Pension expenditure determinants: the case of Portugal

Maria Teresa Medeiros Garcia*

Maria Teresa Medeiros Garcia

Affiliation: Lisbon School of Economics and Management, Universidade de Lisboa, Lisboa, Portugal

0000-0001-8683-2112

Correspondence
mtgarcia@iseg.ulisboa.pt

   

André Fernando Rodrigues Rocha da Silva*

André Fernando Rodrigues Rocha da Silva

Affiliation: Lisbon School of Economics and Management, Universidade de Lisboa, Lisboa, Portugal

0000-0001-8123-6557

Article   |   Year:  2023   |   Pages:  177 - 203   |   Volume:  47   |   Issue:  2
Received:  April 21, 2022   |   Accepted:  January 23, 2023   |   Published online:  June 12, 2023

Download citation

https://doi.org/10.3326/pse.47.2.2

Abstract

1 Introduction

2016

2017

2014

However, a public pension system is expected to experience a pattern of increasing expenditures from the early years of its existence, until it reaches a state of maturity (Plamondon et al., 2002). After a period of 65 to 70 years, under stable conditions, the expenditure of a scheme expressed as a percentage of insured earnings normally stabilizes, since the first generation of young new entrants to the scheme has passed through the various stages of participation.

Pension schemes mature very slowly, that is, over many decades (Cichon et al., 2004). Moreover, increases in pension expenditures are a perfectly normal phenomenon during the maturation phase of national pension schemes, which lasts several decades. Rising pension expenditures are not per se necessarily indicative of a financial sustainability issue.

Therefore, the design of pension financing systems should accommodate this expected growth of pension expenditure. However, pension privatization policies, implemented in a number of countries, in consequence of concern with the pattern of increasing pension expenditure (World Bank, 1994), did not deliver the expected results, as coverage¹ and pension benefits² did not increase, systemic risks were transferred to individuals and fiscal positions worsened (Beattie and McGillivray, 1995; ILO, 2018). Consequently, several countries are reversing privatization measures and returning to public solidarity-based systems.

In addition, recent austerity or fiscal consolidation trends affected the adequacy of pension systems and general conditions of retirement, putting at risk the fulfilment of the minimum standards in social security and, consequently, the contribution of public pension systems to the Sustainable Development Goals (SDGs) (ILO, 2017; 2018).

Few studies are available regarding the factors that influence the evolution of Portuguese pension expenditure, and whether there is a link between pension expenditure as a dependent variable and other relevant explanatory variables, including the most recent developments on relevant variables, covering the current environment and data (Garcia and Lopes, 2009; Garcia, 2017).

2 The Portuguese pension system

2015

In 2007, a sustainability factor was introduced for the calculation of the old age pension benefit, reducing it so that it takes life expectancy into account. Indeed, Portugal’s population is ageing very rapidly and shrinking, due among other things to very low fertility rates (OECD, 2019). A further change came in 2013, with a decrease in the pension benefit, although this only covered early retirement. This reform, whose effects will mainly be felt in the medium and long term, also intended to promote the financial sustainability of public finances, reducing the expected value of future pension expenditure and replacement rates. Simultaneously, as a consequence of the Portuguese bailout in 2011 (European Commission, 2011), an extraordinary solidarity contribution was also introduced which decreased all pension income.

In 2013, the normal retirement age was established at 66 years in 2014, but increased to 66 years and two months in 2015, following the automatic process of adjusting the normal age of retirement by two-thirds of gains in life expectancy from age 65, measured as the average of the previous two years (Garcia, 2017).

In summary, Portugal essentially has a pay-as-you-go pension scheme (World Bank, 2006), which represents the major source of retirement income, with occupational and personal pension funds only existing to a minor extent (Blake, 2006; European Parliament, 2011; Garcia, 2017). The Portuguese system is also a defined-benefit system (European Commission, 2015), offering pensioners more measurable postemployment income benefits (Ramaswamy, 2012). Pensions are indexed to prices and gross domestic product (European Commission, 2015; OECD, 2019).

3 Literature review

2012

2015

2015

2015

2005

2012

Although to limit public expenses, pension benefits might be decreased, retirement income adequacy is a concern (European Parliament, 2011; Chybalski and Marcinkiewicz, 2014). In this context, Orenstein (2011) calls attention to the fact that, from 1981 to 2007, more than thirty countries worldwide fully or partially replaced their pre-existing PAYG pension systems with ones based on individual, private savings accounts in a process often labelled “pension privatisation”.

