Introduction

Jörg Neugschwender

Pension system regulation has grown to a complex field of study. Nowadays, in most pension systems, many actors are setting up the institutional framework securing poverty prevention and income maintenance in old age (cf. OECD 2005; European Commission 2003; World Bank 1994). As a consequence, income security among the elderly is typically provided by an income mix from various pension schemes. In this introductory blog, I will give a brief overview about the conceptual distinction of individual pension schemes within an entire pension system. This section is followed by an argument on the relevance of cross-national studies for evidence-based policy advice. A third section will give an extended conceptual overview which is used throughout Inequality among the Elderly.

Pillars vs. tiers – the core conceptual framework for pension system design

Various typologies help to determine income security in old age. At this stage, I will briefly introduce the traditional concepts of pension pillars and income tiers. The concept of pension pillars distinguishes involved actors in regulation of pension systems: the state (first/public pillar), employers and trade unions (second/occupational pillar), and financial institutes (third/individual pillar). The concept of income tiers clarifies the functions/goals of pension income: poverty prevention and income maintenance. The first tier aims at providing a minimum income protection to prevent poverty (poverty prevention function); the second tier is linked to previous labour market earnings and pension entitlements through contributions. At retirement, these entitlements replace a certain level of the previous earnings (income maintenance function) and thus ensure pension adequacy (European Commission 2003). Third-tier pensions are an individual topping up of second-tier benefits.

What can we gain from cross-national comparisons?

Institutional differences across various pension schemes, such as contribution rates, voluntary vs. mandatory coverage, eligibility criteria and benefit calculation, are essential sources that help to explain cross-national differences in the income mix, financial well-being, and inequality among the elderly. I believe that comparative studies of institutional differences and its outcomes provide an essential understanding about consequences of specific pension system paths of the past and potential future reform needs. For example, public second-tier schemes may crowd out or crowd in further mandated schemes by the state, employers, or trade unions. Thus, occupational pensions can work either as a substitute to public welfare state activity or as complementary welfare programmes. At the same time, occupational equivalents could possibly only provide selective coverage. As a result, cross-national studies of financial well-being among the elderly may reveal country-specific shortcomings in poverty prevention and pension adequacy. Such comparisons are not only informative, but also crucial for evidence-based policy advice.

However, national studies, which aim at adopting other country’s national experiences to their own national context, also need to carefully analyse comparability of the surroundings such as labour markets and individual labour market attachment, living arrangements, and other cultural differences. Blog entries in Inequality among the Elderly seek to also provide appropiate background statistics. Once these differences have been carefully considered, comparative studies on alternative pathways of pension policy and their outcomes are promising for national reform scenarios.

An extended conceptual framework for pension system design

Researchers in the field have systematically analysed the development of multipillar pension systems and the variety of institutional arrangements between the policy actors (Arza and Kohli 2008; Ebbinghaus 2011; Natali et al. 2018). These scholars equally follow the traditional conceptual framework of pension pillars and income tiers. The extended conceptual approach, which is used thoughout blog entries in Inequality among the Elderly, integrates also the typology of Bismarckian vs. Beveridgean systems in this framework. Figure 1 summarises this extended framework for pension system design (see in detail Neugschwender 2016, 14-15).

Figure 1: Classification of Pension Systems

Notes: LIE=low-income earner; MIE=medium-income earner; HIE=high-income earner.

Light grey schemes refer to selective coverage; dark grey schemes refer to comprehensive coverage

First-pillar Bismarckian social insurance systems (for example Germany, Spain, Italy, and the United States) are earnings-related pension schemes that secure individuals with first and second-tier benefits for low to middle-income earners, and partly high-income earners, depending on the (non-)existence of contribution and/or income ceilings. In various contributory pension schemes, minimum pensions are also embedded (for example Estonia, Greece, Hungary, Mexico, Poland, Russia, and Spain); a full amount of the minimum pension is frequently granted only in case of reaching a minimum period of contributions. Contributory pension schemes typically cover the active population.

