Skip to main content

Poverty among the elderly: the role of public pension systems

Abstract

The objective of this paper is to measure the impact of first-pillar public pensions spending on the prevalence of poverty among the elderly. Using data from 27 European countries from 1995 to 2014, we estimate the elasticity of the poverty rate among individuals aged over 65 years to per capita public pensions spending. We show the existence of a nonlinear relationship between these two variables. The elasticity is negative and statistically different from 0 only beyond a level of spending of 685€ per capita. At the average value of 2819€, it is estimated that the elasticity is around − 1.45. This nonlinear relationship is robust to the treatment of possible endogeneity and to different robustness checks like the variation of the poverty line as well as to the inclusion of country-specific differences in public pension plans.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2
Fig. 3
Fig. 4

Notes

  1. 1.

    The 15 OECD countries are Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Sweden, UK.

  2. 2.

    The red curve is the result of a linear regression with expenditures and squared expenditures as explanatory variables. The source and an exhaustive description of the variables are provided in Table 9 of Appendix 1. On the top left of the graph are clustered countries such as Cyprus, Ireland and Latvia, and at the bottom right are Luxembourg and Denmark.

  3. 3.

    Countries are listed in Table 8 in Appendix 1. Note that we excluded Slovakia from our analysis as there were too many missing values for this country.

  4. 4.

    Equivalent disposable income is the total income of a household available to consume or save, after taxes and transfers, divided by the number of individuals in the household. Each individual is transformed into an adult equivalent.

  5. 5.

    About the interest of considering relative poverty rates (i.e. a threshold defined in relation to the median or mean income of each country for a given year) rather than absolute ones, see Bourguignon (2003). This is quite standard in the literature; see Cantillon (2011), Caminada et al. (2012), Caminada and Goudswaard (2012), van Vliet et al. (2012), Been et al. (2016) which also opt for a relative indicator to measure poverty.

  6. 6.

    Using the mean to obtain a poverty rate tends to overestimate poverty since the mean is more sensitive to extreme values. Tables 10 and 11 in Appendix 3 show that the poverty rate is always lower when using the median.

  7. 7.

    See for instance Foster et al. (1984) which compute a poverty measure that is sensitive to income distribution and accounts for the aversion to poverty. Related to this point, Marchand and Smeeding (2016) point out that the poverty rate is not able to account for all poverty dimensions, such as depth and duration of poverty. More recently, researchers (see for instance Lefèbvre et al. 2018) have also drawn attention to the “mortality paradox” and to the fact that standard poverty measures tend to underestimate actual poverty because of a simple selection effect induced by the positive correlation between income and survival (poorer people also die earlier). We abstract here from such considerations that would considerably complicate our computations.

  8. 8.

    For a complete definition, see Eurostat (2014) http://ec.europa.eu/eurostat/statistics-explained/index.php/Glossary:Expenditure_on_pensions.

  9. 9.

    For instance, in some European countries (like Denmark and the Netherlands), contributing to a private pension plan (occupational or not) is mandatory and it may account for an important part of the income of the retirees.

  10. 10.

    For a complete definition of means-tested benefits, see Eurostat (2014): http://ec.europa.eu/eurostat/statistics-explained/index.php?title=Glossary:Means-tested_benefits&oldid=324037.

  11. 11.

    Note that our results are robust to using Purchasing Power Standard instead of constant-2010 Euros.

  12. 12.

    We have also performed a regression adding a cubic term for the log of public pension expenditures. Since the coefficient associated with this term was very small and did not increase the explanatory power of the model, we decided to keep the regressions with only quadratic terms. Regressions including higher-order terms are available from the authors upon request.

  13. 13.

    Note that we also performed regression 3 including also year fixed effects. Since it did not qualitatively change our results, we decided not to report it.

  14. 14.

    Considering a relative measure for the poverty rate may be seen as a drawback of our analysis for this specific reason. However, using absolute measures has other drawbacks such as making more difficult to compare countries with different economic situations.

  15. 15.

    In unreported estimations, we also included the participation rates of individuals between 60 and 64, and of those aged 65 and older. Since the coefficients associated with these variables were found to be nonsignificant, we decided to drop them from the regressions.

  16. 16.

    We vary per capita expenditures from 100€ to 8000€ as the minimum per capita expenditures is 171.75€ and the maximum is 7658.28€ in our sample.

  17. 17.

    Although the value of elasticities is significantly different from zero, confidence intervals are large.

  18. 18.

