of contributions,86 or whether contributions should be credited to the individual during training.87

3. Voluntary and Nonremunerated Workers

Other groups of employees (or those who might be regarded as either employees or officeholders) may also be excluded if they have no earnings or minimal earnings only. For example, members of religious orders or other organizations to which individuals offer their services voluntarily while having their immediate personal needs met, such as voluntary development workers, may be excluded.

4. Recipients of Benefits

A decision must also be taken about the contribution liability, if any, that arises from someone in receipt of benefits. For example, is someone who receives sickness benefits in place of earnings during a period of sickness liable to continue paying contributions toward a retirement pension? This can be a complex question in a state that has separate schemes for separate benefits and also depends on how a benefit is paid. If an individual is ill, her or his earnings may be kept in payment in several ways. The employer may allow the employee a number of days of sickness a year without loss of pay, the employer may have a specific private insurance scheme that makes good the pay, or the pay may be made good by the employer, who is then entitled to recoup from the state the amount paid. Finally, the employee may be entitled to claim a benefit direct from the state. This may or may not come from the same fund as that from which the retirement pension will be paid. Each method of providing the employee’s benefit during sickness may suggest different policy reasons for requiring, or exempting, contribution liability on the sums received. There seems no strong reason to exempt if the funds come from the employer. Whether exemption should apply when sick pay comes from the same fund as a retirement pension may depend on how the real value of the sick pay compares with the pay before sickness.

5. Specific Employments

Approaches differ markedly among countries as to the scope of employments covered by social security schemes. In some countries, all employees are included in one universal scheme.88 The one scheme therefore covers both public servants and private sector employees. At the other extreme are systems where each industry or profession has created its own scheme.89 Decisions must then be taken about the coverage of each scheme. Intermediate positions may be adopted if some employments or industries have special schemes,90 with those not covered by special schemes being within the general scheme.

If separate schemes exist for individual professions or industries, a separate approach may be taken to individual professions. For example, all those who practice a profession may be expected to join their professional scheme whether they are self-employed or are professional employees.91 This approach forms an exception to the usual distinction between employees and the self-employed.

6. Employment Within the Family

As a result of income tax rules, it is convenient in some countries for one spouse to employ the other in order to deduct the salary paid. This may reduce the family’s general liability to income tax. The same is true when parents employ their children or, more generally, when employment takes place within the family. It is to be considered whether such employments should also be regarded as employments for social security purposes. It might be decided that some forms of employment within the family should be ignored, for example, when one member of the family is paid to do the housework. This may technically be an employment, but it is unlikely to be a true commercial contract of employment. Therefore, it may be appropriate to provide for the exclusion from social security schemes of employments within the family other than genuine employment in a family business.

7. Jurisdictional Exclusions

A scheme needs to have rules for the inclusion or exclusion of employees whose work is not solely confined to the jurisdiction of the territory covered by the scheme. This topic is discussed separately below.92

F. Who Are Employers?

It is necessary to identify the employer in respect of any employment because the law imposes contributions directly on employers. The law also requires employers to act as agents for the social security authorities to collect contributions from employees. There is little difficulty in identifying the employer once the employee and the employment are identified. The employer is the other party to the contract of employment with the employee. If a definition is needed, then it might be simply stated that the employer of an individual is the person who employs the individual. Inevitably, there are legal and practical difficulties with this approach. For an officeholder, there is no employer in the legal sense. If officeholders are treated as employees, then the organizations of which they are officeholders must be assumed to be the employers.

Another practical difficulty occurs when an employee appears to have two employers. For example, Val works for company A. Company A instructs Val to work for company B for six months in another part of the state. Company A pays Val’s expenses in moving temporarily to work with company B, but company B pays the salary and costs and instructs Val on the duties to be undertaken. Formally, company A is the employer, but in practical terms company B has become the employer, at least for the time being. Rules are needed to deem the paying company to be the employer. Alternatively, they may allow the paying company to act as agent for the employer, or the employer may recover the relevant social security contribution from the other company.

As a result of the jurisdictional rules,93 an employee will sometimes be within the jurisdiction of a scheme, while the employer is outside the scheme. This problem may be dealt with by imposing the liability to pay the employer’s contribution on the employee (besides the employee’s own contribution). Alternatively, if the employee has been seconded by a foreign employer to work with a business within the state, that business may be deemed to be the employer. The principle is that the person benefiting from the services within the jurisdiction should be treated as the employer although not technically in that position. This prevents abuse of the system and prevents employees from being uninsured when a risk is realized.

G. Allocating Contributions Between Employer and Employee

Almost all schemes place the cost of social security for employees on those employees and their employers. However, agreement has not been reached on the share of the contribution to be borne by the employee rather than the employer; nor do all schemes within a state follow the same pattern.94 Some schemes, as a matter of principle, divide the contribution evenly between the employer and the employee and collect half from each.95 Some share the contributions between employees and employers, with the state—or taxpayers generally—also contributing.96 Several countries, including some economies in transition, share the contribution, but impose nearly all the burden on the employer.97 Others impose the entire burden on the employer.98 It is less usual to find the employee bearing the larger part of the burden, but this also happens.99

Some states have sought to add economic factors to the decision about the level of contributions imposed on employers.100 In effect, this approach introduces an indirect state subsidy to certain employers, the burden of which can be borne by other employers (thereby further increasing the element of subsidy). It runs against both the insurance principle of social security schemes and the trend of reducing the elements of tax expenditures in direct taxes.101

Requirements for imposing contributions on employers and employees are examined below, but two issues of principle need to be addressed at this point. First, are employers and employees liable to pay contributions on the same base? The answer may vary from one scheme to another. In a system in which the state is not involved, and in which contributions are divided evenly between the employer and the employee, principle suggests that the same tax base should apply to both contributors. This is less evident in other cases. However, the answer may relate to a separate issue, that of the linkage—if any—between the contributions paid by or on behalf of an employee and the benefit entitlements of the employee. The linkage for benefits, such as pensions, normally relates to the employee’s contributions only. That being so, any contribution liability of an employer that is not reflected in a matching liability of an employee means that funds are being collected with no related benefit entitlement.102 If that happens, the state should be able to justify this in the context of the general funding and benefits of the scheme.

The other issue is whether the employee can be made liable for the employer’s contribution if the employer does not pay. This relates to enforcement provisions and is covered later. It is similar to the problem, important for income tax, of collecting tax from a taxpayer who has already paid tax by withholding at source, but whose with older has failed to pay the tax to the tax authorities. An additional question in social security law is whether, if contributions are not paid but should have been paid, the employee loses benefit. In a fully contributory system, nonpayment of contributions results in nonreceipt of benefit. It is then in the employee’s own interests to ensure either that the employer pays the contribution or that the employee can make good the default or can be treated as having made it good.

H. Should the State Contribute?

Contrasting views are taken about contributions paid by the state to a social security scheme. One view is that it is appropriate for all three parties to the employment relationship—employee, employer, and state—to contribute to social security funding.103 The opposite view is that the state should be excluded from social security funds, which are financed by employees and employers without outside involvement.104 The choice of approach is a matter for political decision. Recent experience shows that the threat of social security schemes going into deficit has sometimes required temporary state subsidies from general tax revenues.105 The key issue is the extent, if any, to which it is considered appropriate to fund part of the social security scheme from general tax revenues. At the same time, how involved should the state be in running a scheme? The two extreme approaches tend either to involve the state heavily or to exclude it. In practice, the effect of the tax system on a social security scheme means that the position is less straightforward than these extremes suggest. For example, a state may subsidize a social security fund indirectly through tax allowances and exemptions. It may also control social security funds through conditions attached to such allowances and privileges.106

I. Basis of Contribution Liability

The liability to contribute may be decided either on a flat-rate basis or on a basis related to the income and (if they are subject to contribution liability) benefits paid to the employee by the employer. Flat-rate schemes were once common practice among developed countries, but they have largely been replaced by earnings-related contributions and are not discussed further here.107

1. Gender Discrimination

Many schemes used to differentiate between the rates charged for male employees and the rates charged for females. As the working conditions for males and females come closer together, and benefit entitlement rules such as retirement ages are made uniform,108 any economic justification for this form of discrimination is removed. It is often prevented by general principles of law109 or by general labor laws that prevent discrimination based on gender. This and other forms of discrimination may also be subject to human rights equality provisions.110

2. Earnings-Related Contribution Rates

Earnings-related contributions are justified by a need for revenue to finance benefits. They are also justified on the same basis as progressive income tax, by reference to redistribution through the tax system. However, the redistribution does not usually involve progressive rates of contribution. Rather, it occurs through the combined effect of the contributions system and the benefits available. In practice, because of links between contributions and benefits, a progressive rate structure would be irrelevant to many schemes. However, some states have a progressive element in the rate structure.111 The effect of thresholds and caps (see below) can be progressive or regressive, but this is best viewed along with the benefits available to contributors and with the tax position of these individuals to get the full picture.

Income tax systems commonly have thresholds below which payments are not required. This approach is also used in some social security schemes. Income tax laws usually specify an amount that can be earned annually before tax is payable. Alternatively, the law may grant a refundable credit against tax payable. A refundable credit system is not appropriate to social security, as it in effect combines liability and benefit. A social security system with a threshold, therefore, uses an exclusion of liability if income totals less than a set sum during the relevant period.

The adoption of a threshold is a compromise between the insurance role of a social security scheme and the recognition of both the administrative burden of imposing contributions on all earnings112 and the effect this may have on the lowest earners. Insurance considerations suggest that all earnings should be included, while thresholds are justified by both practical reasons and policies of supporting the poorest earners. In many states, the insurance approach prevails and there is no threshold. If thresholds are adopted, the conflict of policies may result in a level of threshold that is lower than that operating for income tax.113

A threshold may work in two ways.114 The most straightforward way is to exclude from liability to contribute anyone earning less than the amount set as the floor. Once that person’s earnings exceed that sum, the floor is ignored and contribution liability is based on total earnings. This method has a distorting effect when earnings are near the level of the floor. The imposition of full contribution liability on someone whose earnings just exceed the floor may appear to penalize that person compared with someone whose earnings are slightly lower and who has no contribution liability. A small increase in gross earnings will result in a reduction of net earnings. This step effect would also exacerbate the poverty trap.115 To avoid this, earnings below the threshold are typically exempted from contribution liability for contributors at all levels of earnings. Alternatively, there may be a special lower rate of contribution for earnings below the floor level. The price of this method of avoiding the step as earnings exceed the floor is a major reduction in the earnings base. This will require a higher rate of contribution on earnings above the floor level.

A threshold to contribution liability has two effects. At the general level, it avoids collecting small amounts and levying contributions on the poorest members of society. In this way, it has broadly the same effect as a personal allowance for income tax purposes. It also excludes a low-paid individual from the social security fund. This may mean that the individual is unable to claim from the fund, either because the individual is not a contributor or because the individual is unable to make enough contribution payments to meet the conditions for a benefit. Whether a threshold represents the best compromise of efficiency and justice then depends on any means-tested or universal benefits available in place of the contributory benefit. Exploration of that issue is beyond the scope of this chapter.

Unlike income tax, social security schemes often have upper limits, sometimes known as caps or ceilings. Capping occurs primarily with benefit entitlement, because schemes need to have an upper level of replacement earnings payable to contributors. This often results in the adoption of a cap or ceiling on earnings covered by contribution liability.116 A cap on earnings is the upper amount of earnings on which contributions are based for the year or other period of liability. Any earnings over the cap are therefore not subject to contributions.117 In a contributions-based and earnings-related benefit scheme, a cap ensures a maximum level of benefits to any individual or from any one source of earnings. This prevents a disproportionate amount of benefits from being paid to those with very high incomes.118 However, if benefits are capped but earnings are not, an overtly redistributional element is included in the scheme.119 Conversely, caps may have a regressive effect on the overall tax/contribution position of employees.120

3. Earnings Bands

An alternative to applying a percentage to earnings is to determine the amount of contributions by using a banding system.121 This scheme sets a series of levels of premiums, each applying to a band of earnings. For example, those earning between, say, $30,000 and $39,999 pay a premium of one level (perhaps based on a given percentage of $35,000). Those earning between $40,000 and $49,999 pay a higher premium (perhaps the same percentage of $45,000). The lowest band may start at a bottom figure other than zero, thus incorporating a threshold into the system. Similarly, the top band will expressly or indirectly build an upper limit into the system.

A banding system may be seen as having advantages in that it avoids an excessive need for accuracy. Earnings of $35,000 and $37,000 both give rise to the same premium. This effect reverses near the limits of the bands and can have very sharp consequences. Someone earning $40,000 will pay a higher premium than someone earning $39,990. Further, the extra amount of premium may be greater than the extra earnings. As a result, the person earning $40,000 may actually receive less than the person earning $39,990. This effect can be reduced by having a larger number of bands and a smaller difference in premium between bands. Any attempt to increase the number of bands will also increase the administrative complexity.

