A Database on ILO’s ‘Cost of Social Security’Mathias Maucher
The following is designed to explain the concept of a joint ILO/EURODATA project. The common intention is to build up a data infrastructure for comparative welfare state research based on ILO’s international inquiry ‘The Cost of Social Security (COSS)’. This survey, started in 1949 and still continued, reports on financial transactions—i.e. categories of receipts and expenditure—of social security institutions (schemes, funds, agencies, programmes). Our main aim consists in producing a CD-ROM comprising both a commented edition of the original questionnaires and a machine-readable version of basic tables already published. It should cover most of the Western and Central European countries.
Serving as a starting point for sketching out the project’s purpose and context, the first section draws on several aspects of social security financing that raise attention and controversies in various contexts. In a second step the article sketches out the design of COSS and tries to illustrate briefly its potential for cross-country research on social security institutions. One paragraph characterizes briefly other comparative databases or information systems on social protection in Europe compared to COSS. In the third section, the idea of the project is developed by describing concepts and contents of the database still under construction. It comprises an overview of already realized and some projected components. The article concludes with an outlook on work planned by the ILO and EURODATA.
1 Challenges Linked to the Financing of Social Security
Debates and controversies related to the already long and widely discussed ‘crisis of the welfare state’ rather frequently deal with or at least touch upon the topic of financing welfare state provisions. In this context, aspects such as volume of financial transactions (either globally dedicated to social protection or reserved to individual programmes) or mode of funding of social security schemes and policies can be identified as common points of dispute in all relevant areas. Four major perspectives and challenges of social security financing are pointed out in the following paragraphs and should facilitate a better comprehension of context coordinates of ILO/EURODATA’s CD-ROM project on COSS as well as options of analyses based on it. Taking into account these different aspects, social security financing clearly constitutes one main thread linking the political, societal and economic debates on welfare state reform and redesign.
Cost Containment Policies and Redefinitions of the Role of the State
As a consequence of continuously increasing quota of government spending related to social protection and indebtedness of public budgets, in all European countries several strategies from a whole set of austerity policies have been on the political agenda, negotiated in the competent arenas and implemented mainly from the mid-1980s onwards. They generally aim at safeguarding the (future) ability to act on all levels of government. Following concepts recommending as the main guideline a restricted state responsibility and a limited role of (central) government regarding societal welfare, retrenchment policies have implemented rules reducing—at least in the medium and long term—the share of state or parafiscal agencies in (co-) financing or administrating social security. The most prominent example for this strategy is undoubtedly the United Kingdom in the 1980s.
When discussing welfare state reform and arguing about the appropriate public-private mix of welfare provision, scholars in the fields of social policy and third sector research as well as expert committees have put on their agenda the role of the state or other semi-public agencies in social security financing, too. Those who are in favour of less state intervention propose two often complementary strategies for the sustained reduction of the government’s involvement in social security financing and offer: 1) legislating reduced obligations of direct benefit or service provision or subsidization by (central) government agencies (e.g. in the case of the unemployment assistance in Germany, 1996) or 2) enacting market mechanisms that enable competition amongst providers who are able to meet fixed quality standards subject to public contol, e.g. implemented for benefits in kind in the German long-term care insurance (Pflegeversicherung), enacted in 1994.
Adapting Systems and Benefits to Changing Context Parameters
Reconsidering institutions of the welfare state in the light of changed family structures, ageing populations or new patterns of paid and unpaid work, social security financing moves to the centre of attention for consultancy boards, planning agencies or scientists tackling challenges of equity, distributional justice and system efficiency on a macro-level. Let’s only name three of them: 1) What could be—in a given period—the appropriate level of total influx and outflow of money and assets dedicated to social security in general? 2) Based on which group- and risk-specific estimations of incidence should this amount be broken down into subsystems or distributed amongst specific programmes of social protection? 3) On which criteria can transfers between schemes, funds or agencies—implying (sometimes ‘hidden’) redistribution amongst groups—be justified as a permanent source of receipts? ‘Re-engineering’ the design of schemes and funds or evaluating the focus on four major possibilities for social security schemes or measures: 1) legal enforcement of membership in scheme or access to fund (compulsory vs. voluntary); 2) modalities of financing schemes or funds dedicated to social security (pay-as-you-go vs. capitalization systems); 3) eligibility conditions delimiting the scope of specific schemes, mainly via rules of adherence or the existence of contribution ceilings; 4) entitlement conditions restricting the eligibility for monetary transfers, social services, benefits in kind or fiscal advantages. Changing these parameters will produce direct consequences for volume and types of social protection financing.
