The Two Dimensions of Housing Inequality in Europe Are High Home Ownership Rates an Indicator of Low Housing Values? Kathrin Kolb, Nora Skopek, Hans-Peter Blossfeld Abstract: Exploring inequalities in home ownership as an important component of household wealth contributes to the understanding of social stratifi cation in mod- ern societies. We argue that inequalities in housing are not only manifested by dif- ferential access to home ownership, but also by differences in housing values, a somewhat neglected aspect in research hitherto. Applying data from the “Survey of Health, Ageing and Retirement in Europe” (SHARE),1 we compare home ownership rates and housing values between 13 European countries. Our results suggest that housing inequality is indeed a two-dimensional phenomenon. Most surprisingly, migration status has a negative impact on the probability of home ownership in Eu- ropean countries, but not on the mean housing value. In addition, we exploratively study the relationship between these two dimensions of housing inequality. Our analyses show a negative though not signifi cant relationship between home owner- ship rates and housing values. Keywords: Housing · Wealth · Social Inequality · Europe · SHARE Comparative Population Studies – Zeitschrift für Bevölkerungswissenschaft Vol. 38, 4 (2013): 1009-1040 (Date of release: 17.12.2013) © Federal Institute for Population Research 2013 URL: www.comparativepopulationstudies.de DOI: 10.12765/CPoS-2013-22en URN: urn:nbn:de:bib-cpos-2013-22en0 1 This paper uses data from SHARE release 2.5.0, as of May 24th, 2011. The SHARE data col- lection has been primarily funded by the European Commission through the 5th framework programme (project QLK6-CT-2001- 00360 in the thematic programme Quality of Life), through the 6th framework programme (projects SHARE-I3, RII-CT- 2006-062193, COMPARE, CIT5- CT-2005-028857, and SHARELIFE, CIT4-CT-2006-028812) and through the 7th framework pro- gramme (SHARE-PREP, 211909 and SHARE-LEAP, 227822). Additional funding from the U.S. National Institute on Aging (U01 AG09740-13S2, P01 AG005842, P01 AG08291, P30 AG12815, Y1-AG-4553-01 and OGHA 04-064, IAG BSR06-11, R21 AG025169) as well as from various na- tional sources is gratefully acknowledged (see http://www.share-project.org for a full list of funding institutions). • Kathrin Kolb, Nora Skopek, Hans-Peter Blossfeld1010 1 Introduction Studying home ownership2 is of major sociological importance, as social inequali- ties are not only defi ned by educational, occupational or income inequalities, but also in terms of real property (Kurz/Blossfeld 2004; Lewin-Epstein et al. 1997). Home ownership is an important purpose in life for many people. Frequently mentioned reasons for the desire to purchase residential property are independence from the landlord, capital investment, the possession of a house as a long-lasting value which can also be transferred to the children, the house as a kind of secure old-age provision3 as well as long-term protection against infl ation and a home of one’s own as a means to achieve a higher quality of life (Faller et al. 2001; LBS Research 2004). Home ownership can also serve as a symbol of status and success (cf. Constant et al. 2007). Various studies have revealed that residential property is an essential factor for wealth accumulation (see Brandolini et al. 2004; Grabka/Frick 2007; Skopek et al. 2012; Sierminska et al. 2007). Yet there are major differences in home ownership rates in Europe, these varying between 35 percent in Switzerland and 83 percent in Spain (Euroconstruct/ifo 2009). The distribution of home ownership in general, but also the analyses of socio-economic determinants that affect the probability of becoming a homeowner, have received the attention of a number of research- ers (e.g. Kurz/Blossfeld 2004; Wagner/Mulder 2000). We argue that the differentia- tion between owners and non-owners is only one dimension of social inequality in housing. Being a homeowner does not necessarily imply that a household is wealthy, as the value of a house heavily depends on the location, the social environ- ment (neighbourhood) as well as the quality of the residential property (e.g. Besley/ Mueller 2012; Li/Brown 1980). All these factors are refl ected in the housing value. Thus, in order to capture social inequality patterns in home ownership in their en- tity, it is unavoidable to also take into account the real estate value, which has been a somewhat neglected aspect in research on housing in the social sciences so far (exceptions: Krivo/Kaufman 2004; Lewin-Epstein et al. 1997). In our paper, we will account for both of the above-mentioned dimensions of social stratifi cation in housing by analysing whether various socio-economic house- hold characteristics differently affect 1) the probability of being a homeowner and 2) the value of housing within different European countries. In addition to that, we are 3) interested in the relationship between these two dimensions of social strati- fi cation in housing. Our contribution is thus twofold: Firstly, we provide a broad in- ternational comparison of home ownership rates and housing values, and secondly we explore the relationship between those two dimensions of housing (inequality). 2 The paper is interested in studying owner-occupied home ownership. The expressions “resi- dential property”, ”home ownership” and “own homes” are used synonymously. The same ap- plies to the expressions “housing value”, “real estate value” and “value of residential property”. 3 Some authors (Castles 1998; Kemeny 1981) also argue that there might be a trade-off between the expansion of home ownership and the generosity of old-age pensions within countries. The Two Dimensions of Housing Inequality in Europe • 1011 The population on which we focus is elderly Europeans because it is at this stage of life that residential property is particularly common in all European societies (Sierminska et al. 2007). As the elderly are generally confronted with a considerable fall in their income when they retire, their socio-economic position can only be ad- equately determined if one additionally takes wealth into account (e.g. Spilermann 2000). The fi nancial position of homeowners whose housing is free from debts is strengthened by the fact that they do not need to invest money to rent a house or fl at, so that they can spend these resources on consumption or savings (Wolff et al. 2005: 1076). Considering the ageing of industrialised societies and the growing importance of private pension provision, we assume that wealth and owner-occu- pied housing as an important part of it will even become more important in future. However, purchasing an apartment or a house can also have negative aspects, es- pecially among the very old (75 years and above). High fi nancial burdens, mobil- ity restrictions and high (transaction) costs when selling residential property are often associated with home ownership (Bourdieu 1998; Häußermann/Petrowsky 1990; Häußermann/Siebel 2000; Sierminska et al. 2007). Moreover, in many cases housing is the only noteworthy wealth component of elderly households. As hous- ing wealth is illiquid wealth, it cannot be directly used for consumption. Therefore, elderly homeowners are sometimes described as housing rich, but cash poor (An- gelini et al. 2009; Venti/Wise 2000). In the following section, we will give an overview of the current research on socio-economic variables affecting the probability of home ownership and housing values. For our statistical analyses, we use the second wave of the Survey of Health, Ageing and Retirement in Europe (SHARE) that was conducted in 13 countries:4 Austria, Belgium, the Czech Republic, Denmark, France, Germany, Greece, Italy, the Netherlands, Poland, Spain, Sweden and Switzerland. We apply logistic regressions to analyse household characteristics affecting the chance of being a homeowner, and linear regressions to analyse household characteristics that infl uence the value of housing among homeowners. We make use of a multi-level model to fi nd out about the relationship between home ownership rates and housing values. Consid- ering housing inequality as a twofold process and analysing it over a broad range of countries will enable us to obtain a better, multidimensional understanding of social inequalities in housing (see Fig. 1). 4 Ireland also took part in the second wave of SHARE. However, as imputations are not available for Ireland, we decided to leave it out. • Kathrin Kolb, Nora Skopek, Hans-Peter Blossfeld1012 2 Current research and expectations The impact of socio-economic characteristics on home ownership rates and housing values Housing is likely to be affected by various individual and household characteris- tics, such as age, household size, children, family status, education, occupation, income, inheritances, migration status and urbanisation (e.g. Krivo/Kaufman 2004; Kurz/Blossfeld 2004; Lewin-Epstein et al. 1997). The impact of these socio-econom- ic characteristics does not necessarily need to be the same for the two dimensions of housing inequality – home ownership rates and housing values. In addition, the national institutional settings, characterised by a country’s welfare regime, should exert a major infl uence on its housing situation (see Kurz/ Blossfeld 2004). Welfare regimes can be expected to affect the individual chances and incentives to acquire property (e.g. through taxation, housing allowances). Dif- ferences in social security systems, especially retirement systems, may affect the need to own a home as a part of private old-age provision (see DeWilde/Raeymae- ckers 2008). Whereas Northern welfare states are known for high state-provided social security also for the elderly, home ownership plays a crucial role for old-age provisions in Southern Europe because there is almost no serious state-provid- Fig. 1: The two dimensions of social stratifi cation in housing Source: Own design Population Low value High value Owners Non-Owners The Two Dimensions of Housing Inequality in Europe • 1013 ed pension scheme. As a result, in Southern European welfare states residential property is the main if not the only wealth component, while in Northern Europe fi nancial assets are of greater importance (Sierminska et al. 2007). In order to cap- ture country-specifi c differences in home ownership rates and values in our analy- ses, we will distinguish between Northern, Central and Southern Europe as well as post-socialist countries, following the welfare state typology of Esping-Andersen (Esping-Andersen 1990). Finally, welfare regimes can also infl uence the patterns and intensities of social inequalities in housing arising from the above-named socio- economic characteristics. The lifecycle hypothesis (Modigliani/Brumberg 1954) states that wealth grows with advancing age, as people accumulate increasing amounts of wealth through- out their working lives by saving parts of their income in order to keep their con- sumption level stable over their life course. When entering retirement, they then start “dissaving” (consuming their wealth). We assume that home ownership rates follow a similar pattern (Artle/Varaiya 1978). However, different studies have shown that the probability of being a homeowner only starts to decrease signifi cantly from the age of 70 onwards (Tatsiramos 2006; Venti/Wise 2000). The reasons for this de- crease are that the elderly put their homes in their children’s names (e.g. for fi scal reasons) or sell them in case of the loss of the partner through death (Chiuri/Jappelli 2010) or to fi nance the move into an old people’s home or into a smaller (rented) fl at (Häußermann/Siebel 2000; Mulder/Wagner 1998). It is nonetheless remarkable that many elderly retain their homes, which means that they have a high stock of illiquid capital that cannot be used directly for consumption (Angelini et al. 2009; Attanasio et al. 2011). Regarding our sample of elderly households, the home ownership rate should already be at its peak (c.f. Scanlon/Whitehead 2004), and should only slightly start to decrease with age. Given the current (market) value of residential property, we do not see any reason to expect differences over age in our sample. However, processes like “ageing in place” (which might result in age-homogenous residential areas) may result in decreasing housing values as people age. As this most often takes place in suburban areas (Frey 2011; Swiaczny et al. 2012), we argue that we can control for this effect by differentiating between urban and sub-urban areas in our analysis.5 The literature unanimously reports that household composition is crucial for the home ownership situation. Couples and families with children in particular live in their own homes more often compared to singles (Davidov/Weick 2011; Lewin- Epstein/Semyonov 2000; Mulder 2006; Wagner/Mulder 2000). In general, the prob- ability of living in one’s own home increases with a growing number of people living in the household, as residential property is often associated with a family-friendly residential area and comfortable living accommodation (Häußermann/Siebel 2000; Mulder/Wagner 1998). The value of residential property is also found to be posi- 5 Another phenomenon that might lead to decreasing housing values over age is the process of “asset meltdown”. Yet, so far there is no empirical evidence that this process is actually taking place (see for example Börsch-Supan et al., 2003). • Kathrin Kolb, Nora Skopek, Hans-Peter Blossfeld1014 tively infl uenced by the number of household members (Lewin-Epstein et al. 1997). When it comes to the property value, we assume that having a partner also has a positive effect, while we have no clear assumptions for parenthood and household size.6 Finally, we expect family-related characteristics to be especially important in Southern European welfare states where families have a central infl uence on the standard of living and therefore also on the home ownership situation (Esping- Andersen 1990). Previous studies have emphasized the signifi cant infl uence of an individual’s ed- ucational and occupational status7 on the transition to home ownership (Kurz/Bloss- feld 2004; Wagner/Mulder 2000). Additionally, the chance of receiving bequests or inheritances increases with higher educational and occupational status, as these individuals often originate from higher-status families (Blau/Duncan 1967; Buch- holz 2008; Szydlik/Schupp 2004). Regarding the impact of different welfare state regimes, Kurz and Blossfeld (2004) were able to show for example that occupational status has a greater impact on the transition to home ownership in liberal welfare states compared to social-democratic regimes. We likewise expect educational at- tainment levels and income to increase the probability of home ownership as well as the value of residential property. The educational level should be particularly important in countries with a highly standardised, stratifi ed education system and a strong vocational specifi city (like Germany and Switzerland), as the impact of formal qualifi cations on the employment career, and therefore on the potential of wealth accumulation, is especially strong in these countries (Müller/Shavit 1998). In addi- tion to that, it we expect that households that received fi nancial gifts or inheritances have a higher probability of being homeowners. Furthermore, if they own a dwell- ing it might be of higher value as intergenerational transfers enhance the house- hold’s wealth position. In the U.S., people of African-American and Latin-American origin are less likely to own residential property, and if they realise home ownership, their houses are often of low value (Krivo/Kaufman 2004; Lewin-Epstein et al. 1997; Parcel 1982). Al- though a higher educational background and growing income weaken the negative impact of a migration background on home ownership, different studies show clear evidence that even when controlling for these variables, discrimination on the hous- ing market can still be observed for migrants (Chiteji/Stafford 1999; Horton/Thomas 1998; Krivo/Kaufman 2004). Possible explanations of these fi ndings are migrants’ disadvantaged labour market position, the fact that migrants less often receive be- quests and inheritances, migrants’ information defi cit on the local housing market, as well as discrimination against migrants in the credit approval process (Charles/ 6 This is because more living space is needed with an increasing number of people living in a household, and at least in multigenerational households more people can help fi nance the property. However, more people and having children cause higher costs that reduce the fi nan- cial resources available to purchase residential property, which could have a diminishing effect on the residential property value as well. 7 Occupational status is not included in our analyses as many people in our sample are already retired and their (former) occupational status is then unknown. The Two Dimensions of Housing Inequality in Europe • 1015 Hurst 2002; Conley 2003; Krivo/Kaufman 2004; Szydlik/Schupp 2004). So far, the question of whether this holds true for European countries as well is however un- resolved. Due to the high real estate prices, greater fi nancial resources are needed in cities in order to gain access to home ownership, thus reducing the probability of own- ing residential property (Kurz/Blossfeld 2004). However, if owner-occupied housing has been realised successfully, it should consequently be of higher value in urban communities. Linking home ownership rates to housing values In addition to the socio-economic factors that have an impact on home ownership rates and housing values, we are also interested in the relationship between these two dimensions. More precisely, we want to fi nd out if high home ownership rates correlate with high housing values (positive relationship) or if they can only be real- ized at the cost of low housing values (negative relationship)? As far as we know, this research question has not yet been adequately addressed in social inequality research. We will exploratively approach this question in our study. Theoretically, both relationships (positive and negative) are possible. Firstly, the socio-economic composition of the group of homeowners may lead to a negative relationship. It might be the case in countries with low ownership rates but high housing values that only a very selective group of better-off households achieves home ownership (Poggio 2006). Those would be households, which can also afford high-value housing. If this is the case, the homogeneous composition of this group could explain the high mean housing values in these countries. Likewise, in coun- tries with high home ownership rates, homeowners might be a rather heterogene- ous group (everybody has access to housing), which should lead to low mean hous- ing values on the aggregate level, given that a large share of those homeowners cannot afford high-value housing. If we fi nd a negative relationship between home ownership rates and housing values in our analysis, homeowners’ socio-economic composition might be a possible explanation for this relationship. Secondly, the scenario of a positive relationship is also possible. If a country’s rental market can be classifi ed as unattractive compared to the home ownership market (e.g. low quality of rented housing or small rental sector), being a home- owner might become an interesting, desirable alternative. If demand for home own- ership is high, housing prices are likely to increase. This however might still not deter individuals from buying residential property, thus pushing up housing prices further. So if we fi nd a positive relationship between home ownership rates and housing values, an unattractive renting market might be a possible explanation. In general, housing