A West German-Dutch Comparison

1 downloads 0 Views 325KB Size Report
This trend is stronger in the Netherlands than in. Germany .... Dutch retrospective life-course data. East ...... regression equations as dummies. Apart from.
Urban Studies http://usj.sagepub.com/

First-time Home-ownership in the Family Life Course: A West German-Dutch Comparison Clara H. Mulder and Michael Wagner Urban Stud 1998 35: 687 DOI: 10.1080/0042098984709 The online version of this article can be found at: http://usj.sagepub.com/content/35/4/687

Published by: http://www.sagepublications.com

On behalf of: Urban Studies Journal Foundation

Additional services and information for Urban Studies can be found at: Email Alerts: http://usj.sagepub.com/cgi/alerts Subscriptions: http://usj.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://usj.sagepub.com/content/35/4/687.refs.html

>> Version of Record - Apr 1, 1998 What is This?

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

Urban Studies, Vol. 35, No. 4, 687± 713, 1998

First-tim e Hom e-ownership in the Fam ily Life Course: A W est Germ an± Dutch Com parison Clara H. Mulder and M ichael W agner [Paper received in ® nal form, April 1997]

Su m m ary. The tran sition to ® rst-tim e hom e-ow nersh ip is m ad e at increasin gly you nger ages in both W est Germ any and the Netherlan ds. This tren d is stron ger in the Netherlan ds th an in G erm any, h ow ever. There are also m ark ed differen ces betw een the two cou ntries in the exten t to which ® rst-tim e hom e-ow nersh ip is con nected with even ts in th e fam ily life cou rse (m arriage and childbirth ) and the availab ility of resou rces from th e paren tal fam ily. These differen ces can be understood in term s of differen ces in hou se prices, housing policy (subsidies and other regu lation s) and other differen ces in the legal and ® nancial system s.

1. Introduction Becoming a hom e-owner is a crucial step in the life course. The effect of this step on people’ s housing situation, accumulation of wealth and disposable incom e can hardly be overstated. Through becom ing a hom eowner, people acquire a maxim um of control over their housing situation and independence of landlords. Not surprisingly, hom eownership is a highly desirable good in W estern societies. Given the im portance of hom e-ownership, scientists of various disciplines and traditions have made it a topic of research. This article is part of a tradition that studies the transition to (® rst tim e) hom e-ownership in the life course. We pay particular attention to the connections between various life-course trajectories and hom e-ownership. Recent contributions to this line of research mainly com e

from geographers, dem ographers and sociologists (see, for example, Henretta, 1987; Courgeau and LelieÁvre, 1992; Clark et al., 1994). Because of the problem s associated with giving a proper longitudinal de® nition of a househo ld (Duncan and Hill, 1985), in life-course research individuals are the units of observation alm ost by de® nition. But it should be acknowledged that the decision to becom e a hom e-owner is in most cases a joint decision of more than one household member, so the factors in¯ uencing this decision also pertain to the level of the household. W hether som e will becom e a hom e-owner, and if so, at what stage of the life course, depends strongly on househo ld members’ preferences and resources. But perhaps of equal im portance are the macro conditions

Clara H. Mulder is in the Urban Research Centre Utrecht, Faculty of Geographical Sciences, Utrecht University, P . O. B ox 80.115, NL-3508 TC Utrecht, The Netherlands. Fax: 1 31-30-2540604. E-mail: [email protected] .nl. Michael Wagner is in the F orschungsinstitut fuÈ r Soziologie, UniversitaÈ t zu KoÈ ln, Greinstrasse 2, 50939 KoÈ ln, BRD, Germany. Fax: (0221) 470 5180. E-mail: mwagner@w iso.uni-koeln.de. The research of M ichael Wagner w as done while he was employed at the M ax P lanck Institute for Human Development and E ducation, B erlin. The research of Clara Mulder w as made possible by a fellowship of the R oyal Netherlands A cademy of A rts and Sciences. She thanks the Max P lanck Institute in Berlin for kindly allow ing her a research stay. Thanks are also due to Monika Albin for preparing the German data, Petra Spengemann for preparing the graphs and Tak Wing Chan for correcting the English. Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

0042-0980/98/040687-27 $7.00

Ó

1998 T he E ditors of Urban Studies

688

CL ARA H . M UL DE R AN D M ICH AE L W A GN ER

facing the househo ld: econom ic, housing market and political conditions. In order to explain the transition into hom e-ownership, it is therefore im portant to try to take into account both micro and contextual factors. In recent studies of the transition into hom eownership, efforts have been made to include contextual factorsÐ by studyin g either changes through historical tim e (Mulder and Hooim eijer, 1995); or differences between countries (Clark et al., 1997) and/or regions (Loikkan en, 1992); or a com bination of temporal and spatial factors (Clark et al., 1994; Deurloo et al., 1994). In this article, we concentrate on the connection between the transition to hom eownership and dynam ics in the family life course in two countries. Our paper is much more detailed than previous studies. W e take into account the roles of both family status and changes in family status. T his gives us the opportu nity to assess both the resources and constraints arising from family status over a long period and the needs for hom eownership arising from events in the family life course. Moreover, we also take into account resources from the family of origin of the prospective hom e-ow ner. W e study the transition to hom eownership in two countries that are organised econom ically, socially and politically in a similar way: W est Germany and the Netherlands. But, astonishingly, these two countries have experienced quite different trends in hom e-ow nership rates. Whereas the Netherlands has witnessed a strong grow th in hom e-ownership rates since the 1950s and has by now almost caught up with other European countries where rates were higher in the 1950s, W est Germany has lagged behind in the growth of ow nership rates. Contrary to the situation in the 1950s and 1960s, it now has a smaller percentage of owner-occupiers than the Netherlands. This intrigu ing difference betw een the two countries makes a com parison interestingÐ arguably more so than a com parison betw een countries which are more different geogra phically, politically and econom ically.

Because the econom ies of the tw o countries are so closely connectedÐ the Dutch econom y mirrors its m uch larger neighbo ur Germany in many respectsÐ it is unlikely that the difference in the evolution of hom eownership in the two countries is caused by different econom ic evolution in general. But there is another differential developm ent: the structure of the rental markets in the tw o countries and the changes therein in the past few decades. Both countries had similar tenure com position s in 1950, with private rental being the largest sector. The private rental sector has retained its prim e positio n in Germany, but it has becom e marginalised in the Netherlands. This has major im plications for the extent to which the rental sector provide s a high-qu ality alternative for owneroccupation. W e ® rst investigate the changes in context in the two countries. Next, we wish to ask to what extent the tim ing of the transition to ® rst-tim e hom e-ownership and the lifecourse factors underlying it differ between the two countries, and how we can relate the observed differences to contextual differences. For the analyses, we use W est German and Dutch retrospective life-course data. East Germany was left out of the study because of its com pletely different political context until 1990. T he samples include people born between 1920 and 1960, questioned in the early 1990s (Netherlands) and in several rounds in a period stretching from 1981 to 1989 (Germany). Section 2 gives the theoretical background s of this paper. Section 3 sketches the context in which the transition to hom eownership takes place in the tw o countries: the housing market, econom ic and policy context. Section 4 describes the data and methods employed. Section 5 reports som e descriptive analyses of the tim ing of ® rsttim e hom e-ownership, both in relation to age and to events and status in the family life course. In section 6, we report the results of multivariate life-course analyses. The paper concludes with a discussion.