First-pillar schemes in a Beverigdean pension system can be categorised in two types: basic pensions and/or targeted means-tested pensions. Basic pensions aim at guaranteeing the first tier of pension income for the entire elderly population. For low-income earners, these benefits possibly could also guarantee their second tier of income, in case the level of the basic pension is relatively high. In contrast to this, minimum pension schemes can also be targeted to the poor (for example Australia, Finland, Italy Spain, the United States) or foresee a means-tested supplement to the basic pension (for example Canada, Denmark, and the United Kingdom). Depending on the generosity of the amount, these schemes provide first and second-tier benefits for low to middle-income earners. Such non-contributory schemes typically cover the whole resident population.

Next to the public first pillar schemes there are also schemes set up by individual employers, employer associations, or trade unions. Inclusion to these occupational second-pillar pensions is strongly determined by the obligation to participate in these schemes. In cases when occupational pensions were kept mostly on voluntary decision basis by employers and individuals, this approach was frequently combined with a Bismarckian social insurance system (for example Germany, Italy, the United States). In contrast to this, countries that designed their public pension scheme to protect the poor only, frequently introduced collective agreements between employers and employees. Thus, most initial Beveridge type pensions have developed towards Beveridge + Bismarck systems (Ebbinghaus and Gronwald 2011). Comprehensive coverage with second-tier earnings-related schemes is reached through ergaomnes regulations in collective agreements, for example in the Netherlands, Denmark, and Sweden, or in the case of Finland by tripartite legislation of employers, trade unions, and the government. Yet a different pathway exists in the United Kingdom and Japan, where mandatory contributions to a state second tier can be contracted out to private plans.

Individual third-pillar pensions can be considered a more recent development that gained further importance in the course of reforms to Bismarckian social insurance systems. Traditionally, personal pensions were mostly voluntary tax favoured savings plans offered by financial institutes. Nowadays, more and more state influence in the design of third-pillar arrangements can be observed. In recent reforms, former contributions to the public system are transferred to private sector pension funds or similar investments. This transformation may include a switch from voluntary to mandatory contributions; the most notable example in Europe is the case of the premium pension in Sweden. However, recent reforms in Latin America and Eastern Europe signify that pension privatisation is not a one-way development; in various countries, previous extensions have been also reversed recently (Ortiz et al. 2018).

References

Arza, C., and Kohli, M. (eds.) (2008a). Pension Reform in Europe. Politics, Policies and Outcomes. London/New York: Routledge.

Ebbinghaus, B. (ed.) (2011). The Varieties of Pension Governance: Pension Privatization in Europe. Oxford: Oxford University Press.

Ebbinghaus, B., and Gronwald, M. (2011). ‘The Changing Public–Private Pension Mix in Europe: From Path Dependence to Path Departure’, in B. Ebbinghaus (ed.), The Varieties of Pension Governance: Pension Privatization in Europe. Oxford: Oxford University Press, 23-53.

European Commission (2003). Adequate and sustainable pensions – Joint report by the Commission and the Council, Luxembourg, Publications Office.

Natali, D., and Pavolini, E. with Vanhercke, B. (2018). Occupational Welfare in Europe: Risks, opportunities and social partner involvement. ETUI, Brussels, OSE, Brussels.

Neugschwender, J. (2016). Pension Systems and Income Inequality among the Elderly in Europe, Mannheim [Dissertation],https://ub-madoc.bib.uni-mannheim.de/41262/.

OECD (2005). Pensions at a Glance. Public Policies across OECD Countries. Paris: OECD Publishing.

Ortiz, I., Durán-Valverde, F., Urban, S., Wodsak, V., Yu, Z. (2018). Reversing Pension Privatization: Rebuilding public pension systems in Eastern European and Latin American countries (2000-18), ESS Working Paper No. 63, Geneva, ILO.

World Bank (1994). Averting the Old Age Crisis: Policies to Protect the Old and Promote Growth. Washington D.C.: The World Bank.