However, pension privatization did not deliver the expected results (ILO, 2018), revealing limited effects on capital markets and economic growth. In fact, coverage rates and pension benefits decreased, the risk of financial market fluctuations was shifted to individuals, and administrative costs increased. Moreover, the high costs of transition created large fiscal pressures. In addition, private pension fund administration did not improve governance as, frequently, the regulatory and supervisory functions were captured by economic groups responsible for managing the pension funds, allowing concentration in pension industry. By using an overlapping generations model with a PAYG pension system, Cipriani (2014) concludes that population ageing due to increased longevity implies a reduction in pension benefits. However, the effects of ageing on pensions may not be negative if the elderly are free to choose their retirement age, while they are always negative in the case of full retirement (Cipriani, 2016). In light of the 2008 economic crisis, Halmosi (2014) emphasises that the study of the pension systems of developed countries is a priority issue. Indeed, Grech (2015) presents evidence that the impacts of the crisis were different for Continental and Mediterranean systems, pension benefits of the latter being cut back significantly. Analysing pension reforms in Greece, Italy, Portugal, and Spain, between, 1990 and 2013, Natali and Stamanti (2014) conclude that all these countries encouraged the spread of private pensions and harmonised their fragmented public schemes. In addition, cost containment was massive, putting future adequacy at risk. In addition, Natali (2015) provides a summary of reforms in Europe since the onset of the Great Recession, providing evidence that austerity has hit both public pay-as-you-go schemes and private pre-funded schemes alike. Indeed, both have been subject to measures to contain costs (e.g., a higher pensionable age, the introduction of automatic stabilisers of future spending, reduced indexation, and higher taxes and/or contributions). In fact, Diamond (1996), much earlier, suggested the indexation of normal retirement age to life expectancy, and the investment of part of the public reserve funds in the private economy as being good measures to solve the social security pension system problem.

The implications of population ageing for economic growth are also a cause for concern. In this context, Bloom, Canning and Fink (2010) conclude that OECD countries are likely to see modest – but not catastrophic – declines in the rate of economic growth, emphasising that policy reforms (including an increase in the legal age of retirement) can mitigate the economic consequences of an ageing population. More recently, Žokalj (2016) examines the fiscal implications of the demographic shift using panel data on 25 European Union countries in the period from 1995 until 2014. The results suggest significant and positive impacts of the elderly share on expenditure for pensions and social protection.

In order to disentangle the macroeconomic impacts on the PAYG Portuguese social security system, Garcia and Lopes ( 2009) conclude that some cumulative measures, such as a changing of indexing rules, a better actuarial match between pensions and contributions, and measures to increase the effective age of retirement, could have a bigger impact on reducing the expected increase in pension expenditure than applying a systemic pension reform. Using a macroeconomic model of the Portuguese economy, the estimations suggest that the elimination of early retirement schemes, combined with an increase in the effective contribution rate could be a good alternative to promote the financial sustainability of the system. Economic growth strengthened by the pension reserve fund (which had an average annual nominal rate of return of 5.17% during the period 1989-2014, and relatively low administrative costs compared with funded systems), brings more advantages to the system when compared with a fully pre-funded system, which has high transition costs, with current tax payers being responsible for paying both their own and the existing pensioners’ benefits (European Parliament, 2011).

4 Methods

4.1 Variables

2015

The second group tries to disentangle the impact of the main pension system laws since 1975 (Garcia, 2017). Therefore, five dummy variables were set, each of which refers to a specific period, that is to say, the variable’s value will be 1 if included in that specific period, and 0 otherwise. The events are: 

4.2 Methodology

2014

(1a)

The data source is PORDATA (Francisco Manuel dos Santos Foundation, 2010) and the descriptive statistics for the variables used in the analysis are presented in the appendix (table A1).

4.3 Data analysis

2009

2014

Non-stationarity may render the potential econometric results statistically invalid. Typically, the linear combination of I(1) variables will be I(1), but it is desirable to obtain I(0) residuals, which are only achieved if the linear combination of I(1) variables is I(0), that is to say, if the variables are co-integrated (Brooks, 2014).