J. What Are Earnings?

A compulsory earnings-related contribution is in all but name an income tax if imposed directly on the employee. It is essentially a payroll tax if imposed directly on the employer. How much should the definition of “earnings” that are liable to an earnings-related contribution vary from the definition used for liability to income tax? There is little reason in principle why the measurement of income used for income tax purposes and that used for contribution purposes should differ. This is so whether the reason for basing the contributions on income is that it reflects ability to pay or that the benefits are earnings related. In both cases, income should include benefits both in cash and in kind. Again, in principle, any deductions permitted should be parallel, save perhaps for deductions for the payments of income tax and contributions themselves.

1. Links with Income Tax

There is a strong practical reason for ensuring a close identity between earnings for income tax purposes and earnings for contribution purposes. It minimizes the legal, administrative, and compliance burdens of collecting two parallel payments from employees. This identity should, if possible, be provided for in the law itself. The legislative terms used for imposing the income tax and the social contributions should be the same unless a difference is intended. Careful consideration should also be given to ensuring that the terms receive consistent interpretation and application by those responsible for administering and supervising the collection process. Failure to ensure this will almost inevitably lead to variations appearing in the operation of the income tax and contribution rules. This is an added level of complexity for both employer and employee, as well as a duplication of official effort. Unless dictated by clear policy reasons, any differences of approach are inefficient.

An additional level of complexity may be introduced by an appeals process. If decisions about the application of income tax legislation and social contribution legislation are dealt with differently, differences in interpretation may appear. One solution is to link the contribution legislation to the income tax on earnings. This link can be comprehensive, with the same rules operating for both income tax and contribution purposes.122 If this link is established, an appeal on an income tax question is automatically an appeal on the contribution question (and the converse is also true). The same position must then be maintained for any official rulings and guidance. That might best be achieved by providing that the guidance offered by the income tax authorities also applies to contributions.

2. Reaching Settlements

Objectives of income tax systems and social security systems are different. If the income tax authorities have power to deal with practical contribution questions in this indirect way, should they be aware of the different objectives and requirements of the contribution system? This issue is important if the income tax authorities can reach settlements and compromises with an employer over the tax liabilities of the employer’s employees. For example, an employer may be found to have misrecorded and underpaid employee income tax and contributions, accidentally or otherwise. The interests of the tax authorities are to make good the loss to the public revenue, for example, through a global settlement of liability with the employer, including a penalty element. The same approach may not be sufficient for contribution purposes because the contributions that should have been paid for or on behalf of the employees can count toward their benefit entitlements. A global settlement does not make any provision for attributing lost contributions to individual employees. That can be done only through a series of individual settlements. It is probably not in the interests of the tax authorities to engage in that level of supervision and negotiation. However, it is in the interests of the individual employees and, therefore, of the social security authorities.123

The example of compromises affecting the integrity of the contribution records of individual employees applies equally to any process of estimation or compromise in the tax system. It may also apply when the tax authorities declare an amnesty or otherwise decide not to enforce the law strictly. It may be in no one’s interest to seek the strict enforcement of the income tax law. Contribution liability has the added dimension of the need to protect the employee’s contribution record. A compromise may therefore involve the employee paying contributions but receiving no benefits. In practice, many such compromises result in little unfairness to individuals because little benefit is lost. The main concern is that the extent of this loss cannot be calculated at the time. The loss of only a small part of a contribution record may have major effects on one individual, but no discernible effects on others. In any system that collects contributions over the lifetime of a contributor, there is no easy solution to this clash of principles. It can be avoided in part by ensuring that the liability rules are effective and consistently applied and that the temptation to cut corners is avoided. However, it must also leave open the position of the social security authorities not to accept compromises of the income tax authorities that are not consistent with the principles of the social security system.

3. Benefits in Kind

Identical treatment of earnings paid in cash for income tax and contribution purposes generally poses no major problems and is desirable. Further attention may be needed to the treatment of benefits in kind and of any allowances and deductions from earnings. In part, this depends on the details of the income tax system, discussed in volume 2 of this book and not repeated here.124 It may be possible to use for contribution purposes the same method whereby benefits in kind are made liable to income tax, including the timing rules.

4. Specific Expenses and Allowances

Different issues of policy may arise for the allowance of deductions for specific expenses and allowances. The personal allowances in income tax are sometimes designed for social purposes, and their replication in social security schemes may serve to defeat rather than reinforce those social purposes.125 Other allowances, such as those for certain forms of saving or expenditure, may be regarded as inappropriate in a social security context. Genuine expenses of employment if allowed for income tax purposes should probably also be allowed for contribution purposes. If deductions are allowed, then efficiency argues for allowance on identical terms and subject to the same procedures.126

One area of difference cannot normally be avoided. This is the income tax treatment of contributions and contribution treatment of income tax.127

5. Practical Effects

The use of different definitions of earnings for tax and contributions purposes either in law or in practice imposes practical problems on employers. This is because any differences result in the need for the employee or employer to keep different sets of records for each employee. They must also provide different returns to the two authorities. This may be only a limited problem for an employer with an automated payroll if the two sets of rules are clear and easily administered. However, it neither avoids the cost associated with the need to supervise a continuing duplicated operation nor deals with the resulting problem of having two teams of officials wishing to audit the employer’s records for their own purposes. The problem of accurate compliance is significantly greater for smaller employers who lack trained staff and facilities for automation.

As a general matter, the importance of achieving uniformity among income tax and contributions, in terms of both legal provisions and administrative arrangements, will vary from state to state depending on the operation of both the income tax and the contributions systems. The discussion in this chapter assumes a substantial overlap in the groups that contribute to the two systems. One can, however, envisage situations where this is not the case, for example, where contributions cover wage earners generally, while income tax may involve only a small number of individuals because of a large zero-rate threshold. Because of differences in local circumstances, the considerations discussed in this chapter may have different implications for different states.

K. How Should Rates of Contribution Be Determined?

A major difference between income tax and contributions lies in the rates of payment, including any allowances. There is no direct correlation between income tax rates and contribution rates. As noted previously, rates of contribution liability are not usually progressive. Their levels will depend above all on how the overall social security scheme is funded. At one extreme are pay-as-you-go schemes that fund existing benefit entitlements from existing contributions. At the opposite extreme are schemes that fund a person’s benefits only from contributions made by or for that person. Between these extremes are funded schemes where contributions are pooled and schemes that rely on reserves held against future liabilities.

Rates are also linked with thresholds or caps to liability. Unlike general taxation, the rate or rates of contribution are based on the revenue needs of the fund to which the contributions are paid. There is normally a direct link between the actuarial estimates of liabilities of the fund and the estimated income to the fund for any given rate of contribution. The rate is dictated by professional advice on the required levels of contribution to fund the outgoings for the period.

Determination of the rate or rates of contribution is therefore not a matter of fiscal policy under a funded benefit scheme. Rather, it should be the result of the exercise of expert judgment of the actuaries, who will base their estimates on relevant demographic data relating both to the contributing membership of the fund and to the benefiting membership. Their decisions will allow those in charge of the funds to identify actual and likely income and expenditure over both the immediate future and the long term. A decision on the precise recovery of costs in any one year is a political decision to be based on actuarial advice.128

L. Earnings Periods

Direct taxes are normally determined annually. However, income tax on earnings is often collected by means of a preliminary or withholding tax. There may or may not be an annual adjustment to ensure that the proper amount of tax is paid over the year.129

1. The Income Tax Year

There are obvious advantages to using the same basis of collection for income tax and for social contributions.130 The advantages apply whether the contribution liability is that of the employer, the employee, or both. When the same base periods are used, income tax and contributions are collected together and follow one set of calculations, thereby reducing administrative and compliance costs of collection. If identical rules are used for income tax and for contribution liability, the annual calculation of social security contribution liability will share most of the other advantages and problems of income tax liability. It is easier to ensure an accurate and fair total assessment of contribution liability for the year. For example, the use of identical rules allows for easier adjustments for benefits in kind provided by the employer, with minimal administrative burdens on either the collection authorities or the employer. It is also easier to provide an accurate and transparent record of the total contributions to be paid and the basis for that total.

These advantages may be offset by a need to relate contribution periods to benefit computation periods. If the benefit year and the income tax year are not the same, administrative problems may arise in calculating benefit entitlement of contributions. It may also be necessary to ensure that contribution liability gives rise to benefit entitlement where entitlement depends on contribution conditions.

If, because of annual calculation, contributions are collected only after an assessment has been made, there are significant disadvantages to all concerned. Collection with each payment of earnings has obvious cash-flow advantages for the scheme. It also minimizes the delay between receipt of earnings and the contributions payable on those earnings and reduces problems of avoidance and bad debts. Furthermore, it ensures that an individual has a full contribution record at any time for claiming a benefit entitlement subject to a contribution condition. This suggests that a preliminary or withholding tax has significant advantages for contribution liability.

2. Separate Earning Periods

An alternative approach is to collect contributions on a final basis as earnings are paid.131 Each earnings period is treated as a separate assessment period, either exactly or by reference to a weekly, monthly, or similar period. If a pure earnings period basis is used, then each payment of earnings is subject to deduction of the appropriate contribution liability. If a flat-rate contribution is used, the weekly or other period for which each contribution is paid can be imposed over the payment pattern used for the employee with little difficulty. When contributions are related to earnings, adjustments are needed. If earnings-related liability to contributions is based on anything other than a straight percentage of the total earnings of the contributor, the liability in respect of any payment of earnings must be calculated by reference to the precise period for which the earnings are paid. Often, payment is by reference to a day, a week, or a month, on a recurring basis, presenting few difficulties. The earnings period basis causes problems when there is no clear pattern of payment and when contribution liability can be avoided through manipulation of the relationships between payments and earnings periods. For example, the existence of lower earnings limits and upper earnings limits creates distortions if uneven payments or earnings periods are used.

The main advantage of an earnings period base for calculation is that payment of contributions for each earnings period is final and not subject to adjustment, eliminating the need for an annual return. A contribution can therefore immediately be counted toward satisfaction of contribution conditions for benefits. The main disadvantage is that it is not possible to make an annual adjustment to reflect uneven patterns of earning or uncertainties. Similarly, if the income tax system contains annual adjustments, these cannot easily be assimilated into a pattern of short earnings periods. One practical result may be the need for antiavoidance provisions to override the finality of contribution payments when it is established that payments are being manipulated for avoidance or evasion.132 This may require an annual review period. In effect, it solves the problem of avoidance by removing the earnings period basis. The same is true if annual adjustments are made to reflect, for example, benefits in kind.

One way of dealing with the conflicting advantages and problems of the annual basis of income tax and an earnings period basis for contributions is to apply two or more kinds of contribution to earnings. Cash earnings can be assessed separately with each payment. Other benefits and adjustments can be collected on an annual basis.

A further complication with the earnings period basis of collection is that many employments have inappropriate earnings periods or, alternatively, no regular basis of payment. A daily basis is unlikely to be practicable; for practical reasons a minimum earnings period of a week should be established. Monthly payments are technically not regular, but may be assumed to be such. A pay pattern that is part weekly and part monthly (e.g., a weekly basic wage with a monthly profits-based bonus) causes practical problems. Also, irregular payments or payments made for regular periods at irregular times cause problems. A scheme adopting the earnings period basis for collection must also adopt a clear set of rules to decide in every case the relevant earnings period in respect of any payment made. Further, these rules must be easy to implement with a minimum of official guidance. It is likely that both antiavoidance rules and a default rule will also be needed.

The complications of operating an earnings period basis of contribution liability are, however, likely to be confined to the unusual case. In the usual case of regular weekly or monthly payments of earnings that are subject to contribution liability on a set percentage basis, the imposition of contribution liability on a final earnings period basis has several advantages. The balance between what is usual and what is unusual must clearly be important in deciding on the balance of advantages for the choice of contribution liability period.

The identity of income for both income tax and contributions purposes is also subject to any practical restraints imposed if there are different timing rules for the two payments. Income tax is usually an annual tax. If contribution liability is based on shorter periods,133 then the system needs methods of determining income that do not involve annual adjustments of the kind found in income taxes. For example, annual personal allowances or credits cannot be operated for shorter periods without some form of amendment or annual adjustment. Similarly, adjustments to take account of benefits in kind must be capable of being made within the contribution period. The timing rules will themselves depend in part on the method, if any, used to link contributions to benefit entitlement. One practical issue is that a benefit entitlement cannot be based on a contribution test that remains indeterminate at the time of benefit claim.

M. Multiple Employments

The traditional pattern of work in most forms of economy has been that individuals are engaged in full-time work with one employer for extended periods. In economies with strong service elements, and economies experiencing high levels of unemployment in recent years, this tradition has weakened. Individuals are more likely to change employments, and areas of employment, than in the past. There are also many more part-time jobs. Both raise issues that need to be considered in formulating contribution policy.

Often, a change of employment by an employee creates no problems. If there is a period of unemployment, then unemployment benefits may be payable. Contributions may not be payable with respect to the unemployment benefits.134 Problems may arise if, because of the change of employment, the individual comes to be employed in a different scheme from the one in which he or she was previously employed. This may involve a need either to transfer funds between schemes or to provide for some overlapping or cross-financing between schemes. The fewer the schemes run by a state, the lower the number of problems caused by such changes.

