Projected population, inequality and social expenditures: The Case of Flanders

Traditionally, in Flanders, household income is among the highest in Europe and inequality as well as poverty are low (Eurostat, 2014). Following the general trend in Western European countries, inequality declined between the 1950s up to the second half of the 1980s. Since then, however, there exists no unequivocal evidence for this equalizing tendency (Van den Bosch, Vandenbroucke, Cantillon, & Pacolet, 2009; Vranken, De Boyser, & Dierckx, 2006), except for Flanders where inequality remains consistently lower than in the rest of Belgium (Cantillon, Horemans, Vandenbroucke, & Van Lancker, 2011; FOD Economie, K.M.O., Middenstand en Energie, 2013).

In this paper, we extend the horizon of analysis, projecting income inequality and poverty indicators, as well as social expenditures for Flanders, up to the year 2031. Moreover, we analyse the underlying demographic elements affecting these evolutions. Traditionally, such trends were mainly attributed to economic factors, such as wages, income, or tax-benefit systems, (Bargain & Callan, 2010; Heathcote, Perri, & Violante, 2010); in contrast, we investigate the role of (projected) demographic evolutions as drivers of future income inequality and poverty trends.

The profound changes in Flanders’ population structure seem to justify our focus. In parallel with the well-documented ageing of the population (Federaal Planbureau en ADSEI, 2011; Studiecommissie voor de vergrijzing, 2006), the Flemish household composition changed quite substantially. From the 1960s onward, driven by new gender and generational relations, divorce rates started to rise, marriage became less popular and, with the exception of the baby-boom years, fertility continued to show an overall decreasing trend (Lesthaeghe & Neels, 2002; Van Bavel & Bastiaenssen, 2006). The onset of this so-called second demographic transition occurred about 20 years later in Belgium, particularly in Flanders, compared to neighbouring countries, while resulting trends were, until recently, less pronounced (Deboosere et al., 2009). After 1980, however, Flanders started to catch-up with international trends and different family formation processes proliferated. As a result, households have become smaller in general (Lodewijcks & Deboosere, 2011) and, especially among the young, unmarried couples and single-headed households have proliferated (Deboosere et al., 2009).

A second major evolution in the population structure concerns increasing educational achievement. Under the impulse of labour market demands and political and societal investments in the educational system, ever more children enter schooling and remain in training for a longer period of time (Vanderstraeten, 1999). Parallel, from the 1960s onward, we observe a democratization and growing gender equity in education, although social inequality remains a critical aspect of Flanders’ educational system at all levels (Pelleriaux, 2000; Verbergt, Cantillon, & Van den Bosch, 2009), the breach between classes has decreased (Groenez, 2010). Moreover, girls outperform boys at all educational levels for some time now (Pelleriaux, 2000). In short, since the 1960s, for each consecutive generation, educational levels are higher compared to earlier generations.

In recent literature, the importance of a clear understanding of the impact of population change on inequality and poverty has repeatedly been stressed (Blank, 1995, 2011; Burtless, 1999; Chen & Föster, 2011; Peichl, Pestel, & Schneider, 2012; Western, Bloome, & Percheski, 2008). This understanding is far from easy, since different components of demographic change are intertwined and may have counteracting effects. In this respect the ageing, in particular the growing share of the population above 65, is presumed to increase inequality, since elderly population’s income is lower and their poverty ratio higher than that of the working-age population. In addition, in Flanders, this generational gap seems to have increased over time (Van den Bosch et al., 2009). However, ageing also implies an increase in life expectancy, and couple’s survival has a positive effect on household income: women depend less on a widows’ pension and also, with the increase in education and female labour force participation, households will often have two pensions or combine a pension with (the younger womans’) income from wages. Consequently, the net effect of ageing on the income distribution remains ambiguous. Mookherjee and Shorrocks (1982) found that, in the UK, the shift in age-income relationship had an impact on inequality, but the population’s compositional change did not. More recently, the latter was confirmed for other European countries by the Social Situation Monitor (European Commission: DG Employment, Social Affairs & Inclusion, 2017).