The Johansen co-integration test allows for the selection of the appropriate lag length and model to choose (table A4). The test result suggests that the number of appropriated lags is two (as stated above), with one co-integrating vector, and the model to adopt consists of the allowance of a quadratic deterministic trend, with intercept and trend in the co-integration equation and intercept in VAR, following Akaike Information Criteria (Brooks, 2014).

2014

Δγt = Πγt-k + Γ1 Δγt-1 + ... + Γk-1 Δγt-(k-1) + ut

(2)

Where   

This VECM contains g variables in first-differenced form on the LHS, and k-1 lags of the dependent variables (differences) on the RHS, each with a Γ short-run coefficient matrix. П consists of a long-run coefficient matrix, as being in equilibrium, all the Δγ_t-1 = 0, and setting E(u_t ) = 0 will leave Πγ_t-k = 0. Π illustrates the speed of adjustment back to equilibrium, that is to say, it measures the proportion of the last period´s equilibrium error that it is corrected for (Brooks, 2014).

The VECM model estimation is depicted in table 1 and encompasses the co-integration equation with dummy variables.

As all inverse roots of the characteristic polynomial are inside the unit circle, the model is stable. The residuals assumptions were tested, and it is possible to conclude that the mean of the residuals is zero (table A5). The White heteroscedasticity test p-value does not allow for the rejection of homoscedastic residuals (table A6). In addition, the covariance between residuals and explanatory variables is zero, thus satisfying the assumption of there being no relationship between them (table A7) and that the residuals are normally distributed (table A8). Finally, the null hypothesis of no residual serial correlation is not rejected at 5% significance level with the use of two lags (table A9).

As such, the estimators are efficient, and the confidence intervals and hypothesis tests using t and F-statistics are reliable.

Table 1

VECM estimation results

DISPLAY Table

5 Results

The presence of a co-integrating vector illustrates an equilibrium phenomenon, as it is possible that co-integrated variables may deviate from their relationship in the short run, but that their association will return in the long run (Brooks, 2014).

The positive long-run coefficient of log apparent labour productivity on pensions to gross domestic product ratio is not in line with European Commission (2015).

Concerning the negative coefficient of the old age dependency ratio, this might be the consequence of the parametric changes introduced to the system since 2000 (Garcia, 2017), especially the change of the normal retirement age (NRA) to 66 years, in 2013, becoming life expectancy-dependent after 2014. Therefore, an increase of the old age dependency ratio does not necessarily imply an increase of pension expenditure as a share of gross domestic product in the long-run. This measure is strongly supported by the literature as a crucial measure to guarantee the financial sustainability of pension systems, smoothing the impact of an ever-increasing number of pensioners (Diamond, 1996; Clements et al., 2015). The introduction of a sustainability factor into the benefit calculation formula, which is related to the evolution of average life expectancy (ALE), also represents a significant decrease in the pension benefit.

Finally, the impulse-response functions were stressed, as well as the variance decomposition for pensions to gross domestic product ratio, which is strongly dependent on the Cholesky ordering, which does not follow a specific requirement (Brooks, 2014). In order to guarantee some consistency and reasonability in the results, the order considered was from the most exogenous variable to the most endogenous one, determined by a VEC Granger Causality Test. The higher the p-value, the greater the exogeneity of the variable. The adopted order is as follows: old age dependency ratio, log unemployment, pensions to gross domestic product ratio and log apparent productivity of labour.

Following the methodology of Brooks (2014), figure 1 gives the impulse responses of the pensions to gross domestic product ratio, regarding unit shocks in the old age dependency ratio, in the log unemployment, and in log apparent productivity of labour, and their impact during 20 periods (years) ahead. Considering the signs of the responses, shocks in the old age dependency ratio have a positive impact until the 5^th year, achieving its peak in the 3^rd year. After this, the impact is negative, although the effect of the shock ends up dying down. A standard deviation shock in log unemployment and log apparent productivity of labour always has a positive impact on pensions to gross domestic product ratio, reaching its peak in the 4^th and 3^rd years, respectively, and stagnating in the long-run. Finally, the own impulse in pensions to gross domestic product ratio registers a similar impact as log unemployment, that is to say, it reaches its peak in the 4^th year, and then stagnation thereafter.

