Abstract
Selected Issues
Household Saving in Portugal1
A. Introduction
1. Portugal's household saving is low by European standards, 3.1 percent of GDP in 2018. Household saving, both as a percent of GDP and of disposable income, has been on a declining trend for the last two decades.
Household Saving, 2017
(Percent of GDP)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Household Saving, 2017
(Percent of GDP)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Household Saving, 2017
(Percent of GDP)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.2. Higher growth over the medium-term will require stronger domestic saving to finance additional private investment. Larger private savings will be necessary to sustain sufficiently high investment rates without creating new external imbalances.
Portugal: Current Account by Sector
(Percent of GDP)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Sources: Eurostat; and IMF staff calculations.Portugal: Current Account by Sector
(Percent of GDP)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Sources: Eurostat; and IMF staff calculations.Portugal: Current Account by Sector
(Percent of GDP)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Sources: Eurostat; and IMF staff calculations.3. Stronger private saving is also needed to mitigate the impact of adverse demographic trends. Portugal's population is aging rapidly, as life expectancy increases and fertility rate declines. The population is already decreasing, 10.3 million in 2017, from peaking at 10.7 million in 2009. According to the 2017 UN projections, the share of working-population (15–64) will decline more than the Europe average in about 10 years, with gap widening further over time. Ageing and decreasing population will weigh on labor's social security contributions, and intensify pressures on social programs, pensions, and health care. As the ratio of workers to the population declines, and as the pension reforms enacted a decade ago gradually complete their transition phase, people will need to save more for retirement.
Working-age Population Projections
(Percent of total population)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: UN World Population Prospects.Working-age Population Projections
(Percent of total population)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: UN World Population Prospects.Working-age Population Projections
(Percent of total population)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: UN World Population Prospects.4. The paper focuses on cross-country differences in savings rates in advanced European countries. It explores a range of demographic, fiscal and financial factors that could explain why household savings are low in Portugal compared to its peers. The paper proceeds as follows. In the next section, we look at cross-country comparison and dynamics of household saving over the last two decades. In section C, we provide a snapshot of savers' profile for EU countries based on the 2010 micro-data compiled by Eurostat. In section D, we discuss the determinants of household savings in Europe based on panel regressions and a literature review. Section E focuses on empirical results for Portugal. Section F discusses policy options and section G concludes.
B. Cross-Country Comparison and Dynamics
5. Despite a notable increase in the last decade, from 10.7 percent of GDP in 2009 to
17.1 percent in 2017, saving in Portugal remains below the euro area and the EU averages. The increase in national saving has been supported mostly by the recovery of corporate and government saving, offsetting a decline in household saving.2 While a shift in the composition of saving away from the household sector and toward the corporate sector is a global trend (Chen, Karabarbounis & Neiman, 2017), the divergence is more striking in Portugal than in other European countries. The private saving composition change in Portugal can be explained by lower total wages, less dividend distribution by firms and tax increases on households (Banco de Portugal (BdP) Economic Bulletin, May 2016).
Private Saving
(Percent of GDP)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Private Saving
(Percent of GDP)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Private Saving
(Percent of GDP)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.
Portugal: Saving by Sector
(Percent of GDP)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Portugal: Saving by Sector
(Percent of GDP)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Portugal: Saving by Sector
(Percent of GDP)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.
Euro Area: Saving by Sector
(Percent of GDP)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Euro Area: Saving by Sector
(Percent of GDP)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Euro Area: Saving by Sector
(Percent of GDP)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.6. Household saving and investment rates in Portugal are among the lowest in Europe. The average gross household saving and investment rates (to disposable income) for the first three quarters of 2018 were 3.6 percent and 5 percent, respectively. Overall, the saving rate in Portugal has exhibited a declining trend since 2002, with a temporary pick up in the crisis years. The investment rate had declined until 2014 and has been modestly increasing since then. The saving-investment balance turned negative in 2017, after being positive during the crisis and marginally positive in the post-crisis years.
Household Saving Rate
(Percent; 2018:Q1–2018:Q4)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Household Saving Rate
(Percent; 2018:Q1–2018:Q4)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Household Saving Rate
(Percent; 2018:Q1–2018:Q4)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.
Household Investment Rate
(Percent; 2018:Q1–2018:Q4)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Household Investment Rate
(Percent; 2018:Q1–2018:Q4)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Household Investment Rate
(Percent; 2018:Q1–2018:Q4)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.7. Portugal experienced a significant increase in household saving from 2009 to 2013 in both saving to GDP and saving to disposable income ratios. This likely, reflected higher precautionary savings due to greater macroeconomic uncertainty, less access to credit and weakening in a social safety net from the government during that period.3 It could also be explained by the decline in income during those years, which was concentrated among households with above-average propensity to consume. The temporary increase in household saving rates during the crisis was more pronounced in Portugal and other countries affected by the euro area sovereign debt crisis. In the recent years, as the economic activity recovered, the saving rate resumed its declining trend.
Household Saving Rate
(Percent; change from 2008 to 2009)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Household Saving Rate
(Percent; change from 2008 to 2009)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Household Saving Rate
(Percent; change from 2008 to 2009)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.
Household Saving Rate
(Percent; 4Q moving average)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Household Saving Rate
(Percent; 4Q moving average)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Household Saving Rate
(Percent; 4Q moving average)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.C. Savers Profile: Snapshot of 2010
8. This section is based on the Eurostat experimental data for 2010 on interaction of household income, consumption, and wealth, calculated using statistical matching and modeling from the combination of different surveys. While Eurostat data helps to better understand the household saving behavior, it is subject to some limitations4 and may not be directly comparable with the empirical results based on national accounts data due to conceptual and measurement differences. When available, we compare the results with studies based on other surveys, including the Household Financial Consumer Survey (HFCS) data and the Household Budget Survey (HBS) for the earlier and later years.
9. Median vs average saving rate. In Portugal the median saving rate (14 percent) exceeded the average (8.2 percent), implying that the saving rate distribution skewed to the lower end, with a longer tail of individuals (or households) that save less or dissave, pulling down the average rate. Indeed, in Portugal the proportion of households that dissave (spend more than they earn) is 39 percent, higher than in many of the european countries. The same result is reported by the OECD, the share of dissaving households in Portugal was 38 percent in 2010, higher than OECD average of 25.4 percent.5 Portugal compares better – closer to the sample average – in terms of the median saving rate.