Denmark – income growth among the elderly from 1995 to 2013

Jörg Neugschwender

Why to look specifically at Denmark? Denmark is a very illustrative case with respect to pension policies and outcomes. In the early 1990s, Denmark has experienced a quite substantial path departure with respect to pension system regulation. Before the 1990s, the Danish public pension system was characterised by a rather generous non-contributory tax-financed universal pension (folkepension – national pension) based on residency in Denmark, and additional mostly voluntary arrangements in the sphere of contributory systems.¹ If any agreements between employers and employees existed, then they provided mostly security for white collar employees in the sense of fringe benefits. Although this institutional structure has been criticised for decades, only in the 1990s, negotiations between employer associations and trade unions came to a closure, resulting in various branch-specific quasi-mandatory occupational (2nd pillar) pension schemes. Thus, after the path departure, most employees were automatically enrolled in additional contributory systems, as they were covered by binding collective agreements. As a result, future retirees could have expected a rather different income mix as compared to the generations which retired before.

In this blog, I am looking at the implication of this path departure. The following analyses refer to results from the Danish data from the Law Model (based on administrative records) which are available through the Luxembourg Income Study (LIS) Database. First, I look at the living standard of the elderly society in Denmark at two points in time, 1995 and 2013. A specific focus is placed on the development of pension income. In a second step, I compare the increase of income among elderly households with the situation among the total society.

Figure 1 splits the elderly population 65+ (here restricted to mostly retired individuals, as for comparability reasons only households with pension income larger than labour income are kept) in 20 population groups sorted by their level of total equivalised² gross household income. Thus, group 1 refers to the 5 % of elderly persons with the lowest level of gross income; group 2 refers to the 5 % with income closest to group 1, and thus represents the elderly persons with incomes that range from above the lowest 5 % to below the lowest 10 % and so on. For both data points, I apply the same exercise. In a next step, each point in the light-blue line in Figure 1 compares the change in the mean gross income for each group, the lowest 5 % of the elderly in Denmark in 1995 against the lowest 5 % in Denmark in 2013 and so on (this procedure is typically referred to as Growth Incidence Curve). That is it? No, not yet – one more detail to be mentioned. The percentages refer to annualised percentage change and are expressed in real values. This means that incomes were price adjusted using PPPs (for more information see previous blog). Now, that is all, right? Yes, almost, but I assume, you might want to understand the other two lines as well. The dark-blue line, pension income, refers to the same grouping for the elderly population sorted by gross household income. The red line, total population gross income, simply refers to the grouping for the whole Danish population.

Now, after this rather dry and theoretical introduction, let me jump to the empirical conclusions. First of all, the living standard among the elderly increased more as it did for the total population. In the light of maturing pension systems this is not astonishing per-se. Younger cohorts have been longer integrated in the new pension schemes which resulted in longer mandatory contribution periods from the early 1990s. Hence, the various occupational schemes helped substantially in maintaining the living standard during working age, as the longer contribution periods led to higher replacement rates of wages. It is particularly the middle income groups which benefitted most from this inclusion to the contributory schemes. Nevertheless, all percentile groups show a rather similar increase with around 1.4 % annualised income growth, with slightly lower rates (around 1.1 %) for the high-income group. In fact, the high-income groups were already well integrated in occupational pension schemes, thus a lower increase as compared to the low and middle-income group for the high-income groups is plausible. It is worth mentioning that particularly at the bottom end of the elderly’s income distribution, continued employment from ongoing labour market transformations towards a later retirement led to a more pronounced relevance of labour income (although not shown here, but confirmed with the data) in the income mix of the more recent observation in 2013.

Figure 1 also reveals another quite interesting insight. As mentioned before, the elderly low to middle-income groups benefitted most from the path departure. Thus, the income gap as compared to the high-income group decreased, and hence inequality decreased for the elderly. Still, the Danish population overall shows a quite opposite trend with respect to inequality trends. Particularly the upper income groups show higher increases. In contrast to this, there is a substantial share of the population that moved closer towards the middle. Guess what, it is the same group of pensioners that benefitted the most from the expansion of the pension system, as reflected by the highest increases for the dark-blue line. How do I know? The ppp-adjusted amounts for the peak in the dark-blue line and the red line are the same, thus the bump in the red line reflects strongly the increased income among the elderly.