When analysing this approach, the main highlight is the fact that the old age dependency ratio registers an almost irrelevant contribution for the evolution of pensions to gross domestic product ratio in the long-run, when compared with the other variables, which is surpassed by the effects of log unemployment and log apparent productivity of labour, this reinforcing the doubts about the impact of ageing on pension expenditure. It is also possible to verify the relevance of unemployment in the presence of a positive shock immediately in the first years (as stressed by European Commission, 2015), over a 20-year forecasting horizon (positive but constant impact), shrinking the contributory base and the economic growth, with a similar pattern in relation to the apparent productivity of labour, guaranteeing higher pension entitlements.

The negative relationship between pensions to gross domestic product ratio and old age dependency ratio supports the hypothesis of a spurious result. Therefore, the Johansen co-integration test with dummy variables was carried out (table A12), although there is a problem in that the critical values may not be valid with exogenous series, such as dummy variables (Johansen, Mosconi and Nielsen, 2000; Giles and Godwin, 2012).

With this test, the old age dependency ratio long-run coefficient becomes positive and the sign of the other two coefficients does not change. However, it is important to take into account the econometric limitations of this change. To derive the VECM p-values, the VECM model with the coefficients as C(1) until C(16) was developed (table 2). C(1) is the coefficient of the co-integration equation (as well as the speed of adjustment back to equilibrium), C(10) is the constant, C(2) up to C(9) are the short-run coefficients of the lagged variables (until the second lag), and C(12) until C(16) are the coefficients of the dummy variables. C(11) is the trend coefficient (Brooks, 2014).

C(1), which is negative and statistically significant at 5%, confirms the long-run relationship between pensions to gross domestic product ratio, log unemployment, log apparent labour productivity, and the old age dependency ratio, as well as the existence of a correction mechanism of deviations (Wooldridge, 2009). When carrying out the Wald tests (table A13), it is not possible to reject the null hypothesis of C(4)=C(5)=0, C(6)=C(7)=0 and C(8)=C(9)=0, and the conclusion that needs to be stressed is the absence of short-run causality running from log unemployment, log apparent labour productivity, and the old age dependency ratio to pensions to gross domestic product ratio.

6 Conclusion

The apparent productivity of labour also seems to have a positive impact on pension expenditure to gross domestic product, which is not in line with European Commission (2015), supporting the assumption that gross domestic product growth is larger than pension expenditure growth in Portugal, because pensions are not fully indexed to wages after retirement.

The most intriguing result concerns the old age dependency ratio. In fact, after the development of the Johansen co-integration tests, both without dummy variables and with dummy variables, the old-age dependency ratio long-run coefficient presents different signs, giving rise to the hypothesis that ageing may not be the most relevant factor jeopardising the financial sustainability of the Portuguese public pension system. This is corroborated by the irrelevance of the influence of the old-age dependency ratio (in the long-run) on the impulse-response-functions, suggesting that the system has reached a state of maturity (Plamondon et al., 2002). Furthermore, this is in line with European Commission projections presented in the 2021 Ageing Report concerning Portugal (European Commission, 2021). Indeed, the country is expected to experience an overall decline in public pension expenditure (-3.2 percentage points) from 2019 to 2070 while the share of the age cohorts above 65 years in the total population is expected to rise from 22% to 33.1% (11.1 percentage points) in the same period.

When designing a pension system policy to reinforce the financial sustainability of the system, policy makers should take these findings into account. In other words, apparently, an increasing demographic strain seems not to impact pension expenditure as critically as unemployment. Therefore, policies to reduce unemployment should be considered as policy options to control pension expenditure, which represents a difficult way to address the financial sustainability of public pension systems. This is even more challenging in a stagflation environment, since actions intended to lower inflation may exacerbate unemployment. Future lines of research should try to do a similar analysis including more recent data and to disentangle the shape of pension expenditures over time (whether it follows an expected a logistical curve) (Cichon et al., 2004).

Appendix

Funding

Notes

2014

2018

* The authors are grateful to two anonymous referees who have contributed to the quality of the final version of the paper.

¹ Coverage (also denoted contribution density) is defined both as the proportion of workers participating in pension schemes and the proportion of the elderly receiving some kind of pension income (OECD/Inter-American Development Bank/The World Bank, 2014).