is considered as an important component of asset-based so- cial security. That is why most welfare states implemented certain housing policies (like interest subsidies and housing allowances) to support and provide social se- curity via home ownership (Elsinga et al. 2007). By running multi-level regressions, we are testing statistically whether country-level variables can explain the variation in individual housing values. • Kathrin Kolb, Nora Skopek, Hans-Peter Blossfeld1016 3 Data and methods Data In our analysis, we make use of the second wave of the SHARE data. This survey is an international, representative panel study of the population aged 50 years and older, it is currently in its fourth wave. The main advantage of the SHARE data is that it provides detailed, internationally comparable information on the fi nancial and housing situation (in waves one, two and four). As the fi rst release of the wave-four data does not yet contain all the variables that are relevant to our analyses, and as waves one and two were conducted in a rather narrow period of time (2004 to 2007), we decided to work with the second wave only, which covers a broader range of countries. The observation that becoming a homeowner in the life course is a slow process with few events confi rms our decision (Venti/Wise 1989). In the second wave, conducted in 2006/2007, 33,281 people in 22,721 households from 13 EU member states (listed above) participated in the survey. After eliminat- ing households with missing or implausible values as well as households where none of the people interviewed were aged 50+ (148 households), our fi nal dataset contains 20,9458 households. A typical problem of questions addressing fi nancial aspects is a high rate of item non-response (Riphahn/Serfl ing 2005). The SHARE team is tackling this problem by applying a multiple imputation strategy for fi lling in missing values (for further information on multiple imputation see Rubin 1987). Five values were estimated for every missing value.9 A more detailed description of the imputation method used in the SHARE can be found in Christelis (2011). All the analyses reported below were run across the fi ve. Moreover, all fi nancial values are adjusted for differences in the purchasing power of money across countries and over time using the exchange rates provided by the SHARE team (see Man- nheim Research Institute for Economics of Aging 2010 for further information). We use cross-sectional calibrated weights that “are calibrated to precisely refl ect each country’s age and gender proportions” (Börsch-Supan et al. 2005: 21) for our de- scriptive analyses. These weights compensate for problems of unit non-response and sample attrition (cf. Mannheim Research Institute for Economics of Aging 2010). Table 1 illustrates the sample size per country (weighted). 8 For our analyses, we eliminated households with missing or implausible values in the follow- ing variables: owner (n=334), family status (n=3), migration status (n=88), educational level (n=60), retirement status (n=266), fi nancial transfers/inheritances (n=260) and residential area (n=1,462). 9 For total household income over all countries in about 60 percent of the households, at least one component (item) of total household income has been imputed. Income is a generated vari- able consisting of a battery of different items (see footnote 13). The Two Dimensions of Housing Inequality in Europe • 1017 Variables10 • Homeowner is the dependent variable in our fi rst analysis. It differentiates between households owning residential property (=owners) and house- holds not owning residential property (=non-owners). • Financial value of residential property is the dependent variable in our sec- ond analysis.11 Financial value stands for the subjective market value esti- mated by the fi nancial respondent.12 It ranges between €013 and €27,950,000 (ppp-adjusted). As the distribution of this variable is very much skewed to the right, we use the variable’s log value in our analyses. Tab. 1: Overview of the dataset Source: SHARE Wave 2, release 2.5.0, weighted data, own calculations Country Total Percent Cum. in % AT – Austria 897 4.29 4.34 BE – Belgium 2,009 9.60 13.89 CH – Switzerland 967 4.62 18.51 CZ – Czech Republic 1,721 8.23 26.74 DE – Germany 1,548 7.40 34.14 DK – Denmark 1,662 7.94 42.08 ES – Spain 1,278 6.11 48.19 FR – France 1,884 8.81 57.00 GR – Greece 2,083 9.96 66.96 IT – Italy 1,786 8.54 75.50 NL – Netherlands 1,709 8.17 83.67 PL – Poland 1,697 8.11 91.78 SE – Sweden 1,723 8.22 100.00 Total 22,924 100.00 10 All the variables mentioned in this section refer to the current state of the household. Unfortu- nately, we have no information on these variables at the time when the household bought or acquired the residential property. 11 As we are interested in the actual value of residential property and not in households’ level of indebtedness, we do not take into account the net but the gross value (market value). The fact that it is likely that the households are in different stages of their repayment and that the meth- od of fi nancing home ownership varies widely between countries makes it even more plausible for us to make use of the gross housing value. 12 The exact question in the SHARE questionnaire was: “In your opinion, how much would you receive if you sold your property today?” 13 20 households were assigned a housing value of zero. We kept them in our sample but for the analysis if the housing value, we added €1 to those households in order to calculate the loga- rithm of those housing values. • Kathrin Kolb, Nora Skopek, Hans-Peter Blossfeld1018 • Age corresponds to the mean age of all household members surveyed. Age ranges between 34 (for households with people aged above and below 50 years) and 104 years. We also calculated age square to test the assumption that the rate of home ownership fi rst increases with age and then starts to decrease. • Household size controls for the number of people living in a household. It ranges between 1 and 14 people. • Family status informs us whether the main respondent is living together with a spouse (family status=1) or as a single person (family status=0). • Children controls for parenthood of the main respondent and his/her spouse, irrespective of whether the child still lives in the parental household. • Migration status informs us whether the main respondent and/or his/her spouse were born abroad (migrations status=1). • Educational level (7 categories, ISCED-coded) equals the highest education- al attainment level of the main respondent and his/her spouse. It ranges be- tween 0 (pre-primary education) and 6 (second stage of tertiary education). • Net equivalent income14 is measured as yearly total household net income divided by the root of the number of people living in this household. It ranges from €0 to €727,000 (ppp-adjusted). • Retirement status differentiates between households where the main re- spondent and/or his/her spouse are already retired vs. households where none of them is retired yet. As a large proportion of household members in our dataset are already retired, we use this variable for control reasons. • Gifts and inheritances15 controls for whether a household has ever received a fi nancial gift, inherited money, goods or property (of at least €5,000). • Residential area informs us whether a household is located in a big city or in the suburbs or outskirts of a big city (residential area=1) or in a small town or a rural area or village (residential area=0). Methods In order to account for socio-economic factors that have an effect on the probabil- ity of being a homeowner, we will apply binary logistic regression models. Subse- quently, trying to fi nd socio-economic factors that affect the fi nancial value of resi- dential property, we use linear regression models. To fi nd out about the relationship between home ownership rates and housing values, we fi nally run linear regression models once more, but this time in a multi-level framework. Our unit of analysis is the household. The binary logistic model that we use aims at estimating the probability of be- longing to the group of homeowners (y=1): 14 See Paccagnella and Weber (2005: 357) for the exact defi nition of income in the SHARE. 15 We pooled information from waves one and two to obtain this information. The Two Dimensions of Housing Inequality in Europe • 1019 We estimate a separate model for each country. For more details on the binary logistic model see Long (1997). To analyse the effect of different socio-economic attributes on the fi nancial value of residential property, we apply a linear regression model, fi rstly in a single-level framework and secondly in a two-level framework, with households nested in countries: where Yij is the dependent variable (housing value) for households i clustered in countries j. Xij are predictors on the household level and εij is the household-level error term. βo+uj is the random intercept that varies across countries. For more details on multilevel regressions, consider for example Hox (2010) or Rabe-Hesketh and Skrondal (2008). Note that there are only 13 cases on level 2 (countries). Al- though there is no consensus in the literature regarding the minimum number of cases for upper levels in multilevel analyses, 13 cases is without doubt very small. Simulation studies on two-level linear models claim that standard errors and vari- ance components tend to be underestimated when the number of cases on the sec- ond level is smaller than 30 (Bell et al. 2008; Hox 2010; Maas/Hox 2005), which is the case in our study. Hence, the statistical power of our country-level effects might be rather small. To account for that, we additionally ran an alternative single-level re- gression with robust and cluster-adjusted standard errors. The results were largely the same. Finally, for substantive reasons, we opted for the multilevel estimation approach, which allows an