    2819.45€ is computed as the average over the 1995–2014 period for each country, of per capita yearly public pensions.

  19. 19.

    In Appendix 5, Figure 7 plots the elasticity of poverty assuming away country fixed effects. We obtain the same decreasing and nonlinear pattern except that it becomes statistically significant around 400€ of per capita public pensions expenditures and that it stabilizes at a lower level, around \(-1.70\) at 8000€.

  20. 20.

    Section 4.4 reports additional robustness checks which, in particular, address the question of how the existence of (non) mandatory occupational plans or of (non) mandatory private plans could affect the baseline value of the elasticity.

  21. 21.

    See in particular Chapter 9, Sect. 9.5.

  22. 22.

    Table 17 in Appendix 7 reports results where instruments are the two-year lagged variables.

  23. 23.

    Recall that the relationship between poverty and per capita public pensions spending is not significant for a poverty line set at 40% of median income.

  24. 24.

    Note that, in some way, some of these differences are already taken into account when we include country fixed effects.

  25. 25.

    The values from Table 6 are obtained using Eq. (4) and the values of the predicted coefficients \((\beta _1,\beta _2,\beta _4,\beta _5)\) from Table 5.

  26. 26.

    15% represents the higher bound, and it is computed as the average level of means-tested expenditures in our sample (\(4.4\%\)) plus one standard deviation (\(10.3\%\)).

  27. 27.

    Note that we could not obtain data over time so that we are not able to know whether some countries took reforms to promote occupational pension plans and/or private pension plans. Hence, we proceed as if these structures had not changed across years. Countries with mandatory or quasi-mandatory occupational pension plans are Belgium, Denmark, Sweden, Estonia, Ireland, Cyprus, Malta, Netherland, Austria, Portugal. Countries with mandatory or quasi-mandatory private pension plans are Bulgaria, Estonia, Croatia, Latvia, Lithuania, Poland, Romania, Slovakia, Sweden. See European Commission, Ageing report (2015), pp. 58.

References

  1. Arellano, M. (1987). Practitioners’ Corner: Computing robust standard errors for within-groups estimators. Oxford Bulletin of Economics and Statistics, 49(4), 431–434.

    Article  Google Scholar 

  2. Barro, R. J. (2000). Inequality and growth in a panel of countries. Journal of Economic Growth, 5(1), 5–32.

    Article  Google Scholar 

  3. Been, J., Caminada, K., Goudswaard, K. P., & Van Vliet, O. P. (2016). Public/private pension mix, income inequality, and poverty among the elderly in Europe: An empirical analysis using new and revised OECD data. Social Policy and Administration, 51(7), 1079–1100.

    Article  Google Scholar 

  4. Bourguignon, F. (2003). The growth elasticity of poverty reduction: Explaining heterogeneity across countries and time periods. Washington, D.C.: World Bank Group. http://documents.worldbank.org/curated/en/503161468780002293/The-growth-elasticity-of-poverty-reduction-explaining-heterogeneity-across-countriesand-time-periods

  5. Caminada, K., & Goudswaard, K. (2012). The relationship between alternative measures of social spending and poverty rates. International Review of Business and Social Sciences, 1(5), 08–25.

    Google Scholar 

  6. Caminada, K., Goudswaard, K., & Koster, F. (2012). Social income transfers and poverty: A cross-country analysis for OECD countries. International Journal of Social Welfare, 21, 115–126.

    Article  Google Scholar 

  7. Cantillon, B. (2011). The paradox of the social investment state: Growth, employment and poverty in the Lisbon Era. Journal of European Social Policy, 21(5), 432–449.

    Article  Google Scholar 

  8. European Commission. (2015). The 2015 Ageing Report: Economic and budgetary projections for the 28 EU Member States (2013–2060). European Economy. https://doi.org/10.2765/877631

  9. Engelhardt, G. V., & Gruber, J. (2004). Social security and the evolution of elderly poverty. No. w10466. National Bureau of Economic Research

  10. Fonseca, R., Kapteyn, A., Lee, J., Zamarro, G., & Feeney, K. (2014). A longitudinal study of well-being of older Europeans: Does retirement matter? Population Ageing, 7, 21–41.

    Article  Google Scholar 

  11. Foster, J., Greer, J., & Thorbecke, E. (1984). A class of decomposable poverty measures. Econometrica, 52(3), 761–766.

    Article  Google Scholar 

  12. Lefèbvre, M., & Pestieau, P. (2006). The generosity of the welfare state towards the elderly. Empirica, 33(5), 351–360.