Summing up, Danish pensioners are still not particularly well off in this story, but the numbers support that pension policies matter quite a bit, and the transformation of voluntary pensions to quasi mandatory ones helped to a large extent gradually closing the income gap among the elderly as compared to the working age population!

¹ The only contributory component of the pension system until then (ATP) was related to working hours, but practically the scheme was quite unimportant for generating income maintenance, as the contributions were very low.

² Equivalised household income refers to the standard assumption in socio-economic analyses that the households pool their income resources and distribute them equally between the household members. At the same time, economies of scale are assumed, as larger households can more effectively share their resources. Here the standard by procedure of dividing the household income by the square root scale is applied. For further information please see an earlier blog entry.

Varieties of pension pathways and gender pension gaps

Jörg Neugschwender

Public pension systems follow different state traditions. In this blog, I look at the implications of these different pathways for gender gaps in the pension income distribution. The recent ‘Pension Adequacy Report 2018‘ by the European Commission revealed that the gender gap in pensions in the EU (36%) is […] twice the gender gap in pay (16.3%). In order to illustrate cross-national differences in pension outcomes, I selected four countries, each representing a rather distinct pathway in pension provision: Finland, Germany, the Netherlands, and the United Kingdom.

Before analysing the different pension outcomes in these four countries, let me briefly highlight the core differences in pension system design. Pension schemes are set up to provide two main goals – poverty prevention and income maintenance, in the sense of replacement of a certain degree of earnings during the retirement phase. A common distinction is analogously made into two ideal types of pension system design: Beveridgean (targeted pensions) and Bismarckian (earnings-related) public pension schemes.

The Beveridge type pension schemes provide only the first-tier of social security among the elderly; these schemes are set up to prevent poverty. Different proto-types of Beveridgean schemes provide minimum pensions either based on means-tested pensions (e.g. Finland), or based on employment years (e.g. the United Kingdom), or based on residency in the country (e.g. Netherlands and Denmark).

The Bismarckian schemes are more generally linked to the employment career (e.g. Germany, Austria, and Italy); own accumulated contributions and contributions paid by employers mostly define pension outcomes. These schemes are similarly known as pay-as-you-go schemes; the higher accumulated contributions, the higher pension annuities. Thus, besides the goal of poverty prevention (first-tier), these schemes also provide income maintenance (second-tier) of previous earnings. There might be redistributive elements included to support specific groups in need, for example credits for taking into account different periods of inactivity such as child raising, invalidity, or unemployment. On the other hand, these contribution-based Bismarckian schemes might also cross-finance payments of minimum pensions for those whose contributions were insufficient to guarantee a self-sufficient protection.

Following these different state traditions, pension provision in public vs. private schemes strongly differs. Bismarckian schemes not only link contributions closely to employment, but also are heavily defined by the level of earnings. Higher earnings lead to higher contributions, as contributions are typically deducted as a percentage from the gross wage. Similarly, less interruptions during the employment career secure longer contribution periods, and hence higher final pensions. As mentioned above, credits for periods of inactivity may scale up contribution years and pension amounts. The latter might particularly be important for women who show substantively shorter contribution years (according to the ‘Pension Adequacy Report 2018‘ average duration of working lives in the EU28: Women: 33.1, Men: 38 years). Private schemes from employers or financial institutes are naturally less relevant in countries following a Bismarckian path, as income maintenance is primarily provided through the public scheme. Nevertheless, private schemes tend to increase selectively the pension provision of men, foremost due to the mechanism that occupational pensions are typically offered to higher paid professions – jobs that are currently still dominated by men.

Countries limiting their public scheme to poverty prevention, even more link accumulated contributions to inequality in labour incomes. Although, at the bottom of the income distribution, basic or guaranteed pensions might create a gender neutral outcome, contributions to private (occupational and individual) schemes may create substantial gender differences. Private schemes lack mechanisms that account for periods of inactivity more generally. Voluntarism in private schemes may additionally create a lack in provision.