On the basis of all 22 countries present in the Luxembourg Income Survey, it has been shown (Tai & Treas, 2008, 2009) that, even in the case of single-headed households, the effects on inequality are not clear-cut. Being less able to pool income resources (combined with the gender pay gap), single-headed households and especially single parents have a lower income compared to the household income of the rest of the population. In addition, no improvement in the situation of single parents is observed over time (Tai & Treas, 2008, 2009). This general phenomenon is also observed in Flanders (Studiedienst Vlaamse Regering, 2013; Van den Bosch et al., 2009). A rise in the prevalence of female single-headed households is therefore often found to trigger an increase in inequality and poverty (Bradley, Huber, Moller, Nielson, & Stephens, 2003; Kollmeyer, 2013). With the increase in education and subsequent increased female labour force participation, however, this tendency might reverse. In addition, Esping-Andersen (2007) indicates that the negative influence of single motherhood is mainly restricted to the United States as the strong European welfare state mitigates its impact.

The rise in educational attainment and the concomitant increase in female labour force participation is also found to increase inequality since double-income households are concentrated at the top of the income distribution (Esping-Andersen, 2007). However, the larger the group of the educated and the lower educational disparity, the smaller the role this factor will play in the determination of income inequality. In the long run, we might well expect the rise in educational attainment to temper inequality. In this respect, Breen and Salazar (2010) found that changes in women’s education and their behavioural consequences account for little if any of the growth in earnings inequality between households, while Kollmeyer (2013) found that it decreases inequality.

This work joins a strand of recent research aiming at quantifying the role of demographic change on inequality (Peichl et al., 2012), but in contrast to former studies, we analyse this relationship prospectively, focusing on the 20 year period from 2011 to 2031. From a policy point of view, this is important; for social policy to go beyond ad hoc answers to structural changes, it needs to be informed on how future population (and economic) characteristics affect inequality. In addition, this prospective view can re-mediate some of the difficulties related to contextual factors faced by retrospective research (Esping-Andersen, 2007). Building on hypothetical scenarios about economic growth and population forecasts, we isolate our analysis from temporal and contextual influences and gain at least partial control over interactions with non-observed variables. Note that we limit our attention to the coming 13 year period, which is the time frame in which the baby-boom generation will reach retirement age. Moreover, beyond this time frame, predictions become increasingly uncertain.

The prospective nature of our endeavour firmly roots it in the microsimulation tradition. A microsimulation model (MSM) is essentially a forecasting device that simulates aggregate and distributional effects of change –such as population change– by applying it to a representative sample of individuals or families, subsequently adding up the results across individual units using population weights (Bourguignon & Spadaro, 2006; Martini & Trivellato, 1997). The results in this paper were obtained using the microsimulation model Mefisto.¹ The baseline model used in this paper investigates the effect of demographic change under the condition of a 1% economic growth attributed to efficiency increases.²

To disentangle economic from demographic growth, we use two decomposition methods. The first is based on the construction of a counterfactual, which involves reweighing the sample at t₀ such that the distribution of population covariate levels becomes identical to the sample at t₁. Comparison of the counterfactual Gini-index with the actual one at t₀ reveals the effect of changing population structure, while the difference between counterfactual and the distribution at t₁ is a residual effect. This method is known as the Blinder-Oaxaca decomposition (Blinder, 1973; Oaxaca, 1973). In the context of distributional analysis it has been applied by DiNardo, Fortin, and Lemieux (1996), Hyslop and Maré (2005), Handcock and Morris (1998), Bargain and Callan (2010), to name but a few. The second method concerns an exact decomposition of the inequality index (Shorrocks, 1980, 1984) to gain insight into the influence on inequality and poverty of shifts in population subgroup prevalences.

In Section 2 we describe the data and methodology. Section 3 presents the results of the population projections. Section 4 describes the forecasted income inequality evolution and its relation to demographic and economic change, with Subsection 4.3 focusing on budgetary effects.

We proceed with a number of distinct steps:

1. We first make multi-state demographic projections at five-year intervals for twenty years (from 2011 up to 2031) by age, sex and household position and by age, sex and educational attainment. The procedure is described in Subsection 2.2.

2. We calibrate the EU-SILC 2008 data to the distribution of the different population subgroups obtained by the demographic projections. In other words, we reweigh the EU-SILC 2008 in such a way that the income data match the population forecast obtained in step 1. This procedure is called static ageing (see Subsection 2.3).

3. We choose a realistic scenario for economic growth to update the income data in EU-SILC 2008 for each five-year interval from 2011 to 2031 (as described in Subsection 2.4).

4. Finally, based on the re-weighted and uprated income data obtained from steps 2 and 3, we estimate inequality and poverty indices, as well as the budgetary implications for each five-year interval (details can be found in Subsection 2.5).

2.4 Economic growth

In order to obtain a realistic impression of future (equivalised) income distributions and concepts derived thereof, such as inequality and poverty, not only demographic changes need to be taken into account, but we need to make some assumptions concerning economic growth as well.