Exonerating Costs of Labour
Possible effects of given levels of protection and existing institutional designs or configurations may be treated as isolated factors. To a far greater extent, however, are impacts on other fields explored, e.g. when interrelations of social protection costs with labour market dynamics and employment volume outcome are discussed. This holds especially for countries that have implemented employment-based social insurance institutions as major pillars in their welfare system, where levies on wages and salaries represent the central source of funding. According to a strand of reasoning generally accepted, the common pattern of growing contributions and taxes burdened on the production factor labour will, in the long run, both increasingly threaten the ability to successfully compete on international markets and favour investments in less labour-intensive production processes.
Not surprisingly, measures aimed at slowing down increases of statutory contributions or even causing their cutback are discussed by the parties to the wage/salary agreement and society. Diminishing the financial burden on labour costs has been—in the last ten to fifteen years—one crucial topic and aim in most of the European countries. This is even more true for countries standing out due to augmented funding by social contributions in a long-term perspective, as it is the case with Germany. Graph 1 clearly reflects this. It is based on values taken from the original documentation, i.e. the questionnaire submitted to the ILO by the German Ministry of Labour and Social Affairs (Bundesministerium für Arbeit und Sozialordnung). Within the last two decades, the focus has been set differently in the different European countries, depending on the prevailing economic paradigma, the balance of power and opportunities for change: 1) As a rule, the accent has been on measures reducing employers’ costs. As a by-product, legislative amendmends increasing the take-home pay of persons in the labour force came into force, too. 2) Instead of extending the contribution basis by enlarging the personal scope of a scheme or fund, an alternative strategy could consist in a shift to a different calculation basis by (partially) replacing employers’ social contributions by value added payments to social security. To our knowledge, this concept, implying a reorientation towards a tax-transfer system—turning away from the concepts of equivalence, reciprocity and social insurance—has, however, to date not been applied with blanket coverage in major social insurance schemes of EU member countries. 3) Finally, the aspect of appropriate shares for apportioning the financial burden ear-marked for social welfare cannot be separated from the question of (implicit) physical or psychological ceilings for individuals and enterprises.
To some extent the tendency of increased funding of the German social protection schemes, funds, agencies and programmes by means of social contributions can, how however, be attributed to inconsistencies with regard to the reported apportionment of forms of receipts. It might differ from the factual repartition of funds. This problem, however, cannot or only partially be assessed by falling back on information provided by COSS. It rather necessitates in-depth knowledge about the financial constitution and management of social security budgets. This becomes evident if one focuses on two categories and the development of their relative share over time: 1) From 1958 onwards, ‘participation of other public authorities’—in the case of Germany this refers to the states (Länder), districts (Kreise) and municipalities (Gemeinden)—has been merged with the category ‘state participation’ for the purpose of COSS. Public welfare agencies on the regional and local level, however, have always co-financed social protection expenditure, especially in the field of social assistance benefits. 2) ‘Transfers from other schemes covered by the inquiry’ as depicted in Graph 1 seem to have lost any significance from 1982 onwards. Divergences with regard to this type of receipts can nevertheless stem from an inconsistent treatment or definition over time. The above-mentioned transfers might be given as unbalanced accounts, i.e. double counts are possible, or stated as balanced accounts, i.e. already cleared up for ‘intra-system’ transfers, i.e. transactions between schemes classified under the same heading. This, e.g., is the case with transfers (mutually) involving the different German pension funds, i.e. the pension insurance for workers (Arbeiterrentenversicherung), salaries (Angestell Angestelltenrentenversicherung), mining personnel (Bundesknappschaft), and persons working in agriculture (Alterssicherung der Landwirte).
Improving Efficiency and Effectiveness, i.e. Welfare Outcomes
Studies screening policy effects (i.e. positive or negative incentives of social protection expenditure or targeting of programmes) and the outcome of eligibility or entitlement conditions and benefit regulations on the level of individuals and households, i.e. the programme effectiveness, are often underpinned by aspects of financing, too. Evaluations of the extent to which the restructuring of a given tax/benefit/service system can guarantee a sufficient disposable household income, minimise unemployment, invalidity or long-term care or avoid social exclusion traps, can hardly be carried out without taking into consideration the distributional implications for the groups or agencies financing them .
2 ILO Inquiry into the Cost of Social Security (COSS)
This section deals with the concept and scope of ILO’s inquiry ‘The Cost of Social Security’. It briefly assesses how a machine-readable version of COSS—including both original documentation and basic tables—could improve the databasis for comparative research on social protection (financing), especially when concentrating on institutions of the welfare state. The paragraph concludes with an overview of publications based on COSS and already realized forms of data dissemination.