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

FIRS T-T IM E HO ME -OW N ERSHIP

2. Theoretical Backgrou nds W e subscribe to Clark et al.’ s (1994) view that the life-course perspective is fruitful for com bining both econom ic and dem ographic variables in the analysis of the transition to hom e-ownership. We want to take their argument a bit further by applyin g the life-course perspective to the conventional econom ists’ view that hom e-ownership is in¯ uenced by preference, resources and the relative costs of owning versus renting (Megbolug be and Linnem an, 1993). As is explicit in most of the econom ic literature and im plicit in most of the lifecourse literature, we conceptualise the decision to becom e a hom e-owner as a (rational) choice between two alternatives: remaining in the rental sector (or in the parents’ hom e) or becom ing a hom e-ow ner. Becom ing a hom e-owner is preferred when the balance betw een the bene® ts and costs of owning exceeds that of not owning. W hether a preference for owning can be realised depends on the availability of resources. It can be argued that the balance between bene® ts and cost of hom e-ownership differs betw een individ ualsÐ and the househo ld they are part of!Ð and that it changes over the life course. W e will use the term `lifecourse differentiation’ to denote this differentiation betw een individ uals and stages in the life course. The availability of resources is also subject to life-course differentiation. Moreover, the bene® ts and cost of owning relative to renting change according to temporal and spatial contexts. This is also true for the availability of resources.

The Bene® ts and Costs of Home-ownership Bene® ts. The bene® ts of hom e-ownership are threefold. First, the quality of ow neroccupied housing is on average better than that of rental accommodation (Megbolu gbe and Linnem an, 1993). Owner-occupied hom es are larger than rented dwellings, have more room s, are more often of the singlefamily type , and more often include amenities such as gardens and garages. For

689

instance in the late 1980s, among owneroccupied dw ellings in Germany and the Netherlands below 10 per cent had 3 room s or less, whereas for rented dwellings the corresponding ® gure was 43 per cent for Germany and 49 per cent for the Netherlands. Secondly, owning a hom e helps in the accumulation of wealth (Megbolu gbe and Linnem an, 1993). Rather than just paying for the hom e, a hom e-ow ner also makes an investm ent which generally keeps its value and can be transferred betw een generations. Thirdly, there are less tangible aspects of hom e-ow nership that make it desirable for many people. With hom e-ownership, people gain control over their own housing situation. It is said to provide a basic form of security (Saunders, 1990; Megbolugb e and Linnem an, 1993). It is, among other things, an im portant status good. As shown by Mayer (1973) , hom e-ownership, like high incom e or a high level of consum ption, is seen by many people as a sym bol of achievement. In short, becom ing a hom e-owner is, in W estern societies, a rather im portant goal in many people’ s lives (Michelson, 1977; Lassarre, 1986; Saunders, 1990). Widespread private hom eownership also tends to be view ed as positive by national governments (Forrest, 1983). This is of course not to say that there is som ething like a universal or `natural’ preference for hom e-ow nership. Costs. There are ® nancial and non-® nancial costs of hom e-ownership. The ® nancial costs are those of the hom e and the land it is built on, the ® nancing costs (except if the hom e is transferred by heritage or gift) and the costs of maintenance, insurance and taxes. Because of the higher quality in the owneroccupied sector, ow ner-occupied dwellings are on average more expensive than rented dwellings. Particularly in the ® rst few years after buying a hom e, the cost of being a hom e-ow ner is often higher than renting even with com parable housing quality. Because of mortgage obligations, becom ing a hom e-ow ner is simply not an option for people with low or unstable incom es. A non® nancial cost is the risk one takes by

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

690

CL ARA H . M UL DE R AN D M ICH AE L W A GN ER

becom ing a hom e-owner. Unlike other moves, moves into hom e-ownership are accom panied by a long-term ® nancial com mitment. Investm ents in owner-occupied housing are not easily undone . Because of high transaction cost, to m ention only one reason, it is highly undesirable to sell a hom e very soon after it was bought. Research has consistently show n that hom e-ow ners are reluctant to move (see, for example, Pickvance, 1974; Courgeau, 1985; W agner, 1989). The wish to buy a hom e, therefore, is likely to arise only among people who do not foresee the need to move again soon. Life-course Differentiation in the Bene® ts and Costs of Home-ownership One could argue that becom ing a hom eowner som e tim e in one’ s life is a preference most people share. But the desire of hom eownership is by no means universal to all people and all stages in the life course. Because owner-occupied housing is often of better quality than rented housing , and owner-occupied housing tends to be more expensive, the desire to move into owner-occupation is often triggered by needs arising from the househo ld or housing situation. Therefore, the preference to becom e a hom eowner is more prevalent among couple s or families than among singles. Moreover, the risk of moving again soon is highest for young singles. They are at the beginning of their occupational and fam ily life courses, and are often uncertain where and with whom they will live even in the near future. For couples and families, the risk of moving again soon is arguably smaller. As postulated by Clark et al. (1994) , hom e-ownership requires a certain stability, not only in terms of incom e, but also in family situation. The ® nancial com mitm ent of hom e-ow nership is apparently strongly connected with com mitments in the family life course: marriage and having children. Com mitm ents are, however, postpon ed by those people still shaping their lives: the young, particularly when they are single, and those in education (Mulder and Manting, 1994).

A low inclination among single s to buy a hom e was found earlier by Kendig (1984) and Clark et al. (1994) . A strong connection between becom ing a hom e-ow ner and getting married was found by Mulder and Manting (1994) . Deurloo et al. (1994) found that the recent birth of a ® rst child positively in¯ uenced the transition to hom e-ow nership in the US. The need for more space might be stronger after the birth of the second child, rather than the ® rst. Furthermore, because hom e-ow nership is arguably often pursued before, rather than after, the birth of a child (Courgeau and LelieÁvre, 1992), im minent childbirth also needs to be taken into account, as well as childbirth itself. From the above, we derive the follow ing hypothe ses about the conne ction between family life-course dyna mics and becom ing a hom e-ow ner. Singles supposedly have a very low probability of becom ing a hom e-owner, even after controlling for socioeconom ic status. W e further expect a connection between hom e-ownership and the events of marriage and childbirth.

Overcoming the Costs of Home-ownership: Life-course Differentiation of Resources Hom e-ownership involve s the inve stment of a large sum of money; inde ed, this is likely to be the largest single investm ent item for most private households. W ith regard to resources, therefore, incom e and wealth are obviously of great im portance (Doling, 1976; Clark et al., 1994; Deurloo et al., 1994). Incom e and wealth are prim arily gained by the occupational careers of house hold members. W hat is relevant is not only the current occupation and the incom e connected with it, but also the occupational history in which savings were made and job experience was accumulated. Future incom e grow th over the life course is also related to occupational history. Because the majority of hom es are bought on mortgage, we need to consider not only the level of incom e, but also its stability. Uncertainty of incom e streams makes people

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

FIRS T-T IM E HO ME -OW N ERSHIP

less inclined to buy a hom e (Haurin and Gill, 1987). W ithin couples, the incom e of the female partner is often view ed as transitory: for both Germany and the Netherlands, it is still true that females often reduce their labour force participation when they have children. Consequently, although the incom e of the female partner helps to make the transition to hom e-ownership (Myers, 1985), we expect current employm ent status to be less im portant for females than males. Expected future incom e grow th, on the other hand, can be a reason for the bank to provide a higher mortgage loan than it would do otherwise. Therefore, among people who are at the same incom e level, one might expect that the better educated will ® nd it easier to obtain a mortgage. The self-employed are by de® nition likely to ow n a farm or ® rm, which increases their probability of owning their hom es. They are also likely to have strong local ties and a corresponding ly low geographical mobility Ð for example, because of a local clienteÁ le. Resources do not necessarily com e from the household itself. They may also be provided by others, often the parents of one of the household members. Parental contributions may com e as ® nancial loans or gifts. Parental wealth is arguably more im portant than incom e, but obviously the two are closely connected. As shown by Lauterbach and LuÈ scher (1996) , wealth tends to accumulate over generations: those with high socioeconom ic status are more likely to inherit or to receive gifts from their parents than others. Even the hom e itself might be transferred from the parents by inhe ritance or gift. Inheritance of a hom e follow ing the death of both parents might not be very com mon in our data, because it usually takes place only after the age of 40 (Lauterbach and LuÈ scher, 1996). But parents who own their hom es are in a much better position to help their children ® nancially or to buy a hom e for them . If they are outrigh t owners, they have low housing costs. They can even take a new mortgage on their own hom e to release wealth for transfer to their children. Further-