² These pension benefits are available to people who have reached pensionable age through: (i) earnings-related contributory pensions (guaranteeing minimum benefit levels, or replacement rates corresponding to a prescribed proportion of an individual’s past earnings – in particular for those with lower earnings); and/or (ii) flat-rate pensions (mostly residency-based and financed by the general budget) and/or means tested pensions (ILO, 2018).

Disclosure statement

References

1. Beattie, R. and McGillivray, W. 1995. A risky strategy: Reflections on the World Bank Report Averting the old age crisis. International Social Security Review, 48(3-4), pp. 5-23 [CrossRef]


2. Blake, D. 2006. Pension Economics. Toronto: John Wiley & Sons.


3. Bloom, D. E., Canning, D. and Fink, G., 2010. Implications of population ageing for economic growth. Oxford Review of Economic Policy, 26(4), pp. 583-612 [CrossRef]


4. Brooks, C., 2014. Introductory econometrics for finance. Cambridge: Cambridge University Press [CrossRef]


5. Chybalski, F. and Marcinkiewicz, E., 2014. How to measure and compare pension expenditures in cross-country analyses? Some methodological remarks. International Journal of Business and Management, 2(4), pp. 44-59.


6. Cichon, M. [et al.], 2004. Financing social protection. Geneva: International Labour Office and International Social Security Association.


7. Cipriani, G. P., 2014. Population aging and PAYG pensions in the OLG model. Journal of Population Economics, 27, pp. 251-256 [CrossRef]


8. Cipriani, G. P., 2016. Aging, Retirement and Pay-As-You-Go Pensions. IZA Discussion Paper, No. 9969 [CrossRef]


9. Clements, B. [et al.], 2015. The Fiscal Consequences of Shrinking Populations. IMF Staff Discussion Note, No. 21 [CrossRef]


10. Diamond, P. A., 1996. Proposals to Restructure Social Security. Journal of Economic Perspectives, 10(3), pp. 67-88 [CrossRef]


11. Eurogroup, 2016. Common principles for strengthening pension sustainability, 16 June 2016.


12. Eurogroup, 2017. Structural reform agenda - thematic discussions on growth and jobs: benchmarking pension sustainability, 144/17, 20 March 2017.


13. ECB, 2015. Box 7 The 2015 Ageing Report: how costly will ageing in Europe be? Economic Bulletin, (4).


14. European Commission, 2011. The Economic Adjustment Program of Portugal. Occasional paper, No. 79.


15. European Commission, 2012. The 2012 Ageing Report. European economy, No. 2.


16. European Commission, 2014. Identifying fiscal sustainability challenges in the areas of pension, health care and long-term care policies. Occasional papers, No 201.


17. European Commission, 2015. The 2015 Ageing Report: Economic and budgetary projections for the 28 EU Member States (2013-2060). European economy, No. 3.


18. European Commission, 2021. The 2021 Ageing Report: economic and budgetary projections for the EU Member States (2019-2070). Institutional paper, No. 148.


19. European Parliament, 2011. Pension Systems in the EU – Contingent Liabilities and Assets in the Public and Private Sector.


20. Francisco Manuel dos Santos Foundation, 2010. PORDATA: the database of contemporary Portugal.


21. Garcia, M. T. M. and Lopes, E. G., 2009. The macroeconomic impact of reforming a PAYG system: The Portuguese Case. International Social Security Review, 62(1), pp. pp. 1-23.


22. Garcia, M. T. M., 2017. Overview of the Portuguese Three Pillar Pension System. International Advances in Economic Research, 23, pp. 175-189 [CrossRef]


23. Giles, D. E. and Godwin, R. T., 2012. Testing for multivariate cointegration in the presence of structural breaks: p-values and critical values. Applied Economics Letters, 19(16), pp. 1561-1565 [CrossRef]


24. Grech, A. G., 2015. Convergence or divergence? How the financial crisis affected European pensioners. International Social Security Review, 68(2), pp. 43-61 [CrossRef]


25. Halmosi, P., 2014. Transformation of the Pension Systems in OECD Countries after the 2008 Crisis. Public Finance Quarterly, (4), pp. 457-469 [CrossRef]


26. ILO, 2017. World Social Protection Report 2017–19: Universal social protection to achieve the Sustainable Development Goals. Geneva: International Labour Office.


27. ILO, 2018. Social protection for older persons: Policy trends and statistics 2017-19. Social Protection Policy Papers, No. 17.