An associated change for which procedures may be necessary involves a person who ceases to be employed and becomes self-employed. This is dealt with in the next part of the chapter, as is the case of a person who is both employed and self-employed at the same time.

Another problem is that of an individual employed in two employments at the same time. This may occur at all levels of the employment market, from the person who is a director of several companies to the casual worker providing his or her services—for example, as a cleaner or temporary office worker—to several different employers during a work week. Some individuals who supply services to several employers at once are properly regarded as self-employed, but this is often not so. For example, someone providing cleaning services will usually be under the direct control of the person paying for the services and is properly regarded as an employee. In principle, an employee with several employments should be separately liable for contributions from each. Consequently, each employment during the week requires separate consideration. This is potentially a heavy administrative burden. It is also likely to lead to avoidance of contributions.

When the employments are all by the same employer (or employers that are connected with each other), the simplest answer may be to treat all the employments as if they are one employment. The problem may also be avoided if one employment is, or is nominated as, the “main” employment, with the obligations of contribution being imposed on the employer for that employment. If the scheme has a cap for contributions, this may be met in full by the main employment. Other employments can then be ignored (but only if the scheme is prepared to lose the employer’s contribution in respect of those employments). The problem may also be avoided if there is a minimum level of earnings from any one employment before contribution liability arises.135

Special rules can be used to deal with those who provide their services through an agency. If individuals, although technically not employed by an agency, provide their services to their employer through an agent, then the agent may be deemed to be the employer and may be required to account for contributions on behalf of all those paying the individuals. This may be effective if the agent (as often happens) collects fees from the employer and is responsible for paying the individuals.

However, these are only partial solutions. They require a complex set of rules, but the liability of employees and employers to contribute in respect of multiple employments of an employee at the same time must also be established to avoid revenue loss and inequity.

N. What Records Are Necessary?

Special consideration needs to be given to the records to be kept of contribution payments.136 Employers are subject to the normal requirement, in parallel with income tax, to keep records of all payments to employees and contributions collected from employees. In addition, central records must be kept of each contributor’s contributions. This may require additional record keeping by the employer. It needs a fully coordinated single record-keeping facility for each scheme and protection of the integrity of the data recorded. These requirements should be reflected in the legislation. The data recorded must reflect all contributions paid by or for individuals that are or may be relevant to benefit entitlement. If notional contributions are added to the individual’s record, these should be recorded as well.

The central recording requirement means that the social security administration needs to have a unique identifying number for each contributor. This is usually achieved by assigning each contributor a number, the individual’s social security number. The number will also be of use to the social security benefit administration. In practice, employees also need identifying numbers for income tax purposes. It is much simpler if the same number is used for both income tax and social security purposes. The number may be assigned by either the social security administration or the tax administration.137 A further alternative is to assign a national identification number to the individual, perhaps when the birth of that individual is registered or when the individual reaches school-leaving age.138

Each individual needs to be assigned his or her number personally. All employees should be required to supply their social security numbers to their employers. The records kept by employers can then be related to the central records with the minimum of problems. The numbers can also be used as a useful cross-check between income tax records and social security records.

The need to protect the integrity and accuracy of the contribution record of individuals over their lifetimes imposes a significant administrative burden on those administering a contribution-based social security system. The level of accuracy required depends on the precise relationship between contributions and benefits. Any close relationship imposes an original and continuing need for accuracy. It means that the data on the record should not be open to alteration or adjustment for any reason save by due process of law.

III. Issues for the Self-Employed

A. What Is Self-Employment?

The self-employed form a significant139 sector of every national workforce unless the national economic system imposes a state monopoly on conducting business. The sector includes both those who are genuinely in business on their own and those who conduct their economic activities in partnership or joint venture with others. Only if the form of business structure used itself has legal personality will the position of the proprietor of a business be converted into that of an employee or officeholder (director) rather than that of an independent worker. The position of the self-employed is particularly important in the service industry. Some professions, for example, lawyers, may prevent their members from becoming employees. In other areas, for example, small farms or restaurants, companies are not the usual form of conducting business. The technical distinction between employment and self-employment in marginal cases has been discussed previously.140

In addition to those who are fully self-employed, there are also those who provide consultancy or other services aside from and in addition to their ordinary employment. This applies equally to anyone who, for example, makes money by writing or engaging in artistic or creative activities or who engages in farming activities in addition to paid employment elsewhere.

The social security risks involved in insuring a self-employed contributor may be different from those for an employee. For example, a self-employed individual cannot become unemployed in the normal way, because he or she cannot be dismissed from or resign a job. Closing a business down is legally a different action and may lead to the realization of a capital gain or loss. However, a self-employed individual whose business fails may be unemployed. There is no set retirement age for a self-employed person. The self-employed frequently carry on working into older age. Further, they may have capital invested, or earned by way of goodwill, in the business that can be realized to provide a personal retirement fund. These considerations have led states to provide more limited schemes for the self-employed or to remove them from the scope of compulsory schemes altogether.141 However, the tendency is to broaden schemes to include them.142

Both the income tax liability and the social security contribution position of an individual who is self-employed are fundamentally different from those of an employee. There is no employment relationship and therefore no employer who can be asked to pay or collect income tax or contributions; nor, in most cases, can an employer be deemed to exist.

As a technical matter, some individuals who are employees may be treated as self-employed for contributions purposes. In schemes that impose some or all of the contribution liability on the employer, some employers may be exempt from tax generally, so that a contribution liability either cannot, or cannot conveniently, be imposed on them. Their employees can therefore be treated as self-employed so that they become responsible for contributions.143

B. What Is the Income of the Self-Employed?

The earnings of a self-employed individual are defined by reference to the actual or assumed net profits of the business. Income tax is payable on these, as defined by the relevant tax laws. At a policy level, there is little reason to define the profits of a self-employed individual differently for contributions purposes from the definition used for income tax purposes, save for the effects of the tax and social security systems themselves.144 A consistent definition is particularly appropriate if the income tax law itself relies on the definition of profits used for commercial accountancy purposes.145 The idea of separate definitions of “income” for social security purposes and for income tax purposes seems unnecessarily complex for the individual, but also devoid of justification by reference to the principles on which profit is normally defined. However, the underlying income of the activity may need adjustment between tax and social security liabilities to deal with specific allowances and provisions.

It follows that the definition of both income included and business deductions allowed under income tax law is entirely relevant for social security purposes, save for three aspects. The most important of these is that the focus on earned income as against investment income of the self-employed in income tax laws varies.146 If social security benefits are designed to be earned income replacements, then the nonearned income of an individual is irrelevant. The profits of an active partner in a business are relevant, but the profits of a sleeping or inactive partner are not. Similarly, the profits of an individual actively renting or leasing accommodations to others (e.g., short-term holiday accommodation) may be regarded as earned. The profits derived from investment property (e.g., an office building leased on the terms that the tenants carry out all repairs and pay all service charges) may be different. These differences may also be relevant for income tax purposes (e.g., they may affect eligibility for capital gain treatment on disposition of the property used, or in a schedular system may affect the schedule under which income is taxed). A clear definition of the scope of income to be included in, and excluded from, the contribution liability of an individual is needed to remove all cases of doubt.

The other distinctions are the treatment, in defining profits, of the payments of contributions, income tax, and any related insurance or pension contributions or premiums, and special allowances for tax purposes that are inappropriate for social security purposes.147

C. How Is Contribution Liability of the Self-Employed Based?

For the reasons just outlined, the social security contribution position of a self-employed individual is often different from that of an employee. Insofar as it depends on the income tax laws, the calculation of earnings must reflect the different income tax rules. The effect in most cases is to delay any contribution liability until the net profits of the individual for each accounting or tax year can be established. The only ways to claim a contribution in advance of this are to demand a flat-rate contribution or to attribute a notional income to the individual.148

Because the entire contribution of a self-employed individual must be paid by that individual, it may seem to be a higher rate than that of employees. It may prove inappropriate to impose full contributions on the self-employed for reasons noted previously. For example, the self-employed may be required to contribute only to basic minimum pensions and not to a full earnings-related scheme. If the self-employed receive the same benefits as employees, any reduction in the contribution of the self-employed may amount to a cross-subsidy by employees and their employers. If the net cost of contributions of the self-employed is reduced by allowing income tax relief, the effect is partly to shift the cost to general taxpayers.

D. What Records Are Necessary?

The same general points about record keeping and registration apply to the self-employed as to employees. The duties to register and to pay the contributions must be placed on the self-employed themselves. The collection of contributions cannot generally be imposed on those making payments to the self-employed.149 Excessive deduction might take place, and the contributions might not find their way to the contribution collection authorities. In addition, it is not incumbent on a customer of a self-employed individual to inquire whether contributions are required from that individual. They will not be required from a company or other legal person in a similar position.

A policy consideration in demanding records from the self-employed for social security purposes is whether this approach is consistent with that followed for income tax purposes. For example, if the tax authorities have decided to dispense with bookkeeping for categories of the self-employed,150 it is inappropriate to demand detailed records for social security purposes. Similarly, simplification of tax administration by adopting a high threshold for income tax purposes may be defeated if those below the threshold still have to meet onerous social security requirements.151

IV. International Aspects

A. What Limits Do Schemes Impose?

All countries find it necessary to impose limits on the jurisdiction of a compulsory social security scheme. A scheme requires two sets of limits: (1) those applying to individuals claiming benefit, and (2) those applying to persons liable to contribute. The limits may not be the same, but both normally require some continued economic activity, or at least some continued presence, in the territory of the state. The comparable direct tax rules are the residence of the employee or self-employed person and the residence of (or permanent establishment by) the employer within the state.152 Labor relations principles do not follow this basis. For labor relations law purposes, the usual rule is that a state claims competence over employment if the place of employment is in the territory of the state. An employee who resides in one state and works (or whose work is based) in another state is treated on these principles as working in the state where the work is based. Income tax is generally charged where the worker lives, but may also be charged where the employment is carried out. Particularly in areas bordering two states or in states with many migrant workers, the tax and labor relations rules may frequently have different and conflicting effects on an employee.

Claims to jurisdiction for social security schemes are usually based on labor relations principles because these schemes have been designed primarily by analogy with employment-based insurance. A scheme therefore aims to cover those who work at a relevant activity within the jurisdiction of the state. Although benefit questions and contribution questions can be decided separately, the rules determining contribution liability are normally coextensive with benefit entitlement. This is because at any one time entitlement to benefit and liability to contribute are normally regarded as coextensive. This follows from the fundamental idea that a benefit is funded by contributions rather than from general tax revenues. Unless a contributor stands to benefit from a scheme, why should there be a contribution? If there is no linkage, is not the contribution a disguised form of general taxation?153

If the state’s social security provision is divided into several different schemes, it may be thought that the jurisdictional rules of schemes should differ. For example, a scheme providing benefits to families in respect of children could define entitlement to benefits by reference to whether a child is present or resident in the jurisdiction. The administrative complexity of having different jurisdictional limits to different funds should not be underestimated. If, as suggested above, there is one composite contribution payable by any contributor, there is a strong argument for having a common set of jurisdictional rules for liability to contribute. Further, if the scheme has common rules for the liability to pay contributions and has links between contributions and benefits, then there should also be common rules for benefits. A decision to widen the scope of benefit—for example, for family benefits—requires a change or removal of the linkage rules. Alternatively, it needs rules providing for arrangements, such as deemed contributions or contribution credits.

States in which contributions and benefits are linked may wish to ensure that a beneficiary can claim benefits only if the beneficiary is within the jurisdiction of the scheme, both as potential beneficiary and as former contributor. Unless the contributor has been within the scope of the scheme and has therefore actually contributed when potentially a contributor, there is no entitlement to benefit. This focuses attention on the continuing importance of the jurisdictional limits of a scheme. For example, if a contributory pension is payable only if an individual has contributed for 25 years, then the individual must, by definition, be within the jurisdictional test for that period to qualify.

The main policy and administrative constraints, therefore, argue for a common set of jurisdictional rules for both benefits and contributions and for all funds. The forms that these rules can take are discussed below, as are the problems of overlapping claims to jurisdiction by schemes in two or more countries.