691

more, the wish to becom e a hom e-ow ner is probably stronger for those people whose parents ow ned a hom e: children are said to desire a socioeconom ic status at least equal to that of their parents (see, for example, Easterlin, 1980; Henretta, 1984). Lauterbach and LuÈ scher (1996) argue that people attach a sym bolic value to living in a hom e that was previou sly owned by their parents. A positive effect of parental hom e-ow nership was found for the US by Henretta (1987) . T o conclude, the transition to hom eownership is hypothe sised to be positively connected with socioeconom ic status, work experience, higher education, self-employment and parental resources (including hom e-ow nership by the parents). Home-ownership and Context Factors Through supply and prices of owneroccupied versus rented dw ellings and variation in ® nancing cost, the spatial and temporal context has a direct in¯ uence on the relative cost and bene® ts of both alternatives. The perceived desirability and risk of ownership may also be context-speci® c. Finally, the context also in¯ uences the distribution of resources in the populat ion and the amount of resources needed to becom e a hom e-ow ner. Supply and prices of ow ner-occupied and rental housing opport unities are determined by the housing market and, mostly indirectly, the land market. The land market is unique in that land is available only in ® xed amounts and is therefore scarce almost by de® nition. Housing also has som e unique attributes, among which are a long life-span and a ® xed location, which im plies lim ited adaptability. The com positio n of local stocks, therefore, to a great extent re¯ ects the local history of dwelling construction and housing policy (Clark and Dieleman, 1996). There is not one housing market, but there are sub-m arkets according to sector and location (Bourne, 1981). Supply and prices of both rented and owner-occupied dw ellings are not only im portant as such, but also in relation to quality. A crucial question is whether renting

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

692

CL ARA H . M UL DE R AN D M ICH AE L W A GN ER

is a realistic alternative to owning in the higher-quality segm ents of the housing market. In relation to differences in housing stock com positio n betw een countries, and also to the relative costs of owning versus renting, national differences in housing policy and ® nancial systems appear to be of major im portance (see, for example, Lundqvi st, 1992; Pow er, 1993). Governm ents can in¯ uence the rental market through new construction of social housing , rent controls and rent subsidies. They can in¯ uence the ow neroccupied market through tax regulations, mortgage guarantees, subsidies and savings programmes. BoÈ rsch-Supan (1993) claims that the differences in propor tions of hom eowners betw een the US, W est Germany and Japan are to a large degree generated by housing policies, particularly by hom e-ownership subsidies and rental housing regulations. W ithin countries, the degree of urbanisation is often a major factor differentiating betw een rental markets and markets dom inated by owner-occupancy. W hereas in small villages supply consists prim arily of ow neroccupied dwellings, the majority of dwellings in large cities are rented; this relationshi p holds in W est Germany and the Netherlands, but also in the US (Clark and Dieleman, 1996). L and is particularly scarce and expensive in cities and especially in their centres (Alonso, 1964). The more extensive types of land use, such as single-family houses, are therefore hardly found in city centres. Rather, cities contain multi-family structures, which are mainly built in the rental sector (Bourne, 1981). People who want to build hom es them selves (or have them built), will hardly ® nd opportu nities to do so in cities, but more so in the countryside. The cost of ® nancing the purchase of a hom e varies with mortgage interest rates, in¯ ation and the relationship betw een the tw o. W ith high in¯ ation, investm ent in hom e-ownership may be used as a protection against future price rises (Deurloo et al., 1994). A com bination of high mortgage rates

and low in¯ ation leads to high ® nancing costs. T he perceived bene® ts of hom e-ow nership might differ betw een countries. From detailed analyses of tenure choice and housing dem and in the US and W est Germany, BoÈ rsch-Supan (1985) ® nds som e evidence suggesting that there is an `American dream’ for hom e-ownership which is not shared, or not as strongly, by W est Germans. In societies with widespread hom e-ownership, the norm to becom e a hom e-owner might be stronger and the efforts made to becom e one might be greater. The perceived risk of hom e-ow nership may also differ through space and tim e. Econom ic crisis and uncertainty might make people hesitant to buy. Besides the relative cost and bene® ts of owning versus renting, contextu al factors also in¯ uence the am ount of resources available to the household (throug h real incom es) and the amount of resources needed to becom e a hom e-ow ner. Individ ual wealth and incom e are particularly im portant where prices are high, banks are reluctant to provide mortgage loans or require a large downpaym ent, and gove rnm ent subsidies are not available. Family suppor t is im portant when it is both needed and available, and when taxes on inheritance and gifts are not prohibitive ly high. Hypotheses about the in¯ uence of contextual factors on hom e-ow nership are derived after a discussion of the relevant contextual factors in the tw o countries. 3. Hom e-ownership in W est Germany and the Netherlands: The Context Hom e-ownership has been on the increase in the W estern world since W orld W ar II, in Europe (Martens, 1985) as in Australia (Neutze and Kendig, 1991) and the US, at least up to 1980 (Chevan, 1989; after 1980 it declined in the US, see Megbolugb e and Linnem an, 1993). W est Germany and the Netherlands conform to the general pattern of rising hom e-ownership. But the growth in the propor tion of hom e-ow ners has developed quite differently in the past few decades

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

FIRS T-T IM E HO ME -OW N ERSHIP

in the two countries. In 1956 the propor tion of owner-occupied dwellings was higher in W est Germany than in the Netherlands, at about 38 and 29 per cent respectively. Since then, the propor tion of owner-occupiers has risen considerably in the Netherlands, but ® rst stagnated and then rose only slightly in W est Germany, resulting in equal percentages in both countries around 1970 (about 35 per cent) and a higher percentage in the Netherlands afterwards (see T able 1). W est German and Dutch Housing M arkets The housing markets of W est Germany and the Netherlands are typical for W estern Europe when com pared with for instance the US, in the sense that the propor tion of hom eowners is lower and government intervention is stronger. But there are signi® cant differences betw een the tw o countries. From the 1950s on, W est Germany’ s housing market has had about equal shares of private rent and owner-occupation, and a rather small share of non-pr o® t (social) rent. More than their Dutch counterp arts, German private landlords are allow ed deductions from their tax incom e for the cost of ow nership and im provem ent of rented dw ellings (Van W eesep and Van Velzen, 1994). Moreover, they have more legal opport unities for rent increases follow ing investm ents in the dwelling. The quality of the private rental stock is relatively high. In the same period, the Dutch housing market developed from a predom inantly private-renter market into a market dom inated by non-pr o® t rent and owner-occupancy alike; the private rental sector has becom e almost marginalised (Feddes, 1995). Although the housing stock more than doubled between 1950 and the mid 1980s, the private rental stock declined even in absolute num bers. The private rental stock now consists of tw o parts of about equal size: an old, cheap, low-quality part owned by private persons and small ® rms; and a newer, more expensive, higher-quality part owned by large investors (such as pension funds). The social rental sector, although by no means stigmatised as it is in the US, is

693

primarily inhabited by the lower-income groups in Germany. The Dutch social rental sector offers housing of reasonable size and quality to the lower- and middle-income groups alike. However, in the Netherlands as in Germany, waiting lists for the social sector are often long and the higher-income groups are not allowed to move into it. Average prices of plots of developed land to be used for housing are higher in Germany than in the Netherlands (Table 1). This is not so much due to differences in prices per square metre (Dieterich et al., 1993; Needham et al., 1993) as to the difference in sizes of plots of land used: 548 sq m in Germany (single-family houses in 1989; Bunde sministerium , 1990) as against only 380 sq m in the Netherlands (unsub sidised single-family houses in 1990; Van W oerkom , 1994). Note that listed plot prices are not exactly com parable betw een Germany and the Netherlands, because the dwelling types are different. The same is true for construction costs and house prices. The differences are probably som ewhat overestim ated. Nevertheless, it can be said that construction costs are also conside rably higher in Germany (Table 1). One reason for this is that legal quality requirements for buildings are much higher than in the Netherlands. All in all, the price of a house is much higher in Germany than in the Netherlands, not only in absolute ® gures but also in relation to househo ld incom e (Table 1). Furtherm ore, house prices have risen faster in Germany than in the Netherlands, even though in¯ ation was on average lower. Rent prices have risen in both countries, but the increase has been faster in the Netherlands since 1980. Housing Policy in W est Germany and the Netherlands The aims of housing policy are rather similar in W est Germany and the Netherlands, as are the types of instrum ent used to achieve these aims. T he amount of public spending on housing differs, however, as does the im plem entation of the instrum ents and the extent to which policy has changed during the post-w ar period.