Since 1949 the ‘International Inquiry into the Cost of Social Security’ has been carried out by the ILO in the form of a questionnaire-based survey. Responses are normally given and/or coordinated by the national ministries in charge of social security and/or the national statistical offices. According to the inquiry’s concept—following a combined functional and institutional approach—all social protection schemes and welfare provisions meeting the definitions given in Graph 2 should, in principle, be included. The purpose is therefore to obtain aggregate data on the financial transactions of national social security programmes for nine risks covered by eleven types of schemes if three conditions are met. The contingencies covered correspond to those defined under the different parts of the ILO Social Security (Minimum Standards) Convention, No. 102, voted in 1952. Explicitly excluded are the four social policy functions active labour market policies, housing, creation of wealth and education, partly covered by other cross-country data and information sources. Nevertheless, a certain deficit is the fact that neither from the published explanations nor the original documentation submitted to the ILO can one normally conclude to which extent it had been possible to realize this ambitious concept for a given country and inquiry or year. However, having screened the available material as well as the already published basic tables for several European countries there can be no doubt that COSS will present a very valuable source for cross-country analysis on financial transactions of the major systems, programmes, agencies and functions of social protection over the last five decades. It can be a key source for comparative research on institutional stablility or innovations including this perspective of social security funding.
Starting with Inquiry XIII (1984–1986), the concept has been extended to incorporate—as a regular component—data on the personal scope of social security schemes. The questionnaires for the ongoing inquiry 1994–1996 for the first time apply an adapted methodology close to the ESSPROS system in order to test its worldwide applicability (cf. ILO 1997a).
The categories for types of receipts and expenditure of the basic questionnaire are presented in Table 1. Except for two amendmends, they have formed the basis of all inquiries since 1949: 1) In a revised version, applied since 1951, the categories ‘transfers from/to other schemes covered by the inquiry’ have been included; 2) Starting with the eighth inquiry, i.e. in 1967, the expenditure category ‘benefits in kind’ has been subdivided into ‘medical care’ and ‘other benefits in kind’. This stability (principally) facilitates comparative analyses over a long span of time. As an illustration of this relative stability, Graph 3 documents the long-term development of social protection expenditure distributed by type for France, referring to data in the basic tables.
However, the breakdown of forms of receipts and expenditure proposed by COSS generates several problems linked to the exact attribution or classification of financial flows. Two examples should illustrate this problem: 1) As a rule, entries in the category ‘contributions from insured persons’ have to be based on estimations if different rates (by wage level, branch of insurance or occupational status, if a scheme for wage and salary earners covers the self-employed or unemployed, too) are applied in respect of different categories of insured persons or if voluntary insurance is possible (because the insured person then has to bear the entire contribution). 2) State participation should include all payments made to finance social security from the government sector, i.e. subsidies, social contributions—in its role as employer—in special schemes for public employees, administrative expenses. In countries with a federal constitution, ‘state participation’ embraces the amounts spent by the federal government, whereas—according to COSS’ logic—receipts from constituent states, provincial governments and municipalities should be entered under the heading ‘participation by other public authorities’. In practice, disentangling contributions made by the government sector may nevertheless not be an easy task or is simply not(totally) applied when filling out the questionnaires, as the German example already demonstrated.
Finally, some conceptual deficits and limitations should be mentioned. A comprehensive analysis of the financial, institutional and organizational features of welfare state development has to remain incomplete if tax reliefs, occupational social protection arrangements or privately contracted social welfare are not taken into account. Whereas the first and last aspect are disregarded completely by COSS, the second pillar of social security is integrated only if social advantages are codified for the gainfully employed in special sectors/branches or certain categories of the labour force. Understandably, efforts to cover enterprise-based schemes have never been undertaken, a sector for which estimations in national surveys are difficult to obtain. Systematic cross-classifications of schemes and risks for all types of receipts and expenditure are generally not possible. Finally, due to its focus on welfare state institutions, COSS is not an appropriate basis for analyses on the mechanisms of income redistribution and their outcome on the level of households or individuals
The results of Inquiries I to XIV, launched between 1949 and 1993, are available in printed form only. This holds both for the basic and comparative tables and the data obtained from questionnaires for the supplementary inquiries. Whereas the basic tables give a detailed account of the financial operations of those social protection schemes covered by the inquiry in each country for (as a rule) every third year, the comparative tables are dressed up with highly aggregated data and central indicators for each consecutive year ranging from 1949 to 1989.