explicit modeling of variance across countries. Housing values can logically only be observed for the group of homeowners. For the analysis of housing values, the most obvious strategy would therefore be to drop non-owners out of the sample. This strategy was applied until the late 1990s (e.g. Horton/Thomas 1998; Myers/Chung 1996; Parcel 1982). The 1997 article by Lewin-Epstein et al. was one of the fi rst to state that restricting the analysis to homeowners can lead to biased estimations, as the sample of homeowners might be a self-selective one. To avoid selection bias, the authors applied tobit regression models (also called censored regression models, see Tobin 1958) instead of linear regressions. Krivo and Kaufman (2004) applied the same strategy. However, neither of the two articles comprehensibly explains that a sample selection problem indeed exists. We argue instead that one cannot diagnose a bias in the analysis due to restricting the sample on homeowners. In fact, in our case, the selection process (the decision to become a homeowner) creates a necessary precondition for our outcome (the housing value) (cf. Rohwer 2012). When applying the tobit regression, Lewin-Epstein et al. (1997) as well as Krivo and Kaufman (2004) are thus modelling a • Kathrin Kolb, Nora Skopek, Hans-Peter Blossfeld1020 very specifi c and very hypothetical choice situation:16 the choice for a certain hous- ing value by an individual who has not yet purchased a house. However, in our pa- per we are explicitly not interested in this hypothetical decision at all. We are rather interested in the realised distribution of housing values of households that actually own a home. Thus, when analysing the housing value we will restrict our sample to homeowners. More precisely, we will carry out a two-part model. Firstly, we regress on the chance of being a homeowner among all households, and secondly we re- gress on the value of housing among those households that own a home. 4 Results Descriptive overview Table 2 shows the distribution of socio-economic household characteristics among homeowners in comparison to the overall population in our sample. About 70 per- cent of the households live in their own real property. Country differences in home ownership rates are also illustrated in Figure 2: Especially in Southern European countries like Spain, Greece and Italy, home ownership is a widespread phenom- enon, which is in line with current research. Also in some Western European coun- tries such as Belgium and France, there are many elderly households who live in owner-occupied property. In contrast, it is, as expected, relatively common not to own residential property in Austria, the Czech Republic, Sweden, Germany and Switzerland. We compared our fi ndings with data collected by Euroconstruct/ifo on the whole adult population. As expected, the rate of homeowners is higher among the elderly.17 Yet the ranking remains similar for all countries except for marginal shifts. Going back to Table 2, we can see that the average value of residential property among homeowners is €260,530; the median is €194,960 (right-skewed distribu- tion). The highest mean housing values can be found in Switzerland (€487,650) and the lowest in the Czech Republic (€105,800).18 These fi ndings are illustrated in Fig- ure 3. Looking at the median value of residential property reveals unambiguous country-specifi c patterns. In Continental Europe, the median is generally very high (greater than €194,000),19 with Switzerland again on the top (€306,210). The median 16 The same holds true for the Heckman selection model often used as an alternative to the tobit regression (Heckman 1979). 17 The Czech Republic and Sweden are the only cases where home ownership for the overall population is higher than for the elderly. 18 The term “housing value” is used in the article as a synonym for the gross housing value (market value). In general, the share of households with a mortgage on their real property is low due to the age structure of the sample (with the exception of Denmark, the Netherlands, Sweden and Switzerland). 19 With the exception of Austria: 153,440 Euros. The Two Dimensions of Housing Inequality in Europe • 1021 T a b . 2 : S o ci o -e co n o m ic h o u se h o ld c h a ra c te ri st ic b y p o p u la ti o n g ro u p a cr o ss c o u n tr ie s A ll C o u n tr ie s N o rt h e rn E u ro p e C o n ti n e n ta l E u ro p e D K S E A T B E C H D E A ll O w n e r A ll O w n e r A ll O w n e r A ll O w n e r A ll O w n e r A ll O w n e r A ll O w n e r O w n e r = y e s (% ) 7 0 .3 0 - 6 6 .6 7 - 5 7 .5 0 5 9 .9 0 - 8 0 .4 0 - 5 7 .4 0 5 8 .2 6 - A g e - ( ) 6 5 .9 4 6 5 .4 4 6 5 .1 1 6 3 .0 0 6 6 .3 4 6 4 .5 1 6 7 .1 6 6 5 .2 8 6 6 .2 2 6 5 .7 2 6 5 .8 5 6 3 .8 8 6 5 .7 2 6 5 .1 8 H o u se h o ld s iz e - ( ) 2 .2 1 2 .3 0 1 .8 0 1 .9 8 1 .9 0 2 .1 1 1 .9 0 2 .0 8 2 .0 1 2 .0 9 1 .9 9 2 .1 8 1 .9 8 2 .1 1 M ar ri e d = y e s (% ) 6 5 .7 8 7 2 .3 8 6 2 .4 6 7 6 .3 1 6 7 .5 3 8 2 .8 6 5 7 .3 8 6 7 .3 3 6 7 .9 2 7 4 .0 2 6 6 .3 3 7 8 .5 5 6 7 .7 9 7 5 .1 3 C h ild (r e n ) = y e s (% ) 8 8 .4 6 8 9 .3 3 8 9 .9 4 9 1 .0 2 9 1 .0 5 9 3 .7 5 8 7 .0 8 8 8 .7 7 8 8 .6 3 8 8 .8 3 8 5 .1 9 8 8 .0 2 8 7 .5 8 8 8 .1 9 M ig ra n t = y e s (% ) 1 0 .6 4 7 .8 8 4 .9 4 4 .4 0 1 0 .9 9 9 .7 3 8 .9 1 7 .0 6 9 .8 6 8 .6 4 1 8 .3 8 1 4 .5 4 2 0 .9 7 1 6 .4 6 E d u ca ti o n ( IS C E D ) - ( ) 2 .5 0 2 .5 2 3 .3 2 3 .5 1 2 .7 1 2 .8 0 2 .8 9 3 .0 0 2 .7 6 2 .8 6 2 .8 4 2 .9 8 3 .3 4 3 .4 6 R e ti re d = y e s (% ) 6 0 .0 4 6 0 .2 6 5 3 .4 3 4 7 .9 6 5 9 .2 1 5 4 .1 2 7 0 .8 4 6 7 .1 7 5 9 .8 8 5 9 .8 3 4 9 .9 3 4 4 .2 3 5 9 .9 9 5 8 .8 2 N e t e q u iv al e n t in co m e * - ( ) 2 0 .2 3 2 1 .6 6 2 2 .1 6 2 4 .6 5 2 2 .8 3 2 4 .7 4 1 9 .4 4 2 0 .8 8 2 2 .2 8 2 3 .2 7 3 0 .1 0 3 4 .1 6 2 4 .1 3 2 7 .5 5 N e t e q u iv al e n t in co m e * - M e d ia n 1 4 .6 8 1 5 .5 9 1 9 .1 0 2 2 .2 8 1 9 .5 6 2 1 .8 8 1 6 .9 9 1 8 .0 1 1 5 .9 9 1 6 .8 3 2 3 .6 2 2 6 .9 5 1 8 .1 2 2 0 .9 6 T ra n sf e r/ B e q u e st = y e s (% ) 2 5 .3 8 2 9 .5 1 4 3 .3 3 4 9 .7 3 4 6 .0 5 5 1 .2 9 2 0 .9 8 2 5 .9 2 4 5 .0 3 4 9 .5 0 4 1 .1 4 4 9 .2 0 2 8 .1 6 3 6 .3 8 U rb an a re a = y e s (% ) 4 4 .5 1 3 9 .2 3 5 2 .2 9 4 4 .3 2 6 4 .5 0 5 1 .9 0 3 8 .1 6 2 6 .4 7 4 0 .9 5 3 6 .3 1 3 0 .7 5 1 7 .9 1 4 2 .1 7 3 3 .3 8 H o u si n g v al u e * - ( ) - 2 6 0 .5 3 - 2 6 1 .5 6 - 2 2 1 .0 2 - 2 0 5 .6 5 - 2 3 1 .2 4 - 4 8 7 .6 5 - 2 3 2 .6 7 H o u si n g v al u e * - M e d ia n - 1 9 4 .9 6 - 1 7 7 .2 1 - 1 4 1 .4 0 - 1 5 3 .4 4 - 1 9 4 .6 6 - 3 0 6 .2 1 - 1 9 6 .9 0 N 2 0 ,9 2 4 1 4 ,8 1 3 1 ,6 6 2 1 ,1 1 3 1 ,7 2 3 1 ,0 2 2 8 9 7 5 2 8 2 ,0 0 9 1 ,5 9 8 9 6 7 5 5 2 1 ,5 4 8 9 3 1 • Kathrin Kolb, Nora Skopek, Hans-Peter Blossfeld1022 T a b . 2 : co n ti n u a ti o n C o n ti n e n ta l E u ro p e S o u th e rn E u ro p e E as te rn E u ro p e F R N L E S G R IT C Z P L A ll O w n e r A ll O w n e r A ll O w n e r A ll O w n e r A ll O w n e r A ll O w n e r A ll O w n e r O w n e r = y e s (% ) 7 4 .6 2 - 6 1 .7 2 - 8 9 .8 7 - 8 6 .1 4 - 7 8 .7 7 - 5 9 .3 4 - 6 4 .1 3 - A g e - ( ) 6 5 .9 9 6 5 .4 4 6 4 .9 5 6 2 .2 0 6 6 .9 4 6 6 .7 0 6 5 .8 7 6 5 .6 0 6 6 .7 0 6 6 .5 4 6 4 .8 0 6 4 .0 0 6 4 .4 4 6 4 .2 7 H o u se h o ld s iz e - ( ) 2 .0 2 2 .0 6 1 .9 8 2 .1 7 2 .4 9 2 .5 0 2 .2 3 2 .2 4 2 .4 4 2 .4 8 2 .0 8 2 .1 7 2 .8 8 2 .9 1 M ar ri e d = y e s (% ) 6 4 .4 5 7 1 .9 7 6 7 .9 4 7 9 .6 1 6 6 .7 5 7 0 .2 7 6 4 .2 2 6 6 .8 7 6 8 .5 0 7 3 .2 3 6 2 .7 1 6 8 .3 7 5 9 .0 0 6 4 .8 7 C h ild (r e n )= y e s (% ) 8 8 .6 6 8 9 .7 8 8 8 .4 6 8 9 .7 4 8 7 .7 6 8 9 .3 3 8 8 .1 1 8 8 .2 1 8 7 .3 8 8 8 .0 9 9 3 .7 2 9 3 .5 5 9 1 .8 3 9 2 .2 5 M ig ra n t = y e s (% ) 1 6 .5 5 1 3 .1 9 6 .9 7 4 .7 2 2 .9 4 2 .1 4 2 .6 5 2 .4 4 1 .9 5 1 .7 1 6 .2 3 4 .8 7 3 .0 0 2 .6 1 E d u ca ti o n ( IS C E D ) - ( ) 2 .4 0 2 .5 6 2 .7 8 3 .1 2 1 .6 4 1 .6 3 1 .9 8 2 .0 0 1 .8 5 1 .9 2 2 .5 4 2 .5 5 2 .2 4 2 .2 9 R e ti re d = y e s (% ) 6 3 .0 1 6 4 .3 8 4 4 .0 3 3 9 .1 9 1 5 .7 4 1 5 .9 0 1 4 .2 5 1 4 .4 9 1 5 .8 5 1 6 .4 0 1 1 .2 0 1 1 .6 5 7 .2 1 7 .6 0 N e t e q u iv al e n t in co m e * - ( ) 2 7 .2 0 2 9 .9 2 2 9 .3 8 3 3 .4 4 1 0 .1 3 1 0 .1 8 1 0 .6 4 1 0 .7 0 1 2 .5 5 1 3 .0 6 9 .1 0 9 .2 1 5 .3 3 5 .7 5 N e t e q u iv al e n t in co m e * - M e d ia n 1 9 .8 7 2 2 .1 8 2 0 .9 0 2 3 .9 1 5 2 .3 4 5 3 .1 2 5 2 .6 6 5 3 .5 0 6 3 .7 2 6 6 .8 9 6 9 .9 9 6 7 .4 9 6 5 .0 4 6 7 .5 0 T ra n sf e r/ B e q u e st = y e s (% ) 2 5 .2 0 2 9 .5 3 2 8 .1 7 3 6 .5 0 2 2 .1 2 2 3 .3 6 2 4 .5 7 2 5 .8 4 2 0 .2 5 2 2 .5 1 2 2 .9 5 2 5 .4 6 1 4 .0 2 1 8 .3 5 U rb an a re a = y e s (% ) 4 3 .1 3 3 8 .2 0 6 7 .5 1 6 3 .0 9 5 5 .6 7 5 4 .5 4 7 5 .1 0 7 3 .9 0 2 5 .6 0 2 2 .8 1 6 5 .2 2 5 9 .1 3 4 6 .3 8 3 9 .4 6 H o u si n g v al u e * - ( ) - 3 3 6 .4 0 - 3 6 6 .9 0 - 3 3 8 .5 2 - 1 4 5 .1 3 - 2 4 8 .2 9 - 1 0 5 .8 0 - 1 1 1 .0 9 H o u si n g v al u e * - M e d ia n - 2 2 1 .0 3 - 2 6 7 .9 7 - 2 0 4 .4 6 - 1 2 0 .5 7 - 1 9 5 .9 0 - 8 4 .9 7 - 4 8 .4 3 N 1 ,8 4 4 1 ,3 4 9 1 ,7 0 9 1 ,0 8 7 1 ,2 7 8 1 ,1 5 3 2 ,0 8 3 1 ,7 9 2 1 ,7 8 6 1 ,4 3 7 1 ,7 2 1 1 ,1 3 0 1 ,6 9 7 1 ,1 2 1 * V al u e s in 1 ,0 0 0 E u ro s, p p p -a d ju st e d , b as e d o n 5 s e ts o f im p u ta ti o n s S o u rc e : S H A R E W av e 2 ( R e le as e 2 .5 .0 .) , d at a w e ig h te d , o w n c al cu la ti o n s The Two Dimensions of Housing Inequality in Europe • 1023 value of owner-occupied housing (at less than €85,000) is lowest in the post-social- ist countries Poland and the Czech Republic. When it comes to age, our fi ndings show that households owning a home are slightly younger on average than the overall population. In line with our expecta- tions, compared to the overall population, owners more often live in a steady part- nership (72.4 percent vs. 65.8 percent), as well as in larger households (2.30 vs. 2.21 people). With respect to parenthood, there are only minor differences between the two groups. Among homeowners, 7.9 percent of the households have a migration background, in contrast to 10.6 percent