Average Saving Rate and Median Saving Rate of the Reference Person
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Average Saving Rate and Median Saving Rate of the Reference Person
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Average Saving Rate and Median Saving Rate of the Reference Person
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.
Proportion of Households with Expenditures Higher than Income
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Proportion of Households with Expenditures Higher than Income
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Proportion of Households with Expenditures Higher than Income
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.10. Median saving rate by income quantile. In most countries, the saving rate increases with income, with the lowest income quantile having negative saving rate. Portugal is broadly in line with the cross-country patterns. Alves and Cardoso (2010) find even stronger inequality by income quantile, using the Household Budget Survey (HBS) for 2004/2005: saving rates increase with income and wealth, with 90 percent of total saving generated by only 20 percent of households. Similarly, using the first wave of HFCS data, Rodriguez-Palenzuela and Dees (2016) document that about 90 percent of total savings are generated by top 20 percent. In Portugal, households in the top income quantiles appear to have lower saving rate compared to the rest of Europe; however, according to the first wave of HFSC data6, the saving rate by top quantile is higher, more comparable with the other euro area countries.
Median Saving Rate by Household Income Quintile
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Median Saving Rate by Household Income Quintile
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Median Saving Rate by Household Income Quintile
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.11. Accordingly, the share of households with negative saving is higher for lower income groups: about two-thirds for the lower income, a third for the middle-income and 15 percent for the upper income groups in Portugal. Also, the share of over-indebted households in Portugal is larger for the middle-income group (23.8 percent) than for the total population (16.8 percent), and it is above the OECD average of 13.1 percent.7 Le Blanc et al (2016) find that during 2008–2011 negative saving was often financed with informal loans in Greece and Portugal.
Share of Households that Spend More than They Earn*, by Income Class
(Percent, 2016 or latest year available)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Sources: OECD, Eurostat, and national sources.* Households that overspend are defined as those whose total expenditures are higher than total household disposable income.Share of Households that Spend More than They Earn*, by Income Class
(Percent, 2016 or latest year available)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Sources: OECD, Eurostat, and national sources.* Households that overspend are defined as those whose total expenditures are higher than total household disposable income.Share of Households that Spend More than They Earn*, by Income Class
(Percent, 2016 or latest year available)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Sources: OECD, Eurostat, and national sources.* Households that overspend are defined as those whose total expenditures are higher than total household disposable income.12. Median saving rate by age. In many countries the saving rate by age is hump-shaped, while in several countries it is increasing with age. In our sample data for Portugal, people aged 60- and older save the most. However, the literature offers less assurances on this point. Rodriguez-Palenzuela and Dees (2016) also show the age group of 75 has the highest median saving rate, followed by the 65 and over group, based on the 2010 HFSC data. In contrast, BdP Economic Bulletin (2016) show that the highest median saving rates are by the age groups of 35–44- and 55–64-years old, based on the 2013 HFCS data. Alves and Cardoso (2010) find that in Portugal the age group of 45–54 had highest saving rate in 2005.
Median Saving Rate by Age of the Reference Person
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Median Saving Rate by Age of the Reference Person
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Median Saving Rate by Age of the Reference Person
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.13. Median saving rate by household type and by education level. Families with children tend to save less. More adults in the households are associated with more savings (see Figure 3). Le Blanc et al (2016) report that for the Euro Area over 2008–2011 household size is significantly and negatively associated with saving for old-age provision. The distribution of saving rates by household type for Portugal is similar to that for other EU countries. Individuals with higher education tend to save more. This is consistent with the HFSC data presented in Rodriguez-Palenzuela and Dees (2016). Alves and Cardoso (2010) and Ares, Lopez, and Bua (2015) find that saving rates are positively related to education. Further, Le Blanc et al (2016) find education to be a significant determinant for home purchase and for precautionary saving. Inequality in saving is similar to inequality in wealth in Portugal, and higher than the EU average. Acording to the 2015 HBS data, inequality in the distribution of expenditure in Portugal is among the highest in the euro area (BdP Economic Bulletin, June 2018).
Household Saving Determinants
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: IMF staff.Household Saving Determinants
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: IMF staff.Household Saving Determinants
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: IMF staff.
Micro-Data Charts
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Micro-Data Charts
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Micro-Data Charts
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Selected Determinants of Household Saving
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Sources: Eurostat and OECD.Selected Determinants of Household Saving
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Sources: Eurostat and OECD.Selected Determinants of Household Saving
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Sources: Eurostat and OECD.14. These micro-data results are broadly in line with findings based on other surveys: the saving rate is positively associated with income and education, it tends to increase with age and then decline for the oldest age group, although not in all countries. Using the survey of Health, Ageing and Retirement for households with at least one person of 50+ age in Portugal and Spain, Ares, Lopez, Bua (2015) find that saving is positively related to education, employment status, home ownership, saving habits, and area of residence, and negatively related to financial risk-aversion. Based on the Household Expenditure Survey for 2005/2006, Alves and Cardoso (2010) find saving rates to be positively related to education and homeownership, and negatively related to unemployment. Using the 2013 HFCS data, BdP Economic Bulletin (May 2016) show the household saving rate declined during the first decade of the euro due to reduced liquidity constraints and lower income inequality. In recent years, household saving has also been affected by macroeconomic uncertainty and lower permanent income expectations.
D. Determinants of Household Saving Rates
15. According to the life-cycle hypothesis (Modigliani and Brumberg, 1954), individuals should plan their consumption and saving behavior to maintain stable lifestyles. This implies that a household is expected to borrow and dissave at a young age, accumulate resources during middle age, and deplete savings after retirement. The extended versions of this theory include the precautionary saving motive (Skinner, 1988, Gourinchas and Parker, 2002), the housing motive (Hayashi, Ito, and Slemrod, 1988), and the bequest motive (Hurd, 1987). As a result, household savings decisions are affected by income and by demographic factors. The literature suggests a wide range of variables that could influence the saving motives along these dimensions. In general, most of the factors work through both income and substitution effects, and the ultimate sign of the relationship depends on which effect dominates.