The national datasets representing the results of Inquiries XV to XVIII (covering the years 1990 to 1993) are—to a varying extent for the different countries—available via the internet (http://www.ilo.org/public/english/110secso/css/). The same holds for the data on financial transactions by type 1960–1989 published in the comparative tables (http://lissy.ceps.lu/compwsp.htm). These data have been made machine-readable within the project ‘The Welfare State in Comparative Perspective: Determinants, Program Characteristics, and Outcomes’, directed by Evelyne Huber, Charles Ragin, and John Stephens, and commonly funded by the Institute for Policy Research at Northwestern University/Chicago, the Department of Political Science at the University of North Carolina and the Luxembourg Income Study (LIS), located at the CEPS, Differdange.
Financial and Structural Innovations Covered by COSS
Since the 1950s it has been possible to witness some considerable developments in the fields of both social security financing and administration. As regards procedures mobilising and allocating financial flows, one could mention progressive refinements of financing techniques, such as a widening of the contribution base or an introduction of new methods of levying taxes or contributions funding social security. Modifications in the structure of sources of financing result from these innovations, irregularly adjusted contribution or tax rates and reorganizational processes in social security administration. Since it is not an infobase on laws and other regulations on social security financing and administration, it goes without saying that COSS does not directly report on the first two points, but indirectly reproduces shifts in all three factors via the reported system structure.
In the last five decades, several forms of restructuring social protection administration took place, too. 1) One consisted in enlarging or even generalizing the personal scope of schemes. This tendency towards universalization can be observed, e.g., for many child benefit/family allowances schemes, well demonstrated by the Finnish (Lapsilisä) or Norwegian (Barnetrygd) child benefit. Introduced in 1948 or 1946 respectively, they both replaced or extended rudimentary benefits for specific categories of families. The Finnish child allowance provided benefits from the first child onwards from the beginning, whereas in Norway this monetary transfer has only been paid to families with one child since 1970; before that time, only single parents with one child had been entitled to benefits. 2) The coverage of contingencies formerly not provided by welfare state institutions or provided in another form has been a second tendency. The establishment of new or independent schemes for long-term care or parenting policies falls under this category. A good example is the introduction of a compulsory long-term care insurance (Pflegeversicherung) in 1994 or the enactment of a federal child raising allowance (Bundeserziehungsgeld) in 1986 in Germany. 3) Thirdly, special funds or compensation mechanisms have been established, which financially relieve groups with structural disadvantages or special needs to a certain extent. Referring to the concept of ‘national solidarity’, this technique—implying a (partial) ‘collectivization’ of costs—is commonly applied with regard to old-age and sickness funds or insurances for farmers, miners, repatriates or poor elderly. Here the French social protection system is an exemplary case if one thinks of Budget Annexe des Prestations Familiales / Sociales Agricoles (established in 1949), Fonds Spécial d’Allocation de Vieillesse (1955), Fonds National de Solidarité (1956) or Fonds d’Allocation Vieillesse aux Rapatriés (1964). 4) Incorporating systems of social protection formerly exclusively administered by employers’ and employees’ organizations into structures open to state control (e.g. for the risk unemployment) has been another option. In Sweden, e.g., the unemployment insurance organized by trade unions and their benefit societies from 1974 onwards has been supplemented with a state-administered unemployment assistance scheme for those not enrolled in a benefit fund or members of benefit societies not entitled to compensation benefits (Olsson, 1986: 22). The insurance societies themselves ‘operate in close connection with the trade unions and under the supervision of the National Labour Market Board (Arbetsmarknadsstyrelsen)’. A related form is the enforcement of legislation framing social security schemes set up by employers’ associations and trade unions. The introduction of a contractual unemployment insurance scheme (Assurance chômage) for France in 1958 or of trade union funds (Caixas Sindicais) in the corporatist social welfare system of Portugal before 1962/1974 are good examples of this. 5) Finally, altered administrative responsibilities implying changes in the welfare mix may also stem from contracts with private insurance companies (as in the case of schemes allowing to opt out the general pension scheme in Great Britain) or from the growing importance of enterprise-based arrangements offering a supplementary compensation.
Institutional Stability and Change Reflected by COSS
The data collected in the framework of ILO’s International Inquiry ‘The Cost of Social Security’ reflect in an appropriate manner important shifts in social protection financing as well as an emerging or declining relative importance of certain risks related to societal change and legislative amendmends. Using data from the basic tables, Graph 3 shows the long-term tendencies toward increased relative shares for old age, sickness and unemployment as well as the decreasing importance of benefits improving the income situation for families in France. These developments result from the interrelated influence of demographic shifts, extended or restricted personal scope of different schemes and altered access conditions (e.g. the duration of insurance prior to claim of benefit; income thresholds excluding an entitlement) and benefit regulations, such as generosity or indexation of benefits.