of the total population. The overall share of migrants is comparably low in Southern Europe and Poland, while it is remarkably high in France, Germany and Switzerland. As expected, owners have a higher level of education than the overall population. The mean educational level is particularly low in Southern Europe. The average net equivalent income of homeowners, at €21,660, is also higher than in the group of the overall population (€20,230). Again, Fig. 2: Percentage of homeowners in European comparison 0 20 40 60 80 100 Spain Greece Belgium Italy France Denmark Poland Netherlands Austria Czech Republic Germany Switzerland Sweden 83% 80% 67% 72% 58% 58% 54% 54% 52% 71% 43% 35% 61% 90% 86% 80% 79% 75% 67% 64% 62% 60% 59% 58% 57% 58% Age 50+ Age 18+ Home ownership rate in % Source: Euroconstruct/ifo (2009) and SHARE Wave 2 (release 2.5.0), weighted SHARE data, own calculations • Kathrin Kolb, Nora Skopek, Hans-Peter Blossfeld1024 notable country differences exist within the total population: Mean income is com- paratively low in Eastern and Southern Europe, while it is rather high in France, the Netherlands and Switzerland. Retirement status does not differ between owners and the overall population. Overall, homeowners appear to benefi t more often from fi nancial gifts or inheritances compared to the total population (29.5 percent vs. 25.4 percent). Finally, homeowners live less often in urban areas compared to the total population (39.2 percent vs. 44.5 percent). Which socio-economic variables can predict home ownership? Table 3 contains the logistic regression models (more detailed models can be found in the Appendix). The likelihood of owning a home rises signifi cantly with age in the Fig. 3: Mean and median values of houses 0,00 50,00 100,00 150,00 200,00 250,00 300,00 350,00 400,00 450,00 500,00 CH NL ES FR DK IT DE BE SE AT GR PL CZ Mean Median Housing value (in 1,000 Euros) Values ppp-adjusted, based on 5 sets of imputations Source: SHARE Wave 2 (release 2.5.0), weighted data, own calculations The Two Dimensions of Housing Inequality in Europe • 1025 T a b . 3: L o g is ti c re g re ss io n ( w it h r o b u st s ta n d a rd e rr o rs ) o n t h e c h an ce o f b e in g a h o m e o w n e r N o rt h e rn E u ro p e C o n ti n e n ta l E u ro p e S o u th e rn E u ro p e E as te rn E u ro p e D K S E A T B E C H D E F R N L E S G R IT C Z P L A g e (+ ) (+ ) (- ) (- ) + + (+ ) (- ) + (- ) + + (+ ) A g e ² (0 ) (0 ) (0 ) (0 ) 0 0 (0 ) (0 ) 0 (0 ) (0 ) 0 (0 ) H o u se h o ld s iz e (+ ) + (+ ) (+ ) (+ ) + - (+ ) (- ) (- ) (+ ) (+ ) (- ) P ar tn e rs h ip = y e s + + (+ ) + + + + + + + + 0 + C h ild (r e n ) = y e s (- ) (+ ) (- ) (- ) + (- ) (- ) (+ ) (+ ) (- ) (+ ) (- ) (- ) M ig ra n t = y e s (- ) (- ) - - - - - - - (- ) - - (- ) E d u ca ti o n ( IS C E D )* + (+ ) + + + + + + (+ ) (+ ) + (+ ) + N e t e q u iv al e n t in co m e * (+ ) (+ ) + (+ ) + + + (+ ) (0 ) (+ ) (0 ) (+ ) (+ ) R e ti re m e n t = y e s (+ ) (- ) (- ) (+ ) (- ) (- ) (+ ) (- ) (+ ) (+ ) + (- ) (+ ) T ra n sf e r/ B e q u e st = y e s + + + + + + + + + + + + + U rb an c o m m u n it y = y e s - - - - - - - - - - - - - C o n st an t (- ) - (+ ) (+ ) - - (- ) (+ ) (- ) (+ ) - - (- ) P se u d o R ² 0 .2 1 0 .1 9 0 .1 5 0 .1 5 0 .1 8 0 .3 5 0 .1 5 0 .2 0 0 .1 2 0 .0 5 0 .1 0 0 .0 5 0 .0 7 N 1 ,6 6 3 1 ,7 2 5 8 9 7 2 ,0 2 2 9 6 7 1 ,5 5 0 1 ,8 4 4 1 ,7 1 0 1 ,2 7 9 2 ,0 8 3 1 ,7 8 6 1 ,7 2 2 1 ,6 9 7 * V al u e s ar e b as e d o n 5 s e ts o f im p u ta ti o n s + p o si ti v e e ff e ct ( p 0 .0 5 ); - n e g at iv e e ff e ct ( p 0 .0 5 ); ( + )/ (- ) - n o t si g n ifi c an t S o u rc e : S H A R E W av e 2 ( re le as e 2 .5 .0 ), u n w e ig h te d d at a, o w n c al cu la ti o n s • Kathrin Kolb, Nora Skopek, Hans-Peter Blossfeld1026 Czech Republic, Germany, Italy, Spain and Switzerland. The quadratic age term has no effect on the probability of home ownership in our sample, i.e. in the countries analysed, home ownership rates do not appear to decline as people become older, thus contradicting the life-cycle hypothesis (Modigliani/Brumberg 1954). Parenthood only signifi cantly affects the likelihood of belonging to the group of homeowners in Switzerland. One explanation for this fi nding might be that child- lessness is a rather rare phenomenon in the population studied (Dorbritz 2005), as already indicated in Table 2. As expected, households that are experiencing a steady partnership have a signifi cantly higher chance of belonging to the group of homeowners compared to single households (except for Austria and the Czech Re- public). In Germany and Sweden, household size has a positive impact on the prob- ability of being a homeowner. However, there is not much variation in household sizes across countries. Looking at the migration status, we fi nd a negative impact on the probability of being a homeowner in all countries analysed, though the effect is not signifi cant in the Northern countries, Greece and Poland. Thus, the U.S. fi ndings also apply to Europe. As expected, education has a positive impact on the probability of being a home- owner in all countries studied, though it is not signifi cant. Moreover, our fi ndings highlight the strong infl uence of education in Central Europe, which is in line with previous research. Income has a strong impact on the probability of being a home- owner, particularly in Continental Europe. The control variable “retirement status” does not have a signifi cant effect on the likelihood of home ownership in most countries. This could be because the decision to acquire a home is mostly made before entering retirement. It is only in Italy that being retired positively affects the likelihood of home ownership. Financial gifts and inheritances have a signifi cantly positive effect in all countries. As expected, owner-occupied home ownership is less probable in urban areas. Summing up, the results of the logistic model for the probability of being a homeowner are basically consistent with our expectations presented above. Which socio-economic factors can predict the housing value? Table 4 contains the results of the regression analyses on the impact of socio-eco- nomic household characteristics on the value of housing. The impact of age is posi- tive in almost all countries studied. Again, the quadratic term has no infl uence on our dependent variable. Our analyses demonstrate – as expected – that household size and partnership status have a positive impact on the housing value. Household size has a signifi cant positive effect in the Southern and Eastern European countries as well as in Germany. The infl uence of parenthood is only signifi cant (positive) in Italy and the Czech Republic. A very interesting and somewhat astonishing fi nding is that migration status has no signifi cant effect on the value of housing in all countries except for Austria. This contradicts the fi ndings from previous studies, mostly conducted in traditional im- migration countries like the USA and Israel. The Two Dimensions of Housing Inequality in Europe • 1027 T a b . 4: L in e a r R e g re ss io n o n lo g ( h o u si n g v a lu e) N o rt h e rn E u ro p e C o n ti n e n ta l E u ro p e S o u th e rn E u ro p e E as te rn E u ro p e D K S E A T B E C H D E F R N L E S G R IT C Z P L A g e (- ) (+ ) (+ ) (+ ) (+ ) + (+ ) (+ ) (- ) (+ ) (+ ) (+ ) (+ ) A g e ² (0 ) (0 ) (0 ) (0 ) (0 ) 0 (0 ) (0 ) (0 ) (0 ) (0 ) (0 ) (0 ) H o u se h o ld s iz e (+ ) (+ ) (+ ) (+ ) (+ ) + (+ ) (+ ) + + (+ ) + + P ar tn e rs h ip = y e s (+ ) (+ ) (+ ) + (+ ) (0 ) (+ ) + (+ ) (+ ) (+ ) + (+ ) C h ild (r e n ) = y e s (- ) (+ ) (+ ) (+ ) (+ ) (- ) (+ ) (- ) (+ ) (- ) + + (+ ) M ig ra n t = y e s (+ ) (+ ) - (- ) (+ ) (+ ) (+ ) (- ) (+ ) (+ ) (- ) (0 ) (- ) E d u ca ti o n ( IS C E D )* + + + + + (+ ) + + + + + + + N e t e q u iv al e n t in co m e * + + + (0 ) + 0 0 0 0 0 + (0 ) (+ ) R e ti re m e n t = y e s (- ) - (- ) - (- ) (- ) (+ ) - (- ) (+ ) (+ ) (- ) (- ) T ra n sf e r/ B e q u e st = y e s (+ ) + (+ ) + (+ ) + (+ ) + (- ) (+ ) (+ ) (+ ) + U rb an c o m m u n it y = y e s + + (- ) - (+ ) (+ ) (+ ) - + + + (- ) (+ ) C o n st an t + (+ ) (+ ) + (+ ) (+ ) + + + + + + (+ ) R ² 0 .1 8 0 .1 7 0 .1 2 0 .1 4 0 .0 9 0 .1 1 0 .1 1 0 .1 2 0 .1 4 0 .1 9 0 .1 6 0 .0 8 0 .1 1 N 1 ,1 1 4 1 ,0 2 4 5 2 8 1 ,6 0 8 5 5 2 9 3 2 1 ,3 4 9 1 ,0 8 7 1 ,1 5 3 1 ,7 9 2 1 ,4 3 7 1 ,1 3 0 1 ,1 2 1 * V al u e s ar e b as e d o n 5 s e ts o f im p u ta ti o n s + p o si ti v e e ff e ct ( p 0 .0 5 ); - n e g at iv e e ff e ct ( p 0 .0 5 ); ( + )/ (- ) - n o t si g n ifi c an t S o u rc e : S H A R E W av e 2 ( re le as e 2 .5 .0 ), u n w e ig h te d d at a, o w n c al cu la ti o n s • Kathrin Kolb, Nora Skopek, Hans-Peter Blossfeld1028 Education is not only important for the probability of home ownership, but also for the value of residential property. Higher education (except for Germany) as well as higher income signifi cantly increase the housing value in all countries studied. Retirement status has no impact on the property value in most countries. Transfers and bequests contribute to an increase in the value of residential property. Espe- cially in Belgium, Germany, Poland, the Netherlands and Sweden, inheritances and fi nancial gifts are important for the probability of being a homeowner as well as for the value of housing. Consistent with our expectations, living in an urban area has a positive effect on the housing value, particularly in Southern and Northern Europe. A result that needs to be explored in more detail is that living in an urban area has a signifi cant negative impact on the housing value in Belgium and the Netherlands. Is there a relationship between home ownership rates and housing values? To explore the relationship between home ownership rates and housing values on the country level, in a fi rst step, we plotted all countries in a two-dimensional coor- dinate system (Fig. 4). No clear relationship however becomes evident in the emerg- ing picture. Two lines divide