Ok, let us now proceed with the empirics. While presenting these findings, I will elaborate the characteristics of the national pension scheme in each country, explaining the findings. The following graphs visualise pension income distributions. The figures plot Kernel density estimates, which are practically nothing else than a smoothing technique for histograms. The four selected pension schemes provide a broad variation towards their approach to income security among the elderly. In order to minimize the presence of partial pension payments respectively invalidity pensions before switching to retirement, I selected only individuals aged 66 and above. The pension amounts are ppp-adjusted to 2011 International Dollars in order to compare the level of pension amounts across countries. There are two vertical lines in these charts. Each refers to the average pension amounts received by men respectively women aged 66 or above. Besides the legend, I show the level of the gender pension gap, which is derived by 1- (average men – average women) divided by average pension for men.

Germany represents the case of a truly earnings-related system, pension amounts closely relate to contribution-based pension schemes. The pension income distribution reveals a rather smooth density, pension amounts do not cluster around specific values, however, pensions for men respectively concentrate within specific income ranges. The typical range for women is much lower as compared to men, mirrored by the high pension gender gap. The case of Finland represents a combination of targeted minimum pensions – in the past universal minimum pensions based on residency – and tripartite mandated contribution-based schemes. In the graph mostly the shape for women looks different. The access to minimum pensions in case of too low own entitlement from employment, pushes pension amounts for women closer together, reducing the gender gap. Most men have collected sufficient own pension amounts from the various mandated contribution-based schemes, hence the shape is rather similar to the one in Germany. This pattern is even more striking in the case of the Netherlands, where minimum pensions are still generally paid according to years of residence in the country. However, due to the interplay of entitlements from the public scheme, women are left far behind average pensions as compared to men; partly this mirrors the high relevance of part-time employment in the Dutch context which is mostly common among women. In addition, the Dutch scheme does not offer public credits for child raising. Last, let us have a quick look at the British elderly population. The system is known for comparatively low benefits, and the chart reveals a concentration with pensioners receiving low pensions for both sexes, but women in particular. The United Kingdom provides employment-based Beveridge type minimum pensions supplemented by means-tested pension credits. Income maintenance is provided through a mixture of public and contracted out occupational and individual pensions. Mandatory contributions are regulated at a relatively low level in the European context. If people were paying higher voluntary contributions throughout their working career, then it may be linked mostly to their stable employment. Thus, not surprisingly men receive frequently much higher benefits as compared to women, resulting in the highest gender pension gap in this comparison.

This blog post revealed that pension system design matters a great deal for the persistence of gender pension gaps. One closing remark remains. This contribution kept pension income defined broadly, i. e. pension income is analysed jointly as an income mix from all three pension pillars. More detailed analyses, which look the interplay of public and private pension schemes, will provide a more nuanced picture with respect to implications for financial well-being among the elderly. In this sense, stay tuned for more!

Relative vs. absolute income development among the elderly between 2000 and 2010/13

Jörg Neugschwender

During the last decades, younger cohorts of the elderly population strongly benefitted from the continued maturing of more generous contribution-based pension systems. At the same time, many countries further introduced non-contributory minimum pension protection systems, resulting in a rather steep decrease in poverty rates among the elderly throughout this period. However, besides this positive development, one might wonder what has happened to income inequality among the elderly? Who are the best performers with respect to increased income standards? And is it actually true that the elderly are better off in all advanced economies?

Foremost, the argument in this article concentrates on the technical side, i.e. the choice of measurement techniques for analysing income levels in a comparative perspective. A specific focus is placed on equivalisation, ppp-conversion and the distinction between relative and absolute income measures. Lastly, this article attempts to briefly respond to some of the questions raised with respect to the nuances of income inequality and income growth among the elderly over time.