Clearly, economic growth is not independent of demographic changes. It can be split up into the following components:

1. An increase in workforce participation rate. We assumed that this component remains unchanged, conditional on the variables used in the demographic projection, that is sex, age, household position and education. Any changes in this component are thus mainly driven by ageing and educational changes.

2. Productivity growth, which breaks down into two components:

   1. an endogenous productivity growth component, caused by the changing distribution of individual and household characteristics over time (considering the projections we use, the main source of this component is an increase in educational attainment). Whenever a (projected) population stabilizes, this component tends to zero. Keep in mind that we assume that the returns to education remain at their present level.

   2. an efficiency growth component, conditional on individual and household characteristics, induced by the fact that, people produce ever more, hence more efficiently, without any change to said characteristics, including hours worked.

Since changes in workforce participation rate and endogenous productivity growth are completely determined by the demographic projections, we name the sum of both the demographic growth. It is only the varying component of productivity growth, conditional on the demographic assumptions, for which we need to make additional assumptions, and which we termed efficiency growth. In the remainder of this paper, we assume that the efficiency growth amounts to a yearly real-term increase of 1%.⁹

Following additional assumptions are made:

• economic growth is proportionally shared among the factors of production, which entails that (self-)employment income and income from capital have the same growth rate.

• benefits grow according to the assumptions described in Dekkers, Desmet, Fasquelle, and Weemaes (2013): the minimum income will grow at an annual rate of 1% in real terms, all other benefits at 0.5%.

To ensure that we do not overestimate inequality, we allow pensions to grow with the efficiency growth rate. Let us justify the pension growth rate by providing an example. Consider a pensioner in the observed sample who retired five years prior to observation at time t₀. In the philosophy of the reweighing methodology, this observation will represent a number of pensioners at time t₁, who will have retired in t₁ − 5. Exact calculation of their pension would involve calculating each observation’s last wage at t₀ − 5, applying the observed real wage growth between t₀ − 5 and t₀, applying the hypothesized real wage growth between t₀ and t₁ − 5, and finally, applying an hypothesized real pension growth between t₁ − 5 and t₁. Simply up-rating pensions by the hypothesized real wage growth rate will produce accurate results if:

1. real pension growth rates between t₁ − 5 and t₁ are identical to those between t₀ − 5 and t₀,

2. the real wage growth rate between t₀ − 5 and t₀ was close to the hypothesized real wage growth rate.

It seems reasonable to make assumptions concerning incomes in terms of the efficiency growth rate, since collective wage agreements pertain to individual increases in wages. Assuming the educational attainment of one firm, or even a whole sector of the economy, remains quite stable over time, wage increases will be accorded conditional on age and education of employees. Likewise, an increase in benefits is stated as a percentage increase of the individual’s benefit. Consequently, wage and benefit rises seem not directly connected with any demographic evolution. The demographic growth component is not observed by individual firms, but only appears ex-post at the aggregate level.

The resulting factors by which incomes and pensions are uprated can be found in Table 2. As mentioned above, this represents only the efficiency growth component of the total economic growth, the second component arising from the fact that increasingly educated cohorts grow older. The combined effect of efficiency growth and growth induced by demographic change is depicted in Figure 1. It is obtained by calculating the growth rate of aggregated primary income per capita, under the combined demographic and economic assumptions, which should be a close proxy to projected per capita economic growth. The total growth rate gradually decreases to the annual efficiency growth rate of 1%. Although this boundary is not fully reached, the projected evolution is consistent with what we theoretically expect: once we reach a stable population (keeping boundary conditions on fertility, migration,... constant), economic growth will settle at the level of what we termed the efficiency growth component.

Table 2

year	factor
2011	1.030
2016	1.083
2021	1.138
2026	1.196
2031	1.257

Figure 1

Download asset Open asset

As small word of caution might be in order: the endogenous productivity growth component is driven by the fact that younger cohorts are more educated, leading to an increasing educational average over time. Our results partly depend on one of the many implicit ceteris paribus assumptions we have made: the returns to education remain fixed at their present level. However, the returns to education can fluctuate for each education level depending on the relative demand and supply in each labour market segment. Jobs for which there are shortages of qualified candidates, will typically be more highly remunerated. On the other hand, if the proportion of highly educated increases faster than the labour demand for this type of workers, we might witness a decrease of the return to education over time for the highest educated. Such a detailed level of labour market scenarios is, however, beyond the scope of this paper.