Empirical Analysis
16. We analyze a panel of 14 European advanced economies over 1999–2017. The country set includes: Austria, Belgium, Denmark, Finland, France, Germany, Ireland, Italy, Luxemburg, Netherlands, Portugal, Spain, Sweden, and the United Kingdom. We show that a combination of macroeconomic, demographic, fiscal and financial variables are significant and economically important determinants of household saving rates in European countries. The chart below provides a summary of the determinants, with selected bivariate relationships illustrated in Figure 3, and Table 1 presents the regression results.8
17. We focus on factors that help to explain saving rates (the dependent variable) as well as their cross-country variation. As some of the variables do not exhibit strong time variation, we relax country fixed effects (regressions 4–6 in Table 1) to better capture determinants of crosscountry differences. It comes at the expense of the explanatory power of these regressions but helps to throw light on slow-moving factors such as fertility, which is statistically significant only when country fixed effects are relaxed.
Income. Household saving rates are strongly associated with household disposable income, especially in per capita terms. This is consistent with the evidence from the HBS and HFCS data on saving rates in different income quantiles.
Demographic factors also play an important role, with population ageing negatively affecting saving rates. In some cases, this effect is offset by longer duration of the working life and higher old-age employment rate, which could explain why, according to some micro data, older people tend to save more. Our empirical results indicate that increased duration of the working life has positive and statistically significant coefficients across different specifications. Old-age dependency is negatively associated with saving rates, but it is statistically significant only when country fixed effects are relaxed. Employment of people over 65 years old have positive coefficients, but statistically significant only in some specifications. Fertility appears to be negatively associated with saving rates if country fixed effects are relaxed, consistent with the survey data evidence that households with more children tend to save less. However, lower fertility rates (together with longer life expectancy) also has an adverse impact on the country population growth and ageing, possibly captured in the regression analysis via old-age dependency.
Fiscal variables. Several fiscal policy variables are strongly associated with saving rates. Consistent with Ricardian equivalence, both public saving to GDP and fiscal balance to GDP ratios have negative and statistically significant coefficients, and so does the share of direct taxes. Government spending on pensions and on social protection benefits tend to have negative and statistically significant impact. The negative relationship between the saving rate and the aggregate replacement ratio is weak and not statistically significant in most regressions. The coverage of private pension schemes is positively associated with household saving, but data are available only for 2016, and therefore not used in our regression analysis.
Financial variables. Reflecting the income effect channel, household assets and financial net worth are positively related to saving rates, and the household debt ratio (both lagged and contemporaneous) is negatively related.9 Home ownership appears to be a strong factor negatively associated with the saving rate. The relationship between the saving rate and housing prices tends to be positive but not statistically significant.
Precautionary motive. We find unemployment to have positive and statistically significant coefficients across specifications. Unemployment can be positively associated with saving due to precautionary motives. For an individual, becoming unemployed could result in a depletion of savings to finance consumption, but this effect appears empirically weaker.
Social indicators. Education is positively associated with saving if country fixed effects are relaxed, consistent with the micro-data findings. Inequality appears to be negatively related to saving, but statistically significant only when country fixed effects are not included.
Portugal: Regressions Results
Portugal: Regressions Results
| Dependent var: Saving rate, log | eq1a | eq1b | eq2 | eq3 | eq4 | eq5 | eq6a | eq6b | ||
|---|---|---|---|---|---|---|---|---|---|---|
| Income | ||||||||||
| Real disposable income per capita, log | 1.41*** | 1.33*** | 2.29*** | 2.19*** | 2.32*** | 2.22*** | 2.42*** | 2.44*** | ||
| Demographics | ||||||||||
| Life expectancy, log | -9.80** | -9.81** | ||||||||
| Duration of working life, log | 1.61*** | 1.56*** | 4.57*** | 5.16*** | 1.41*** | 1.10*** | 1.12*** | |||
| Employment of people with age 65+, | 0.16*** | 0.17*** | 0.12*** | |||||||
| Fertility, log | -0.43*** | -0.48*** | -0.48*** | |||||||
| Fiscal variables | ||||||||||
| Public saving to GDP | -0.04*** | |||||||||
| Fiscal balance to GDP | -0.02*** | -0.02*** | -0.02*** | |||||||
| Pension to GDP, log | -0.80*** | -0.96** | -0.39*** | -0.38*** | ||||||
| Social Protection Benefits for old to GDP, log | -0.98*** | -0.62** | -0.20*** | -0.44*** | ||||||
| Direct Taxes, share, log | -0.25*** | |||||||||
| Fiancial variables | ||||||||||
| Financial assets to income, log, 1 year | 0.52*** | 0.55*** | 0.51*** | 0.64*** | 0.56*** | 0.57*** | ||||
| Net financial worth to income, log, 1 year lag | 0.46*** | 0.46*** | ||||||||
| Household debt to income, log, 1 year | -0.31*** | -0.33*** | -0.36* | -0.35* | -0.34*** | -0.56*** | -0.47*** | -0.47*** | ||
| Homeownership rate, log | -0.92** | -0.93* | -0.85*** | -0.63*** | -0.71*** | -0.75*** | ||||
| Housing prices index, real, log | 0.23* | |||||||||
| Housing prices index, nominal, log | 0.24** | |||||||||
| Precautionary motive | ||||||||||
| Unemployment, log | 0.39*** | 0.34*** | 0.64*** | 0.57*** | 0.59*** | 0.57*** | 0.62*** | 0.62*** | ||
| Country fixed effects | yes | yes | yes | yes | no | no | no | no | ||
| Time fixed effects | yes | yes | yes | yes | yes | yes | yes | yes | ||
| const | -16.67*** | -16.67*** | 10.38 | 8.81 | -19.75*** | -23.22*** | -24.73** | * -251*** | ||
| R-squared | 0.83 | 0.83 | 0.89 | 0.88 | 0.78 | 0.76 | 0.75 | 0.75 | ||
| Adj R-squared | 0.80 | 0.80 | 0.86 | 0.85 | 0.75 | 0.73 | 0.71 | 0.71 | ||
| Observations | 231 | 231 | 172 | 172 | 171 | 172 | 172 | 172 | ||
| countries | 14 | 14 | 14 | 14 | 14 | 14 | 14 | 14 | ||
| time periods | 17 | 17 | 14 | 14 | 14 | 14 | 14 | 14 | ||
Portugal: Regressions Results
| Dependent var: Saving rate, log | eq1a | eq1b | eq2 | eq3 | eq4 | eq5 | eq6a | eq6b | ||
|---|---|---|---|---|---|---|---|---|---|---|
| Income | ||||||||||
| Real disposable income per capita, log | 1.41*** | 1.33*** | 2.29*** | 2.19*** | 2.32*** | 2.22*** | 2.42*** | 2.44*** | ||
| Demographics | ||||||||||