A standard approach in inequality research is to equivalise household income. This means two things: First equivalisation assumes full sharing of income sources across all household members. Second equivalisation also assumes that larger households share resources and thus need less resources compared to smaller households, they achieve economies of scale (Jenkins and Van Kerm, 2009). This approach has some limitations, which should be acknowledged at this point. In Latin American countries as well as in most other developing countries in the LIS Database, every second person aged 65 or older is living in a multi-generation/family situation. This has two implications: First, the pooling of income sources and equal sharing between all household members is a strong assumption to make for the elderly. And second, the financial well-being of the elderly is heavily defined by the reallocation of other income sources (particularly employment) from other family members. Therefore, these countries are excluded from this comparative overview, the narrative becomes more a story of generosity of pension systems¹. For the remaining countries, the common equivalisation approach by LIS is applied, which divides the disposable household income (DHI) by the square root of number of household members. Note the remaining numbers have to be understood as individual level income measure.

Both figures refer to the same underlying income thresholds. First, the median of equivalised disposable household income (DHI) of the total population was calculated. Then for the reduced sample of the elderly population (defined as persons aged 65+), three thresholds were calculated: the median of equivalised DHI of the 65+, the mean equivalised DHI of the four lower income deciles (bottom 40) of the 65+, and the mean equivalised DHI of the six upper income deciles (upper 60) of the 65+². For each country two points in time have been selected, one around 2000 and one around 2010/13 depending on data availability.

Figure 1 shows two relative measures applied on both time points (repeated cross-section): the mean equivalised DHI of the bottom 40 respectively the upper 60 of the 65+ each divided by the median equivalised DHI by the total population. The countries are ranked clockwise by starting with the lowest level of median equivalised DHI for the 65+.

Australia’s position in the graph is unique, but acknowledged, as major parts of the contribution-based superannuation schemes were introduced only in the 1980s; those systems need to further mature, persons who started their working career in the 1980s are not yet retired. Hence the current retirees receive only partial benefits from these schemes. Moreover, many pensioners received their pension rather as a lump sum than as an annuity (King et al., 2001). Also, Denmark ranks only third on median equivalised DHI of the 65+. Similarly, like in Australia, the pension system concentrated on providing basic pensions, and only in the 1990s introduced major quasi mandated contribution-based pensions (Andersen, 2011).

More generalised, the higher the distance between the bottom 40 and upper 60, the more stratification in elderly income there is in a country. This is partly related to primary market inequality which is translated to the retirement phase (Ebbinghaus and Neugschwender, 2011). The best examples in this respect are the United States and Israel, both countries compare reasonably well with respect to median equivalised DHI of the 65+. However, both countries (as also Australia and Estonia) remained in 2010/13 at a level of only 40 percent of median equalised income for the bottom 40 of the 65+³. On the other hand, the United States and Israel are also among the countries with the highest relative levels for the upper 60 group.

For Figure 2, all amounts were converted to real amounts (ppp-adjusted), applying Consumer Price Indices (CPI) and conversion factors to adjust to 2011 country’s currency as calculated by the World Bank Development Indicators for 2011⁴. Figure 2 takes these ppp-adjusted and equivalised mean DHI values for each group, the bottom 40 respectively the upper 60 of the 65+, and looks at the overall percentage increase between the two points in time. As the interval between the country years varies, the numbers are converted to annualised growth rates by dividing the overall increase by the number of years between the earlier and the later year. Data points below the diagonal signify decreases in the distance between the bottom 40 and the upper 60 measure; in these countries, over the whole period, income levels among the bottom 40 have grown more as compared to the upper 60.

A broader research agenda is needed for carefully interpreting Figure 2. Reasonably high numbers in growth might be linked to high wage growth, depending on existing indexation rules in the various pension schemes. Also, the countries below the diagonal may have concentrated rather on better inclusion or higher generosity of targeted pensions to the income poor elderly; hence these countries have undertaken a relatively higher investment in increasing income of the poor elderly as compared to increasing payments from contribution-based schemes. Cost containment in pension systems might particularly affect the upper 60 group in future, when e. g. partly privatised pension provision for future retirees does not result in a replacement of previous public provision. On the other hand, cross-national and intertemporal variation in continued employment beyond the legal retirement age among the elderly across this period may heavily affect this narrative. Future research may aim at better capturing the joint analysis (multivariate modeling) of institutional characteristics of pension systems, living arrangements and labour market attachment, and its overall implication for pension outcomes and inequality trends among the elderly.