B. What Jurisdictional Rules Are Used?

There is no required international set of provisions deciding the jurisdictional limits of social security schemes. Principles and practice have been developed through the activities of the ILO and ISSA, but these amount to best practice only. The adoption of an international approach to coordination of schemes has occurred only in the European Economic Area (EEA), as addressed in detail below. The steps taken in the EEA are important because they follow the forms of best practice accepted more generally. Aside possibly from nondiscrimination provisions, double taxation conventions do not apply to social security contributions.154

1. Nationally Determined Limits

Each scheme establishes its own jurisdictional coverage. For reasons set out previously, the rules for jurisdictional coverage may reflect the jurisdictional rules adopted by the state to define when an individual is employed within the jurisdiction of the state and when the employee’s employer is within the state. For employees, the individual must be within the jurisdictional reach of the state for any rules applied by the state to operate realistically. This is because, without international agreements, the contribution authorities are unlikely to have either information or powers to enforce contribution liability beyond the territory covered by the national laws. Absent specific agreements, there are no information powers equivalent to those on which tax authorities rely.155

The practice of states is to set jurisdictional limits within these practical limits and without direct reference to any overlap with any other social security system. The key test is whether an employee is employed within the jurisdiction. For this to be shown to be the case, the employee’s economic activities must occur, or primarily occur, within the territory covered by the scheme. The employer does not have to be within the jurisdiction of the state for this to occur. The issue of where an employment occurs is primarily a question of fact, but will also take the requirements of the contract of employment into account. Employment within the jurisdiction will require the presence of the employee in the jurisdiction and probably his or her residence. However, for the reasons noted previously, the employee may be regarded as resident for this purpose even though the employee is regarded as resident elsewhere for income tax purposes.156

2. Specific Problems

Most employments cause no jurisdictional problems, because there is no doubt about where the employment occurs. Doubts are likely to occur only in a few particular cases. One group is frontier workers, those who work in one country and live in another, crossing between the two frequently and perhaps daily. Another is migrant workers, those whose long-term homes are in one country but who are absent from those homes for limited periods for work reasons only. Others are transport workers (those whose jobs involve traveling on international transport) and government employees (e.g., diplomats and members of the armed forces on foreign tours of duty). Best practice has been to evolve special rules for each of these groups of employees as exceptions to the general jurisdictional rules. Separate provision is made for them in bilateral agreements. States may wish to restrict their claims to jurisdiction in this way by agreement. It is suggested that they should also provide for an underlying claim to jurisdiction over these groups of employees in case of doubt or in absence of an agreement.

3. Migrant Workers

It may be appropriate to follow a general practice regarding migrant workers, that is, those who come into the country to work or who leave it to work. Most countries have workers in both categories. Rules can be applied in the same way to both immigrants and emigrants. The simplest rule is to ignore immigration and emigration unless the change of country lasts more than a set time, for example, one or two calendar years. A time limit rule will exclude immigrants from liability to and entitlement from the state’s schemes until they have been present in the state for the defined time. A year is likely to be the shortest practical time for a limit of this kind. One year may prove to be too short for many employees, and a period of two years, or even longer, is being adopted by agreement between countries or in individual cases. Because the rule is exclusionary, it may be appropriate to allow the period to be extended when an employee is in the country temporarily, but for more than one year.

The time limit rule also applies to those leaving the country. Here, an employee remains within a scheme for, say, one year after leaving the country to work elsewhere. If the employee returns within the year, then the absence is ignored. The employee’s liability to contribute and entitlement to benefit remain in place throughout the period, subject to any additional requirements, such as presence imposed on any claim to benefit.

4. Rules for Employers

Inclusion of an employer within the jurisdiction of a scheme is separate from inclusion of an employee of the employer. The presence of the employer needs to continue for the scheme to be applied. In practice, this may require a degree of presence similar to that of a permanent establishment for other tax purposes.157 This requires a continuing presence of an economically active part of the business of the employer, normally a branch or an agency. The concept of permanent establishment may not be enough in itself. For example, the activities of the employee in question may amount to a permanent establishment of the employer, but with no other active presence in the state. To regard the employee’s presence as establishing jurisdiction over the employer for social security purposes may be ineffective.

Special rules may be needed when employees are within the scope of a scheme but their employer is not. It may be necessary to provide that the employees are responsible for meeting the employer’s obligations to pay the employees’ contributions. It may be difficult to go further and impose the employer’s contributions on that employee. Instead, if the employee is associated with any other employer in the state, that employer might be deemed to be the employee’s employer for contribution purposes. This is an appropriate approach when the actual employer and the deemed employer are associated. It is particularly so when it is believed that the precise employment arrangements are designed to avoid contribution liability.

5. Rules for the Self-Employed

Separate jurisdictional rules are needed for the self-employed. As with employers, the analogy with the income tax rules is more valuable than for employees. The rules require that the individual conducting a profession have a fixed base. A permanent establishment is required for an individual trading directly. For income tax purposes, the rules decide how much, if any, of the profits of the individual are earned within the jurisdiction. For contribution liability, the key question is whether the individual is carrying out economic activities in his or her own name within the territory of the scheme to the extent that he or she should be within the scheme. The nationality of the individual will normally not be relevant to that decision, nor will the level of economic activities or their success (as measured by profitability). At the practical level, if the individual is not making any profits within the territory of the scheme, then no income-related contributions are due in any event.

Finally, a scheme may allow those who fall outside the obligatory contribution provisions to continue to contribute voluntarily. This may apply even if the individual is compelled to contribute to another scheme. The voluntary extension of a scheme in this way raises no issues of jurisdictional law.

C. Rules in the European Economic Area

The countries within the European Economic Area (the members of the European Union and some neighboring countries) have common rules to deal with international aspects of the jurisdictions of social security schemes. Although the common rules do not apply as such outside the EEA, the rules follow the recommended approach of the ILO. They provide a comprehensive approach to the problems of contribution liability of both employees and the self-employed when two or more countries are involved in determining liability. They must also influence the approach of candidate members of the EU or the EEA. In each case, they will also affect the policy choices of those countries in reaching bilateral agreement with other countries outside the EEA.

The main principles and rules for all EU member countries are laid down in Council Regulation 1408/71, as amended frequently since adoption.158 Because the rules are contained in a regulation, its text applies directly in all members. It is also subject to official interpretation by the European Court of Justice. The Oporto Agreement establishing the European Economic Area applied these rules (including court decisions at the date of adoption) also to the other members of the EEA. Consequently, the regulation (and its supporting administrative regulation, 542/72)159 applies throughout the 15 members of the EU (and some of their dependent territories), along with Norway and Iceland. The regulation covers all key issues relating to both contribution liability and benefit entitlement. The extensive rules about benefits are not discussed here, but provide the context of the rules relating to contribution liability. They allow states to ask each other for a transfer of contributions in appropriate cases.

The general rule, applicable to both employees and the self-employed, is that an individual is liable to contribute to the social security scheme of only one state at any one time. That state is the state in which he or she is employed, even if his or her residence is elsewhere and the employer is based elsewhere.160 This is consistent, as noted previously, with the general principles of labor relations law. For this purpose, the individual is employed where he or she normally carries out the duties of the employment. If those duties are normally carried out in two or more countries, a tiebreaker rule applies. This allocates the individual to the country in which he or she is habitually resident, if some duties of the employment are carried out there.161 The habitual residence test also applies to multiple employments. The liability of the employer follows that of the employee. The employer must comply with the employer’s obligations under the laws of that country in which the employee is found to be employed.

The regulation provides six specific exceptions to the general rule, as follows:

(1) Employees leaving one state for another for a temporary assignment remain within their home state’s schemes if the absence is a year or less. States have power to extend this period and are encouraged to do so for longer temporary assignments.

(2) If the undertaking of the employer itself crosses a frontier, the employment occurs in the state in which the employer is registered.

(3) Public servants and members of the armed forces of a state remain subject to the jurisdiction of that state.

(4) Diplomats and consular staff are entitled to the usual immunities from local jurisdiction and remain potentially subject to their home state jurisdictions. EU staff may elect to be covered either by the state where they work or by the state of nationality.

(5) International transport workers are treated as being within the jurisdiction of the state where they are based. However, if they are principally employed where they reside, then they fall within the jurisdiction of that state even if they are based elsewhere.

(6) Members of the merchant marine based on a ship registered in a member state are subject to the social security system of that state, unless an individual resides in a state and the employer is also registered in that state. In that case, the individual’s state of residence has jurisdiction.162

Regulation 1408/71 also contains rules dealing with refugees and stateless persons. More generally, the regulation works within the framework of the fundamental and general principles of EU law. The two most important general rules are that states and public authorities may not discriminate within the EU either on grounds of nationality of any member state or on grounds of gender. These rules therefore apply alike to females and males and to all nationals of all member states of the EEA.

D. What Other Treaty Rules Exist?

Save for the EEA provisions, there is no general international agreement dealing with social security contribution liability. Some agreements make specific provision for particular cases. Specific agreements have been adopted to deal with international transport workers, with diplomats, consuls, and employees of international organizations, and with refugees and stateless persons. Apart from these provisions, agreement is usually by bilateral convention.163

The outline of the EU provisions also serves as an outline of the contribution provisions in the usual form of bilateral agreement. Within the scope of a bilateral agreement, the key principle is that the law of only one of the two states applies on a compulsory basis. The rules for deciding which of the two states has jurisdiction are usually similar to those outlined for the EEA and are based on the same approach. In practice, the time limit rule may have a period longer than one year; two years is increasingly common, and the United States is seeking a five-year period. The pattern of bilateral agreements is far from complete and falls far short of the otherwise similar patterns of double tax conventions or of trade and investment agreements. Within the European Union, bilateral agreements have been superseded by the Council Regulation in situations where the two texts overlap.

V. Interaction of Income Tax and Social Security Schemes

A. General Principles

A funded social security system is designed to collect funds for disbursal within a closed group of people, usually those who contribute and their dependents. Income tax is an open system, taking funds from all those within the scope of the law and releasing them to be spent as general public expenditure. Any transfer between the funds collected by income tax and the funds collected in the social security system is therefore a transfer between two disparate groups. As a matter of general principle, it is assumed that these transfers should not happen except as a result of explicit policy decisions.

Interactions between income tax and social security are unavoidable. They can also be expensive in terms of funds forgone or transferred because of the large numbers involved in both schemes at any time. In practice, the rules of income tax systems and social security systems often interact in unintended ways. These interactions may result in transfers either from the open system to the closed system or the reverse. This section is designed to draw attention to these interactions so that policymakers and drafters have them in mind in considering the effects of their laws.

Interactions occur between income tax and contributions as competing methods of collecting funds. They also occur between income tax and benefit entitlement. The income tax treatment of private social security arrangements (e.g., occupational schemes and retirement benefit plans) must also be noted.

B. Interactions Between Contributions and Income Tax

The income tax and contributions systems are independent, even if both are collected by the same means. Without explicit provisions in either law, it may appear that the two systems therefore do not interact. In practice, they do.

1. Rules for Deductibility

The simplest case is that of income tax and contributions being collected under the same rules for defining liable income and for the same periods. This assumes that there are no differences in liability or timing between contribution liability and income tax liability. Each applies at a given percentage rate on the earnings. Will each be collected on the total amount of earnings or on the net amount after deduction of the other? This depends on the deduction rules for the two taxes.

There is no strong reason to grant employees a deduction for either of the payments against the other if they are both at reasonable overall levels. The total system is simpler and easier to administer if no deductions are allowed. It is also more transparent and has lower marginal rates. If a deduction for contributions is allowed against income tax, then the total income tax collected from most individuals is reduced. Making good that loss can only be achieved by raising either the rate of income tax or some other tax. It is better to ensure that neither system allows a deduction for the employee’s payment to the other.164

For example, the rate of income tax in state D is 25 percent at the level of average earnings, and the employee’s contribution is also 25 percent. If no deductions are allowed, the employee pays $50 on marginal earnings of $100. If state D allowed a deduction for social security contributions to be made against income tax, then, on earnings of $100, $25 is paid as contributions. The amount of earnings for income tax liability is $75. If state D wishes to collect $25 from the employee, the rate of income tax must be increased to 33 percent. The amount collected is the same, but the total marginal rate appears to be not 50 percent but 58 percent. If state D does not raise the rate to 33 percent, there may be hidden transfers between those paying contributions and other taxpayers. This is because, in effect, the income tax relief for the contributors has to be paid for elsewhere in the system. Some of it is therefore likely to be paid by non-contributors, who are essentially being asked indirectly to subsidize the contributions. What if state D adopted the converse rule? Instead of allowing a deduction for income tax, it allows it for contribution purposes. Here again the lost revenue would have to be made good. This would force up contribution rates because there is no other way of meeting the income loss. Further, unless care is taken in the way in which the income tax is relieved, the result could be unfair between different contributors. Those paying more income tax could pay lower contributions. Why should that be so?

2. Measurement and Timing Issues

The assumptions made in the above example are that the measurement and timing of the income tax and contribution liabilities on earnings are identical. That is frequently not the case. There may be timing differences in the collection procedures, and there may be differences in definition in the two laws. Each difference may cause a distortion between the two systems. For example, a particular benefit in kind is taxed to income tax but not to contributions.165 The overall burden of tax on that benefit is therefore less than on other forms of income. It is therefore likely that employers will use that benefit when they can to pay employees. This will not affect the amount of income tax collected, but will reduce total contributions. The effect is to cause a shortfall in contributions. This may require a rate rise that will, of course, not fall on those receiving the nonliable benefit in kind. What has happened is a shift of liabilities from some contributors to others caused by the mismatch of income tax and contribution rules. This problem is compounded if allowances are made in the income tax system for contribution payments (or the converse).

These complexities can be avoided for contributions by employees. They are harder to avoid for employer’s contributions because the employer’s contribution is usually regarded as a cost of labor. To prevent the employer from deducting the contribution is to increase the post-tax cost to the employer of that contribution compared with the employee’s pay, which is difficult to justify in income tax terms. The result may be different if considered in terms of social security funding. For example, state D requires employers to pay an employer’s contribution of 30 percent of earnings. This is deductible as an expense for business income tax purposes. One effect is that the employer recovers part of the cost of the contribution from general taxation. It may be argued that it is better, therefore, to deny employers a deduction for contributions, although this is not common practice.