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

39.3

157 469

b

53 270

51 273

52 038 50 234

48 937 43 979 49 124

36 990 47 751 49 498

37 416

43 104

89 746

74 627

68 837 68 194

53 719

43 426 58 866

32 167 39 243 45 034

25 027 28 011

Netherlands

336 900

319 900

280 900 298 700

258 100 261 700 271 300

247 100 247 800 249 400

255 100

252 900

235 800

215 200

180 100 198 800

169 800

157 200 162 500

121 500 136 500 152 300

93 100 106 900

W est Germ any

147 370

109 220

161 550

94 130

50 160

Netherlands

Construction cost c

416 309

393 357

360 395 365 369

284 428 289 955 324 102

308 214 287 241 291 515

318 793

321 818

304 558

272 666

220 483 250 370

221 745

222 000 216 250

170 555 195 789 202 142

138 421 144 836

W est Germany

212 400

194 800

174 494 180 400

153 940 160 092 171 570

139 578 140 094 146 974

142 072

138 030

153 424

171 054

198 746 187 222

184 126

102 598 131 838

84 194 89 354 98 126

76 884 78 346

Netherlands

House price d

6.0 6.0

6.5 6.4

7.5

7.0

7.7

7.1

W est Germany

4.9

4.6

4.3 4.3

4.3 4.4 4.6

4.1

4.1

4.6

5.9

6.5

4.1

4.4

5.0

Netherlands

House price in years of incom e e

16

16

13

13

13

W est Germany

19.7

18.0

15.8

Netherlands

R ent as a percentage of monthly income f

f

existing owner-occupied dwellings (Netherlands). House price divided by household income (as in Table 2). Sources: B undesm inisterium (1990); DGVH (1993). R enters with average income (W est Germany); all renters (Netherlands).

e

Sources: B undesm inisterium (1990); M inisterie VR OM (1994). Sources: Golland (1996). Detached houses (W est Germ any); unsubsidised single-family dwellings (Netherlands). Prices in DM (W est Germany) and HFL (Netherlands); DM 1 is about HFL 1.10.cSources: Bundesministerium (1990); C BS (1996). Single-family dwellings (W est Germany); all unsubsidised dwellings (Netherlands).dSource: Golland (1996). E xisting detached houses (W est Germ any); all

a

148 097

41.9

131 571 138 717

1993

45

104 806 113 742 122 230

1992

1990 1991

1987 1988 1989

119 000 111 230 115 551

43

127 038

1984 1985 1986

116 172

1982

1983

101 640

1981

42

100 173

65 142 77 181

57 611

53 400 53 857

52 928 61 600 52 750

44 464 48 862

W est Germ any

Plot price b

Table 1. Housing m arket indicators for W est Germ any and the Netherlands

1980

37.5

41

1978 1979

1977

35

Netherlands

39

35.8

W est Germ any

1975 1976

1972 1973 1974

1970 1971

Year

Percentage of dwellings owner-occupied a

694 CL ARA H . M UL DE R AN D M ICH AE L W A GN ER

FIRS T-T IM E HO ME -OW N ERSHIP

Two major aims of housing polic y are com mon to both countries: to provide affordable housing for lower-incom e groups , and to encourage ow ner-occupancy. In Germany, the policy of encouraging hom e-ow nership is speci® cally directed towards families with children. The ideal of the family living in its own hom e has been strongly advocated by the German government. In the Netherlands to a much larger extent than in Germany, there has been a major shift between the two policy goals. The encouragement of ow neroccupancy has gained emphasis over tim e, whereas buildin g progra mmes in the social sector have been reduced. The main instrum ents to achieve housing policy aim s in both countries are rent control, subsidies and tax bene® ts. In addition, Germany has a government-supported savings programme and the Dutch government gives mortgage guarantees. Rents were frozen after W orld W ar II in both countries. Later, lim its were set to annual rent increases and to the amount of rent due in relation to the quality of the dw elling. In the Netherlands, the year 1980 was a turning -point. Before that tim e, annual rent increases were permitted at rates below in¯ ation and renting a dwelling of a certain quality was often cheaper than buying. After 1980, the allowed rent increases exceeded in¯ ation. Consequently, renting lost much of its attractiveness com pared with hom e-ownership. Both countries have subsidies for low erincom e renters as well as subsidies for housing construction. Furthermore, the German states gives subsidies to hom e-owners. These subsidies are higher for hom e-ow ners with children than those without. The German government has spent less on housing subsidies than has the Dutch. Rough estimates show that 0.3 per cent of the W est German gross dom estic product was spent on housing subsidies in 1988 as against 2.3 per cent in the Netherlands (Van Weesep and Van Velzen, 1994). Feddes (1995) calculated that, in the period betw een 1950 and 1985, 56 per cent of the total num ber of dwellings constructed in W est Germany were built with

695

government subsidies, com pared with 82 per cent of those in the Netherlands. In the Netherlands, the social sector was strongly subsidised and expanded rapidly until the 1970s, but there were severe budget cuts from 1985. On the other hand, Dutch subsidy programmes for ow ner-occupied housing were extended substantially in 1979, which stim ulated the construction of ow ner-occupied dwellings at prices affordable to the middle incom e groups. German hom e-ow ners are not allowed tax deductibility of mortgage interest. Instead, the hom e-ow ner is allow ed tax deductions based on accelerated depreciation of the value of the hom e, but only in the ® rst few years after becom ing a hom e-owner and only once in a lifetim e. On the other hand, in Germany there is no taxable im puted rent (a ® ctitious rent value owner-occupiers have to add to their tax incom e). In the Netherlands, tax deductibility of mortgage interest, although not deliberately meant as an instrument to stim ulate ow ner-occupancy, has in practice worked out as a strong stim ulus. Its ® scal advantage by far outw eighs the taxation of im puted rent. Com pared with other W estern European countries, the Netherlands is quite exceptional in not lim iting tax deductibility of paid interest (Van W eesep and Van Velzen, 1994). T he German government supports savings speci® cally meant for the ow n hom e (B ausparen) with prem ium s and by not taxing the interest. Bausparen com bines a savings program me with loans. The loans can only be obtained after a certain num ber of years of saving. Bausparen is quite popular in Germany. The num ber of contracts is estim ated at one per household (Papa, 1992). The Netherlands has no com parable saving scheme. M ortgage guarantees by Dutch local governm ents became available from 1956 for newly construc ted subsidised dwellings, and from 1973 for most dw ellings below a certain price and for people earning an incom e below a certain level (Van der Schaar and Hereijgers, 1991). T he guarantees facilitated the availability of mortgages up to a high and