| Life expectancy, log | -9.80** | -9.81** | ||||||||
| Duration of working life, log | 1.61*** | 1.56*** | 4.57*** | 5.16*** | 1.41*** | 1.10*** | 1.12*** | |||
| Employment of people with age 65+, | 0.16*** | 0.17*** | 0.12*** | |||||||
| Fertility, log | -0.43*** | -0.48*** | -0.48*** | |||||||
| Fiscal variables | ||||||||||
| Public saving to GDP | -0.04*** | |||||||||
| Fiscal balance to GDP | -0.02*** | -0.02*** | -0.02*** | |||||||
| Pension to GDP, log | -0.80*** | -0.96** | -0.39*** | -0.38*** | ||||||
| Social Protection Benefits for old to GDP, log | -0.98*** | -0.62** | -0.20*** | -0.44*** | ||||||
| Direct Taxes, share, log | -0.25*** | |||||||||
| Fiancial variables | ||||||||||
| Financial assets to income, log, 1 year | 0.52*** | 0.55*** | 0.51*** | 0.64*** | 0.56*** | 0.57*** | ||||
| Net financial worth to income, log, 1 year lag | 0.46*** | 0.46*** | ||||||||
| Household debt to income, log, 1 year | -0.31*** | -0.33*** | -0.36* | -0.35* | -0.34*** | -0.56*** | -0.47*** | -0.47*** | ||
| Homeownership rate, log | -0.92** | -0.93* | -0.85*** | -0.63*** | -0.71*** | -0.75*** | ||||
| Housing prices index, real, log | 0.23* | |||||||||
| Housing prices index, nominal, log | 0.24** | |||||||||
| Precautionary motive | ||||||||||
| Unemployment, log | 0.39*** | 0.34*** | 0.64*** | 0.57*** | 0.59*** | 0.57*** | 0.62*** | 0.62*** | ||
| Country fixed effects | yes | yes | yes | yes | no | no | no | no | ||
| Time fixed effects | yes | yes | yes | yes | yes | yes | yes | yes | ||
| const | -16.67*** | -16.67*** | 10.38 | 8.81 | -19.75*** | -23.22*** | -24.73** | * -251*** | ||
| R-squared | 0.83 | 0.83 | 0.89 | 0.88 | 0.78 | 0.76 | 0.75 | 0.75 | ||
| Adj R-squared | 0.80 | 0.80 | 0.86 | 0.85 | 0.75 | 0.73 | 0.71 | 0.71 | ||
| Observations | 231 | 231 | 172 | 172 | 171 | 172 | 172 | 172 | ||
| countries | 14 | 14 | 14 | 14 | 14 | 14 | 14 | 14 | ||
| time periods | 17 | 17 | 14 | 14 | 14 | 14 | 14 | 14 | ||
18. The foregoing results are broadly consistent with the main results in the literature— with some previous findings worth highlighting. Based on the panel analysis of 21 OECD countries over 1975–1995, Callen and Thimann (1997) find household saving is positively affected by income growth, and negatively affected by old dependency ratio, public saving, higher reliance on direct taxes (as a share of total taxes), and higher government transfers to households. Studying the euro area in the period 2008–2011, Le Blanc et al (2016) show that the gross replacement rate from the first (public) pillar remarkably decreases the importance of saving for old-age provision, suggesting a substitution effect between public and private pension savings. Amaglobeli et al (2019) show that pension generosity and sustainability (when interacted with old-age dependency and with life expectancy) are negatively associated with private saving.
19. According to the 2013 HFSC, Portuguese households mainly save to protect themselves against unexpected events (BdP Economic Bulletin, May 2016). Using the first wave HFCS that covers the years 2008–2011, Le Blanc et al (2016) find that for the euro area countries precautionary saving is the most commonly reported motive, followed by saving for old-age. From a cross-country view, they show that saving for home purchase and for old-age is more common in the Netherlands, Portugal, and Malta than in Germany.10
20. Based on a study of EU 15 countries over 2007–2013, Bouyon (2016) finds household saving positively affected by disposable income per capita and unemployment and negatively related to nominal housing prices. Analyzing household saving in Portugal during 1985–2009, Alves and Cardoso (2010) find that in the long-run the saving rate is positively related with the nominal interest rate and GDP growth, and negatively to the budget balance.
21. Inflation and real interest rate could influence the opportunity cost of savings as well as the borrowing costs. The empirical evidence is mixed, though, with some studies find a positive relationship, while others show no significant association. We find these variables not to be statistically significant in the regressions.
E. Empirical Results: Implications for Portugal
22. In this section, we examine to what extent the determinants discussed in the previous section affect the saving rate in Portugal, compared to the sample average. The strong negative contributors are disposable income per capita, financial assets to disposable income ratio, public saving to GDP, and pension to GDP. The positive contributors include employment of elderly, duration of working life, household debt, and unemployment. The country fixed effect coefficients for Portugal tend to be negative. The charts below illustrate the impact of those variables (based on the regression 1a from Table 1).
Income. Portugal has the lowest disposable income per capita among the 14 countries in the sample over the estimation period. The lower disposable income is a significant contributor to the difference between the saving rates in Portugal and the sample average.
Demographics. Portugal's old age dependency ratio is one of the highest in the sample, with the gap increasing over time, while life-expectancy is below the sample average. The duration of working life and employment of the elderly is above the sample average, positively affecting the household saving rate in Portugal compared to the other countries.
Fiscal variables. Portugal has lower public saving and lower fiscal balance compared to the sample average for much of the estimation period. Portugal also has one of the most generous pensions (measured against own average earnings), and one of the highest government spending on pensions and on social protection benefits (for elderly population) compared to the sample average. Both public saving to GDP and pension to GDP explain part of the difference in saving rates between Portugal and the country sample.
Financial variables. Portugal has lower financial assets and net wealth ratios to disposable income than the sample average. So along with the disposable income per capita, the financial assets ratio or the financial net wealth ratio (either lagged or contemporaneous) can explain a significant part of the difference in saving rates between Portugal and the sample average. Portugal's ratio of household debt to disposable income is slightly below the sample, even though the share of indebted households is above the European averages. As such, the household debt (lagged one-year) offsets some of the difference in the saving rates between Portugal and Europe.
Unemployment (“precautionary motive”). Compared to the sample average, Portugal had lower unemployment rates in the early 2000s, with the difference reversing in mid-2000 and peaking during the crisis years; this difference has shrunk in recent years. So, unemployment in Portugal moderated the relative shortfall in household saving from 2005.