This blog post has been earlier published in Inequality Matters – LIS Newsletter, Issue 4, December 2017.

¹ see Chapter 3 on the living arrangements, labour market attachment, and the income mix of the elderly for LIS countries in Neugschwender (2016). The remaining country group is characterised by a rather high share of single/couple elderly households, hence a substantial share of income is drawn from pension income in these countries.

² The breakdown in Bottom 40 and Upper 60 has been borrowed from the ‘Poverty and Shared Prosperity’report by the World Bank (2016).

³ This measure mirrors the highest at-risk-of-poverty for the elderly in these countries in these four countries, see LIS Key Figures.

⁴ PPP deflators (Consumer Price Index (2011 = 100), and conversion factor to adjust 2011 country’s currency into 2011 international USD) adjusted to the LIS country years can be found on the following LIS webpage.

References

Andersen, J. G. (2011). ‘Denmark: The Silent Revolution toward a Multipillar Pension System’, in B. Ebbinghaus (ed.), The Varieties of Pension Governance: Pension Privatization in Europe. Oxford: Oxford University Press, 183-209.

Ebbinghaus, B., and Neugschwender, J. (2011). ‘The Public-Private Mix and Old Age Income Inequality in Europe’, in B. Ebbinghaus (ed.), The Varieties of Pension Governance: Pension Privatization in Europe. Oxford: Oxford University Press, 384-422.

Jenkins, S. P., and Van Kerm, P. (2009). ‘The Measurement of Economic Inequality’, in W. Salverda, B. Nolan and T. M. Smeeding (eds.), The Oxford Handbook of Economic Inequality. Oxford: Oxford University Press, 40-67.

King, A., Baekgaard, H., Harding, A. (2001). ‘Pension Provision in Australia’, in R. Disney & P. Johnson (eds.) Pension Systems and Retirement Incomes Across OECD Countries, Cheltenham, UK, Edward Elgar, 48-91.

Neugschwender, J. (2016). Pension Systems and Income Inequality among the Elderly in Europe, Mannheim [Dissertation],https://ub-madoc.bib.uni-mannheim.de/41262/.

World Bank (2016). Poverty and Shared Prosperity 2016: Taking on Inequality. Washington, DC: World Bank. doi:10.1596/9781464809583.

Occupational pensions – data evidence of gender gaps

Jörg Neugschwender

Nowadays, in many societies pension entitlements are built up from various schemes that are regulated and administered by the state (first pillar), employers and trade unions (second pillar), and/or financial institutes (third pillar). Researchers in the field have systematically analysed the development of multipillar pension systems and the variety of institutional arrangements between the policy actors over time (Arza and Kohli 2008; Ebbinghaus 2011; Natali et al. 2017). Building on national country case studies, in comparative contributions the scholars grouped together countries with similar institutional approaches in order to better contrast the main pathways and implications for the redistributive outcomes of such systems.

Besides the common analytical approaches of pension pillars and income tiers, two main archetypes of pension systems have been conceptualised: Bismarckian social security systems were founded mostly on one main public system that secure the elderly against poverty (first tier of income) and maintain the living standard (second tier of income). In contrast to the Bismarckian approach, in Beveridgean systems the state restricted its role mostly to protect the elderly against poverty. Additionally, the state might be involved in joint regulation with employers and trade unions setting up complementary pension schemes. Alternatively, the state might increase incentives to take up personal pension plans with financial institutes through favourable tax treatment and subsidised contributions. Typically Beveridgean systems are also classified as multipillar systems.

Ebbinghaus and Gronwald (2011) analysed the main pathways of pension system development. In Bismarckian systems, occupational (second-pillar) and personal (third pillar) pensions have been typically crowded out, as the contribution-based pension income by the state already served to maintain the living standard. It is particularly the multipillar pension systems which crowd in a variety of alternative solutions that may even partly or fully substitute the public system. Following up on the policy brief by Natali and Pavolini in this newsletter issue, this article will try to shed some more light on the relevance of occupational pensions in the pension income mix around the world.