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

696

CL ARA H . M UL DE R AN D M ICH AE L W A GN ER

growing percentage of the house price. Mortgages up to 110 per cent of the house price are possible now adays. This is not true for Germany, where the amount of dow npayment required to obtain a m ortgage has remained unchanged at 20 to over 30 per cent (Tom ann, 1996). Parent± Child Transfers in W est Germany and the Netherlands Parental wealth, and hom es and land in particular, are more easily transferred to children in Germany than in the Netherlands. The German tax system for inhe ritance and gifts is very friendly to children inheriting or receiving gifts from parents, and even more so where ow ner-occupied hom es or ® nancial suppor t for owner-occupancy is concerned. The amount that can be transferred tax-free from parents to children is considerable. In addition, transfers of land, dw ellings, or money used to buy these, are taxed not at market value but at a much lower value, based on a legally ® xed E inhe itswert. The taxed value is mostly not even one-quarter of the market value. T he Netherlands has a low er tax-free parent±child transfer, particularly for inheritance. The Netherlands also has higher tax rates. The inheritance of DM 190 000 (or HFL 210 000), on which a German inheriting child pays DM 4 000, is taxed at HFL 22 500 (or DM 20 500) for a Dutch inheriting child. Moreover, transfers of dwellings are taxed at market value (or 60 per cent of it, if the child already lives in the inherited dw elling). To conclude, inheriting is m uch more expensive in the Netherlands than in Germany. Furthermore, whereas in Germany transferring a dwelling or the money to buy it is an im portant way to avoid the taxation of parent±child transfers, this is not true in the Netherlands. The Econom ic Context in West Germany and the Netherlands Both countries experienced strong econom ic growth after World W ar II and particularly before 1973; W est Germany som ewhat more

so than the Netherlands. Likewise, real disposable incom es grew in the long term (Table 2). The growth of household incom es was further strengthened by a grow th in female labour force participation (Table 2). This growth was particularly strong in the Netherlands, where female labour force participation was low in the ® rst few post-w ar decades. Since W orld W ar II, both countries have had serious in¯ ation only in the early 1970s (Table 2). In the Netherlands, mortgage interest rates did not even exceed in¯ ation during this in¯ ation peak (T able 2), which made it very advantageous to buy a hom e at the tim e. A peak in mortga ge interest rates around 1980 coincided with a tem porary collapse of the Dutch owner-occupied m arket: num bers of transactions fell after a boom in the 1970s, as did house prices (see also T able 1). The German market witnessed a similar, but som ewhat less spectacular, downturn in the mid 1980s. From the differences in context, the following hypothe ses can be derived. First, because of higher house prices and the higher amount of downpaym ent needed for mortgages in Germany, we expect the Germans to make the transition to hom e-ownership after more years of saving, so later in the life course than the Dutch. For the same reason, but also because of cheaper parent±child transfers, we expect saving and the use of parental resources to be both more needed and more advantageous in Germany than in the Netherlands. The better opportu nities for inheritance and gifts might even make Germans postpon e hom e-ownership until parental resources becom e available. The connection between ownership subsidies and family situation may result in a stronger in¯ uence of family life-course dynam ics in Germany. Because of real incom e growth, we expect successive cohorts to make the transition to hom e-ow nership at increasingly younge r ages. In the Netherlands, active government intervention in the housing market has made the private rental sector less and less a highquality alternative to owner-occupation,

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

44.4 45.4 47.0 48.2

38.8 40.8

28.9

19.6

West Germ any

35.9 36.4 37.6 40.2 41.6

34.3

34.5

33.5

31.7

28.3

24.7

20.2

15.3

Netherlands

57

53

53

51

49

West Germany

56

41

36

31

30

Netherlands

Percentage of females in b the labour force

2

3.3 5.4 5.5 7.0 7.0 5.9 4.3 3.7 2.7 4.1 5.5 6.3 5.2 3.3 2.4 2.0 0.1 0.2 1.3 2.8

West Germany

Inflation

c

3.6 7.5 7.8 8.0 9.6 10.2 8.8 6.4 4.1 4.2 6.5 6.7 5.9 2.7 3.3 2.3 0.1 2 0.7 0.7 1.1 2.4

Netherlands 100.0 103.3 108.9 114.9 122.9 131.5 139.3 145.3 150.6 154.7 161.0 169.9 180.6 190.0 196.3 201.0 205.0 204.8 205.2 207.9 213.7

West Germany 100.0 103.6 111.4 120.1 129.7 142.1 156.6 170.4 181.3 188.7 196.6 209.4 223.5 236.6 243.0 251.1 256.8 257.1 255.3 257.1 259.9 266.1

Netherlands

Cumulative d inflation

Table 2. E conom ic indicators for W est G erm any and the Netherlands

8.6 8.5 8.3 9.9 10.5 8.7 7.8 7.1 6.4 7.7 9.6 11.1 10.4 8.5 8.3 7.8 6.9 6.4 6.4 7.6

West Germ any

e

8.0 8.3 7.9 8.0 9.6 9.3 8.8 8.7 8.3 9.1 10.2 10.9 10.0 8.3 8.3 7.8 7.0 7.0 6.9 7.6 6.3

Netherlands

Mortgage interest

Sources: Bundesm inisterium (1990) (W est Germ any; annual incom e per head in 1000 DM ; recalculated towards income per household); C BS-Statline on Internet (Netherlands; annual incom e per b c incom e-earning person in 1000 HFL; recalculated towards incom e per household). Sources: Bosveld (1996). Percentage of wom en aged 15±64 years who are in the labour force. Sources: d e Bundesministerium (1990); OE CD. Calculated from 3. Sources: Bundesm inisterium (1990); DGVH (1993).

a

1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991

Year

Annual household a incom e (1000s)

FIRS T-T IM E HO ME -OW N ERSHIP

697

698

CL ARA H . M UL DE R AN D M ICH AE L W A GN ER

whereas in Germany the private rental sector has remained strong. W e therefore expect the trend tow ards hom e-ow nership at younge r ages to be stronger in the Netherlands than in Germany. 4. Data, Operationalisation and M ethods Retrospective life-history data were used for both countries. Com pared with the socioeconom ic panel data also available for both countries, a great advantage of the retrospective data is that they cover a much longer tim e-span. T his is true not only for historic al tim e, but also for individ ual tim e: all respondents are observed from youth onwards, if not always until old age. The family form ation process, which is crucial because we want to focus on com mitm ents in the family life course, is very well covered in the data. A disadvantage, how ever, is that (reliable) incom e estimates are not available and the inform ation about househo ld members other than the respondent is lim ited or missing. Another disadva ntage, present in the German data, is that not all periods were observed for all age-groups because the observation was restricted to birth cohorts spaced 10 years apart. T he incorporation of period effects or, alternatively, indicators for the econom ic or housing market context at a certain tim e, is not feasible because the effect estimates would be biased due to age selectivity. Another potential proble m, arising from the fact that people from distinct birth cohorts were observed at distinct ages, is selectivity with respect to birth cohort membership for older ages. In order to avoid this problem, the upper age lim it of observation was set to 40 years for the multivariate analyses. The German data were derived from the W est German samples in the German L ife History Study (GLHS). (GLHS is located at the Centre for Sociology and the Study of the Life Course of the Max Planck Institute for Hum an Developm ent and E ducation (Berlin).) The GLHS consists of three studies, each representative of W est Germans born in one or more three-year periods. The ® rst study was conducted in 1981±83 am ong