28. Johansen, S., Mosconi, R. and Nielsen, B., 2000. Cointegration analysis in the presence of structural breaks in the deterministic trend. The Econometrics Journal, 3(2), pp. 216-249.


29. Natali, D. and Stamati, F., 2014. Reassessing South European pensions after the crisis: Evidence from two decades of reforms. South European Society and Politics, 19(3), pp. pp. 309-330 [CrossRef]


30. Natali, D., 2015. Pension Reform in Europe: What has happened in the Wake of the Crisis? CESifo DICE Report, No. 2/2015, pp 31-35 [CrossRef]


31. OECD, 2014. OECD Economic Surveys: Portugal 2014. Paris: OECD [CrossRef]


32. OECD, 2015. Pensions at a Glance 2015. Paris: OECD [CrossRef]


33. OECD, 2019. OECD Reviews of Pension Systems: Portugal, OECD Reviews of Pension Systems. Paris: OECD [CrossRef]


34. OECD/Inter-American Development Bank/The World Bank, 2014. Policy issues: Coverage and adequacy. In: Pensions at a Glance: Latin America and the Caribbean. Paris: OECD [CrossRef]


35. Orenstein, M. A., 2011. Pension Privatization in Crisis: Death or Rebirth of a Global Policy Trend? International Social Security Review, 64(3), pp. 65-80 [CrossRef]


36. Plamondon, P. [et al.], 2002. Actuarial practice in social security. Geneva: International Labour Office/International Social Security Association, 2002.


37. Ramaswamy, S., 2012. The Sustainability of Pension Schemes. BIS Working Papers, No 368 [CrossRef]


38. Roach, B. and Ackerman, F., 2005. Securing Social Security: Sensitivity to Economic Assumptions and Analysis of Policy Options. GDAE Working Paper, No. 05-03.


39. Wooldridge, J. M., 2009. Introductory Econometrics, A Modern Approach. 4th ed. Boston: South-Western.


40. World Bank, 1994. Averting the Old Age Crisis: Policies to Protect the Old and Promote Growth. Oxford University Press.


41. World Bank, 2006. Pension Reform and the Development of Pension Systems: An Evaluation of World Bank Assistance. Washington, DC: Independent Evaluation Group, World Bank.


42. Zokalj, M., 2016. The impact of population aging on public finance in the European Union. Financial Theory and Practice, 40(4), pp. 383-412 [CrossRef]

Beattie, R. and McGillivray, W. 1995. A risky strategy: Reflections on the World Bank Report Averting the old age crisis. International Social Security Review, 48(3-4), pp. 5-23 [CrossRef]

Blake, D. 2006. Pension Economics. Toronto: John Wiley & Sons.

Bloom, D. E., Canning, D. and Fink, G., 2010. Implications of population ageing for economic growth. Oxford Review of Economic Policy, 26(4), pp. 583-612 [CrossRef]

Brooks, C., 2014. Introductory econometrics for finance. Cambridge: Cambridge University Press [CrossRef]

Chybalski, F. and Marcinkiewicz, E., 2014. How to measure and compare pension expenditures in cross-country analyses? Some methodological remarks. International Journal of Business and Management, 2(4), pp. 44-59.

Cichon, M. [et al.], 2004. Financing social protection. Geneva: International Labour Office and International Social Security Association.

Cipriani, G. P., 2014. Population aging and PAYG pensions in the OLG model. Journal of Population Economics, 27, pp. 251-256 [CrossRef]

Cipriani, G. P., 2016. Aging, Retirement and Pay-As-You-Go Pensions. IZA Discussion Paper, No. 9969 [CrossRef]

Clements, B. [et al.], 2015. The Fiscal Consequences of Shrinking Populations. IMF Staff Discussion Note, No. 21 [CrossRef]

Diamond, P. A., 1996. Proposals to Restructure Social Security. Journal of Economic Perspectives, 10(3), pp. 67-88 [CrossRef]

Eurogroup, 2016. Common principles for strengthening pension sustainability, 16 June 2016.

Eurogroup, 2017. Structural reform agenda - thematic discussions on growth and jobs: benchmarking pension sustainability, 144/17, 20 March 2017.

ECB, 2015. Box 7 The 2015 Ageing Report: how costly will ageing in Europe be? Economic Bulletin, (4).