During the second half of the last century, intense modifications in family formation and dissolution were observed (Deboosere et al., 2009): a postponement of first marriage, a reduction of marriage intensity and an increment in divorce rates. These evolutions, in combination with the endurance of fertility decline and the increase of life expectancy, will induce a profound change in Flanders demographic structure over the next twenty years.

Figure 2 shows the population distribution by age and sex in 2011 and the projection result for 2031. As the baby-boom generations at working and reproductive age in 2011 grow older, the bottom of the pyramid shrinks and the top becomes heavier. Parallel, this ageing process gives rise to changes in the population’s household composition as the younger generations of 2011 grow older and their family formation behaviour is reflected in successive age groups in each subsequent projection year. This is demonstrated in Figures 3–5.

Figure 2

Download asset Open asset

Figures 3–5 depict the proportional distribution of household positions, by age for each five-year projection period between 2011 and 2031. The x-axes represent five-year age groups and the y-axes, the proportion of individuals from each age group in single-headed households (Figure 5) and in couples with and without children (Figures 3 and 4, respectively). Each curve, from light to dark grey, represents a consecutive projection year. A first important evolution is the decrease of the share of individuals that live in couples with children. This is related to the decrease and postponement of couple formation and to fertility decline. The latter development explains the larger decrease after the age of 40, as low fertility leads to a shorter total time span that children are present at their parent’s household. Higher divorce rates add to this evolution and also explain the decreasing prevalence of couples without children at more advanced adult age. However, among the elderly population, above the age of 75 for women and 85 for men, the prevalence of individuals living in a couple increases in time. This is easily understood as the result of growing partner’s survival rates since we projected women’s and especially men’s life expectancy to grow. The decrease in the share of couples in the population over time results in the rise of single-headed households, most eminent between the age of 40 and 70.

Figure 3

Download asset Open asset

Figure 4

Download asset Open asset

Figure 5

Download asset Open asset

Figure 6 depicts the proportional distribution by age of individuals with only primary or less education, lower and higher secondary education and higher education for each five year projection period between 2011 and 2031, for women (left panels) and men (right panels). The Figure shows that parallel to ageing and changes in household position, for all age groups and each gender, we foresee a considerable advancement in educational attainment. That is, for each age group the share of the population with higher secondary and higher education becomes larger over time. This process is almost entirely due to the ageing of the population as from each projection year to the next, the educational profile of the younger generations is progressively spread to all ages. Therefore, in the first part of the projection period, especially the 40 − 64-age group’s educational attainment increases; in the second half of the projection period, the 65+ educational attainment rises most while below 60, practically no change is observed any more. Within the population under the age of 35, practically no change is observed since we assumed only minor adaptations in educational behaviour (see Table 1). Consequently, in the long run, an equalizing tendency between generations and gender is observed.

The following comparison demonstrates this tendency: in 2011, the part of the population having at least a higher secondary degree below the age of 40 was about 80% while this was only the case for about 22% of the group above 65. In 2031, 80% of the whole population under 65 will have a higher secondary degree, and this is also the case for almost 60% of the 65+ group. Furthermore, in 2011, the younger generations of women had already exceeded the education level of men; in 2031 this will be the case for all age groups, except the most advanced ones.

In this paper, we forecasted changes in inequality, poverty and social expenditures in Flanders between 2011 and 2031, and analysed their relationship to population change. We used multi-state population projections (LIPRO-projections) based on Census and Register data from 2001 and 2006. We reweighed the EU-SILC 2008 survey data to construct the necessary income data for each five year period between 2011 and 2031. Assumptions about population change were conservative: we presumed only changes in mortality, fertility and migration following the current official Flemish hypotheses, whereas household formation processes and educational behaviour were kept (nearly) constant to the projection base-year information of 2006. In addition, we assumed a modest economic efficiency growth of 1% annually.

Our population projections foresee a pronounced ageing of the Flemish population between 2011 and 2031. Along with the ageing process, characteristics of the current younger generations are spread to older age groups. Consequently, the projection results demonstrate an increase of single-headed households to the detriment of couples, especially those with children. Educational attainment levels increase and intergenerational and gender differences attenuate.

Under the assumptions mentioned, inequality and poverty show unequal forecasted evolutions and are differently affected by demographic and economic change. Inequality, measured by the Gini and Theil indices, is lower in 2031 than in 2011, but it exhibits a non-monotonous pattern over time, reaching a maximum around 2020. This pattern is mainly driven by population change, while economic growth has a small but increasing impact over time. Poverty steadily declines from 2011 onward. This is due to a growth effect, reinforced by population change.