2171 respondents born in 1929±31, 1939±41 and 1949±51 (Mayer and BruÈ ckner, 1989). The second study, conducted in 1985±88, included about 1400 mem bers of 1919±21 birth cohort (BruÈ ckner, 1993). The third study was conducted in 1989 and com prised som e 2000 members of the 1954±56 and 1959±61 birth cohorts (BruÈ ckner and Mayer, 1995). The six cohort groups will be referred to as the 1920, 1930, 1940, 1950, 1955 and 1960 cohorts. The pooled sample of six cohorts contains 5588 respondents. These responden ts contributed just over 84 000 tim e-units (person-years). T his am ounts to about 15 years per respondent. T he GLHS data contain inform ation about residential, educational, family and job histories. For the 1955 and 1960 cohorts, information about non-m arital co-habitation is also available. The inform ation in the data essentially pertains to the individ ual respondents, but includes their situation in the family of origin and, therefore, a quite extended measurement of parental resources. Only very lim ited inform ation was gathered about the current (or last) marital partner, and no inform ation at all about previous partners. W e therefore decided not to use partner inform ation but, instead, to run som e analyses separately for males and females. T he Dutch data were taken from tw o retrospective life-history studies with som ewhat different designs and focuses, but both containing the core variables needed for our analyses: the SSCW survey and the Dutch Family Survey (DFS). Both samples are (more or less) representative of the Dutch populat ion aged 18 years and over (SSCW ) or 21±64 years (DFS; Ultee and Ganzeboom , 1995). The data from these two samples were pooled. T he SSCW survey (som etimes denoted as ESR/T elepanel survey) was conduc ted in 1993 among a sample of som e 3000 members in about 1600 households. It was a single-round retrospective survey among the respondents of a longer-lasting panel answering questions weekly about a wide variety of topics. Response varies betw een the distinct topics, because the data were collected in

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

FIRS T-T IM E HO ME -OW N ERSHIP

several rounds and several tens of respondents left or entered the sample in between. The Dutch Family Survey was conduc ted in the period 1992±93 among a sample of 1000 prim ary respondents and, if present, their current partners. Both sets contain data about the respondents’ residential, educational, household and labour market histories as well as som e data on their families of origin. In the descriptive analysis, the prim ary respondents of the DFS and all the respondents of the SSCW survey were included, because these were the sets of responden ts representative of the Dutch popula tion. The multivariate analyses were perform ed on the prim ary respondents of the DFS and a random ly chosen member of each househo ld in the SSCW survey. In the models estimated for males and females separately, all respondents of the SSCW survey were include d. The dependent variable is whether or not the respondent moves to the ® rst dw elling he or she owns (and inhabits independently from the parents) in a certain tim e unit. The respondents are observed from the year they reach the age of 18 until this transition. After the transition has taken place, the respondent is removed from the analysis. Otherw ise, the respondent remains under observation until either the year of interview or the year in which the age of 41 is reached. It is not clear in all of the data sets how the transition to hom e-ownership is made: through buying the hom e, throug h inheritance or gift, or by moving in with a partner who already ow ns the hom e. Also, the data sets do not contain inform ation as to whether transitions to hom e-ownership are made by buying a dwelling which the respondent was already inhabiting. Those independent variables whose values may change during the life course were updated each tim e-unitÐ except education, which was simply measured as the highest level of general education reached. Our emphasis on the connection of hom eownership with com mitm ents in the family life course requires a precise operationalisation not only of family status, but also of events in the family life course: marriage and

699

having children. By measuring change as well as status, we want to examine how far the com mitm ent to becom e a hom e-owner is `synchronised’ with com mitm ents in the family life course (M ulder and W agner, 1993). T he variable consists of a series of mutually exclusive categories entering the regression equations as dum mies. Apart from obvious categories such as `unm arried’ , `married without children’ , `with one child’ and `with tw o children’ , there are also categories for `marrying, not having a child in same or next year’ , `marrying and having a child in same or next year’ , `expecting or having ® rst child’ , and `expecting or having second child’ . People are placed in `marrying’ during their year of marriage, and in `having or expecting ® rst/second child’ in the year they have a child and the year before. Because m arriage is often quickly follow ed by childbirth, the `marrying’ category is split up further into marrying without having a child in the near future, and marrying while having a child soon. W ith the latter category, we can examine the cum ulation of com mitments: marriage, childbirth and hom e-ow nership. Only ® rst marriages are included in the analyses. Respondents are rem oved from the observation (censored) upon the end of their ® rst marriage. Cohabitation is fully covered only for the last tw o cohorts in the German data. In the Dutch data, inform ation about cohabitation is available for all cohorts, but only in the post-w ar cohorts do people cohabit in any num bers. No distinction is made between cohabiting and married couples with children, because in both countries children are rarely born outside marriage. W ealth or savings are not measured longitudinally in the data. As a proxy measure, we use the amount of tim e the respondent has had the opportu nity to saveÐ that is, the num ber of years spent in paid employm ent from the ® rst job. Because of the high correlation of age with this variable, and because age has less theoretical meaning, we did not also include age. A third reason not to do so is that age is strongly connected with situation in the family life course.

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

700

CL ARA H . M UL DE R AN D M ICH AE L W A GN ER

The GLHS and DFS data do not contain direct retrospective measures of househo ld incom e. The SSCW survey did measure incom e, but the reliability of retrospective incom e measurement is questionable and partial non-response is high. We do have, how ever, measures of social status, earning potential and labour market participation. Our ® rst measure is socioeconom ic status, measured according to Ganzeboom et al.’ s (1992) international socio-econom ic index. W here possible, the status of the job held in the year of observation was used. If the respondent did not work, the status of the previous job was used or, if there was no previous job, that of the next job. For those who never worked during the period of observation, the partner’ s prestige was taken if available. In this way, the percentage of person-years for which no socioeconom ic status measurement was available at all was reduced to 3 per cent (German data) and 0.4 per cent (Dutch data). T hese respondents were given the average status. Besides occupational prestige, we included level of education as an additional measure for earning potential. A dum my variable indicates whether the respondent was self-employed and, therefore, likely to own a farm or ® rm. Another variable distingu ishes those who have paid employm ent in the year of observation from those who have not. People who are observed as not employed may of course have a partner who is employed . This holds particularly for females, because one-earner families in which the female partner is a housew ife are still rather com mon in both countries. Employm ent, therefore, has a different meaning for females than for males. For females, employment is often connected with dual earnership. Because we have no adequate measurement of dual earnership in the German data (for the Netherlands, such measurements exist for only som e responden ts), we examine the in¯ uence of the employm ent variable separately for males and females and regard it as a proxy for dual earnership for the female respondents.

E ducation was measured as the highest level of general education reached, in three categories for Germany and two for the Netherlands. T wo measures for parental resources were used. The socioeconom ic status of the parental family was operationalised by means of the education of the father. Two categories distingu ish between higher and lower levels of education of the father; a third denotes the absence of the father or his education being unknow n. Another im portant measure is whether the parents owned a hom e; in the GLHS and DFS (but not SSCW ) data this inform ation was obtaine d for age 15 of the respondent. T he com parison between the tw o national contexts was made by carrying out the analyses for the tw o countrie s separately. Furthermore, the signi® cance of the differences between the two countries was tested by examining the contribution of the interactions between country and all other independent variables to a model including the respondents of both countries. Degree of urbanisation was included as an indicator for the local housing market. In the German data, degree of urbanisation was measured subjectively: the respondents were asked to classify the places where they lived as a house outside a village, in a village, in a small tow n (up to 30 000 inhabitants), in a mid-sized town (30 000±100 000 inhabitants), or in a large city (100 000 or more inhabitants). In the Dutch data, the respondents were asked to mention the nam es of the municipalities where they lived. The municipalities were then coded according to their degree of urbanisation. T he num ber of inhabitants is less appropriate for typifying Dutch municipalities, because municipalities have often been joined precisely in order to reach a minim um num ber of inhabitants. T he temporal context is expressed in cohort differences. Surviv al analyses are presented for separate cohorts. Cohort, rather than period, was also taken as an independent variable in the multivariate analyses in order to ensure com patibility with the descriptive analyses. Furthermore, in the