Occupational welfare solutions set the scope for various outcomes: How employers and employees share contributions to occupational pension plans. Whether or not, individuals in case of job change can transfer their rights to the new employer. But also whether or not, individuals are allowed to cancel their accumulated rights from occupational welfare at any point in time during their working career. In contrast to this, personal pensions with financial institutes might protect two very distinct groups, the ones earning above income ceilings, and the ones who do not have sufficiently access to occupational pensions respectively for whom personal pensions are more suitable than employer-based solutions. Therefore, it is important that during the data collection phase a clear distinction in the income sources public vs. occupational vs. personal pensions respectively nature and type of current pension savings contracts is made.

In general, the Luxembourg Income Study (LIS) Database tries to distinguish between various pension income sources/LIS variables:

employment-related public pensions,

old-age/disability/survivors universal pensions,

old-age/disability/survivors assistance pensions,

occupational pensions,

voluntary individual pensions.

For various LIS datasets from the early 2010s a split in all three pension income sources public vs. occupational vs. personal is available. The presented graph concentrates on the relevance of occupational pensions in the pension income mix and its cross-national variation. The numbers below the graph show the weighted percentage of elderly persons (here defined as persons aged 65 or older) receiving occupational pension income. In order to focus on the pension income mix of retirees and to reduce the influence of partial pensions in the pension income mix, for the presented income shares, the sample of the elderly (65 and older) has been further restricted to those elderly, whose individual pension income is the main income source (pension income larger than 50 % of total individual labour income). Further breakdowns by three income groups and by gender offer additional insights in the spread of occupational pension income and its redistributive impact among the elderly in current societies.

Among the analysed economies, in the Bismarckian countries such as Luxembourg, Italy, Greece, and mostly in Germany, occupational pensions are barely contributing to the pension income mix. Although Germany and the United States show a similar pattern in recipient rates, occupational pensions are relatively more important in the United States due to the lower generosity of the public earnings-related pension system. In general, in the Beveridgean pension systems of Ireland, the United Kingdom, and the Netherlands, there is a strong variation of relevance of occupational pension in the income mix across the income groups. As occupational pensions, particularly for high-skilled workers, function at the same time as fringe benefits, a comparatively higher relevance of occupational pensions in the upper end of the income distribution could be expected.

The separate analyses by gender reveal in most of the countries a strong gender gap. Ireland shows the highest difference between men and women regarding the spread of occupational pensions among the elderly population. The gender divide is also particularly high in the Netherlands, where women due to a high relevance of part-time employment careers collect substantially less contribution-based entitlements; furthermore, the public residence-based pension income is intertwined with the payment from occupational systems. Note that the various Finnish occupation-based pension schemes are a hybrid between public and occupational pensions, as they are legislated by tripartite agreements; for this overview the various Finnish contribution-based pension schemes have been reclassified from public to occupational pensions.

This blog post has been earlier published in Inequality Matters – LIS Newsletter, Issue 1, March 2017.

References

Arza, C., and Kohli, M. (eds.) (2008a). Pension Reform in Europe. Politics, Policies and Outcomes. London/New York: Routledge.

Ebbinghaus, B. (ed.) (2011a). The Varieties of Pension Governance: Pension Privatization in Europe. Oxford: Oxford University Press.

Ebbinghaus, B., and Gronwald, M. (2011). ‘The Changing Public Private Pension Mix in Europe: From Path Dependence to Path Departure’, in B. Ebbinghaus (ed.), The Varieties of Pension Governance: Pension Privatization in Europe. Oxford: Oxford University Press, 23-53.

Natali, D. and Pavolini, E. with Vanhercke, B. (2017), Occupational Welfare in Europe. Risks, opportunities and social partner involvement in pensions and unemployment protection, ETUI, Brussels, forthcoming.