The evolution of public finances is forecasted as follows. On the revenue side, income taxes increase from 22% to almost 30% of gross income earned by all families. On the expense side, pensions increase from 23% to 34% of gross income earned by all families. However, this spectacular increase in pension volume is completely paid by the modest growth we assume and by keeping the tax system constant in real terms. These results are obtained without altering the retirement age, nor the real level of pensions.

In order to gain insight into the relationship between population and inequality change, we decomposed the Theil index and the poverty rate into the effects of shifts in particular subgroups defined by age and household composition on the one hand, and age and educational level on the other. The impact on inequality of changes in population subgroups remain similar over time, but depending on the observation period the effect of one predominates over the other. In addition, the results show the counteracting effects of the population ageing process. With ageing, the elderly population increases but educational levels also rise and generational differences attenuate. Over the total observation period between 2011 and 2031 the latter two processes play the larger role: the overall decrease in inequality is mainly due to a declining between-group income variation reinforced by a decline of the population with no or only primary education. Also the drop in prevalence of couples with children reduces inequality. During the period between 2011 and 2021, in contrast, the impact of the growing elderly population, especially single-headed households, outgrows that of educational attainment growth and between-group inequality reduction, producing a temporary increase in inequality. The reduction in poverty between 2011 and 2031 can to a large extent be attributed to the decrease of the population share with only or less than primary education. Among the elderly population this effect is enhanced by a a prominent reduction in within-poverty levels. Educational increase counterbalances the effect of the rise of single-headed households to the detriment of couples, that would have increased overall poverty if their poverty levels would not have declined due to educational improvement.

1. 1
   Analysing the effects of tax-benefit reforms on income distribution: A decomposition approach

   1. O Bargain
   2. T Callan

   (2010)

   Journal of Economic Inequality 8:1–21.

   • Google Scholar
2. 2
   Changes in inequality and unemployment over the 1980s: Comparative cross-national responses

   1. RM Blank

   (1995)

   Journal of Population Economics 8:1–21.

   • Google Scholar
3. 3
   Changing inequality

   1. RM Blank

   (2011)

   University of California Press.

   • Google Scholar
4. 4
   Wage discrimination: Reduced form and structural estimates

   1. AS Blinder

   (1973)

   Journal of Human Resources 8:436–455.

   • Google Scholar
5. 5
   Microsimulation as a tool for evaluating redistribution policies

   1. F Bourguignon
   2. A Spadaro

   (2006)

   Journal of Economic Inequality 4:77–106.

   • Google Scholar
6. 6
   Distribution and redistribution in postindustrial democracies

   1. D Bradley
   2. E Huber
   3. S Moller
   4. F Nielson
   5. JD Stephens

   (2003)

   World Politics 55:193–228.

   • Google Scholar
7. 7
   Has increased women’s educational attainment led to greater earnings inequality in the United Kingdom? A multivariate decomposition analysis

   1. R Breen
   2. L Salazar

   (2010)

   European Sociological Review 26:143–157.

   • Google Scholar
8. 8
   Effects of growing wage disparities and changing family composition on the U.S. income distribution

   1. G Burtless

   (1999)

   European Economic Review 43:853–865.

   • Google Scholar
9. 9
   Inkomen en armoede in Vlaanderen en Europa

   1. B Cantillon
   2. J Horemans
   3. P Vandenbroucke
   4. W Van Lancker

   (2011)

   131–160, Eds., De sociale staat van Vlaanderen 2011, Studiedienst Vlaamse Regering, http://www.deverenigdeverenigingen.be/images/stories/SSV_2011-publicatie.pdf.

   • Google Scholar
10. 10
    http://www.bmask.gv.at/cms/site/.../0/7/.../oecd_divided_we_stand_2011.pdf

    1. W-H Chen
    2. M Föster

    (2011)

    Divided we stand: Why inequalities keep rising.
11. 11
    The impact of demographic change on policy indicators and reforms (FLEMOSI Discussion Paper No. DP25)

    1. R De Blander
    2. I Schockaert
    3. A Decoster
    4. P Deboosere

    (2013)

    CES - KU Leuven.