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

FIRS T-T IM E HO ME -OW N ERSHIP

German data cohorts are spaced 10 years apart, which makes a period approach less feasible because in each period different agegroups are observed. T he spacing of the German cohorts is also a reason why the direct incorporation of indicators for the housing market or econom ic situation is less feasible. The method used is logistic regression, applied as discrete-time event history analysis. As Allison (1982) and Yamaguchi (1991) have show n, this method gives a satisfactory approxim ation for continu ous-tim e models. More im portantly, discrete-time models are particularly appropriate when one wants to include independent variables denotin g a synchronisation (see above). The units of analysis are person-years. T he logistic regression model is speci® ed as follow s:

l (ti ; X) log 5 1 2 l (t i; X )

ai 1

O

bkX k k

where l (t i;X) is the conditional probability of moving into hom e-ow nership at tim e t i for a given covariate vector X 5 (X 1, ¼ , X k ); b k, k 5 1, ¼ , K are parameters, and ai 5

log

12

l 0 (ti ) l 0(ti )

which is the logarithm of the odds (log-odds) of the event’ s occurrence for a baseline respondent with value 0 for all independent variables. Odds are ratios of the proba bilities of occurrence and non-occurrence of an event within a certain tim e-unit, and are used in discrete-time analysis instead of a hazard (instantaneous rate of event occurrence) in continu ous-tim e models. The parameter’ s exponent can be interpreted as an odds ratio: the amount by which the odds are estimated to be multiplied when the independent variable increases with one unit. First, models were estimated for Germany and the Netherlands separately, containing all respondents and all explanatory variables. The parameter for ow nership of the parents in the model for the Netherlands was derived from a model including the DFS respondents only; the other parameters were derived from a model including all Dutch respondents and not including ownership of the parents.

701

Cohabitation is not included in the family status variable in these models. For the younge r cohorts, similar models were estimated with cohabitation included. In order to test the signi® cance of the differences between the two countries, the aforementioned models were also estimated for the two countries together. Besides all inde pendent variables, these models also contained a dum my variable for country and the interaction between all independent variables and country . The signi® cance of the contribution of these interaction terms to the models was assessed using a likelihood ratio test. Furthermore, separate models for males and females were estimated for both countries. 5. The Transition to Hom e-ownership: Descriptive Measures The tim ing of the transition to ® rst hom eownership in the life course is adequately represented by means of survival functions by age (Figures 1 and 2). T he survival curves dem onstrate a marked difference between Germany and the Netherlands in the age of becom ing a hom e-ow ner for successive cohorts. Cohort differences are considerably larger in the Netherlands than in Germany. In the Netherlands, 12 per cent had made the transition at age 35 years in the cohort born around 1920, as against 55 per cent in the 1950 cohort and 58 per cent in the 1960 cohort. For the German cohorts born around 1920 and 1955, these ® gures amount to 26 per cent and 50 per cent respectively. An obvious acceleration is seen between the Dutch 1920 and 1950 cohorts. The 1960 cohort made the transition only slightly earlier than the 1950 cohort. For Germany, any acceleration between cohorts is seen for cohorts 1920±40 only. The German pre-war cohorts became hom e-owners at younge r ages than the Dutch. T his difference had nearly disappeared for cohort 1940. Dutch cohorts born after W orld W ar II bought at younge r ages than their German counterparts. Hazard rates of the transition to hom eownership according to age are depicted in

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

702

CL ARA H . M UL DE R AN D M ICH AE L W A GN ER

F igure 1. Hom e-ow nership by age: survival function s, W est Germ any.

F igu re 2. Hom e-ow nership by age: survival function s, N etherlan ds.

Figure 3 for all cohorts together. The transition is obvious ly more concentrated at younge r ages in the Netherlands. This conform s with the idea that the Germans need more tim e to save before buying a hom e. The event of marriage and the transition to hom e-ownership frequently occur in the same year in both countries (Figures 4 and

5). The concentration around marriage of the transition to hom e-ownership is stronger in Germany than in the Netherlands, particularly in the older cohorts. The ® gures also show that transition rates around marriage continu ed growing in the Netherlands, whereas they declined after the 1940 cohort in Germany.

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

FIRS T-T IM E HO ME -OW N ERSHIP

703

F igure 3. Rates of transitio n into hom e-ow nership by age, W est Germ any and the N etherlan ds.

Figure 4. Rates of transitio n into hom e-ow nership around m arriage, W est Germ any.

6. The Transition to Hom e-ow nership: Results of M ultivariate Analyse s Descriptive measures of the dependent and independent variables are given in Table 3.

The results of the discrete-time event history analyses of the transition to ® rst hom e-ow nership are presented in Tables 4±7. Table 4 gives the results of the models for all German respondents (left panel) and all Dutch

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

704

CL ARA H . M UL DE R AN D M ICH AE L W A GN ER

F igure 5. Rates of transitio n into hom e-ow nership around m arriage, Netherlan ds.

respondents (right panel). In Table 5, the model including cohabitation is given (for the German cohorts born around 1955 and 1960 and the Dutch cohorts born 1945±64). Only the parameters for house hold status are shown. Table 6 shows the likelihood ratio and its signi® cance of leaving out interactions with a dum my for country from models including the respondents of both countries. The parameters of these models are not shown. The separate models for males and females are shown in Table 7. From these models only the variables concerning education and work are shown. Fam ily and Household Status As expected, never-married people have a signi® cantly low er likelihood of becom ing a hom e-owner than married people without children (T able 4): 0.25 tim es lower in Germany and 0.15 tim es lower in the Netherlands. From Table 5, it can be seen that cohabiting couples becom e hom e-ow ners much more frequently than singles. Their

likeliho od of making the transition to hom eownership is more similar to that of married couple s than to that of singles; for Germany, the difference between cohabitors and married couples without children is not even signi® cant at the 0.05 level. T he event of marriage strongly increases the odds of becom ing a hom e-ow ner. People who marry in a certain year are by far the most likely hom e-buyers in both countries, but particularly in Germany. T here is a marked difference between the two countries in the in¯ uence of childbirth and the presence of children in the household. Anticipa ted childbirth coinciding with the event of marriage elevates the odds of becom ing a hom e-ow ner in the year of marriage in Germany, but not in the Netherlands (Table 4). Expecting or having a ® rst child has no positiv e in¯ uence whatsoever in either country when it takes place outside the year of marriage. After the ® rst child is born, becom ing a hom e-ow ner remains about as likely in Germany but becom es signi® cantly less likely in the Netherlands. In Germany,

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

a

1412

1985±88 14.1 76.2 Ð 66.6 8.2 18.0 16.9 36.8 39.2 28.7 707

1981±83 28.0 86.1 Ð 70.7 6.4 17.6 16.8 32.9 44.7 18.0

1930

728

1981±83 33.8 87.2 Ð 76.6 6.1 17.9 17.5 39.3 38.7 28.5

1940

733

1981±83 34.0 79.3 Ð 62.5 14.1 18.8 18.0 44.8 47.3 27.1

1950

1955

1007

1988±89 35.2 71.6 51.6 55.1 20.9 19.4 18.8 46.7 53.4 25.0

West German birth cohorts

For respondents who were under 30 at the date of interview: age at interview.