the coordinate system in Figure 4: The mean housing rate over all countries (70.3 percent) divides the y-axis, and the mean housing value (€260,530) divides the x-axis. What we can see is that countries are evenly distrib- uted across all four quadrants. There are as many countries showing a negative relationship between home ownership rates and housing values (GR, BE, IT, DK, NL, CH), as countries showing a positive relation (PL, CZ, AT, SE, DE, FR, ES). Thus, in a next step we will carry out a multilevel regression to statistically test whether, controlling for individual characteristics of homeowners, the country con- text has a discrete impact on the housing value. The results are presented in Ta- ble 5. Model 0 (M0) shows that we have an intra-class correlation of 0.23, meaning that the households within countries are not independent from one another. The multilevel framework thus seems to be appropriate. The variance between coun- tries is 0.23; the variance between households is 0.77. If we introduce the home ownership rate into the model (M1), we can see that it has a negative but not signifi - cant20 impact on individual housing values. The variance between countries does not change at all from M0 to M1. In Model 2, we further include the country clusters. We can see that this explains a large share of the variance between countries. The Northern and Eastern European countries show signifi cantly lower mean housing values compared to Continental countries. Finally, we include our household-level factors (the demographic and socio-economic household characteristics) into Mod- el 3. This model is able to explain part of the variance between households. Overall, home ownership rates and housing values are negatively related to one another, also if one controls for the socio-economic composition of homeowners. Whereas 20 The small number of cases and thus the low number of degrees of freedom on the context level (countries) might explain why this effect does not become statistically signifi cant. The Two Dimensions of Housing Inequality in Europe • 1029 the home ownership rate itself cannot help to explain the differences in housing values, homeowners’ socio-economic composition contributes to explaining them, as does the welfare state context to a considerable degree. In our theoretical part, we argued that if we fi nd a negative relationship between home ownership rates and housing values the mechanism behind that relationship might be that homeowners are a selective group compared to the overall population in countries with low home ownership rates (like in Switzerland, Germany and Swe- den) and vice versa in countries with a high homeownership rate (like in Belgium, Greece and Spain). To test for the validity of this argument, we carried out a Heck- man selection test (Heckman 1979). The detailed results of this analysis are avail- able from the authors upon request. We found the group of homeowners to differ signifi cantly in their composition from the overall population (including homeown- ers, non-homeowners and owners of houses that are rented out) in seven countries; four of them with comparatively low home ownership rates – Denmark, Germany, Sweden and Poland – but three of them with rather high rates – Belgium, Greece and Spain. Overall, the composition of the group of homeowners does not seem to be a helpful explanation for the negative though not signifi cant relationship between home ownership rates and housing values. Summing up, the relationship between home ownership rates and housing values is far from being clear-cut. Fig. 4: Home ownership rates and housing values AT DE SE NL ES IT FR DK GR CH BE CZ PL 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0,00 50,00 100,00 150,00 200,00 250,00 300,00 350,00 400,00 450,00 500,00 H om e on w er sh ip r at e (in % ) Housing Value (in 1,000 Euros) II III I IV Values in 1,000 Euros, ppp-adjusted, based on 5 sets of imputations Source: SHARE Wave 2 (release 2.5.0), weighted data, own calculations • Kathrin Kolb, Nora Skopek, Hans-Peter Blossfeld1030 5 Conclusion and discussion In times of demographic ageing and less generous public pensions, home owner- ship will probably gain in importance for the fi nancial well-being of the elderly. To determine the socio-economic position of households, it is therefore important to consider not only income, but also the stock of wealth (and especially home owner- Tab. 5: Multilevel regression on log (housing value) with households (level 1) clustered in countries (level 2), robust standard errors M0 M1 M2 M3 Constant 5.07 5.18 5.53 3.49 Level 1 variables Age + Age² 0 Household size + Partnership = yes + Child(ren) = yes + Migrant = yes (+) Education (ISCED)* + Net equivalent income* + Retirement = yes (-) Transfer/Bequest = yes + Urban community = yes + Level 2 variables Home ownership rate (-) (-) (+) Continental Ref. Ref. North - - South (-) (-) East - - N (level 1) 14,827 14,827 14,827 14,827 N (level 2) 13 13 13 13 ICC 0.23 0.23 0.07 0.21 Variance components Variance between households 0.77 0.77 0.77 0.70 Variance between countries 0.23 0.23 0.05 0.20 * Values are based on 5 sets of imputations + positive effect (p0.05); - negative effect (p0.05); (+)/(-) - not signifi cant Source: SHARE Wave 2 (release 2.5.0), unweighted data, own calculations The Two Dimensions of Housing Inequality in Europe • 1031 ship as a central wealth component), as wealth is especially important for retirees’ socio-economic position (Modigliani/Brumberg 1954). Previous international comparative studies mainly focused on the distribution of home ownership across countries, while neglecting its value. The value of housing, however, signifi cantly determines the wealth position of households and therefore patterns of social inequality. Thus, we emphasized the importance of going beyond the approach of considering only access to home ownership by including the analy- sis of the value of residential property as we assume that housing inequality is a two-dimensional phenomenon. Our analyses were indeed able to show that the infl uence of socio-economic factors on the probability of being a homeowner, on the one hand, and the value of housing, on the other, can differ considerably within a country. For example, the effect of education and income is signifi cant for both di- mensions in most countries, whereas the infl uence of family-related characteristics varies: Having a partner seems to be especially important for being a homeowner, while household size mainly affects the value of housing. Especially the results on migration status are astonishing: The main obstacle for migrants appears to be ac- cess to home ownership. Once they obtained residential property, no difference in the housing value could be found between migrants and the total population of homeowners. Thus, in contrast to the USA and Israel, migration status only affects the fi rst dimension of housing inequality among elderly Europeans. In the second part of our work, we focused on the link between these two di- mensions of social inequality in housing. In our theoretical considerations, we were able to fi nd arguments for both situations: a negative and a positive relationship between home ownership rates and housing values. While our descriptive analysis did not show any clear-cut relationship between these two measures, the results of the multi-level regression revealed a negative though not signifi cant relationship between home ownership rates and individual housing values. We assumed the composition of homeowners to be a possible mechanism behind the negative rela- tionship. Our further analyses did not confi rm this assumption, however. In addition to the infl uence of socio-economic household characteristics, many other factors, such as the overall demographic and economic situation or cultural attitudes or patterns of behaviour towards home ownership, also play an important role in determining a country’s home ownership situation. Due to the diversity and complexity of the country-specifi c jurisdictions, considering these factors would have gone beyond the scope of this article. Especially for countries like Spain and France, whose real estate markets were hit by the fi nancial crisis in 2008 (Ball 2010), further analysis of the housing value before and after the fi nancial crisis would also be interesting to look at. One restriction of our study was that we had no retrospec- tive information on socio-economic household characteristics at the time of the transition to home ownership. Retrospective, longitudinal data would be necessary to trace back developments and analyse complex processes of property acquisition (Kurz/Blossfeld 2004). Nevertheless, a study carried out with longitudinal data for Western Germany by Davidov and Weick (2011) supports our central fi ndings on the infl uence of socio-economic factors on the probability of owning a home. Consider- ing the impact of the country context on home ownership rates and housing values, • Kathrin Kolb, Nora Skopek, Hans-Peter Blossfeld1032 it would be necessary to replicate our analyses with a larger number of countries and longitudinal data. This would be a more robust approach to examine causal relationships. Yet, by now, no such data is available. We therefore have to leave this task for future research. Summing up, despite some limitations, our article offers an innovative approach of an internationally comparative, two-dimensional analysis of housing inequalities. Acknowledgements We would like to thank the German-Israeli Foundation for Scientifi c Research and Development (G.I.F) for its fi nancial support. References Angelini, Viola; Brugiavini, Agar; Weber, Guglielmo 2009: Ageing and unused capac- ity in Europe: is there an early retirement trap? In: Economic Policy 59: 463-508 [doi: 10.1111/j.1468-0327.2009.00227.x]. Artle, Roland; Varaiya, Pravin 1978: Life cycle consumption and ownership. In: Journal of Economic