    • Google Scholar
12. 12
    http://statbel.fgov.be/nl/binaries/mono_200104_nl[1]_tcm325-92942.pdf

    1. P Deboosere
    2. R Lesthaeghe
    3. J Surkyn
    4. D Willaert
    5. P-M Boulanger
    6. A Lambert
    7. L Lohlé-Tart

    (2009)

    Huishoudens en gezinnen in België (No. 4).
13. 13
    The social and budgetary impacts of recent social security reform in Belgium. Technical Report

    1. G Dekkers
    2. R Desmet
    3. N Fasquelle
    4. S Weemaes

    (2013)

    The social and budgetary impacts of recent social security reform in Belgium. Technical Report, Federal Planning Bureau, Belgium, https://lirias.kuleuven.be/bitstream/123456789/442606/4/Dekkers_Desmet_Fasquelle_Weemaes_impactofreform_IMA_2013.pdf, Paper presented at the IMPALLA-ESPANET International Conference“Building blocks for an inclusive society: empirical evidence from social policy research”, Luxembourg, April 18–19, 2013.

    • Google Scholar
14. 14
    Calibration estimators in survey sampling

    1. J-C Deville
    2. C-E Sarndal

    (1992)

    Journal of The American Statistical Association 87:376–382.

    • Google Scholar
15. 15
    Labor market institutions and the distribution of wages, 1973-1992: A semiparametric approach

    1. J DiNardo
    2. NM Fortin
    3. T Lemieux

    (1996)

    Econometrica 64:1001–1044.

    • Google Scholar
16. 16
    Sociological explanations of changing income distributions

    1. G Esping-Andersen

    (2007)

    American Behavioral Scientist 50:639–658.

    • Google Scholar
17. 17
    website. Retrieved from

    1. European Commission: DG Employment, Social Affairs & Inclusion

    (2017)

    Research findings – social situation monitor – decomposition of inequality by population subgroups, website. Retrieved from, http://ec.europa.eu/social/main.jsp?catId=1050&intPageId=1882&langId=e.

    • Google Scholar
18. 18
    http://epp.eurostat.ec.europa.eu/portal/page/portal/income_social_inclusion_living_conditions/legislation

    1. European Union

    (2003)

    Regulation (EC) 1177/2003 of the European parliament and of the council concerning community statistics on income and living conditions (EU-SILC).
19. 19
    http://epp.eurostat.ec.europa.eu/-portal/-page/-portal/-region_cities/-regional_statistics/-data/-database

    1. Eurostat

    (2014)

    Regional database.
20. 20
    http://www.plan.be/admin/ uploaded/201112190816070.bevpop2011_nl.pdf

    1. Federaal Planbureau en ADSEI

    (2011)

    Bevolkingsvooruitzichten 2010-2060 (Vooruitzichten).
21. 21
    Fiscale statistiek — evolutie van de inkomensongelijkheid in België voor en na belastingen

    1. FOD Economie, K.M.O., Middenstand en Energie

    (2013)
22. 22
    A class of decomposable poverty measures

    1. J Foster
    2. J Greer
    3. E Thorbecke

    (1984)

    Econometrica 52:761–766.

    • Google Scholar
23. 23
    Marry your like: Assortative mating and income inequality

    1. J Greenwood
    2. G Guner
    3. G Kocharkov
    4. C Santos

    (2014)

    American Economic Review 104:348–533.

    • Google Scholar
24. 24
    Onderwijsexpansie en – democratisering in Vlaanderen

    1. S Groenez

    (2010)

    Tijdschrift Sociologie 3–4:199–238.

    • Google Scholar
25. 25
    Relative distribution methods

    1. MS Handcock
    2. M Morris

    (1998)

    Sociological Methodology 28:53–97.

    • Google Scholar
26. 26
    Unequal we stand: An empirical analysis of economic inequality in the United States, 1967-2006

    1. J Heathcote
    2. F Perri
    3. GL Violante

    (2010)

    15–51, Review of Economic Dynamics, 13, 1, Special issue: Cross-Sectional Facts for Macroeconomists.

    • Google Scholar
27. 27
    Methods of weighting for unit non-response

    1. D Holt
    2. D Elliot

    (1991)

    333–342, Journal of the Royal Statistical Society. Series D (The Statistician), 40, 3, Special Issue: Survey Design, Methodology and Analysis, 2.

    • Google Scholar
28. 28
    Understanding New Zealand’s changing income distribution, 1983-1998: A semi-parametric analysis

    1. DR Hyslop
    2. DC Maré

    (2005)

    Economica 72:469–495.

    • Google Scholar
29. 29
    Weighting methods

    1. G Kalton
    2. I Flores-Cervantes

    (2003)

    Journal of Official Statistics 19:81–97.