Even if to a somewhat lesser extent, this data collection may also serve as a starting point for tracing back innovations in social security organization and administration. A rather complete comprehension can, however, only be achieved by referring to other comparative or national sources of information. Within the database project we try on the one hand to provide at least some basic information on central institutional characteristics, including administration, by using comparative typologies, cf. Graph 4. Additionally, we intend to document mergers or splits of institutions and their establishment or dissolution via coloured markings.
Focusing on Institutions in (Comparative) Social Policy Analysis
A pattern of social policy research fitting well with COSS is, obviously, the institutional approach. It concentrates on institutional and organizational features of welfare states, stressing variations in structures in a comparative perspective. Assuming a long-term perspective it aims at explaining them by linking earlier national or intra-scheme divergences with general developments on a macro-level. Hereby one can account for country-specific lines of development and path dependencies. In the context of an institution-oriented analysis of the welfare state, global and institution-specific volume and mode of social protection financing and consumption are central issues at stake. First, with regard to their interrelation with context influences from different sectors (demography, economics, labour market, political system, regulative framework). Second, regarding their redistributive effects, cutting across income groups and shifting burdens horizontally amongst groups in a comparable financial situation (e.g. from single persons to families) or intertemporally between generations (e.g. in pension insurances having recourse to pay-as-you-go mechanisms as a mode of capital accumulation). This second aspect touches upon the question of how welfare state institutions structure group-specific and individual life chances with regard to two important dimensions: central mechanisms of scheme design and benefit regulations. A data infrastructure based on COSS, comprising aggregate information on financial transactions of social security schemes and funds, therefore promises to be a valuable instrument of cross-country analysis on welfare state institutions compared to the time-series datasets already available.
COSS and other Cross-country Sources for Comparative Welfare State Research
‘The Cost of Social Security’ is one in a number of other comparative collections of time-series or information systems on social protection institutions or benefits in Europe listed in Table 2. These data- or infobases—mainly collected or produced by supranational or intergovernmental organizations or their subdepartments—may share to a varying extent a body of common or similar approaches and contents. However, clear differences exist with regard to concept, variables, geographical scope, and technical realization. In-depth juxtapositions of two or three different concepts respectively are given, e.g., in ILO 1997a, US DHHS: SSA/ISSA 1997 or OECD 1996.
* The Council of Europe in his publication Comparative Tables of Social Security Schemes in Council of Europe Member States Not Members of the European Union (1982–1996/biannually) applies the concept of MISSOC (overviews on nine risks and chapters on organization and financing) to report on social protection schemes of European countries not being EU members
A better understanding of the financial transactions and the institutional structure of social protection schemes, agencies and funds requires an exploitation of qualitative cross-country and national sources, too. Table 2 lists five cross-country information systems describing main features of system organization and funding. Except for the ESSPROS methodology they all incorporate central institutional regulations, i.e. eligibility criteria applied by a given scheme, entitlement conditions for specific programmes and benefit regulations for individual measures. Apart from NOSOSCO, no infobase, however, includes tax benefits, a deficit impeding a more comprehensive analysis of social policy institutions and outcomes if one knows about the prominent role of fiscal advantages in many welfare states. Several classifications applied by these information systems will find their way into the categories in order to describe the institutions covered by COSS, cf. Graph 4.
3 COSS on CD-ROM
Design and Scope of the Database
Referring to Graph 4 we will describe design and functionality of the ILO/EURODATA database under construction.
The project’s central aim consists in producing a comprehensive machine-readable edition of sources connected to COSS for the period 1949 to 1993. For this purpose both the orginal documentation submitted by the national ministries in charge of social protection and time-series data already published in form of basic tables will be processed and documented. We intend to include all EU member countries plus Norway, Switzerland, Czechoslovakia (including the succession states Czech and Slovak Republic), Hungary and Poland.
All parts arranged above the broken line in Graph 4 make up the core of the database on COSS. First priority is given to reproducing the time-series data contained in the original documentation and the basic tables, framed by a comprehensive documentation apparatus. We aim at making machine-readable all data and information covered by the components listed in the three rectangles on the top right-hand corner of the diagram. Additionally, parts of the documentation on comparative datasets as well as some background information on COSS—listed in the three boxes in the upper left part of the overview—have already been set up. In a second phase efforts will be made to give a detailed characterization of important institutions figuring in original documentation and basic tables. The components to be worked on during this phase are graphically represented by the three rectangles below. In a third step we would then be able to reach beyond COSS by working on comparisons with other cross-country sources or by supplementing—for a limited number of countries—ILO’s inquiry design for COSS with classification systems for national social budgets/accounts.