Theory 18: 35-58. Attanasio, Orazio; Leicester, Andrew; Wakefi eld, Matthew 2011: Do house prices drive consumption growth? The coincident cycles of house prices and consumption in the UK. In: Journal of the European Economic Association 9: 399-435 [doi: 10.1111/j.1542- 4774.2011.01021.x]. Augustin, Baudelaire; Sanga, Dimitri 2002: Income and wealth. In: Perspectives: 15-19. Ball, Michael 2010: European housing review. Europäischer Wohnungsbericht. Zusam- menfassung. Brüssel. Bell, Bethany A.; Ferron, John M.; Kromrey, Jeffrey D. 2008: Cluster size in multilevel models: The impact of small level-1 units on point and interval estimates in two level models. Proceedings of the American Statistical Association, Social Statistics Section [CD-ROM], Alexandria, VA: American Statistical Association. Besley, Timothy; Mueller, Hannes 2012: Estimating the peace dividend: The impact of violence on house prices in Northern Ireland. In: The American Economic Association 102,2: 810-833 [doi: 10.1257/aer.102.2.810]. Blau, Peter M.; Duncan, Otis Dudley 1967: The American occupational structure. New York: Wiley. Börsch-Supan, Axel; Heiss, Florian; Ludwig, Alexander; Winter, Joachim 2003: Pension Reform, Capital Markets, and the Rate of Return. In: German Economic Review 4,2: 151-181[doi: 10.1111/1468-0475.00077]. Börsch-Supan, Axel; Alcser, Kirsten H.; Mannheim Research Institute for the Economics of Aging (Eds.) 2005: Health, ageing and retirement in Europe: First results from the survey of health, ageing and retirement in Europe. Mannheim: MEA . Bourdieu, Pierre 1998: Der Einzige und sein Eigenheim. Hamburg: VSA-Verlag. Brandolini, Andrea; Cannari, Luigi; D’Alessio, Giovanni; Faiella, Ivan 2004: Household wealth distribution in Italy in the 1990s. Temi di discussione (Economic working pa- pers) 530, Bank of Italy, Economic Research Department. The Two Dimensions of Housing Inequality in Europe • 1033 Buchholz, Sandra 2008: Die Flexibilisierung des Erwerbsverlaufs. Eine Analyse von Ein- stiegs- und Ausstiegsprozessen in Ost- und Westdeutschland. Wiesbaden: VS Verlag für Sozialwissenschaften. Castles, Francis G. 1998: The really big trade-odd: home-ownership and the welfare state in the new world and the old. In: ActaPolitica 33,1: 5-19. Charles, Kerwin Kofi ; Hurst, Erik 2002: The transition to home ownership and the black- white wealth gap. In: The Review of Economics and Statistics 84,2: 281-297 [doi: 10.1162/003465302317411532]. Chiteji, Ngina S.; Stafford, Frank P. 1999: Portfolio choices of parents and their children as young adults: Asset accumulation by African-American families. In: American Eco- nomic Review 89,2: 377-380. Chiuri, Maria Concetta; Jappelli, Tullio 2010: Do the elderly reduce housing equity? An international comparison. In: Journal of Population Economics 23: 643–663 [doi: 10.1007/s00148-008-0217-4]. Christelis, Dimitris 2011: Imputation of missing data in waves 1 and 2 of SHARE. [http:// www.share-project.org/t3/share/fi leadmin/pdf_documentation/Imputation_of_Miss- ing_Data_in_Waves_1_and_ 2_of_SHARE.pdf, 21.08.2011]. Conley, Dalton 2003: Wealth and poverty in America: A Reader. Malden, MA: Blackwell. Constant, Amelie; Roberts, Rowan; Zimmermann, Klaus F. 2007: Ethnic identity and immigrant home ownership. CEPR Discussion Papers, No.6490, Centre for Economic Policy Research, London. Davidov, Eldad; Weick, Stefan 2011: Transition to home ownership among immigrant groups and natives in West Germany, 1984-2008. In: Journal of Immigrant & Refugee Studies 9,4: 393-415 [doi: 10.1080/15562948.2011.616807]. DeWilde, Caroline; Raeymaeckers, Peter 2008: The trade-off between home-ownership and pensions: individual and institutional determinants of old-age poverty. In: Ageing & Society 28: 805-830 [doi: 10.1017/S0144686X08007277]. Dorbritz, Jürgen 2005: Kinderlosigkeit in Deutschland und Europa – Daten, Trends und Einstellungen. In: Zeitschrift für Bevölkerungswissenschaft 30,4: 359-408. Elsinga, Marja; De Decker, Pascal; Teller, Nóra; Toussaint, Janneke 2007: Home owner- ship beyond asset and security perceptions of housing related security and insecurity in eight European countries. Amsterdam: Housing and Urban Policy Studies. Esping-Andersen, Gøsta 1990: The three worlds of welfare capitalism. Princeton, NJ: Princeton University Press. Euroconstruct/ifo 2009: Wohneigentum in Europa. [http://de.statista.com/statistik/ daten/studie/155734 /umfrage/wohneigentumsquoten-in-europa, 20.01.2011]. Eurostat 2010: Erwerbstätigenquote in der EU27 fi el auf 64,6% im Jahr 2009. Pressemit- teilung. 117/2010 – 4. August 2010. Faller, Bernhard; Braun, Reiner; Hey, Timo; Pfeiffer, Ullrich 2001: Hemmnisse der Woh- neigentumsbildung. Forschungsbericht im Auftrag des Bundesamts für Bauwesen und Raumordnung. Bonn. Frey, William H. 2011: The Uneven Aging and ‘Younging’ of America: State and Metro- politan Trends in the 2010 Census. In: Brookings: State of Metropolitan Area Series 32. Grabka, Markus M.; Frick, Joachim R. 2007: Wealth inequality on the rise in Germany. In: DIW Berlin Weekly Report 5,10: 62-73. • Kathrin Kolb, Nora Skopek, Hans-Peter Blossfeld1034 Häußermann, Hartmut; Petrowsky, Werner 1990: Lebenszyklus, Arbeitslosigkeit und Hauseigentum. In: Bertls, Lothar; Herlyn, Ulfert (Eds.): Lebenslauf und Raumer- fahrung. Opladen: Leske + Budrich: 103-122. Häußermann, Hartmut; Siebel, Walter 2000: Soziologie des Wohnens. 2nd ed. Wein- heim/München: Juventa Verlag. Heckman, James 1979: Sample selection bias as a specifi cation error. In: Econometrica 47,1: 153-161. Horton, Hayward Derrick; Thomas, Melvin E. 1998: Race, class, and the family struc- ture: Differences in housing values for black and white homeowners. In: Sociological Inquiry 68,1: 114-136. Hox, Joop J. 2010: Multilevel analysis: Techniques and applications, Second Edition. Hove, East Sussex: Routledge. Kemeny, Jim 1981: The myth of home ownership. London: Routledge. Krivo, Lauren J.; Kaufman, Robert L. 2004: Housing and wealth inequality: racial-ethnic differences in home equity in the United States. In: Demography 41,3: 585-605 [doi: 10.1353/dem.2004.0023]. Kurz, Karin; Blossfeld, Hans-Peter 2004: Home ownership and social inequality in com- parative perspective. Stanford, Calif.: Stanford University Press. LBS Research 2004: Wohnungsmarkt. 57 Prozent der Mieter streben Wohneigentum an [http://www.lbs.de/bw/presse/infodienste/wohnungsmarkt/mieter, 24.11.2010]. Lefebure, Stijn; Mangeleer, Joris; Van den Bosch, Karel 2006: Elderly prosperity and home ownership in the European Union: New evidence from the SHARE data, paper to the 29th general conference of the international association for research in income and wealth. Joensuu, Finland. Lewin-Epstein, Noah; Semyonov, Moshe 2000: Migration, ethnicity and inequality in home ownership. In: Social Problems 47,3: 425-444. Lewin-Epstein, Noah; Elmelech, Yuval; Semyonov, Moshe 1997: Ethnic inequality in home ownership and the value of housing: The case of immigrants in Israel. In: Social Forces 75,4: 1439-1462. Li Mingche M.; Brown, H. James 1980: Micro-neighborhood externalities and hedonic housing prices. In: Land Economics 56,2: 125-141. Long, J. Scott 1997: Regression models for categorical and limited dependent variables. Advanced quantitative techniques in the social sciences number 7. Sage Publications: Thousand Oaks, CA. Maas, Cora J. M.; Hox, Joop J. 2005: Suffi cient sample sizes for multilevel modeling. In: Methodology 1,3: 86-92 [doi: 10.1027/1614-2241.1.3.86]. Mannheim Research Institute for Economics of Aging 2010: Share – guide to release 2.5.0: Waves 1 & 2. [http://www.share-project.org/t3/share/fi leadmin/pdf_documenta- tion/SHARE_release_guide.pdf, 06.01.2012]. Modigliani, Franco; Brumberg, Richard 1954: Utility analysis and the consumption func- tion: An interpretation of cross-section data. In: Kurihara, Kenneth K. (Eds.): Post- Keynesian economics. London: Routledge: 383-436. Mulder, Clara H. 2006: Home-ownership and family formation. In: Journal of Housing and the Built Environment 21: 281-298 [doi: 10.1007/s10901-006-9050-9]. Mulder, Clara H.; Wagner, Michael 1998: First-time home-ownership in the family life course: A West German-Dutch comparison. In: Urban Studies 35,4: 687-713. The Two Dimensions of Housing Inequality in Europe • 1035 Müller, Walter; Shavit, Yossi 1998: Bildung und Beruf im institutionellen Kontext – Eine vergleichende Studie in 13 Ländern. In: Zeitschrift für Erziehungswissenschaft 4: 501- 533. Myers, Samuel L.; Chung, Chanjin 1996: Racial differences in home ownership and home equity among preretirement-aged households. In: The Gerontologist 36,3: 350- 360. Paccagnella, Omar; Weber, Guglielmo 2005: Income imputation. In: Börsch-Supan, Axel et al. (Eds.): Health, ageing and retirement in Europe. First results from the survey of health, ageing and retirement in Europe. Mannheim: MEA: 357-358. Parcel, Toby L. 1982: Wealth accumulation of black and white men: The case of housing equity. In: Social Problems 30,2: 199-211. Poggio, Teresio 2006: Different patterns of home ownership in Europe. Workshop Home ownership in Europe: Policy and research issues: Delft, The Netherlands. Rabe-Hesketh, Sophia; Skrondal, Anders 2008: Multilevel and longitudinal modeling us- ing Stata, Second Edition. College Station, Texas: Stata Press. Riphahn, Regina T.; Serfl ing, Oliver 2005: Item non-response on income and wealth questions. In: Empirical Economics 30,2: 521-538 [doi: 10.1007/s00181-005-0247-7]. Rohwer, Götz 2012: Statistical Methods in sociological research of education. University of Bamberg [http://www.stat.rub.de/papers/dsre.pdf, 15.03.2012]. Rubin, Donald B. 1987: Multiple imputation for nonresponse in surveys. New York: Wiley. Skopek, Nora; Kolb, Kathrin; Buchholz, Sandra; Blossfeld, Hans-Peter 2012: Einkommen- sreich – vermögensarm? Die Zusammensetzung von Vermögen und die Bedeutung einzelner Vermögenskomponenten im europäischen Vergleich. In: Berliner Journal für Soziologie 22,2: 163-187 [doi: 