European Commission, 2011. The Economic Adjustment Program of Portugal. Occasional paper, No. 79.

European Commission, 2012. The 2012 Ageing Report. European economy, No. 2.

European Commission, 2014. Identifying fiscal sustainability challenges in the areas of pension, health care and long-term care policies. Occasional papers, No 201.

European Commission, 2015. The 2015 Ageing Report: Economic and budgetary projections for the 28 EU Member States (2013-2060). European economy, No. 3.

European Commission, 2021. The 2021 Ageing Report: economic and budgetary projections for the EU Member States (2019-2070). Institutional paper, No. 148.

European Parliament, 2011. Pension Systems in the EU – Contingent Liabilities and Assets in the Public and Private Sector.

Francisco Manuel dos Santos Foundation, 2010. PORDATA: the database of contemporary Portugal.

Garcia, M. T. M. and Lopes, E. G., 2009. The macroeconomic impact of reforming a PAYG system: The Portuguese Case. International Social Security Review, 62(1), pp. pp. 1-23.

Garcia, M. T. M., 2017. Overview of the Portuguese Three Pillar Pension System. International Advances in Economic Research, 23, pp. 175-189 [CrossRef]

Giles, D. E. and Godwin, R. T., 2012. Testing for multivariate cointegration in the presence of structural breaks: p-values and critical values. Applied Economics Letters, 19(16), pp. 1561-1565 [CrossRef]

Grech, A. G., 2015. Convergence or divergence? How the financial crisis affected European pensioners. International Social Security Review, 68(2), pp. 43-61 [CrossRef]

Halmosi, P., 2014. Transformation of the Pension Systems in OECD Countries after the 2008 Crisis. Public Finance Quarterly, (4), pp. 457-469 [CrossRef]

ILO, 2017. World Social Protection Report 2017–19: Universal social protection to achieve the Sustainable Development Goals. Geneva: International Labour Office.

ILO, 2018. Social protection for older persons: Policy trends and statistics 2017-19. Social Protection Policy Papers, No. 17.

Johansen, S., Mosconi, R. and Nielsen, B., 2000. Cointegration analysis in the presence of structural breaks in the deterministic trend. The Econometrics Journal, 3(2), pp. 216-249.

Natali, D. and Stamati, F., 2014. Reassessing South European pensions after the crisis: Evidence from two decades of reforms. South European Society and Politics, 19(3), pp. pp. 309-330 [CrossRef]

Natali, D., 2015. Pension Reform in Europe: What has happened in the Wake of the Crisis? CESifo DICE Report, No. 2/2015, pp 31-35 [CrossRef]

OECD, 2014. OECD Economic Surveys: Portugal 2014. Paris: OECD [CrossRef]

OECD, 2015. Pensions at a Glance 2015. Paris: OECD [CrossRef]

OECD, 2019. OECD Reviews of Pension Systems: Portugal, OECD Reviews of Pension Systems. Paris: OECD [CrossRef]

OECD/Inter-American Development Bank/The World Bank, 2014. Policy issues: Coverage and adequacy. In: Pensions at a Glance: Latin America and the Caribbean. Paris: OECD [CrossRef]

Orenstein, M. A., 2011. Pension Privatization in Crisis: Death or Rebirth of a Global Policy Trend? International Social Security Review, 64(3), pp. 65-80 [CrossRef]

Plamondon, P. [et al.], 2002. Actuarial practice in social security. Geneva: International Labour Office/International Social Security Association, 2002.

Ramaswamy, S., 2012. The Sustainability of Pension Schemes. BIS Working Papers, No 368 [CrossRef]

Roach, B. and Ackerman, F., 2005. Securing Social Security: Sensitivity to Economic Assumptions and Analysis of Policy Options. GDAE Working Paper, No. 05-03.

Wooldridge, J. M., 2009. Introductory Econometrics, A Modern Approach. 4th ed. Boston: South-Western.

World Bank, 1994. Averting the Old Age Crisis: Policies to Protect the Old and Promote Growth. Oxford University Press.

World Bank, 2006. Pension Reform and the Development of Pension Systems: An Evaluation of World Bank Assistance. Washington, DC: Independent Evaluation Group, World Bank.

Zokalj, M., 2016. The impact of population aging on public finance in the European Union. Financial Theory and Practice, 40(4), pp. 383-412 [CrossRef]