Data and Documentation
Commented and partly restructured datasets (1.) as well as the documentation on data, inquiries, questionnaires and institutions (2.) make up for the core part of the database on COSS 1949–1993. All of them normally report on a specific year. Additionally, efforts are made to grasp major lines of development, underlining COSS’ character of a repeated cross-sectional inquiry by documenting shifts in the institutional structure of a country’s social protection system (3.). This holds for ‘real’ changes as well as for those going back to redefinitions of the aggregate categories by either the national agencies submitting the questionnaires or the ILO itself, an aspect partly covered by the correspondence tables (4.). Additionally, the database contains overviews of the history and concept of COSS as well as of the publications issued on the basis of this international inquiry.
1. Machine-readable datasets will be made available in two forms: First, by reproducing, documenting and annotating the data and explanations contained in the original documentation and the basic tables—this will be our major task; seconod, by restructuring them—at a later stage in the work process—, especially with regard to the order in which the social protection schemes are presented. These reorganized datasets should on the one hand enable to better link country-specific data over time. On the other hand they could improve their comparability when cross-country typologies are applied.
2. A documentation apparatus has been designed to improve the understanding of data and institutions. It consists of the components ‘General Country Information’, ‘Questionnaires’, ‘Datasets’, and ‘Institutions’. Additionally, background information on the inquiry and technical aspects (cf. part ‘Methodology, Terminology, Technical Aspects’) is provided.
3. Reporting on institutional stability and change over time: A non-appropriate format prevents us from providing a detailed documentation of the institutional diversity in a given year or to retrace change and stability of social security schemes, funds, agencies or programmes in this article. Combining elements of a universal and a fragmented welfare system, nevertheless continuously progressing towards the first pattern since World War II, e.g. The Netherlands could, however, serve as an interesting example in this context. With a high consistency over time, the original documentation reports on about 30 different schemes, many of which functioning for decades. This, nonetheless, does not preclude the fact that major legal amendments, modifying the system structure, are clearly reflected in the (reported) institutional structure. Two good examples are 1) the general disability benefits act (Algemene Arbeidsongeschiktheidswet), coming into effect in 1976 and resulting in the establishment of a new universal scheme; 2) the general family allowances act (Algemene Kinderbijslagswet), laying the foundation for converging five different child benefit provisions into one scheme in 1980.
4. Correspondence tables serve the aim of juxtaposing the institutional structure given in the basic tables or in the original documentation respectively. Their purpose consists of sketching out differences with regard to the degree of detail to which the different schemes, funds, agencies or programmes are documented. Where possible, they should also facilitate a tracing back of the rules applied by the ILO when either publishing the institutional structures unchanged or when restructuring the original tables by attributing institutions to newly aggregated systems. Reflecting a snapshot of Portugal in 1966, Table 3 gives an example of correspondence or divergence of institutional structures. It compares those published in the basic tables on the one hand with those reported on in the original questionnaires on the other hand.
Users of cross-country databases and information systems on social protection share the knowledge that it can be hard to retrace back data or qualitative information in national sources. This difficulty can partly be attributed to the application of typologies and procedures of standardization developed predominantly for comparative analyses. Given the basic tension between optimal comparability and the need for case-oriented studies, they tend to cut across country specifics. Aggregation mechanisms that are not made completely explicit are another common obstacle when one needs to refer back to national institutions. Finally, a last challenge is the terminology, with concepts or classifications having a different meaning in different national contexts. A cross-national characterization of social protection institutions using comparative categories with without disregarding country specifics therefore seems to be a strategy for enhancing the comprehensibility of data.
All the problems mentioned turn up when reproducing and documenting COSS, too. One task will therefore be to collect and combine—to a realizable extent—information not contained in COSS, but relevant for a better understanding of financial transactions of major systems and country-specific institutional structures or configurations. This ‘information system’ could then contribute to making data on social protection more context-embedded.
Graph 2 only sketches out the context of COSS. The major institutions reported on, however, are also influenced by long-term macro-developments and reshaped by social, labour and family law. This regulatory framework institutionalizes (provisional) agreements along lines of conflict in society, politics and on the labour market at a given point of time. The description of social policy institutions intends to frame the time-series in a comparative perspective, allowing, however, for an integration of country-specifics, too. It could consist of a triangle comprising the aspects financing, organization/administration and personal scope of institutions. Given the ‘historical’ orientation of the database project, we will always be confronted with the challenge of handling institutional instabilities adequately due to establishments, dissolutions, mergers and splits of schemes, funds, agencies or programmes. A ‘solution’ for this challenge has still to be worked out.