Country fixed effects. Unsurprisingly, country fixed effect estimates for Portugal are negative in most specifications. The negative country fixed effect possibly captures some of the slow-changing variables such as homeownership, education, inequality, and fertility. It might also capture variables that were not included in the analysis due to data availability, such as private pension coverage. The country fixed effects could also reflect cultural and institutional differences not discussed in this paper.
Homeownership ratio is significantly higher in Portugal than in Europe, contributing to the Portugal's relatively low saving rate. Homeownership is typically considered beneficial for the economy, as it is associated with wealth accumulation and social benefits, but it is also found to restrict labor mobility.
Education. The gap in the education level (a share of employment with upper-secondary, post-secondary, and tertiary education levels) compared to the sample average is significant but is decreasing over time.11
Inequality is also above the sample average (measured by both the GINI coefficient and the income quantile share ratio), with the difference decreasing over time.
Fertility rates in Portugal are below the sample average. While households with fewer children tend to save more (the effect that is captured in some regressions), low fertility has an adverse impact on the country population growth and contributes to ageing, possibly captured in the regression analysis via old-age dependency.
Old Dependency Ratio
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Old Dependency Ratio
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Old Dependency Ratio
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.
Employment Rate of the Elderly (65+)
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Employment Rate of the Elderly (65+)
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Employment Rate of the Elderly (65+)
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.
Pension to GDP Ratio
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Pension to GDP Ratio
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.Pension to GDP Ratio
(Percent)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Source: Eurostat.
Contributions to the Difference of the Portugal Saving Rate from the Sample Average
(Percentage points)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Sources: OECD, WEO, Eurostat, and IMF staff calculations.Contributions to the Difference of the Portugal Saving Rate from the Sample Average
(Percentage points)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Sources: OECD, WEO, Eurostat, and IMF staff calculations.Contributions to the Difference of the Portugal Saving Rate from the Sample Average
(Percentage points)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Sources: OECD, WEO, Eurostat, and IMF staff calculations.
Contributions to the Difference of the Portugal Saving Rate from the Sample Average
(Percentage Points; cumulative)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Sources: OECD, WEO, Eurostat, and IMF staff calculations.Contributions to the Difference of the Portugal Saving Rate from the Sample Average
(Percentage Points; cumulative)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Sources: OECD, WEO, Eurostat, and IMF staff calculations.Contributions to the Difference of the Portugal Saving Rate from the Sample Average
(Percentage Points; cumulative)
Citation: IMF Staff Country Reports 2019, 222; 10.5089/9781498325196.002.A001
Sources: OECD, WEO, Eurostat, and IMF staff calculations.F. Policy Options
23. Broader policies, pursued for their own reasons, could have positive side-effects on private saving, as suggested by the econometric analysis. Policies generally aimed at increasing per capita income should generate stronger household saving as a by-product. Fiscal sustainability arguments recommend containing age-related public spending over the medium-term. Reducing accrual rates in the mandatory (defined-benefit) pillar of the social security for top earners would help curb the growth in social security deficits, and as a by-product, it should also nudge these individuals to increase their discretionary saving.
24. More specific policy options could include measures to encourage private retirement saving. Policies that encourage employment of the elderly (65–74) and increase in the duration of working life, including by ensuring that unemployment schemes do not encourage early retirement, are particularly important to mitigate the impact [on saving] of the declining working age population in the longer term.
25. Promoting private pension plans would help to improve household savings. Some of the challenges facing Portuguese private pension schemes include limited development, low penetration, home bias in asset allocations, and low portability of pension schemes across the border (see the chapter on “Private Pension Schemes in Portugal: An Overview and Policy Options” in this SIP). One of the main recommendations to encourage retirement savings is gradually to move from a Taxed-Exempt-Taxed regime for employees—in which their contributions and benefits are taxed, and the investment income is exempted—to an Exempt-Exempt-Taxed regime, which is used in most European countries.
OECD recommendations:
Tightening rules that allow early withdrawals from Retirement Savings Plans and aligning retirement age rules with the statutory retirement ages.
Supporting the growth of occupational plans to increase coverage and encouraging steady contributions to pension plans.
Improving funding rules for defined benefit plans by developing Portuguese mortality tables for funding ratio calculations and updating the minimum funding scenario assumptions.
Introducing financial knowledge programs that focus on retirement income planning and decision-making.
26. According to the OECD, a more effective taxation of personal capital income could lower the tax burden on labor, providing tax-relief for middle-income households and encouraging employment. The OECD report on Household Taxation finds that the tax system in Portugal often favors the savings of households that are financially better off. While their main residence represents a large share of wealth for lower income households, the poorest households generally do not own residential property. The favorable treatment of owner-occupied housing could therefore provide a greater tax benefit to those in the middle and the top of income distribution.
G. Conclusion
27. Portugal's household saving rate is lower than those of the average European country. This difference can be explained by Portugal's lower disposable income, lower financial net wealth, higher old-age dependency ratio, higher government spending on pensions and on social protection benefits, and higher homeownership ratio, as suggested by a comparison against another 14 European countries conducted with the aid of panel regressions. Other factors that could underlie Portugal's low household saving are the country's lower education levels, fertility rate, and private pension coverage. Many of these factors are not amenable to simple or direct policy interventions, although some policy initiatives aimed at higher level objectives, such as promoting economic growth, could have positive side effects on household saving, our analysis shows. More specific policy options to boost household saving include measures to promote private occupational and personal plans, including some changes in taxation, and developing incentives to work past age 65.