3. Treatment of Employer’s Contributions

Should the employer’s contribution be treated as part of the earnings of the employee? The usual approach adopted by states is that it is not. If the employer’s contribution is regarded as part of the earnings of the employee, the social security fund is not affected. However, an income tax problem arises. Income tax becomes due on the employer’s contribution as well as on that of the employee. There may be no net gain to the system from this tax if, as a result, the rates of tax are lowered. There is an implicit tax privilege that benefits contributors compared with other taxpayers.

4. Treatment of the Self-Employed

If the self-employed are to be treated as receiving profits for income tax purposes, there is an argument that contributions should be deducted against profits for income tax purposes. Against that, comparison with the position of the employee might argue for nondeduction. These arguments will apply to the alternative treatments for the self-employed contribution. The most appropriate treatment may depend on whether the self-employed are paying a flat-rate contribution or an earnings-related one. It also depends, as noted above, on the total intended cost to the self-employed of their contributions.

C. Should Benefits Be Subject to Tax and Contributions?

The interaction between social security benefits and the income tax and contribution treatment of those benefits involves several complex issues.

1. Making Benefits Subject to Contributions

The first issue is whether benefits are subject to contribution liability. When benefits and contributions are related to the same fund, the simplest approach is to treat benefits as not subject to contributions. The fairness of this policy depends on how the rate of benefits compares with the income being replaced by the benefit. If, for example, the benefit fully replaces the income, the beneficiary will gain by having saved the contribution. The easiest approach to avoid overcompensation is to reduce the level of benefits by the amount saved. If the benefit only partially replaces the income, then exempting the benefit from contribution increases the net value of the benefit.

If the benefits and contributions do not relate to the same fund, broader policy issues arise. A failure to collect contributions amounts to preferential treatment of a benefit funded by other means. If the benefit is a safety net benefit, then contribution may not arise because the level of benefit is below the lower earnings level for contribution liability. That raises another policy question. How should the threshold for contribution liability relate to the level of social assistance benefits? Should the relationship be taken for each earnings period or be calculated to include a full year at a time? If the lower earnings level is below that of benefits received, then the state must decide if these benefits are to be subject to contribution liability in the same way as earnings. It may be argued that income-replacement benefits should be treated in this way. One reason for this is that exempting benefits from contribution liability may mean that the beneficiary accumulates no contribution record while receiving a benefit. Because of this, the beneficiary might appear to gain in the short term, but will lose in the long term.

2. Notional Contributions

A solution to the problem of imposing contributions on benefits is to use notional contributions. Benefits can be treated as subject to contributions, but the beneficiary is not expected to pay the contributions. Instead, he or she is granted a credit of a notional contribution. In this way, the beneficiary continues to accumulate a contribution record but does not have to pay for it directly. This raises another policy question. Who pays for the benefit receivable because of the credit? If the cost is met from within the general social security fund, then it amounts to a cross-subsidy, which is entirely appropriate if the benefits are paid from the fund. It may not be appropriate if the benefits are paid from other funds. Generally, it is desirable to subject benefits to either actual or notional contributions, both as a matter of administrative simplicity and to protect the long-term position of individual beneficiaries.

3. Imposing Income Tax on Benefits

With respect to the income tax treatment of benefits, there are two issues of principle. The first is whether benefit income should be liable to income tax. The second is the legal nature of the benefit income if it is to be subject to income tax. It is not earnings, although it will normally be of an income nature. It may or may not be paid to replace earnings.

A broad-based income tax should, as a general rule, be imposed on benefits.166 This is the simplest and least distortive approach and is consistent with the view that problems of social security and social assistance should be left to the social security system. It also prevents overcompensation of beneficiaries. If the income tax system has personal allowances or credits, the effect may be to remove those receiving minimum benefits from the charge to income tax. This depends, of course, on the level at which allowances or credits are set. If a personal allowance is set at a level above that of the safety net benefit entitlement of an individual, then no tax is payable unless the beneficiary has other sources of income. For instance, if a person is sick or unemployed for part of the year, but at work and earning during the rest of the year, then income tax is calculated on the total income of the individual for the year. The cumulative value of the benefits is therefore taxable. Imposing tax on benefits in this way prevents overcompensation. That will occur, for example, if a taxpayer receives sick pay free of tax in a year when income tax is payable on total earnings. In some cases, however, it may be appropriate to exclude benefits from income tax for administrative reasons, particularly when the value of benefits is low.

Assuming that benefits are to be taxed as income, how should they be taxed? Strictly, benefit income does not fall within the usual main forms of income. It may be most appropriate to treat it in the same way as earnings, although the beneficiary is unlikely to have any deductible expenditure in “earning” the income. This is because most forms of benefit are designed to replace income, particularly retirement and survivor’s benefits, unemployment pay, and sickness benefits. Family benefits and benefits to the severely disabled are different but are most easily taxed in this way. For the sake of clarity, the income tax legislation should specifically authorize the taxation of taxable benefits.

These arguments suggest that benefits should be taxable and also subject to actual or notional contribution liability, although this policy may conflict with the policy toward private social security provision. Countries often give preferential treatment to private pension and insurance schemes. For example, there are three ways in which some countries may treat private retirement funds preferentially. First, the contributions of the employee (or self-employed individual) to the fund are deductible from earnings before the imposition of income tax. This can be argued to be a tax deferral, because the pension, when paid, is treated as taxable (and usually as deferred earnings).167 Second, the employee’s contribution is not treated as part of the employee’s income.168 Third, the income of the pension fund is itself exempt from tax.169 A treatment of public social security funds inconsistent with this treatment of private funds creates a distortion between the state system and the private systems. The desirability of such a distortion is a policy issue. The hidden state support for retirement schemes in this income tax treatment, measured as tax forgone, is very expensive. The level of support is compounded if exemption is extended also to public social security schemes. The generosity of the tax treatment of private pensions is being reviewed in many places because of its expense, and it is suggested that there is little reason to increase the problem by any extension to the public sector.

D. Implicit Tax Rates on Benefit Programs

The complex patterns of interaction of contributions, income tax, and benefits rules can cause hidden traps in the general system of tax and support.170 Two forms of trap may exist. The first is the poverty trap. This is the situation where an individual is trapped in poverty because he or she is unable to increase his or her net earnings through extra effort. Linked with this is the unemployment trap. This happens when an individual is unable to obtain a job that pays enough to make the claimant better off than receiving benefits. Both these effects deter individuals from seeking to improve their position and can be argued to trap them into receiving benefits. The argument is that the individual has no incentive to do anything other than continue to receive benefits. Other policies are also involved. Income tax is often regarded as a progressive tax. The tax system may also be regarded as partly designed to redistribute income. Poverty and unemployment traps distort the rate structure of the tax system as a whole and cause redistribution from the wrong individuals.

For example, state F has a progressive income tax. The lowest rate of income tax is 15 percent. It has a contributory social security system, with a set rate of contribution, the current employee’s rate being 20 percent. It provides benefits in two ways. First, the social security fund provides a replacement income for the sick and the unemployed. Second, the state finances a rent benefit that meets the rent of those whose income is below the state poverty level. Those with incomes above that level receive no benefit.

Ben is currently unemployed. Ben receives $150 a week benefit from the state fund, and has the rent of $100 paid by the rent benefit. The state benefit of $150, and the rent benefit, are both free of income tax and contributions. To replace current benefits, Ben must earn enough so that, after deduction of income tax at 15 percent and contribution liability at 20 percent, the amount left is at least $250. Ignoring any contribution thresholds and income tax allowances, the minimum replacement income Ben needs is $392. To escape from the unemployment trap, Ben must therefore increase income from the $100 paid in cash by $292. Can that be done? Unless Ben earns more than $392, there is no point in his seeking work. Ben is caught in the unemployment trap.

Por has a part-time job because of the need to look after her two young children. Por’s total income is below the state poverty level for a parent and two children. Consequently, Por receives the rent benefit, which is worth $150. Por has just been asked to work one more day a week, which will earn her $50 more. However, Por notes that this extra $50 will take her earnings over the state poverty level. She will therefore lose the state rent benefit and will not receive the full $50 because of income tax and contribution liability. At present rates, Por will receive only $33 of the $50. The extra day’s work will therefore leave Por $117 a week worse off. Por refuses to work the extra day. Por is caught in the poverty trap.

E. Can High Implicit Tax Rates Be Avoided?

Can these traps be avoided? The examples given are deliberately very simple. In practice, examples are often considerably more complex. However, a benefit like the rent benefit in the example always creates a trap for those who cannot afford to lose the benefit by increasing their earnings. One answer is for the allowance to be tapered. The allowance could be granted as a maximum of $150, with lower benefits to those with earnings in the “trap” area.

Suppose that the rent benefit had a “taper” of 70 percent. That is, for each $100 over the poverty level a beneficiary earns, $70 of the benefit is lost. This would still trap Por. Por pays income tax of 15 percent and contributions of 20 percent on the extra earnings, and loses 70 percent of the rent benefit under the taper rule. This amounts to a composite tax rate of 105 percent. It is not worth paying. Ben may be assisted. Tapering lowers the sum Ben must earn to improve his total income. If the taper were 60 percent and not 70 percent, then Por would gain by taking the extra work, although not by very much. Alternatively, the taper could be based on Por’s income after deduction of income tax and contributions. The taper would then be 70 percent of the net increase in earnings, not of the full increase. Of course, a taper in either form also means that benefit must be paid to those above the poverty level. This may have a considerable cost. The income tax and contribution rates may also be tapered. The allowances may also be set so that they cut the trap effect on low earners. Removal of any threshold for social security contributions will further dampen the effect of changes. Support to low earners from the income tax system could be given through credits rather than through allowances, thereby changing the effect of an increase in earnings. Sometimes traps may be made less significant by removing a tax exemption, but increasing the underlying level of benefit so that the most deserving beneficiaries do not lose. Further up the income scale, the value of the benefit will decrease because of the tax. At the lower end, there is no net cost or benefit to the state.

How can traps like this be avoided? The most important issue is to ensure that all tax, contribution, and benefit provisions are examined to ensure that trapping effects are identified and, if possible, minimized. Sometimes this can be done through the way a benefit is worked, rather than through a simple increase in value. For example, a taper based on net income rather than gross income keeps the total effective tax rate from exceeding 100 percent. This shows that a solution to a trap may lie in the way a benefit is paid rather than the way it is taxed or made subject to contribution liability. It may be that the nature of the benefit has to be changed. Alternatively, it may help to pay a larger benefit that is taxable rather than a smaller benefit that is exempt from tax, so that the beneficiary receives the same amount, but the distortions are avoided. There is no simple solution.171

VI. Conclusion

The issues outlined in this chapter are receiving considerably more attention than in the past, largely for demographic reasons—the adverse shift in most countries in the dependency ratio of those receiving state benefits compared with those paying for them. This has forced schemes to increase contributory funding and to look for more effective ways of raising funds. The discussion shows that there is no clear consensus of how this should be done. Important issues are clearly regarded as based on tradition and history to a greater extent than with other taxes. Nonetheless, there is a need in many countries to address the issues raised in this chapter and to seek best practices both in the forms of contribution liabilities adopted and in the methods by which they are collected and enforced.

Note: The writer is indebted to Warren McGillivray of the International Social Security Association, and to Stanford Ross, former United States Social Security Commissioner and Public Trustee, for making available some of their extensive knowledge of social security and for reviewing drafts of this chapter. My thanks are also due to Antoine Delarue, Panit Dhirapharbwongse, Victor Thuronyi, and Bertil Wiman.


For further information on this subject, see Janet Stotsky, Payroll Taxes and the Funding of Social Security Systems, in Tax Policy Handbook 177 (Parthasarathi Shome ed., 1995).

Much useful literature is published by the International Social Security Association. See infra note 4. Other current summaries of national social security systems are available in the following publications: Coopers & Lybrand, 1995 International Tax Summaries (summarizing the position in most states) [hereinafter C&L 1995]; Organization for Economic Cooperation and Development, The Tax/Benefit Position of Production Workers, Annual Report Information Covering 1990–93 (l994)(containing details for each of the OECD member countries) [hereinafter OECD Tax/Benefit Report]; OECD, Revenue Statistics of OECD Member Countries 1965–93 (1994)(published annually) [hereinafter OECD Revenue Statistics]; K.C. Messere, Tax Policy in OECD Countries 167–84 (1993). For more detailed summaries of the position of most European countries, see 6 International Bureau for Fiscal Documentation, Guides to European Taxation, Taxation of Individuals in Europe (looseleaf) [hereinafter 6 IBFD European Taxation]. The European Commission has also published summaries of the systems operating in the European Union. Comparative Tables of Social Security Schemes (5th ed. 1990).


Larger countries without any social security taxes include China, Australia (at federal level), Indonesia, and Thailand. See OECD Tax/Benefit Report, supra note 1, at 110 (regarding Australia). For a full comparison of coverage in 1990, see World Bank, Averting the Old Age Crisis: Policies to Protect the Old and Promote Growth, tbls. A.1 to A.6 (1994).