At the moment we are only able to dress up a draft concept for the purpose of characterizing these schemes, funds, agencies or programmes of social protection covered by COSS. The following headwords therefore only propose some central features which have to be transformed into variables and variable categories for the purpose of establishing the infobase on social security institutions. a) Financing: 1) origin of resources; 2) apportionment of financing obligations (percentages; initial and final distribution); 3) mode of capital accumulation (pay-as-you-go-system; capitalization); 4) regulations concerning social contributions, taxes allocated to social security or state subsidies (the basis funds are levied on; calculation mode; ceilings; exemptions; indexing); b) Administration: 1) contractual basis (law; agreement between employers/employees’ associations and workers or employees/trade unions; individual decision); 2) legal status; 3) administrative structure and autonomy (existence of regional or local agencies; right to self-government; body with decision-making competences); 4) body with controlling power, monitoring the functional adequacy and legislative correctness of financial management; c) personal scope: 1) eligibility criteria for coverage (gainful employment or assimilated activity; residence; nationality); 2) modalities of adherence (legal obligation; insurance ceiling; co-insurance for family members; voluntary membership); 3) the way entitlement to benefits and services is established (realization of risk; neediness; special needs); 4) level of protection (basic; supplementary).
Data Handbook on Social Security
Building up a database on COSS 1949–1993 is linked to the project ‘The Societies of Europe—European Society’, aiming at producing a series of historical data handbooks. Topics covered in the eight volumes edited by Peter Flora, Franz Kraus and Franz Rothenbacher (Macmillan Publishers, 1999–2002) are demo-graphy, labour force, social security, trade unions/labour relations and elections. The two latter, both including a CD-ROM, are just being produced and should be published within the next months. The whole series is designed to enlarge the systematical empirical basis for studying long-term developments in European societies. All data handbooks focus on descriptions and analyses of their differences and similarities, whereas main cross-country divergences and convergences will be studied with regard to their internal dynamics.
The planned data handbook on social security 1950–1990 could contain selected national time series already included on the CD-ROM and additionally present comparative data and indicators over time. We intend to ‘frame’ this bulk of quantitatitive material with country chapters on the development of central welfare state institutions and overviews of core legislation in the fields of social policy covered by COSS.
Outlook on Work Planned by ILO
ILO plans to publish via the WWW all national datasets from 1990 or at least 1994 onwards, when the revised concept was implemented for the first time. The definitive form will, however, be subject to sectoral budget decisions. In any case, ILO and EURODATA share the idea of mutually coordinating their projects in order to improve access to and more extensively document COSS.
For most of the European countries a large majority of basic tables covering the period 1949–1989 has already been made machine-readable by EURODATA. Joining forces, the social security department intends to make them available online. This should be done countrywise from spring 2000 onwards, probably starting with France, Germany, The Netherlands and Portugal.
Adema, Willem (1999), Net Social Expenditure. OECD Labour Market and Social Policy Occasional Papers, No. 39. Paris: OECD.
Ahrens, Ulrike (1998), Alterssicherung in der Europäischen Union IV. Alterssicherung in Portugal. Eine institutionelle und empirische Analyse. Berlin: Duncker & Humblot.
Bureau International du Travail (BIT) / Michel Voirin (1991), L’organisation administrative de la sécurité sociale. Un enjeu social et politique. Genève: BIT.
Commission des Comptes de la Sécurité Sociale (1989), Les Comptes de la Sécurité Sociale. Prévisions 1988–1989. Rapport Janvier 1989. Paris: La Documentation Française.
Commission of the European Communities, Directorate-General Employment, Industrial Relations and Social Affairs/MISSOC-Mutual Information System on Social Protection in the Member States of the Community (several years), Social Security in the Member States of the European Communities. Situation on 1st of July 19.. and Evolution. Brussels/Cologne: MISSOC. 
Döring, Diether, Richard Hauser, and Winfried Schmähl, eds. (1999), Soziale Leistungen und ihre Finanzierung. Länderstudien zu Frankreich, Italien und den Niederlanden. Berlin: Duncker & Humblot.
EUROSTAT (1992–1994), Digest of Statistics on Social Protection in Europe. 8 vols. Luxembourg: EUROSTAT. 