Appendix I. Panel Unit Root Test Results (Levin, Lin, and Chu Test)
| Levin, Lin & Chu t-statistic | prob. | |||
|---|---|---|---|---|
| Saving rate, log | level, intercept | -2.47085 | 0.00670 | |
| Income | ||||
| Real disposable income per capita, log | level, intercept | -6.47057 | 0.00000 | |
| Demographics | ||||
| Old-age dependency, log | difference | -2.17877 | 0.01470 | |
| Life expectancy | level, intercept | -4.69734 | 0.00000 | |
| Duration of working life, log | level, intercept & trend | -2.40748 | 0.00800 | |
| Employment of people with age 65+ | level, intercept & trend | -1.76284 | 0.03900 | |
| Fertility | level, intercept | -2.13285 | 0.01650 | |
| Fiscal variables | ||||
| Public saving to GDP | level | -4.27344 | 0.00000 | |
| Fiscal balance to GDP | level | -5.77090 | 0.00000 | |
| Pension to GDP, log (full sample 2000–2016) | level, intercept & trend | -1.22649 | 0.11000 | |
| Pension to GDP, log (2000–2015) | level, intercept & trend | -1.89939 | 0.02880 | |
| Replacement ratio | level, intercept | -2.46239 | 0.00690 | |
| Social Protection Benefits for old to GDP, log | level, intercept & trend | -4.58288 | 0.00000 | |
| Direct Taxes, share, log | level, intercept | -1.39151 | 0.08200 | |
| Fiancial variables | ||||
| Financial assets to income | level, intercept & trend | -3.26474 | 0.00050 | |
| Net financial worth to income | level, intercept & trend | -2.43269 | 0.00750 | |
| Household debt to income | level, intercept | -2.70432 | 0.00340 | |
| Homeownership rate | level | -1.77258 | 0.03810 | |
| Housing prices index, real, log | -1.63940 | 0.05060 | ||
| Housing prices index, nominal, log | level, intercept | -2.17778 | 0.01470 | |
| Labor market and social indicators | ||||
| Unemployment | level | -2.89867 | 0.00190 | |
| Education | level, intercept | -5.10655 | 0.00000 | |
| Inequality S80/S20, log | level, intercept | -8.34901 | 0.00000 | |
| Levin, Lin & Chu t-statistic | prob. | |||
|---|---|---|---|---|
| Saving rate, log | level, intercept | -2.47085 | 0.00670 | |
| Income | ||||
| Real disposable income per capita, log | level, intercept | -6.47057 | 0.00000 | |
| Demographics | ||||
| Old-age dependency, log | difference | -2.17877 | 0.01470 | |
| Life expectancy | level, intercept | -4.69734 | 0.00000 | |
| Duration of working life, log | level, intercept & trend | -2.40748 | 0.00800 | |
| Employment of people with age 65+ | level, intercept & trend | -1.76284 | 0.03900 | |
| Fertility | level, intercept | -2.13285 | 0.01650 | |
| Fiscal variables | ||||
| Public saving to GDP | level | -4.27344 | 0.00000 | |
| Fiscal balance to GDP | level | -5.77090 | 0.00000 | |
| Pension to GDP, log (full sample 2000–2016) | level, intercept & trend | -1.22649 | 0.11000 | |
| Pension to GDP, log (2000–2015) | level, intercept & trend | -1.89939 | 0.02880 | |
| Replacement ratio | level, intercept | -2.46239 | 0.00690 | |
| Social Protection Benefits for old to GDP, log | level, intercept & trend | -4.58288 | 0.00000 | |
| Direct Taxes, share, log | level, intercept | -1.39151 | 0.08200 | |
| Fiancial variables | ||||
| Financial assets to income | level, intercept & trend | -3.26474 | 0.00050 | |
| Net financial worth to income | level, intercept & trend | -2.43269 | 0.00750 | |
| Household debt to income | level, intercept | -2.70432 | 0.00340 | |
| Homeownership rate | level | -1.77258 | 0.03810 | |
| Housing prices index, real, log | -1.63940 | 0.05060 | ||
| Housing prices index, nominal, log | level, intercept | -2.17778 | 0.01470 | |
| Labor market and social indicators | ||||
| Unemployment | level | -2.89867 | 0.00190 | |
| Education | level, intercept | -5.10655 | 0.00000 | |
| Inequality S80/S20, log | level, intercept | -8.34901 | 0.00000 | |
References
Amaglobeli, M. D., Chai, H., Dabla-Norris, M. E., Dybczak, M. K., Soto, M., and Tieman, A. F., 2019, The Future of Saving: The Role of Pension System Design in an Aging World, International Monetary Fund.
Banco de Portugal Economic Bulletin, May 2016, Special Issue: “An Interpretation of Household Saving Rate Developments in Portugal”.
Banco de Portugal Economic Bulletin, June 2017, Special Issue: “Saving and Investment Dynamics of Portuguese Firms”.
Banco de Portugal Economic Bulletin, June 2018, Special Issue: “Household Consumption Inequality in Portugal”.
Bouyon, S., 2016, Policy Options for European Household Saving, Journal of Economic Integration, pp.134–165.
Callen, M.T. and Thimann, M. C., 1997, Empirical Determinants of Household Saving: Evidence from OECD Countries (No. 97–181), International Monetary Fund.
Cardoso, F. and Alves, N., 2010, Household Saving in Portugal: Micro and Macroeconomic Evidence, Economic Bulletin and Financial Stability Report Articles and Banco de Portugal Economic Studies.
Chen, P., Karabarbounis, L., and Neiman, B., 2017, The Global Rise of Corporate Saving, Journal of Monetary Economics, 89, pp.1–19.
Costa, S. and Farinha, L., 2012, Households Indebtedness: A Microeconomic Analysis Based on the Results of the Households Financial and Consumption Survey, Financial Stability Report, Banco de Portugal, May.
Costa, S., 2016, Financial situation of the Households in Portugal: An Analysis Based on the HFCS 2013, Economic Bulletin and Financial Stability Report Articles and Banco de Portugal Economic Studies.
Le Blanc, J., Porpiglia, A., Teppa, F., Zhu, J., and Ziegelmeyer, M., 2016, Household Saving Behavior in EA, International Journal of Central Banking, 12(2), pp.15–69.
OECD Report, 2018, “Taxation of Household Savings”.
OECD Reviews of Pension Systems: Portugal, 2019.
OECD Report, 2019, “Under Pressure: The Squeezed Middle Class”.
Rodriguez-Palenzuela, D. and Dees, S., 2016, Savings and Investment Behavior in the Euro Area, ECB Occasional Paper, (167).
Rey-Ares, L., Fernández-López, S., and Vivel-Búa, M., 2015, The Determinants of Privately Saving for Retirement: The Cases of Portugal and Spain, European Journal of Applied Business and Management, 1(1).
Prepared by Koralai Kirabaeva (EUR). I am grateful to Alfredo Cuevas for guidance and advice and to the participants of the discussion at the Banco de Portugal, especially to Luisa Farinha, for their useful comments.
It is possible that part of the decline in household saving could be related to the increase in corporate saving as it could be difficult to distinguish statistically saving by household and by micro firms. In any case, Banco de Portugal Economic Bulletin (June 2017) finds that large firms have driven the aggregate developments of corporate saving, with smaller firms decreasing their savings.