Detailed treatment of the problem of levels and trends in social security expenditure is beyond the scope of this chapter. The main problem is that of the growing world population of older people. The World Bank recently provided an updated analysis of the problem in Averting the Old Age Crisis: Policies to Protect the Old and Promote Growth (1994). See World Bank, supra note 2. The key conclusion is that the number of people over 60, measured as a percentage of the total world population, will double between 1990 and 2030, and that this growth would be reflected in all parts of the world. For a critique of the report by authors from the International Labor Organization and the International Social Security Association, See Roger Beat & Warren McGillivray, A Risky Strategy: Reflections on the World Bank Report “Averting the Old Age Crisis,” 48 Int’l Social Security Rev. 5 (1995).


The ILO, based in Geneva, is a specialized agency of the UN. ISSA is a nongovernmental body unofficially linked with the ILO, and also based in Geneva. The members of ISSA are representatives of individual social security schemes (national, industry-based, or specific). ISSA has a regular conference and publication program on contribution and benefit issues.


Of the OECD member states, only Australia and New Zealand have no, or minimal, social security taxes. OECD Tax/Benefit Report, supra note 1, at 110, 186; see Messere, supra note 1, tbl. 8.1, at 183. Denmark also used to have minimal contributions, but in a reform taking effect between 1994 and 1997, it is introducing a “labor market contribution” payable by employees, employers, and the self-employed. 6 IBFD European Taxation, supra note 1, at 63; OECD Tax/Benefit Report, supra note 1, at 131. Korea also has no social security contributions, although it has a less common education tax. KOR BNTA, part XIV.


OECD Revenue Statistics, supra note 1, shows that 45 percent of total French tax revenues come from social security contributions, 39 percent of total Dutch revenues, and 38 percent of total German revenues. Id. at 21–22. The definition of “public revenues” is that used in OECD Revenue Statistics. See Id. at 28 et seq. This publication monitors both taxes and other contributions to schemes operated within the government sector. Id. at 10. For example, Finland (like most Nordic countries) has high compulsory contributions to a general scheme, but does not regard this as taxation, even though the contributions use tax laws for assessment purposes. 6 IBFD European Taxation, supra note 1, at 55.


The subject was extensively reviewed recently in ISSA, Restructuring Social Security in Central and Eastern Europe (1994). See also George Kopits, Social Security, in Fiscal Policies in Economies in Transition 291 (Vito Tanzi ed., 1992).


For an account of recent reforms in Argentina, Colombia, and Peru, see Monika Queisser, Chile and Beyond: The Second-Generation Pension Reforms in Latin America, 48 Int’l Social Security Rev. 23 (1995).


The contrast in approaches in OECD states is sharp. In France and the Netherlands, lower-paid workers pay far more in social security contributions than in income tax. In Australia and New Zealand, there are no separate social security taxes, although a tax on fringe benefits is imposed on employers in addition to the income tax. For an annual survey of the position in each OECD member state, see the annual volumes of OECD, The Tax Benefit Position of Production Workers; see also note 1 supra.


As in Australia, which adopted a fringe benefits tax that imposes an additional charge on employers who grant fringe benefits to their employees. New Zealand, which also has no direct contributions to social funding, adopted a similar tax. C&L 1995, supra note 1, at A-39 to A-40.


This is one factor in the high VAT rate in Denmark, and the reason why Denmark attempted to adopt its employment levy and has since adopted its labor market contribution scheme. See infra note 22.


This happened in New Zealand upon the introduction of the broad Goods and Services Tax. See supra ch. 6. New Zealand, however, does not impose social security contributions.


The issue of socially sensitive exemptions from value-added tax is discussed in ch. 6, supra, where it is argued that such exemptions should be minimal.


See also sec. II(A).


This is true in the English language. Terminology in other European languages is similar and poses few terminological problems.


Analysts have categorized the approaches that states may take in providing or encouraging provision of benefits for individuals with a social effect. The six accepted, basic kinds of approach are (1) social insurance; (2) employer mandates, or compulsory provision of employee welfare benefits by employers; (3) individual mandates, or compulsion on individuals to provide for their own welfare benefits; (4) tax-supported voluntary arrangements; (5) social assistance, or means-tested welfare payments financed from general funds; and (6) universal schemes, with entitlement for all citizens or residents funded from general funds. This chapter focuses on schemes that fall in category (1), with less attention to categories (2) and (3). Category (4) is relevant in the context of the interaction between tax and social security. For a recent critique of these approaches, see Lawrence H. Thompson, The Advantages and Disadvantages of Different Social Welfare Strategies, Int’l Social Security Rev. 59 (1995).


This framework has long been the basis for discussion and action by the ILO, the international body with prime responsibility for this area of activity. See also Messere, supra note 1, at 170 (commenting that contributions for other benefits in the OECD states “are of negligible revenue importance”).


That is, there is no special compensation for unemployed individuals, as opposed to welfare payments for those without income.


See supra note 16, in which this is category (1). Social insurance has been defined as consisting of schemes with the following attributes: compulsory membership extending beyond government employees; compulsory contributions payable by members (or their employers); government regulation or support; prescribed benefit entitlements; benefit entitlement deriving from contributions, but not directly related to them; and separate scheme accounting and financial planning. See Robert J. Myers, Social Security 877 (4th ed. 1993)(setting forth the definition developed by the Committee of Social Insurance Terminology of the American Risk and Insurance Association).


See supra note 16, categories (5) and (6).


An example is the U.K. national health service, which is available to all residents without contribution or charge. It is funded largely from general taxes, but also from a (largely hidden) levy on social security contributions. The hidden levy is authorized by the Social Security Administration Act, 1992, ch. 5, § 162. Known officially as the national health service allocation, the levy is between 10 percent and 15 percent of total contributions, but is not separately identified in any way to contributors. Free national health services were established by the National Health Service Act, 1946, ch. 81, § 1.


For example, Denmark has significant social security benefits financed from general taxes, although it has only recently started introducing specific contributions. The specific contributions are imposed by the Labor Market Fund Law. Lovbekendtgorelse nr. 837 af 28.9.1994 am arbejdsmarkedsfonds. The rate was 6 percent in 1995 and will be raised to 8 percent by 1998. See also Danish Labor Market Supplementary Pension (ATP), Annual Report (1995).


ISSA produces a significant amount of literature on these issues and also publishes a regular international bibliography. See World Bibliography of Social Security (semiannual); Catalogue of ISSA Publications (semiannual).


This is the method throughout Central and Eastern Europe. See Restructuring Social Security in Central and Eastern Europe, supra note 7; Kopits, supra note 7.


This approach is used in many Western European countries. In the United States, the Social Security Administration was recently made an independent agency in the executive branch of the Government. Social Security Independence and Program Improvements Act, Pub. L. No. 103-296, 103d Cong., 2d Sess., 108 Stat. 1464 (1994). This is one of many institutional reforms that have taken place to the structure of social security schemes in recent years. Institutional changes are summarized in the ISSA newsletter, Trends in Social Security (published quarterly).


This is offered as a working definition. It reflects the OECD working definition of a tax in its Revenue Statistics, published annually. This defines taxes as compulsory general payments for public purposes and includes social security contributions. For further discussion, see the introduction to each year’s volume. This is not offered as a formal definition because the wide variety of arrangements that have been adopted make it almost impossible to generalize. For example, in Peru, employees are required to insure through either the state scheme or a private scheme. Insurance is mandatory, but public insurance is not. Therefore, the individual can choose between the social insurance scheme and private schemes.


For example, social security coverage of the self-employed is sometimes voluntary (e.g., Germany). 6 IBFD European Taxation, supra note 1, at 41.


For example, a state may impose contributions on most citizens, but may allow those not required to pay to be voluntary contributors. This happens in the United Kingdom, where “Class 3” contributions are authorized by the Social Security (Contributions and Benefits) Act, 1992, ch. 4, § 13. The Irish scheme contains a similar provision. Social Welfare Act, 1993, § 21.


This approach is often adopted in France, where the institutions running social security schemes are much nearer the market model than in neighboring countries, such as the United Kingdom and Ireland, which have uniform state-run systems.


Generalization again is difficult because some states have hybrid schemes. Compulsory insurance (e.g., that required of employers who are liable to employees for accidents at work) is widespread. It may occur through a state scheme or by obliging companies to take out private insurance, which may or may not be regulated. Premiums may be individual to companies, generalized across industries, or spread throughout all similar workers or across the entire working community.


This is true of some features of the French system where contributions are often closely related to risk because of the many different funds within the French compulsory system.


See Treas. Reg. § 1.901-1 (as amended in 1987)(USA).


For example, the limited provisions in Australia and New Zealand.


As in the United Kingdom, where the National Insurance Fund was created under the authority of the Social Security Administration Act, 1992, ch. 5, § 161.


As in Russia. Restructuring Social Security in Central and Eastern Europe, supra note 7, at 242.


As in France, where control is exercised through the Commission des comptes de la sécurité sociale. The French system pools contributions to fund current beneficiaries, but others, such as the Singapore system, hold individual funds for individual contributors. Under the Singapore arrangements, a contributor may only benefit to the extent of contributions made by or for the contributor. This leads the Singapore authorities to argue that their system is a savings scheme, not a social security scheme. See Letters, The Economist, May 11, 1996, at 10.


This is the term usually used by the ILO and ISSA. It is used in the United Kingdom, Ireland, and Canada. However, the United States refers to social security or payroll taxes (for constitutional reasons). The French equivalent term is cotisation. It may be noted that the U.K. income tax (one of the world’s first) used to refer to the income tax as a “contribution”! In the United Kingdom, the term “contributions under the Social Security Acts.” is sometimes used; see Social Security Contributions and Benefits Act, 1992, ch. 4, § 1, but the term “national insurance contributions” is popularly used by government departments, even though this ceased to be the technical term in 1973.


For example, the payroll tax in the United States. However, the formal name in the U.S. legislation is “contribution,” and the schemes are often referred to by the initials of the enabling legislation as FICA (Federal Income Contributions Act) and SECA (Self-Employed Contributions Act). The terminology used in the United States was adopted in 1935 specifically to avoid possible historical constitutional problems. This terminology raises a further issue in tax theory, namely, the identification of these taxes as direct or indirect. One relevance of this classification is whether the taxes are within the scope of double tax conventions. Article 2 of the OECD Model Tax Convention on Income and on Capital of 1992, reprinted in Philip Baker, Double Taxation Conventions and International Tax Law (2d ed. 1994), states that “payroll taxes” are within the scope of the model, but paragraph 3 of the commentary on that article makes it clear that it does not extend to social security taxes if there is a direct link between the contributions and the individual advantages received from the contributions. Alternatively, if they are indirect taxes, they may be within the scope of the General Agreement on Tariffs and Trade. See Richard A. Musgrave, Fiscal Systems 174 (1969)(noting that many considered payroll taxes to be indirect, in his view wrongly).


As in the United Kingdom. See Social Security (Contributions and Benefits) Act, 1992, ch. 4.


As in the Netherlands. See Gerrit te Spenke, Taxation in the Netherlands ch. 6 (1995).


As in the United States. See Steward Machine Co. v. Davis, 301 U.S. 548 (1937).


This approach has been considered by a number of states of the former Soviet Union.


This happens in Canada and Switzerland. See International Tax Program, Harvard Law School, World Tax Series: Taxation in Switzerland 94 (1976).


In Canada, both income tax and social security contributions are levied at both federal and provincial levels.


For two recent general policy discussions, see Kopits, supra note 7; Messere, supra note 1, at 167–84.


As in the United Kingdom. Although as noted in note 16, supra, a part is transferred to help fund health costs.


Or, alternatively, for groups of benefits. Multiple schemes are the most common form of system adopted in developed countries with full social security systems.


This is reflected in the way contributions are set in many European states, for example, in France, Germany, and Russia.


This can result in quite complicated patterns of contribution. For example, in Austria, in 1995, there are five elements in an employee’s contribution, four of which are also elements in an employer’s contribution. Of these, two are shared equally between employer and employee; one is heavier on the employer; one is heavier on the employee; and one (as noted) is borne only by the employer. In addition, employers make two other separate contributions to social funds. C&L 1995, supra note 1, at A-55.


In France, the contributions are to separate funds with individual powers to set contributions; no mechanism exists to unite the contributions.


As in the Netherlands. See supra note 40.


This is seen, for example, in France and Russia. In Russia, each area of funding is independent of the others.


Against this argument, the point must be made that settled, but separate, systems, such as those in Germany, can operate most efficiently. In part, this is because the necessity of the social security funds is widely accepted by those paying contributions, so that there is a high level of voluntary compliance and a low level of disputes. These factors may more than compensate for the absence of other forms of efficiency.


This is the most common pattern. If the contributions are not covered by the groups affected, there is an argument that they are not “proper” social security contributions, but are general taxes. See supra note 38 (noting the working definition in the OECD Model Tax Convention).


As in the Netherlands. See supra note 40.


Flat-rate contributions used to be common, but have tended to be replaced by income-related contributions. This chapter therefore assumes that contributions are income related rather than flat rate.