    • Google Scholar
30. 30
    Family structure, female employment, and national income inequality: A cross-national study of 16 Western countries

    1. C Kollmeyer

    (2013)

    European Sociological Review 29:816–827.

    • Google Scholar
31. 31
    From the first to the second demographic transition: An interpretation of the spatial continuity of demographic innovation in France, Belgium and Switzerland

    1. R Lesthaeghe
    2. K Neels

    (2002)

    European Journal of Population 18:325–360.

    • Google Scholar
32. 32
    Households and families: Stability and fast development go hand in hand (GGP Belgium Paper Series No. 6)

    1. E Lodewijcks
    2. P Deboosere

    (2011)

    Households and families: Stability and fast development go hand in hand (GGP Belgium Paper Series No. 6), Generations & Gender Programme Belgium, GGP Belgium Paper Series - No. 6. Retrieved from, http://www.ggps.be/doc/GGP_Belgium_Paper_Series_6-EN.pdf.

    • Google Scholar
33. 33
    The role of survey data in microsimulation models for social policy analysis

    1. A Martini
    2. U Trivellato

    (1997)

    Labour 11:83–112.

    • Google Scholar
34. 34
    A decomposition analysis of the trend in UK income inequality

    1. D Mookherjee
    2. A Shorrocks

    (1982)

    Economic Journal 92:886–902.

    • Google Scholar
35. 35
    Male-female wage differentials in urban labor markets

    1. R Oaxaca

    (1973)

    International Economic Review 14:693–709.

    • Google Scholar
36. 36
    sreweight: A Stata command to reweight survey data to external totals

    1. D Pacifico

    (2014)

    Stata Journal 14:4–21.

    • Google Scholar
37. 37
    Does size matter? the impact of changes in household structure on income distribution in Germany

    1. A Peichl
    2. N Pestel
    3. H Schneider

    (2012)

    Review of Income & Wealth 58:118–141.

    • Google Scholar
38. 38
    (TOR Working Paper 2000/41)

    1. K Pelleriaux

    (2000)

    Het ontstaan van een demotiecultuur in het secundair onderwijs. een cultuursociologische analyse van de nieuwe vormen van ongelijkheid en van de nieuwe sociale kwestie, Doctoral dissertation, Vrije Universiteit Brussel, Faculteit Economsiche, Sociale en Politieke Wetenschappen, Vakgroep Sociologie, Onderzoeksgroep TOR, Brussels, http://www.vub.ac.be/TOR/cgi-bin/publicatiea.cgi?Command=Opzoeken&opdracht=auteur&id=11, (TOR Working Paper 2000/41).

    • Google Scholar
39. 39
    Labor market discrimination against hispanic and black men

    1. CW Reimers

    (1983)

    Review of Economics and Statistics 65:570–579.

    • Google Scholar
40. 40
    Snapshots of our future families (FLEMOSI Discussion Paper No. 12). VUB

    1. I Schockaert
    2. J Surkyn

    (2012)

    Snapshots of our future families (FLEMOSI Discussion Paper No. 12). VUB, http://www.flemosi.be/uploads/125/FLEMOSIDP12PopulationProjectionbyHH.pdf.

    • Google Scholar
41. 41
    Technical report on the LiPro-projections (FLEMOSI Working Paper). VUB

    1. I Schockaert
    2. J Surkyn

    (2013)

    Technical report on the LiPro-projections (FLEMOSI Working Paper). VUB.

    • Google Scholar
42. 42
    Earnings inequality and the changing association between spouses’ earnings

    1. C Schwartz

    (2010)

    1524–1557, American Journal of Sociology, 115, 5, doi: 10.1086/651373.

    • Google Scholar
43. 43
    The class of additively decomposable inequality measures

    1. AF Shorrocks

    (1980)

    Econometrica 48:613–625.

    • Google Scholar
44. 44
    Inequality decomposition by population subgroups

    1. AF Shorrocks

    (1984)

    Econometrica 52:1369–1385.

    • Google Scholar
45. 45
    http://www.plan.be/admin/uploaded/200611090953120.OPVERG200601nl.pdf

    1. Studiecommissie voor de vergrijzing

    (2006)

    Jaarlijks verslag (Tech. Rep.). Hoge Raad van Financiёn.
46. 46
    SVR-internet page. Retrieved from

    1. Studiedienst Vlaamse Regering.

    (2011)

    SVR-projecties van de bevolking en de huishoudens voor Vlaamse steden en gemeenten, 2009–2030 (Tech. Rep.), Brussels, Studiedienst Vlaamse Regering, SVR-internet page. Retrieved from, http://data.gov.be/nl/dataset/svr-projecties-van-de-bevolking-en-de-huishoudens-voor-vlaamse-steden-en-gemeenten-2009-2030.