For more discussion on the role of precautionary motives in savings recovery see the Special Issue: “An interpretation of household saving rate developments in Portugal”, Banco de Portugal Economic Bulletin, May 2016 and ECB Occasional Paper on “Savings and investment behavior in the euro area”.
For example, due to the hypotheses underlying the matching process. For more details see the methodological note on “Measuring Income, Consumption and Wealth jointly at the micro-level”.
Costa (2016) shows that according to the 2013 HFCS the share of indebted is about 45, similar to that in the euro area, although the median value of debt is higher, reflecting the higher participation in mortgages.
Rodriguez-Palenzuela and Dees (2016).
OECD report “Under Pressure: The Squeezed Middle Class”.
The variables used in the regressions are stationary in level, see Appendix Table 1.
The relationship between saving and indebtedness and financial assets/wealth are multifaceted (see Costa and Farinha 2012) and possibly endogenous. The Granger causality test indicates the Granger causality from the saving rate only to financial net wealth but not to financial assets or household debt. The Hausman test on the balanced panel (2007–2016) does not reject the assumption that random effects are uncorrelated with the explanatory variables.
This finding is consistent with the homeownership ratio in these three countries being higher than in Germany.
Educational attainment in Portugal remains one of the lowest in Europe, although it has been improving rapidly. The education gap is much smaller for younger cohorts.
Appendix I. The 2007 Pension Reform and the Recent Measures
In 2007, the Portuguese general pension scheme was reformed by Law 4/2007 and Decree-Law 187/2007 to contain budgetary pressures and ensure the long-term sustainability of the SS. The main measures introduced were:
The introduction of a “sustainability factor” in the pension formula, which is a demographic adjustment factor that links the legal retirement age to life expectancy at the time of retirement, and reduces pension benefits for early retirement in line with improvements in life expectancy (i.e., the pension was reduced by the sustainability factor and the distance between the actual and legal retirement age times 0.5 percent). At the same time, a bonus to retire after 65 years of age was introduced.
The pension formula was changed. Starting in 2007, pension levels are calculated on the basis of a reference remuneration that would also consider the earnings of the entire career. Since 1994, pension levels were calculated on the basis of the average earnings of the ten best years of the final fifteen.
The rule for indexing the benefits after retirement was changed. Before the reform all benefits were adjusted to inflation; after the reform, only pensions below a certain level determined by the “Indexante de Apoios Sociais” (IAS) were protected against inflation.
More recent measures include:
In 2011, pensions above €5,000 per month were subjected to an “extraordinary solidarity contribution.” This contribution was then extended to lower pensions and increased progressively from 3.5 percent for pensions above €1,000 to a marginal rate of 40 percent for pensions exceeding €7,126. Pensioners with the lowest pensions were not affected by any pension reductions. In addition, the pension update regime was suspended.
In mid-2012, early retirement was suspended.
In 2013, the sustainability factor was redesigned by changing the reference year of the average life expectancy at 65 from 2006 to 2000. Since 2014, the sustainability factor has only been applied only to early-retirement pensions. At the same time, the “normal retirement age” is linked to life expectancy gains at 65 years of age (Decree-law 67-E/2013).
In 2015, early retirement at the age of 60 or higher was reinstalled with penalties (i.e., sustainability factor and distance between actual and legal retirement age).
In March 2016, the new government partially suspended the new early retirement regime at the age of 60 or higher to reassess the applied penalties.
In 2017, Decree-Laws 3/2017 and 4/2017 expanded the convergence initiated in 2007 between the CGA and the SS system by standardizing the assignment requirements and calculation rules for retirement and old-age pensions of the police and military personnel covered by the CGA or the SS. In addition, Decree-Law 126-B/2017 further encouraged convergence by requiring contributory periods to be based on the entire contribution career, regardless of the scheme.
In August 2017, an extraordinary uprating was introduced for pensions equal or lower than €631.98 per month (1.5 times the social support index) to offset the purchase power loss due to the suspension of the pension uprate regime during 2011–15 and to increase the income of pensioners with lower pensions.
In October 2017, the early-retirement regime was modified to benefit very long careers (at least 60 years old and 48 years or more of contributions, or at least 60 years old with 46 years or more of contributions and started paying contributions before the age of 15). These workers will not be subject to any penalty, including no application of the sustainability factor.
In August 2018, an additional extraordinary increase for low pensions (1.5 times the social support index) was granted.
In 2018, the extraordinary solidarity contribution was eliminated.
In 2019, the authorities further relaxed the early retirement rules for very long careers by eliminating the sustainability factor penalty for early retirement of citizens with 60 years of age and at least 40 years of contributions to the SS or the CGA. These individuals get a bonus of four months less than the normal retirement age per each additional year of contributions and they can retire at their “personal retirement age” without any financial penalty. If they retire before this “personal” age they have a penalty of 0.5 per month of anticipation. (DL 118/2018, entry in force in 2019 (with a transitional period from January to September applying only to those with at least 63 years of age).
References
Ambachtsheer, K. and J. McLaughlin, 2015, How Effective is Pension Fund Governance Today? And Do Pension Funds Invest for the Long Term? Findings from a New Survey, (Toronto: KPA Advisory Services).
Amaglobeli, D., H. Chai, E. Dabla-Norris, K. Dybczak, M. Soto, and A. F. Tieman, 2019, “The Future of Savings: The Role of Pension System Design in an Aging World,” IMF Staff Discussion Note No. 19/01, (Washington: International Monetary Fund).
Autoridade de Supervisão de Seguros e de Fundos de Pensões, 2018a, Relatório de Atividade e Contas Anuais 2017, (Lisbon: Autoridade de Supervisão de Seguros e de Fundos de Pensões).
Autoridade de Supervisão de Seguros e de Fundos de Pensões, 2018b, Relatório do Setor Segurador e dos Fundos de Pensões 2017, (Lisbon: Autoridade de Supervisão de Seguros e de Fundos de Pensões).
Autoridade de Supervisão de Seguros e de Fundos de Pensões, 2018c, Estatísticas de Fundos de Pensões 2017, (Lisbon: Autoridade de Supervisão de Seguros e de Fundos de Pensões).
Bradley, D, C. Pantzalis, and X. Yuan, 2016, “The Influence of Political Bias in State Pension Funds,” Journal of Financial Economics, Vol. 199, pp. 69–91.