This is common practice in most Western European states. For surveys of the practices in Western Europe, see the works cited in note 1 supra.


The extreme case is perhaps the Netherlands, where the main rate of income tax on most taxpayers is the social security tax, and the “proper” income tax is at a much lower rate on those taxpayers, reaching higher rates only on the taxpayers earning the highest incomes. See supra note 40. The rates in 1994 were 31.075 percent social security tax and 7.05 percent income tax on the first f. 43,267, with no social security tax but a 50 percent marginal income tax rate applying to income immediately above that level. C&L 1995, supra note 1, at N-3.


See infra secs. II(A), III.


In the Netherlands, the social security tax is levied on most forms of income, including investment income. See supra note 40. In France, a special levy was placed on several forms of investment income and capital gains to provide extra revenues to support social security funds running into deficit. The Solidarity Contribution is imposed at 0.1 percent of turnover of companies whose turnover exceeds F 3 million. C&L 1995, supra note 1, at F-32.


A “social VAT”—in effect an extra levy on VAT—has been discussed recently in France. However, a compulsory Danish employment levy enacted in 1987 was ruled to be in breach of EU law by the European Court of Justice. Case C-234/91, Commission v. Denmark (Dec. 1, 1993), summarized in Proceedings of the Court of Justice and the Court of First Instance of the European Communities, No. 34/93, at 8 (Nov. 29-Dec. 3, 1993). The tax was essentially on the same basis as a VAT but did not comply with the general prohibition on EU states against introducing other forms of turnover tax. This effectively stops the adoption of social taxes in this form in the European Union.


As in Brazil, where companies have to pay a 10 percent levy (23 percent for financial companies) to the solidarity fund. C&L 1995, supra note 1, at B-56.


States sometimes choose shorter periods than this or have no periods at all. For example, until 1994, the Finnish system had no qualifying period of residence for Finnish citizens. Since 1994, the period is five years after age 16. Further, the period of residence now affects the level of benefit, with a citizen receiving one-fortieth of the total pension in respect of each year of residence. Previously, there was no such requirement. See Finland: Changes in Benefits, Trends in Social Security (ISSA), Nov. 1995, No. 9, at 10 (referring to the Finnish National Pensions Act, 1993). Similar moves have also taken place under the Italian scheme. The French scheme increased the period for a full pension entitlement from 150 quarters to 160 quarters (40 years) on an incremental basis. See France: Measures for Safeguarding Social Protection, Trends in Social Security (ISSA), Nov. 1995, No. 9, at 12 (referring to decree of Aug. 27, 1993).


This is the form of state scheme most commonly found in Western European countries, as well as in economies in transition. For a comparative survey, see Emmanuel Reynaud, Financing Retirement Pensions: Pay-As-You-Go and Funded Systems in the European Union, 48 Int’l Social Security Rev. 41 (1995).


This approach is used in the United Kingdom and Ireland. The U.K. contribution conditions are set out in the Social Security (Contributions and Benefits) Act, 1992, ch. 4, sched. 3.


A single payment may be enough to ensure coverage from the beginning of employment. Many states sidestep even this minimal problem by imposing the requirement to contribute on the employer alone.


The French scheme depends on a wide range of individual schemes. The linkage is achieved in individual schemes by reference to periods of insurance of individuals, which is directly related to their having employment that is insured through the appropriate fund. See supra note 63 regarding the recent change in the linkage rules.


The approach adopted, for example, in Denmark and Finland. See id.


There is no standard vocabulary. However, the U.S. Department of Social Security has published glossaries of terms in its Handbook of Social Security Around the World (annual).


See infra sec. II(L).


This used to be the case in countries such as China and the Soviet Union.


For example, those engaged in agriculture and fisheries are given separate treatment in Iceland. C&L 1995, supra note 1, at I–4; See also for Norway, 6 IBFD European Taxation, supra note 1, at Norway 43. In Belgium and Luxembourg, manual workers are treated separately from office workers. 6 IBFD European Taxation, supra, at Belgium 46–47; C&L 1995, supra, at B-27, L-44. In both cases, the rates for industrial workers payable by employers are higher than those for office workers. Spain has different minimum and maximum levels of earnings within which contributions have to be paid for different activities. 6 IBFD European Taxation, supra, at Spain 65–66.


See vol. 2, ch. 14.


See supra ch. 6, sec. III(E).


The laws of different nations view the key terminology as linked in different ways. The U.K. definition of “employed earner” is “a person who is gainfully employed … either under a contract of service, or in an office....” Social Security (Contributions and Benefits) Act, 1992, ch. 4, § 2(1)(a). Similarly, in French law, an employee is defined as “a person working for another person called the employer who supervises and controls his or her activity in exchange for which the employee receives a salary.” Code de sécurité social art. L 311.2. Both definitions rely on undefined references to general labor relations law.


See vol. 2, ch. 14.


The terminology used for VAT purposes is different, and the reference is to a person engaged in independent economic activities. See supra ch. 6, sec. IV(B).


This approach has been adopted in the United Kingdom and the United States.


See infra sec. II(C-N).


This has been experienced widely in the economies in transition, where there has been a sharp growth in alleged self-employment and a reduction of employment by large employers.


This approach has been adopted in some states of the former Soviet Union and has been discussed in the United States.


International practice is inconsistent. Countries such as the United States and the United Kingdom treat officeholders in the same way as employees. See supra note 75. Some states, for example, Belgium, Germany, and Ireland, exclude directors (in Ireland, controlling directors) from the schemes for employees, but may bring them within the schemes available for the self-employed. 6 IBFD European Taxation, supra note 1, at Belgium 45, Germany 42, Ireland 49. This recognizes the problem in the text in the most general way.


See supra note 82. The United Kingdom has a more specific rule under which director’s fees paid to the director’s firm (and not to the director personally) are not regarded as imposing contribution liability on the director, although for general purposes a director is treated in the same way as an employee. See Social Security (Contributions) Regulations, 1979, Regulation 19B.


This is recognized in Sweden, where no contributions are required from or for pensioners, but where instead a salary equalization tax is payable. 6 IBFD European Taxation, supra note 1, at Sweden 49, 50, 52.


With some exceptions; see, e.g., Fair Labor Standards Act, § 13 (as amended), 12 U.S.C. § 213(c) (USA)(in certain circumstances, exempting from the child labor provisions children working on farms or as actors).


An individual who has a deficient contribution record because no liability to contribute arose (e.g., because the individual was on an extended holiday or had private income) is sometimes allowed to contribute voluntarily in order to make good the deficiencies in the contribution record. For example, the United Kingdom has a category of contributions (Class 3) payable on a voluntary basis in these cases. Social Security (Contributions and Benefits) Act, 1992, ch. 4, § 13. There may be advantages to the state in allowing this form of additional contribution as it may remove any obligation to provide means-tested benefits for an individual.


If benefit entitlement in a scheme is subject to contribution conditions, consideration should be given to the crediting of notional contributions to the contribution record of an individual if the individual, for good reason, is not contributing. One example is the crediting of notional contributions during approved periods of education and training. If a contribution test for a benefit requires a minimum level of contributions in any period, or a minimum period of contribution, before benefit entitlement arises, then it may be necessary to allow for notional contributions in a scheme. Otherwise, for example, an individual who has just started work after an extended period of education or training may not be entitled to benefits. Similar problems arise for those who have been receiving benefits, for example, for illness or unemployment, for mothers on maternity leave, and for those who give up work to look after sick relatives. The provision of a notional contribution replaces the requirement that the individual pay a contribution either compulsorily or voluntarily.


For example, Ireland and the United Kingdom.


This is the French approach.


For example, agricultural workers or members of the armed forces. See supra note 72.


For example, Portugal, where the self-employed may choose which professional scheme is appropriate for them (on the basis of their activities). 6 IBFD European Taxation, supra note 1, at Portugal 61.


See infra sec. IV.


See infra sec. IV(B).


See supra note 49 (regarding the example of Austria).


This is the practice in the United States.


This approach was originally pioneered by Bismarck in Germany in the last century and has been adopted by a number of other European states. It was also the approach in the United Kingdom, confirmed in the Beveridge Report in 1946, Social Insurance and Allied Services, 1942, Cmd 6404, although the Government began to phase out the public contribution in 1979, only to reintroduce it recently to deal with a pending deficit in the National Insurance fund.


For example, Russia imposes all but 1 percent of the contribution on the employer, C&L 1995, supra note 1, at R-9. In Hungary, social security contributions are paid by employees (at 10 percent of their gross salary) and by employers (at 44 percent of gross salaries) up to certain limits. Id. at H-15. OECD members that impose the main contributions on employers include Canada and Finland. Id. at C-9, F-15-16. Current rates for each country are set out in 6 IBFD European Taxation, supra note 1, and C&L 1995, supra.


For example, Bulgaria and Poland. This is also widely used as an approach for schemes to compensate workers for injuries at work.


For example, this happens in Ireland and in the United Kingdom (in both of which the balance between employer’s contributions and employees contributions has varied from time to time). A particularly interesting example is the Netherlands, where the main burden was recently shifted from employers to employees. To compensate, employers were required to make an extra transitional payment to each employee. This payment is, for general purposes, treated as extra earnings, although its amount is laid down by the Government.


For example, Norway is for this purpose divided into five geographical areas, with different employer contributions in each area (0 percent in the far north). Argentina has abatements of contribution for employers in specific industries. The United Kingdom recently introduced a reduction in the employer’s contribution when the employer takes on someone who has been unemployed for over two years.


It is possible that such distortions might also amount to state aids or subsidies for international trade purposes. Treaty Establishing the European Community, art. 92. This may happen if the combined effect of the contributions and the benefits payable from those contributions involves a hidden subsidy of a national industry against foreign competitors or amounts to nationality-based discrimination. Such discrimination could be in breach of EU law. It is less clear whether it would fall to be reviewed under a nondiscrimination provision in a double taxation convention.


An example is the recent imposition in the United Kingdom of a social security contribution on employers, but not on employees, related to the provision by employers of company cars to their employees. This reflects the absence of a general charge on benefits in kind in the United Kingdom. The employer contribution gives rise to no benefit entitlement. The charge is imposed by a Class 1A contribution. Social Security Contributions and Benefits Act, 1992, ch. 4, § 10.


This is the original German model of social security, adopted by a number of other European countries.


This is the approach adopted for the U.S. federal social security funds.


In France, this was done through a new old-age solidarity fund levy. France: Measures for Safeguarding Social Protection, Trends in Social Security (ISSA), Nov. 1995, No. 9, at 13 (referring to law of Dec. 30, 1993). In the United Kingdom, it was done by reintroducing the subsidy to the National Insurance Fund from general taxation that had been abolished a few years before. See supra note 96.


These issues are raised in the final part of this chapter. See infra sec. V(C).


Denmark is the only OECD member now using a major flat-rate scheme for its Labor Market Supplementary Pension (ATP). The rate is not related to pay levels, but reflects the length of the work week of the individual. Those working over 27 hours pay the full premium (in quarterly installments), while those working from 18 to 27 hours pay two-thirds, those working 9 to 18 hours pay one-third, and those working fewer than 9 hours pay nothing. Public sector employees pay a lower contribution than others. ATP, Annual Report 4 (1993).


One justification for the distinction could have been the lower retirement age of women in many countries. However, in practice, men were required to pay higher contributions despite having the higher retirement age, indicating that this was not the reason for the discrimination. Another factor could be policies to encourage employment of females. If so, that policy aim has now been fulfilled in several developed countries.


For example, under the constitutions of individual countries. It is prohibited in the EU member states, inter alia, by the Treaty Establishing the European Community. EEC Treaty art. 119. The equal treatment policy is more fully implemented by EC Council Directive 79/7 of December 19, 1978, on the Progressive Implementation of the Principle of Equal Treatment for Men and Women in Matters of Social Security, 1979 O.J. (L 6) 24. Article 4 deals explicitly with contribution liability. Id. art. 4.


General statements about nondiscrimination in social security are set out in the European Social Charter, Oct. 18, 1961, art. 12, 529 U.N.T.S. 89; the United Nations International Covenant on Economic and Social Rights, Dec. 16, 1966, art. 9, 993 U.N.T.S. 3; and similar agreements.


Colombia imposes on the highest-paid employees an extra 1 percent contribution to the state Social Security Institute. C&L 1995, supra note 1, at C-61. More unusual is the approach of the United Kingdom, which has a single rate for most employees (with a lower rate for earnings below the threshold), but has five rates of contribution (including zero) for employers, dependent on the level of total earnings of each employee. Social Security (Contributions and Benefits) Act, 1992, ch. 4, § 9 (as amended from year to year).


The burden is less important if contributions are collected entirely from employers (or through employers) and are based on payroll levels. In such schemes, the introduction of an individual threshold adds to employer compliance costs.


In the United Kingdom, the income tax threshold and contributions threshold were set on different bases in 1975. Their levels are determined from year to year mainly on the basis of indexation, so the difference has been maintained.


Another alternative is the provision of a credit against income tax, which may be related to a greater or lesser extent to the amount of social security contributions. See USA IRC § 32 (regarding the earned income credit).