    Article  Google Scholar 

  13. Lefèbvre, M., Pestieau, P., & Ponthiere, G. (2018). FGT old-age poverty measures and the mortality paradox: Theory and evidence. Review of Income and Wealth, 64(2), 428–458.

    Article  Google Scholar 

  14. Milligan, K. (2008). The evolution of elderly poverty in Canada. Canadian Public Policy, 34(4), S79–S94.

    Article  Google Scholar 

  15. Marchand, J., & Smeeding, T. (2016). Poverty and aging. Handbook of the Economics of Population Aging, 1, 905–950.

    Article  Google Scholar 

  16. OECD. (2017). Pensions at a Glance 2017: OECD and G20 Indicators. Paris: OECD Publishing.

  17. Orenstein, M. A. (2011). Pension privatization in crisis: Death or rebirth of a global policy trend? International Social Security Review, 64(3), 65–80.

    Article  Google Scholar 

  18. Smeeding, T. (2006). Poor people in rich nations: The United States in comparative perspective. The Journal of Economic Perspectives, 20(1), 69–90.

    Article  Google Scholar 

  19. Smeeding, T. M., Williamson, J. (2001). Income maintenance in old age: What can be learned from cross-national comparisons. Center for Retirement Research Working Papers, 45.

  20. van Vliet, O. (2010). Divergence within convergence: Europeanization of social and labour market policies. European Integration, 32(3), 269–290.

    Article  Google Scholar 

  21. van Vliet, O., Been, J., Caminada, K., & Goudswaard, K. (2012). Pension reform and income inequality among older people in 15 European countries. International Journal of Social Welfare, 21, S8–S29.

    Article  Google Scholar 

  22. Wooldridge, J. M. (2002). Econometric Analysis of Cross Section and Panel Data. Cambridge: MIT Press.

    Google Scholar 

Download references

Acknowledgements

We thank Philippe De Donder, Raquel Fonseca, Mathieu Lefebvre, Florian Mayneris, Pierre Pestieau as well as participants to the 2017 CRDCN Conference in Montreal and participants to the Economic Department internal seminar at UQAM, for their comments and suggestions. We also thank an anonymous referee as well as the editor for their comments.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Marie-Louise Leroux.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendices

Appendices

Appendix 1: Country list and abbreviations

See Table 8.

Table 8 List of countries

Appendix 2: Definition and source of variables

See Table 9.

Table 9 Definition and source of variables

Appendix 3: Descriptive statistics

See Tables 10, 11, 12 and 13.

Table 10 Descriptive statistics (1)
Table 11 Descriptive statistics (2)
Table 12 Descriptive statistics (3)
Table 13 Descriptive statistics (4)

Appendix 4: Common trend in public pensions expenditures and in poverty rates

See Figures 5 and 6.

Fig. 5
figure5

Evolution of per capita pension expenditures (in constant-2010 euros) by country from 1995 to 2014

Fig. 6
figure6

Evolution of poverty rates (in %) by country from 1995 to 2014

Appendix 5: Baseline regression without country-fixed effects

See Tables 14, 15 and Fig. 7.

Table 14 Baseline regression results with no country-fixed effects
Fig. 7
figure7

Elasticity of the poverty rate to public pensions expenditures using the baseline regression with no fixed effects

Note The shaded areas represent a 95% confidence interval, derived using the delta method

Table 15 Baseline regression results with restricted definition of public pensions expenditures

Appendix 6: Robustness: Leave-out countries and group of countries

See Table 16.

Table 16 Elasticity to the mean by excluding successively the above (group of) countries

Appendix 7: Robustness: 2SLS with two-year lagged instruments

See Table 17.

Table 17 Two-stage least square

Appendix 8: Robustness: Changing the definition of the poverty line

The following Table 18 presents specification (3) of the baseline scenario (see Table 1) when we vary the poverty line at 40, 50 and 60% of the median or of the mean income for the computation of the poverty rate.

Table 18 Robustness analysis: variation of the poverty line

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Jacques, P., Leroux, ML. & Stevanovic, D. Poverty among the elderly: the role of public pension systems. Int Tax Public Finance 28, 24–67 (2021). https://doi.org/10.1007/s10797-020-09617-2

Download citation

Keywords

  • Ageing
  • Poverty
  • Income inequalities
  • Public pension systems
  • Panel data

JEL Classification

  • H55
  • I32
  • I38