    • Google Scholar
47. 47
    http://ebl.vlaanderen.be/publications/documents/44799

    1. Studiedienst Vlaamse Regering

    (2013)

    Vlaamse armoedemonitor 2013 (Tech. Rep.). Studiedienst Vlaamse Regering.
48. 48
    Poverty, household composition, and welfare states: A multi-level analysis of 22 countries(Luxembourg Income Study Working Paper Series No. 492) Luxembourg Income Study. Retrieved from

    1. T-o Tai
    2. J Treas

    (2008)

    Poverty, household composition, and welfare states: A multi-level analysis of 22 countries(Luxembourg Income Study Working Paper Series No. 492) Luxembourg Income Study. Retrieved from, http://s3.amazonaws.com/zanran_storage/www.lisproject.org/Content[a80]ages/53940035.pdf, Presented at 2009 Annual Meeting - Population Association of America, Detroit, MI, United States, 30 April - 2 May.

    • Google Scholar
49. 49
    Does household composition explain welfare regime poverty risks for older adults and other household members?

    1. T-O Tai
    2. J Treas

    (2009)

    777–787, Journal of Gerontology: Psychology and Social Sciences, 64B, doi: doi:IO.IO93/geronb/gbp039.

    • Google Scholar
50. 50
    De recente evolutie van de vruchtbaarheid in het Vlaamse Gewest tussen 2001 en 2005 (Interface Demography Working Paper No. 2006-1)VUB. Retrieved from

    1. J Van Bavel
    2. V Bastiaenssen

    (2006)

    De recente evolutie van de vruchtbaarheid in het Vlaamse Gewest tussen 2001 en 2005 (Interface Demography Working Paper No. 2006-1)VUB. Retrieved from, http://www.vub.ac.be/SOCO/demo/papersonline/IDWP2006-1.pdf.

    • Google Scholar
51. 51
    Inkomen, verdeling en armoede: over groei, stabiliteit en de kloof tussen werkenden en uitkeringstrekkers

    1. K Van den Bosch
    2. P Vandenbroucke
    3. B Cantillon
    4. J Pacolet

    (2009)

    De sociale staat van Vlaanderen.
52. 52
    LiPro 2.0: An application of a dynamic demographic projection model to household structure in the Netherlands

    1. E Van Imhoff
    2. NW Keilman

    (1991)

    Amsterdam: Swets and Zeitlinger B.V.

    • Google Scholar
53. 53
    Educational expansion in Belgium: A sociological analysis using system theory

    1. R Vanderstraeten

    (1999)

    Journal of Educational Policy 14:507–522.

    • Google Scholar
54. 54
    http://www.centrumvoorsociaalbeleid.be/sites/default/files/socialeongelijkhedeninhetvlaamseonderwijs.pdf

    1. G Verbergt
    2. B Cantillon
    3. K Van den Bosch

    (2009)

    Sociale ongelijkheden in het Vlaamse onderwijs: tien jaar later.
55. 55
    Armoede en sociale uitsluiting. jaarboek 2006

    1. J Vranken
    2. K De Boyser
    3. D Dierckx

    (2006)

    Armoede en sociale uitsluiting. jaarboek 2006, Leuven, Acco.

    • Google Scholar
56. 56
    Inequality among American families with children, 1975 to 2005

    1. B Western
    2. D Bloome
    3. C Percheski

    (2008)

    American Sociological Review 73:903–920.

    • Google Scholar
57. 57
    http://databank .worldbank.org/data/reports.aspx?source=2&series=NY.GD[a80].MKTP.KD.ZG&country=

    1. World Bank

    (2017)

    World Development Indicators..

Author details

1. Rembert De Blander

   Department of Economics, KU Leuven, Belgium

   For correspondence

   Rembert@DeBlander.eu

2. Ingrid Schockaert

   Interface Demography, Vrije Universiteit Brussel, Belgium

   For correspondence

   ingrid.schockaert@vub.ac.be

3. André Decoster

   Department of Economics, KU Leuven, Belgium

   For correspondence

   andre.decoster@kuleuven.be

4. Patrick Deboosere

   Interface Demography, Vrije Universiteit Brussel, Belgium

   For correspondence

   Patrick.Deboosere@vub.ac.be