See infra sec. V(D).


Two examples are the U.S. federal scheme and the German schemes. The German schemes have higher upper earnings levels for the former west German states than for the former east German states. C&L 1995, supra note 1, at G-10; see also Messere, supra note 1, tbl. 8.1, at 183 (listing caps in the following OECD members: Austria, Canada, France, Germany, Greece, Japan, Luxembourg, Turkey, and the United States); see also Messere, supra note 1, at 177–78.


A variant on this is to exclude entirely those employees whose earnings exceed the cap, on the ground that they do not need compulsory insurance cover. For example, in Malaysia, employers are required to pay to the state scheme for employees whose earnings are below a set level and are required to take out local liability insurance for those above the level. C&L 1995, supra note 1, at M-28 (citing the Malaysian Employees’ Social Security Act, 1969); for current rates, see id.


For example, if directors of small but profitable private companies are regarded as employees, they may be in this position, as may highly paid expatriates working in low-wage economies. These problems are sometimes dealt with by excluding such groups entirely from the schemes. For example, Belgium treats directors as self-employed. 6 IBFD European Taxation, supra note 1, at Belgium 45. Ireland excludes controlling directors from the social insurance scheme. Id. at Ireland 49.


Caps may sometimes apply to employers but not to employees (as in the United Kingdom) or the reverse (as in Ireland).


This is expressly dealt with in the Netherlands scheme, which applies higher rates of income tax to those not paying the social security tax, to produce a combined rate of payment of both income tax and contributions on a progressive basis. 6 IBFD European Taxation, supra note 1, at Netherlands 43.


For example, in the Philippines. For current Philippine bands, see C&L 1995, supra note 1 at P-58-59. The banding system is analogous to the use of tax tables for individual income tax. See vol. 2, ch. 14.


This is subject to specific differences caused by the different natures of the two payments. This approach is used, for example, in Sweden. 6 IBFD European Taxation, supra note 1, at Sweden 49. This applies in the United Kingdom to contributions paid by the self-employed, where the amounts are expressly based on the assessed profits for income tax purposes. Social Security Contributions and Benefits Act, 1992, ch. 4, § 15 (regarding Class 4).


These questions have recently been the subject of unpublished, interdepartmental review in the United Kingdom. It was concluded that the scope for further combining these aspects of the tax and contribution schemes was limited.


See vol. 2, ch. 14.


For example, a disabled person may receive additional income tax allowances. If the person also receives a reduction in contribution liability, then the individual may also receive reduced benefits when it would be better for the individual to pay the contributions and receive the full benefits.


For example, an employee may be allowed to deduct any genuine employment costs, such as special clothing or travel costs. It is clearly most efficient to check on the extent that such claims should be allowed in the same way for both income tax and contribution liability.


See infra sec. V.


This has two awkward effects for those collecting contributions. One effect is regular changes in contribution rates. The other effect is that contribution rates are often expressed to one or two decimal points. For example, in Sweden, the 1994 rate for employees was 31.36 percent, and that for the self-employed was 29.75 percent. It also leads to small changes of rate from year to year.


See vol. 2, ch. 15.


Many countries, including the United States, Germany, and Sweden, base contributions on annual earnings (often linked to the calendar year).


This approach is used in the United Kingdom.


For example, directors are treated in the United Kingdom as employees. However, because they can often control the way in which their companies pay them (e.g., by loans, rather than earnings, or by erratic pay practices), they are assessed on an annual basis and not by relation to earnings periods.


See supra text accompanying note 131.


See infra sec. V(C).


However, a consequence may be that the individual is left with a reasonable level of total earnings, but with no obligation to contribute and possibly no entitlement to benefit.


ISSA has produced a series of useful booklets on these topics. See Collection of Contributions (1994), Enforcement and Compliance (1994), Maintenance of Records (1994), and Registration Procedures (1994). ISSA also provides training courses and support materials on these issues.


For example, it is assigned by the social security administration in the United Kingdom and the United States. Because of the social security and health needs of children, it is likely that the social security authorities will need a number for an individual before any requirement arises for income tax purposes. Further, it is unlikely that an individual will get by without needing a social security number at some stage in his or her life. Many individuals may never become income tax payers. The use of a single identification number may raise privacy concerns in some countries, as it has done in Hungary according to a recent ruling of the supreme court. In such cases, special protections for privacy may have to be provided in the legislation to meet these concerns.


Many of the issues are the same as those for income tax, VAT, and other official numbers, also discussed elsewhere in this book. See supra ch. 6, sec. II(D).


It is a growing sector in many economies and is large in many developing economies, particularly if account is taken of those engaged in farming and fishing. For example, Korea, in 1995, decided to bring farmers and fishermen within its compulsory schemes for the first time, increasing the number of scheme members by 2 million. The Republic of Korea: Proposal to Extend Compulsory Coverage, Trends in Social Security (ISSA), Apr. 1995, No. 7, at 9.


See supra sec. II(B).


For example, the Netherlands requires the self-employed to be members of some schemes but not others (health and disability insurance are excluded, as is unemployment insurance). Spenke, supra note 35, at 87–89. The self-employed who were previously wage earners in Denmark may, on a voluntary basis, continue their membership of the state supplementary pension scheme. ATP, Annual Report, supra note 107, at 4.


As, for example, in Korea. See supra note 139.


See, e.g., USA IRC § § 1402(c)(2)(c), 3121(b)(15) (regarding employees of international organizations).


For example, when a self-employed individual claims allowances for tax purposes that do not form part of the commercially determined profit of the business, such as a tax holiday.


For example, in Germany. See vol. 2, ch. 16, appendix.


See USA IRC § 1402(a), (b) (defining self-employment income). Compare the approach of IRC § 911(d)(2) (definition of earned income), which limits earned income to 30 percent of net profit in the case of a business in which capital is a material income-producing factor.


For example, tax holidays.


See, e.g., USA IRC § 6513 (regarding the timing of advance payments).


Some payments to the self-employed may be subject to withholding for income tax purposes. It may be possible to coordinate this with social security by allowing taxpayers to credit amounts withheld against their contribution obligation.


For instance, by using presumptive income or by accepting estimates or very simple records.


This is a reason why, as noted previously, the self-employed are only brought within voluntary social security schemes in some states.


See vol. 2, ch. 18.


The situation is less clear when the scope of benefit entitlement is wider than contribution liability. If all contributors are entitled to benefit, and in addition other persons are also entitled to benefit, then the link exists alongside a form of cross-subsidy. The cross-subsidy may be justified on another basis, for example, to ensure that, on the ground of solidarity, the few children not covered by the general scheme are not deprived of support. Other general issues also arise when there are significant numbers of migrant workers who regularly work outside their “own” state, but whose families remain at home, or who are likely to return at several points during their working lives and on retirement.


See OECD Model Tax Convention, reprinted in Baker, supra note 38, at Commentary on Article 2 (Taxes Covered), ¶ 3 (providing “[s]ocial security charges, or any other charges paid where there is a direct connection between the levy and the individual benefits to be received, shall not be regarded as [within the scope of the convention].”) Although the “direct connection” is less apparent in some schemes than was the case when this wording was first adopted in 1963, the general proposition has not been questioned. However, article 24 of the OECD model, unlike the rest of the text, is not subject to article 2 and applies to all taxes. Otherwise, the provisions of the model and conventions that follow the model apply only to income taxes. The author is not aware of any practical application of, or specific argument for, any part of article 24 applying to social security taxes and contributions.


The information exchange powers traditionally included in double taxation conventions do not apply to social security contributions. See supra note 154.


Residence for income tax purposes will, in the case of disputed resident status, be determined under the rules of the relevant double taxation convention. This makes the home of the individual the residence rather than the place of work.


Permanent establishment is defined in article 5 of the OECD and UN Model Tax Conventions in similar but not identical terms. OECD Model Tax Convention on Income and on Capital of 1992, art. 5, reprinted in Baker, supra note 38; UN Dep’t of Int’l Economics & Social Affairs, UN Model Double Taxation Convention Between Developed and Developing Countries, art. 5, UN Doc. ST/ESA/102, UN Sales No. E.80.XVI.3 (1980). The same pattern of definition occurs in most individual tax conventions. A similar concept may also apply for other taxes. See supra ch. 6, sec. II(A).


Regulation 1408/71 of June 14, 1971, on the Application of Social Security Schemes to Employed Persons and Their Families Moving within the Community, 1971(II) O.J. Special Edition (Dec. 1972), as amended. For the text and analysis, see David W. Williams, The National Insurance Contributions Handbook, FT Law and Tax at ch. A5, part D.1 (loose-leaf). For an account of the rules in the context of relevant EU law, see Stephen Weatherill & Paul Beaumont, EC Law at ch. 18 (1993).


Council Regulation 574/72 of March 21, 1972, Fixing the Procedure for Implementing Regulation 1408/71 on the Application of Social Security Schemes to Employed Persons and Their Families Moving within the Community, 1972(i) O.J. Special Edition (Dec. 1972), as amended.


Regulation 1408/71, supra note 158, art. 13.


Id. art. 14. The concept of habitual residence is one of EU law and is therefore subject to determination by the European Court of Justice. See, e.g., Case 76/76, Di Paolo v. Office national de l’emploi, 1977 E.C.R. 315.


Regulation 1408/71, supra note 158, arts. 13–17.


Networks of bilateral agreements are growing slowly, but are far smaller than the equivalent networks of double tax conventions. For example, the United Kingdom, which has the largest network of double tax conventions (over 100), has one of the most extensive networks of social security conventions, covering reciprocal arrangements with the other 16 members of the EEA, and bilateral agreements relevant to contributions with the following other states: Barbados, Bermuda, Cyprus, Gibraltar, Israel, Jamaica, Malta, Mauritius, Philippines, Switzerland, Turkey, the United States, and Yugoslavia (plus British territories such as Northern Ireland and the Channel Isles where internal jurisdictional limits require agreements). It has a number of other social security agreements (e.g., with Australia), which have no provisions dealing with contributions. Poland and other countries with economies in transition are now negotiating these conventions. Poland has concluded several new conventions, including those with Belgium, France, and Germany.


Practice varies. In 1990, social security contributions could be fully deducted as expenses against income tax for employees in the following OECD members: Austria (subject to a ceiling), Belgium, Denmark, France, Germany (subject to a ceiling), Greece, Italy, Japan, Luxembourg, Spain, Switzerland, and Turkey. Partial deduction was allowed in Canada, Ireland, and Portugal. Deduction was not allowed in Finland, Iceland, the Netherlands, Norway, the United Kingdom, and the United States. See Messere, supra note 1, tbl. 10.7 at 272–73.


For example, a deduction may be denied to the employer for the cost of certain fringe benefits provided to employees. Although not ideal, this may be more or less acceptable for income tax purposes (see vol. 2, ch. 14) but, as the text indicates, will create a problem for the contributions base. This illustrates the importance of thinking about the income tax and contributions together.


National practice varies (as does treatment of different benefits within a state). In the OECD, most countries tax public pensions, while treatment is inconsistent with unemployment pay. Most countries do not tax invalidity payment, and almost all refrain from taxing family benefits. Some countries tax them in part, e.g., USA IRC § 86. Canada and the Netherlands have few exemptions from tax, while Turkey has a considerable number. See Messere, supra note 1, tbl. 10.1, at 264–65.


This is common, but not universal practice. Id. at 168. Instead, in some states the pensions are not taxed (e.g., Turkey). A problem arises when the contributions are made to a pension scheme based in one state by a contributor liable to income tax in another state. In such cases, the tax authorities often refuse to recognize the contribution as deductible. The European Court of Justice, in Case C-80/94, Wielockz v. Inspecteur der Directe Belastingen, 1995 Simons Tax Cases 876, ruled that in some cases this could amount to unlawful discrimination under EU law.


This is national practice throughout the OECD. See Messere, supra note 1, tbl. 10.1, at 264–65.


National practice is not consistent.


The terminology used in this section is derived from English public welfare economics literature. An early account is A.R. Prest, Social Benefits and Tax Rates (1970). An excellent series of essays on the topic based on a seminar bringing together officials and commentators is in Taxation and Social Policy (Cedric Sandford, et al., eds., 1980). The term “poverty trap” is sometimes used as a generic term to cover all forms of effect of the kinds described here. The reference to trap indicates a value judgment. The adoption of the convenient phraseology of that literature in this account is not meant to imply an adoption of those value judgments by the author. A neutral description of the trapping effect is that the interaction of each of the separate rates and thresholds for taxes, social security contributions and benefits, and social welfare receipts affecting an individual with little or no earned income produces a net negative marginal increase of total income to, or benefit for, that individual from all sources when adjustments are made to all receipts and deductions to take account of a unit of increase of earnings of the individual (poverty trap) or for a unit of replacement of benefit or other income with earnings (unemployment trap).


The discussion under this head is based on the U.K. experience, in which the rent benefit (called housing benefit) and the help for poor families (known as family credit) are tapered in the way illustrated. This has reduced the highest implicit tax rates in the system to below 100 percent, but still leaves them far higher than the highest express income tax rate (40 percent). It is still widely regarded as unsatisfactory.