10.1007/s11609-012-0185-0]. Scanlon, Kathleen; Whitehead, Christine 2004: International trends in housing tenure and mortgage fi nance. London: CML. Sierminska, Eva; Brandolini, Andrea; Smeeding, Timothy M. 2007: Cross-national com- parison of income and wealth status in retirement: First results from the Luxembourg Wealth Study (LWS). Chestnut Hill, MA: Center for Retirement Research at Boston College. Spilerman, Seymour 2000: Wealth and Stratifi cation Process. Annual Review of Sociol- ogy 26, 497-524. Swiaczny, Frank; Graze, Philip; Schlömer, Claus 2012: Spatial impacts of demographic change in Germany. In: Zeitschrift für Bevölkerungswissenschaft 33,2: 181-205 [doi: 10.1007/s12523-009-0009-2]. Szydlik, Marc; Schupp, Jürgen 2004: Wer erbt mehr? Erbschaften, Sozialstruktur und Alterssicherung. In: Kölner Zeitschrift für Soziologie und Sozialpsychologie 56,4: 609- 629 [doi: 10.1007/s11577-004-0106-0]. Tatsiramos, Konstantinos 2006: Residential mobility and housing adjustment of older households in Europe. IZA Discussion Paper Series, No. 2435. Tobin, James 1958: Estimation of Relationships for Limited Dependent Variables. In: Econometrica 26,1: 24–36. Venti, Steven F.; Wise, David A. 1989: Aging, moving, and housing wealth. In: Wise, David A. (Eds.): The economics of aging. Chicago: University of Chicago Press: 9-54. Venti, Steven F.; Wise, David A. 2000: Aging and housing equity. NBER Working Paper Series, No. 7882, Cambridge. • Kathrin Kolb, Nora Skopek, Hans-Peter Blossfeld1036 Wagner, Michael; Mulder, Clara H. 2000: Wohneigentum im Lebenslauf. Kohortendyna- mik, Familiengründung und sozioökonomische Ressourcen. In: Zeitschrift für Sozio- logie 29,1: 44-59. Wolff, Edward N.; Zacharias, Ajit; Caner, Asena 2005: Household wealth, public con- sumption and economic well-being in the United States. In: Cambridge Journal of Economics 29: 1073-1090 [doi: 10.2139/ssrn.447101]. A German translation of this reviewed and authors’ authorised original article by the Federal In- stitute for Population Research is available under the title “Zwei Dimensionen der Wohneigen- tumsungleichheit in Europa – Sind hohe Wohneigentumsquoten ein Indiz für niedrige Immobilien- werte?”, DOI 10.4232/10.CPoS-2013-22de or URN urn:nbn:de:bib-cpos-2013-22de7, at http://www. comparativepopulationstudies.de. Date of submission: 18.01.2012 Date of Acceptance: 17.10.2012 Dr. Kathrin Kolb (). Nora Skopek, Prof. Dr. Hans-Peter Blossfeld. Bamberg University, Department of Sociology, Chair of Sociology I. Bamberg, Germany. E-Mail: kathrin.kolb@uni-bamberg.de, nora.skopek@uni-bamberg.de, hans-peter.blossfeld@uni-bamberg.de URL: http://www.uni-bamberg.de The Two Dimensions of Housing Inequality in Europe • 1037 A p p e n d ix T a b . A 1: L o g is ti c re g re ss io n ( w it h r o b u st s ta n d a rd e rr o rs ) o n t h e c h an ce o f b e in g a h o m e o w n e r N o rt h e rn E u ro p e C o n ti n e n ta l E u ro p e D K S E A T B E C H D E F R N L A g e 0 .0 5 0 .1 5 -0 .0 3 -0 .0 3 0 .2 6 ** 0 .1 9 * 0 .0 5 -0 .0 7 A g e ² 0 .0 0 0 .0 0 0 .0 0 0 .0 0 0 .0 0 ** 0 .0 0 * 0 .0 0 0 .0 0 H o u se h o ld s iz e 0 .2 2 0 .2 8 * 0 .2 6 0 .0 9 0 .2 1 0 .2 6 * -0 .2 1 * 0 .0 0 P ar tn e rs h ip = y e s 1 .2 0 ** * 1 .1 6 ** * 0 .2 2 1 .2 8 ** * 0 .7 1 ** * 0 .5 6 ** 1 .3 1 ** * 1 .0 8 ** * C h ild (r e n ) = y e s -0 .2 8 0 .1 1 -0 .0 7 -0 .2 4 0 .4 2 * -0 .1 7 -0 .0 1 0 .1 9 M ig ra n t = y e s -0 .5 2 -0 .3 4 -0 .5 7 * -0 .4 9 ** -0 .6 1 ** -0 .5 4 ** * -0 .5 9 ** * -0 .9 0 ** * E d u ca ti o n ( IS C E D ) 0 .2 0 ** * 0 .0 4 0 .2 1 ** 0 .1 6 ** * 0 .1 9 ** 0 .1 9 ** 0 .2 1 ** * 0 .4 5 ** * N e t e q u iv al e n t in co m e 0 .0 2 0 .0 1 0 .0 2 ** 0 .0 1 0 .0 1 * 0 .0 2 * 0 .0 1 ** 0 .0 1 R e ti re m e n t = y e s 0 .2 4 -0 .0 7 -0 .0 1 0 .0 2 -0 .0 6 -0 .2 1 0 .2 1 -0 .1 9 T ra n sf e r/ B e q u e st = y e s 0 .6 1 ** * 0 .2 4 * 0 .7 6 ** * 0 .8 0 ** * 0 .6 5 ** * 0 .9 2 ** * 0 .8 3 ** * 1 .0 0 ** * U rb an c o m m u n it y = y e s -1 .1 2 ** * -1 .6 9 ** * -1 .5 0 ** * -0 .8 6 ** * -1 .3 6 ** * -0 .9 6 ** * -0 .7 9 ** * -0 .6 7 ** * C o n st an t -1 .9 7 -4 .7 2 1 .5 3 1 .1 2 -9 .9 8 ** -7 .9 6 ** -1 .8 0 2 .1 2 P se u d o -R ² 0 .2 1 0 .1 9 0 .1 5 0 .1 5 0 .1 8 0 .1 3 0 .1 5 0 .2 0 W al d C h i² 3 0 8 .4 4 3 2 7 .0 7 1 3 8 .4 0 2 4 8 .6 8 1 8 8 .2 4 2 0 2 .5 6 2 3 3 .5 9 3 1 3 .2 3 N 1 ,6 6 3 1 ,7 2 5 8 9 7 2 ,0 2 2 9 6 7 1 ,5 5 0 1 ,8 4 4 1 ,7 1 0 • Kathrin Kolb, Nora Skopek, Hans-Peter Blossfeld1038 T a b . A 1: co n ti n u a ti o n S o u th e rn E u ro p e E as te rn E u ro p e E S G R IT C Z P L A g e 0 .1 9 * 0 .0 1 0 .1 7 * 0 .2 5 ** 0 .1 3 A g e ² 0 .0 0 * 0 .0 0 0 .0 0 0 .0 0 ** 0 .0 0 H o u se h o ld s iz e -0 .1 4 -0 .1 5 0 .0 9 0 .0 6 -0 .0 4 P ar tn e rs h ip = y e s 1 .4 0 ** * 0 .7 6 ** * 0 .8 3 ** * 0 .4 5 ** * 0 .6 6 ** * C h ild (r e n ) = y e s 0 .5 5 -0 .0 2 0 .0 5 0 .0 0 -0 .0 7 M ig ra n t = y e s -1 .7 6 ** * -0 .5 7 -0 .7 8 * -0 .6 7 ** -0 .5 1 E d u ca ti o n ( IS C E D ) 0 .0 5 0 .0 6 0 .3 3 ** * 0 .0 1 0 .1 4 ** N e t e q u iv al e n t in co m e 0 .0 0 0 .0 1 0 .0 0 0 .0 1 0 .0 1 R e ti re m e n t = y e s 0 .0 1 0 .2 5 0 .4 9 ** -0 .2 9 0 .2 4 T ra n sf e r/ B e q u e st = y e s 0 .6 8 * 0 .6 4 ** * 0 .6 7 ** * 0 .5 4 ** * 0 .9 5 ** * U rb an c o m m u n it y = y e s -0 .4 4 * -0 .4 5 ** -0 .8 9 ** * -0 .5 7 ** * -0 .9 0 ** * C o n st an t -5 .2 4 2 .7 8 -6 .4 8 * -7 .3 7 ** -4 .1 0 P se u d o -R ² 0 .1 2 0 .0 5 0 .1 0 0 .0 5 0 .0 7 W al d C h i² 9 4 .7 4 7 7 .3 6 1 6 2 .5 8 1 0 4 .0 7 1 1 9 .3 6 N 1 ,2 7 9 2 ,0 8 3 1 ,7 8 6 1 ,7 2 2 1 ,6 9 7 A ll an al y se s b as e d o n 5 s e ts o f im p u ta ti o n s *p 0 .0 5 , ** p 0 .0 1 ,* ** p 0 .0 0 1 . N = 2 0 ,9 4 5 S o u rc e : S H A R E W av e 2 ( re le as e 2 .5 .0 ), u n w e ig h te d d at a, o w n c al cu la ti o n s The Two Dimensions of Housing Inequality in Europe • 1039 T a b . A 2 : L in e a r re g re ss io n ( w it h r o b u st s ta n d a rd e rr o rs ) o n lo g (h o u si n g v a lu e) N o rt h e rn E u ro p e C o n ti n e n ta l E u ro p e D K S E A T B E C H D E F R N L A g e -0 .0 1 0 .1 0 0 .0 9 0 .0 3 0 .0 9 0 .1 6 ** * 0 .0 4 0 .0 7 A g e r² 0 .0 0 0 .0 0 0 .0 0 0 .0 0 0 .0 0 0 .0 0 ** * 0 .0 0 0 .0 0 H o u se h o ld s iz e 0 .0 9 0 .0 7 0 .0 2 0 .0 1 0 .0 5 0 .1 0 ** * 0 .0 7 0 .0 1 P ar tn e rs h ip = y e s 0 .1 5 * 0 .1 0 0 .0 4 0 .1 8 ** * 0 .0 7 0 .0 0 0 .0 5 0 .2 5 ** * C h ild (r e n ) = y e s -0 .1 0 0 .0 2 0 .0 9 0 .0 7 0 .1 2 -0 .0 4 0 .0 7 -0 .0 7 M ig ra n t = y e s 0 .1 0 0 .0 6 -0 .2 6 * -0 .0 4 0 .0 5 0 .0 2 0 .0 9 -0 .0 5 E d u ca ti o n ( IS C E D ) 0 .1 4 ** * 0 .1 4 ** * 0 .1 3 ** * 0 .1 0 ** * 0 .0 7 * 0 .0 6 0 .1 0 ** * 0 .0 7 ** * N e t e q u iv al e n t in co m e 0 .0 1 ** * 0 .0 1 * 0 .0 1 ** * 0 .0 0 0 .0 0 ** 0 .0 0 ** 0 .0 0 ** 0 .0 0 * R e ti re m e n t = y e s -0 .0 2 -0 .1 7 * -0 .0 8 -0 .1 0 ** -0 .0 5 -0 .0 4 0 .0 8 -0 .1 1 * T ra n sf e r/ B e q u e st = y e s 0 .0 1 0 .2 2 ** 0 .0 0 0 .1 3 ** * 0 .1 0 0 .2 6 ** * 0 .0 9 0 .1 0 * U rb an c o m m u n it y = y e s 0 .2 1 ** * 0 .3 3 ** * -0 .1 0 -0 .0 9 ** 0 .1 3 0 .0 3 0 .0 4 -0 .1 7 ** * C o n st an t 4 .4 3 ** * 0 .3 9 1 .5 2 3 .7 1 ** * 2 .3 8 -0 .5 5 3 .3 6 * 2 .9 0 ** * R ² 0 .1 5 0 .1 3 0 .1 2 0 .1 4 0 .0 9 0 .0 9 0 .0 9 0 .0 7 N 1 ,1 1 4 1 ,0 2 4 5 2 8 1 ,6 0 8 5 5 2 9 3 2 1 ,3 4 9 1 ,0 8 7 • Kathrin Kolb, Nora Skopek, Hans-Peter Blossfeld1040 T a b . A 2 : co n ti n u a ti o n S o u th e rn E u ro p e E as te rn E u ro p e E S G R IT C Z P L A g e -0 .0 4 0 .0 2 0 .0 2 0 .0 1 0 .0 2 A g e ² 0 .0 0 0 .0 0 0 .0 0 0 .0 0 0 .0 0 H o u se h o ld s iz e 0 .0 7 * 0 .1 1 ** * 0 .0 4 0 .0 5 * 0 .0 8 ** * P ar tn e rs h ip = y e s 0 .0 4 0 .0 1 -0 .0 6 0 .1 9 ** 0 .1 3 C h ild (r e n ) = y e s 0 .0 3 -0 .0 1 0 .2 0 ** 0 .2 4 ** 0 .1 9 M ig ra n t = y e s 0 .1 8 0 .1 7 -0 .1 5 0 .0 0 -0 .0 4 E d u ca ti o n ( IS C E D ) 0 .1 1 ** * 0 .1 2 ** * 0 .1 3 ** 0 .1 0 ** 0 .1 3 ** * N e t e q u iv al e n t in co m e 0 .0 0 ** 0 .0 0 ** * 0 .0 1 ** * 0 .0 0 0 .0 1 R e ti re m e n t = y e s -0 .0 2 0 .0 2 0 .0 9 -0 .0 2 -0 .0 6 T ra n sf e r/ B e q u e st = y e s -0 .0 8 0 .0 5 0 .0 6 0 .0 7 0 .2 2 ** U rb an c o m m u n it y = y e s 0 .3 9 ** * 0 .2 3 ** * 0 .2 8 ** * -0 .0 7 0 .0 4 C o n st an t 6 .1 8 ** * 3 .5 9 ** * 4 .0 5 ** 3 .5 9 ** 2 .4 3 R ² 0 .0 8 0 .1 9 0 .0 7 0 .0 8 0 .1 1 N 1 ,1 5 3 1 ,7 9 2 1 ,4 3 7 1 ,1 3 0 1 ,1 2 1 A ll an al y se s b as e d o n 5 s e ts o f im p u ta ti o n s *p 0 .0 5 , ** p 0 .0 1 ,* ** p 0 .0 0 1 . N = 1 4 ,8 2 7 S o u rc e : S H A R E W av e 2 ( re le as e 2 .5 .0 ) u n w e ig h te d d at a, o w n c al cu la ti o n s © Federal Institute for Population Research 2013 – All rights reserved Published by / Herausgegeben von Prof. Dr. Norbert F. Schneider Federal Institute for Population Research D-65180 Wiesbaden / Germany Managing Editor / Verantwortlicher Redakteur Frank Swiaczny Assistant Managing Editor / Stellvertretende Redakteurin Katrin Schiefer Language & Copy Editor (English) / Lektorat & Übersetzungen (englisch) Amelie Franke Copy Editor (German) / Lektorat (deutsch) Dr. Evelyn Grünheid Layout / Satz Beatriz Feiler-Fuchs E-mail: cpos@bib.bund.de Scientifi c Advisory Board / Wissenschaftlicher Beirat Jürgen Dorbritz (Wiesbaden) Paul Gans (Mannheim) Johannes Huinink (Bremen) Marc Luy (Wien) Clara H. Mulder (Groningen) Notburga Ott (Bochum) Peter Preisendörfer (Mainz) Board of Reviewers / Gutachterbeirat Martin Abraham (Erlangen) Laura Bernardi (Lausanne) Hansjörg Bucher (Bonn) Claudia Diehl (Konstanz) Andreas Diekmann (Zürich) Gabriele Doblhammer-Reiter (Rostock) Henriette Engelhardt-Wölfl er (Bamberg) E.-Jürgen Flöthmann (Bielefeld) Alexia Fürnkranz-Prskawetz (Wien) Beat Fux (Zürich) Joshua Goldstein (Rostock) Karsten Hank (Köln) Sonja Haug (Regensburg) Franz-Josef Kemper (Berlin) † Michaela Kreyenfeld (Rostock) Aart C. Liefbroer (Den Haag) Kurt Lüscher (Konstanz) Dimiter Philipov (Wien) Tomáš Sobotka (Wien) Heike Trappe (Rostock) Comparative Population Studies – Zeitschrift für Bevölkerungswissenschaft www.comparativepopulationstudies.de ISSN: 1869-8980 (Print) – 1869-8999 (Internet)