—— (1996), ESSPROS Manual 1996. Luxembourg: EUROSTAT. 
Flora, Peter, ed. (1986), Growth to Limits. The Western Welfare States Since World War II. Vol. 1: Sweden, Norway, Finland, Denmark. Berlin/New York: De Gruyter.
Guibentif, Pierre (1993), ‘Der portugiesische Sozialstaat: Allgemeine Darstellung’. In Detlef Merten, and Rainer Pitchas, (eds.), Der Europäische Sozialstaat und seine Institutionen. Berlin: Duncker & Humblot, 32–58.
International Labour Office (ILO) (several years), The Cost of Social Security. Basic Tables. Geneva: ILO. [1a]
—— (1984), Financing Social Security: The Options. An International Analysis. Geneva: ILO.
—— (1996), The Cost of Social Security. Fourteenth International Inquiry, 1987–1989. Comparative Tables. Geneva: ILO. [1b]
—— (1997a), The Cost of Social Security. Nineteenth International Inquiry. Geneva: ILO.
—— (1997b), Social Security Financing. Geneva: ILO.
—— (1998a), Administration of Social Security. Geneva: ILO.
—— (1998b), Social Security Principles. Geneva: ILO.
International Social Security Association (ISSA) (1997), Interactions of Social Security and Tax Systems. Social Security Documentation. European Series No. 25. Geneva: ISSA.
—— (1988), Economic and Social Aspects of Social Security Financing. Social Security Documentation. European Series No. 14. Geneva: ISSA.
—— (1993), Organisation and Structure of Social Security Systems. Social Security Documentation. European Series No. 19. Geneva: ISSA.
Julliot, Jacques (1991), La sécurité sociale. Organisation–Assujettissement–Cotisations–Prestations–Contrôle–Contentieux. Paris: La Villeguérin Editions.
Ministerie van Sociale Zaken en Werkgelegenheid (1990), Social Security in the Netherlands. Deventer/Boston: Kluwer Law and Taxation Publishers.
Ministère de l’Economie, des Finances et de l‘Industrie (1994), Projet de loi de finances pour 1995. Etat retraçant l’effort social de la nation. Résultats 1991–1992–1993. Paris: Ministère de l’Economie, des Finances et de l‘Industrie.
MIRE (Mission Recherche)/Ministère des Affaires Sociales, de la Santé et de la Ville (1996), Comparing Social Welfare Systems in Europe: Vol. 1: Oxford Conference–France–United Kingdom. MIRE Rencontres et Recherches. Paris: Ministère des Affaires Sociales, de la Santé et de la Ville.
Nordic Social Statistical Committee (NOSOSCO) (1998), Social Protection in the Nordic Countries 1996. Scope, Expenditure and Financing. Copenhagen: NOSOSCO. 
Olson, Sven (1986), ‘Sweden’. In Peter Flora, ed., Growth to Limits. The Western Welfare States Since World War II. Vol. 1: Sweden, Norway, Finland, Denmark. Berlin/New York: De Gruyter, 1–116.
Organisation de Coopération et de Développement Economique (OCDE) (1988), L'avenir de la protection sociale. Social policy studies, 6. Paris: OCDE.
Organisation for Economic Co-operation and Development (OECD) (1996), Social Expenditure Statistics of OECD Member Countries. Labour Market and Social Policy Occasional Papers, No. 17. Paris: OECD. 
Schmähl, Winfried, ed. (1985), Versicherungsprinzip und soziale Sicherung. Tübingen: Mohr.
—— (1997), Financing of Social Security: Two Papers on the Instruments and Methods of Financing Social Insurance Schemes. ZeS-Arbeitspapier 1997, 21. Bremen: Zentrum für Sozialpolitik.
Schulte, Bernd (1980), ‘Reformen der sozialen Sicherheit in Westeuropa 1965–1980’. Vierteljahrsschrift für Sozialrecht 8: 323–61.
Tegtmeier, Werner (1997), Zur Finanzierung und Finanzierbarkeit sozialer Sicherung bei veränderten Rahmenbedingungen. ZeS-Arbeitspapier 1997, 11. Bremen: Zentrum für Sozialpolitik.
U.S. Department of Health and Human Services, Social Security Administration/International Social Security Association (ISSA) (1997), Social Security Programs Throughout the World, 1997. Research Report #65. Washington, DC: SSA, Office of Research, Evaluation and Statistics. 
Mathias Maucher has graduated in administrative sciences and works as researcher at the EURODATA Research Archive. He is manager of the MZES Family Policy Database. Recently he started work on a data base on social security statistics.