Brown, J. R., J. M. Pollet, and S. J. Weibernner, 2015, “The In-State Equity bias of State Pension Plans,” NBER Working Paper No. 21020, March.
Darvas, Z. and D. Schoenmaker, 2017, “Institutional Investors and Home Bias in Europe’s Capital Markets Union,” Bruegel Working Paper No. 2, (Brussels: Bruegel).
DiSalvo, D., 2018, The Politics of Public Pension Boards, September (New York: Manhattan Institute).
European Commission (2018a), Pension Adequacy Report, Vols. 1 and 2, (Brussels: European Commission).
European Commission (2018b), “The 2018 Aging Report,” Institutional Paper No. 079, May (Brussels: European Commission).
European Insurance and Occupational Pensions Authority, 2017, Market Development Report on Occupational Pensions and Cross-Border IORPs, EIOPA-BOS-18/13 (Frankfurt: European Insurance and Occupational Pensions Authority).
Financial Stability Board Regional Consultative Group, 2017, Report on European Private Pension Schemes: Functioning, Vulnerabilities, and Future Challenges (Basel: Financial Stability Board).
Ghilarducci, T., J. Saad-Lessler, and E. Fisher, 2012, “The Macroeconomic Stabilisation Effects of Social Security and 401(k) Plans,” Cambridge Journal of Economics, No. 36, pp 237–251.
Impavido, G, 2013, Pension Funds in Caprio, G., Handbook of Key Global Financial Markets, Institutions, and Infrastructure, Vol. 1, pp. 523–532.
Lindbeck, A. and M. Persson (2003), The Gains from Pension Reform, Journal of Economic Literature, Vo. XLI(March), pp. 74–112.
Moreira, A., A. B. Azevedo, L. P. Manso, and R. Nicola, 2019, Financial and Social Sustainability of the Portuguese Pension System, (Lisbon: Fundação Francisco Manuel dos Santos).
Munnell, A. H., J.-P. Aubry, J. Hurwitz, and L. Quinby, 2011, “A Role for Defined Contribution Plans in the Public Sector,” State and Local Pension Plans No.16(April), (Boston: Center for Retirement Research at Boston College).
Oakley, D. and J. E. Brown, 2016, Preserving Retirement Security for Public Sector Employees, (Washington: National Institute on Retirement Security).
Organization for Economic Cooperation and Development, 2017, Pension at a Glance, (Paris: Organization for Economic Cooperation and Development).
Organization for Economic Cooperation and Development, 2018a, Pension Markets in Focus 2018, (Paris: Organization for Economic Cooperation and Development).
Organization for Economic Cooperation and Development, 2018b, Financial Incentives and Retirement Savings, (Paris: Organization for Economic Cooperation and Development).
Organization for Economic Cooperation and Development, 2019, Reviews of Pension Systems: Portugal, (Paris: Organization for Economic Cooperation and Development).
Oliver Wyman, 2015, CRSB Retirement Program Financial Feasibility Analysis Results,” (Hartford: Oliver Wyman).
Overture Financial, 2016, Final Report to the California Secure Choice Retirement Savings Investment Board, (New York: Overture Financial).
Oxera, 2013, Study on the Position of Savers in Private Pension Products, (Oxford: Oxera).
Patzner, A., J. Nagler, and D. Mann, 2018, EU Claim: Refund of (Dividend) Withholding Tax (Frankfurt: KPMG).
PensionsEurope, 2016, Position Paper on the Withholding Tax Refund Barriers to Cross-Border Investment in the EU, (Brussels: PensionsEurope).
Rocha, R., R. Hinz, and J. Gutierrez, 1999, “Improving the Regulation and Supervision of Pension Funds: Are there Lessons from the Banking Sector?”, Social Protection Discussion Paper Series, No. 9929, (Washington: World Bank).
Scharfstein, D., 2018, Pension Policy and the Financial System, (Philadelphia: Presidential Address to the American Finance Association).
Seguridade Social, 2017, Fundo dos Certificados de Reforma – Folheto Informativo, (Lisbon: Seguridade Social).
Sercu, P. and R. Vanpee, 2012, “The Home Bias Puzzle in Equity Portfolios” in Kaker, H. K. and L. A. Riddick, International Finance: A Survey (Oxford: University Press).
Vidlund, M., H. Ramberg, A. Mielonen, and E. Lappalainen, 2012, “The Structure of Pension Provision and the Significance of Occupational Pensions in Different Countries,” Finnish Centre for Pensions Surveys, No. 01, (Helsinki: Finnish Centre for Pensions).
Valerio Crispolti and André Oliveira Santos.
Impavido (2013) contains a detailed discussion on the nature of pension schemes, institutions, and policy issues.
Pension schemes and plans refer to the same concept of a legally binding contract. The former expression is common in Europe while the latter one is used in the US.
Taxonomy based on European Commission (2018b).
Some degree of annuitization of DB payouts is common in advanced economies.
For a list of risk factors affecting financial stability, see Financial Stability Board Regional Consultative Group (2017).
Based on EC (2018) and Moreira, Azevedo, Manso, and Nicola (2019).
EC (2018b) projects an increase in public pensions from 13.5 percent of GDP in 2016 to 14.7 percent in 2040 and a decline after 2040 to 11.7 percent in 2070.
Data on assets under management by European (excluding Portugal) pension schemes are based on the survey reported in Financial Stability Board Regional Consultative Group (2017), complemented with data for German pension schemes in 2013 from Oxera (2013).
Scharftstein (2018) also found a negative relationship between gross replacement rates in public pension schemes and total asset pensions as a percentage of GDP.
Data from the Autoridade de Supervisão de Seguros e Fundos de Pensões (ASF) on pension fund asset allocations do not provide a breakdown of UCITS by underlying assets, in particular equity and debt.
Neither ASF nor EIOPA data on pension fund asset allocations provide a geographical breakdown of UCITS by underlying assets, in particular foreign and domestic equity and debt.
OECD (2019) recommends that, in the benefit calculation, Portugal should update past wages with wage growth rather than a combination of price inflation and wage growth while lowering accrual rates.
20 percent of employee contributions to private pension schemes are tax deductible up to maximum of €400 per year if the employee is less than 35 years old, €350 if the employee is more than 35 years old but less than 50 years old, and €300 if the employee is less than 50 years old.
OECD (2018b) emphasizes that the deductions of pension contributions from taxable income encourage contributions to pension schemes by middle-to-high income earners because they are responsive to the upfront tax relief on contributions that help reduce their tax liabilities.