N

Year of interview a Percentage ever hom e-owner, age 30 a Percentage ever m arried, age 30 a Percentage ever cohabited, age 30 a Percentage ever had child, age 30 Percentage highly educated Median age ® rst job (m ales) Median age ® rst job (females) Mean ISE I index ® rst job Percentage parents owned hom e, age 15 Percentage large town, age 15

1920

1001

1988±89 27.1 56.7 59.5 42.3 24.9 19.8 19.6 44.2 55.8 24.4

1960

135

1992±93 5.4 74.0 1.5 57.7 34.2 17.3 18.1 45.9 Ð 40.9

1900±24

363

1992±93 14.4 80.4 2.5 61.5 30.6 18.0 17.7 42.7 40.6 29.7

1925±34

Table 3. Descriptive statistics of W est G erman and Dutch data, by birth cohort

549

1992±93 32.0 86.9 5.5 72.6 33.0 17.3 17.1 44.3 42.6 24.8

827

1992±93 42.7 84.9 19.0 64.6 38.6 18.1 17.2 45.0 46.9 25.6

1945±54

Dutch birth cohorts 1935±44

921

1992±93 43.2 63.7 48.7 54.2 48.8 20.0 19.0 45.6 56.5 24.6

1955±64

FIRS T-T IM E HO ME -OW N ERSHIP

705

706

CL ARA H . M UL DE R AN D M ICH AE L W A GN ER

Table 4. Logistic regressio n of the transitio n to hom e-ow nership, W est Germ any and the Netherlan ds G erm any

Family status M arried, no children Unm arried M arrying, not expectin g M arrying, expectin g E xpecting /having ® rst child W ith ® rst child E xpecting /having second child W ith second child

2

2 2

Educatio n L ow M iddle High Years w orked Not em ployed ISEI index Self-em ployed Parents owned a

2

Educatio n of father L ow M iddle/high Not present/ unknow n

2

Urbanisat ion Village Sm all tow n M id-sized tow n L arge city

2 2 2

Cohort 1920 (Netherlan ds 1900±24) 1930 (Netherlan ds 1925±34) 1940 (Netherlan ds 1935±44) 1950 (Netherlan ds 1945±54) 1955 1960 (Netherlan ds 1955±64)

b

Standard error

Exp ( b )

0 1.40* 1.47* 1.85*

0.10 0.10 0.10

1 0.25 4.37 6.38

0.03 0.05

0.12 0.10

0.97 0.95

0.29* 0.12

0.11 0.10

1.34 1.12

0 0.26* 0.46*

0.06 0.09

1 1.30 1.59

0.03* 0.15* 0.04* 0.35* 0.63*

0.01 0.06 0.02 0.09 0.05

1.03 0.86 1.04 1.42 1.87

0 0.03 0.06

0.07 0.08

1 1.03 0.94

0 0.30* 0.81* 1.10*

0.06 0.07 0.07

1 0.74 0.44 0.33

0

2

Constant Initial-2 log likeliho od M odel-2 log likelihoo d Improvem ent N person-y ears ( N transitio ns)

Netherlan ds Standard error

Exp ( b )

0 1.87* 1.23* 1.10*

0.12 0.12 0.15

1 0.15 3.44 3.01

2

0.23 0.59*

0.14 0.15

0.80 0.55

2

0.29* 0.53*

0.15 0.11

0.75 0.59

2

b

2 2

0

2

2 2 2

1

0.27*

0.08

1.31

0.01 0.46* 0.14* 0.18 0.34*

0.01 0.09 0.02 0.13 0.10

1.01 0.63 1.15 1.20 1.40

0 0.12 0.02

0.11 0.07

1 1.13 1.02

0 0.17* 0.78* 1.08*

0.08 0.10 0.10

1 0.84 0.46 0.34

1

0

1

0.32*

0.07

1.38

0.71*

0.27

2.03

0.63*

0.07

1.87

1.17*

0.26

3.23

0.71* 0.84*

0.08 0.07

2.03 2.31

1.45*

0.25

4.27

0.75*

0.09

2.12

1.72*

0.25

5.59

4.22*

0.13

4.22*

0.31

20 242.45 17 358.18 2 884.27, df 5 24, p 5 82 190 (2196)

2

0.00

a

9 663.21 8 067.76 1 595.45, df 5 21, p 5 31 259 (1118)

0.00

For the Netherlan ds: param eter for `parents owned’ from a m odel with only respond ents from D FS data (N person-y ears: 10 830); cohorts 1925±34 and later; control variable s: fam ily status, educatio n, years w orked, not em ployed, IS EI index, urbanisa tion, cohort. Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

707

FIRS T-T IM E HO ME -OW N ERSHIP

Table 5. Logistic regressio n of the transitio n to hom e-ow nership, m odels includin g cohabita tion Germ any Standard error

E xp ( b )

0 1.23* 0.32 1.22* 1.33*

0.16 0.17 0.17 0.20

1 0.29 0.73 3.37 3.77

0.05 0.08

0.19 0.17

1.05 1.08

0.49* 0.09

0.20 0.18

1.63 1.09

b Married, no children Single Co-habi ting Marrying , not expectin g Marrying , expectin g Expecting /having ® rst child W ith ® rst child Expecting /having second child W ith second child

2 2

N etherlan ds

Initial-2 log likeliho od M odel-2 log likelihoo d Improvem ent N person-y ears ( N transitio ns)

6 290.88 5 515.16 7 75.71, df 5 18, p 5 22 603 (707)

Standard error

E xp ( b )

0 2.55* 0.37* 1.17* 0.55*

0.16 0.15 0.13 0.21

1 0.08 0.69 3.24 1.73

2

0.21 0.48*

0.17 0.19

0.81 0.62

2

0.29 0.59*

0.18 0.14

0.75 0.55

2

2

2

2

b

6 293.56 5 189.07 1 104.5, df 5 16, p 5 17 253 (770)

0.00

0.00

Germ any: cohorts 1955 and 1960. Control variable s: educatio n, years worked, not em ployed, ISE I index, self-em ploym ent, degree of urbanisa tion, ownership parents. Netherlan ds: cohorts 1945±64. Control variable s: see Germ any, except ownership parents.

Table 6. L ikelihoo d ratio tests for interacti on term s of independ ent variable s with a dum m y for country : m odels as in Tables 4 and 5, but for tw o countrie s together Indepen dent variable (interact ion with country ) Fam ily status Household status Education Years worked Not em ployed ISE I index Self-em ployed Ownership parents Degree of urbanisa tion Cohort

we see a signi® cantly positiv e effect of the event of second childbirth. In the Netherlands, the effect of second childbirth is negativeÐ when com pared with the status of being married without children. Com pared with couples living with one or two children, however, those who expect or have their second child have higher odds of buying a hom e. The difference between Germany and the Netherlands in the in¯ uence of family and household status is signi® cant (Table 6).

L ikelihoo d ratio Chi squared

df

p

31.10 79.88 1.42 5.49 8.49 7.20 2.51 7.12 5.61 19.63

7 8 1 1 1 1 1 1 3 4

0.00 0.00 0.23 0.02 0.00 0.01 0.11 0.01 0.13 0.00

In Germany, the transition to ® rst hom eownership is strongly conne cted with events in the fam ily life course: marriage, ® rst childbirth when it occurs close to marriage, and second childbir th. In the Netherlands, the connection with the event of marriage and particularly childbirth is weaker. T his difference betw een the countries is signi® cant even after controlling for socioeconom ic status, num ber of years worked, employm ent and education. The conne ction of hom e-

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

708

CL ARA H . M UL DE R AN D M ICH AE L W A GN ER

Table 7. Logistic regressio n of the transitio n to hom e-ow nership, m ales and fem ales separatel y M ales

b Germ any L ow educatio n M iddle educatio n High educatio n Years w orked Not em ployed ISEI index

2

0 0.16 0.45* 0.03* 0.67* 0.06*

Initial 2-log likeliho od M odel 2-log likelihoo d Improvem ent N person-y ears ( N transitio ns) Netherla nds High educatio n Years w orked Not em ployed ISEI index Initial 2-log likeliho od M odel 2-log likelihoo d Improvem ent N person-y ears ( N transitio ns)

Fem ales

S tandard error

E xp ( b )

0.10 0.13 0.01 0.13 0.03

1 1.17 1.57 1.03 0.51 1.07

9 422.06 8 087.46 1 334.61, df 5

21, p 5

b

2

0 0.37* 0.68* 0.05* 0.05 0.03

2

0.33* 0.00 0.82* 0.14*

0.08 0.12 0.01 0.08 0.03

1 1.45 1.98 1.05 0.95 1.03

0.00

21, p 5

0.00

42 115 (1183)

0.09 0.01 0.13 0.03

6 479.60 5 417.70 1 061.90, df 5

E xp ( b )

10 813.26 9 222.50 1 590.76, df 5

40 072 (1009)

2

Standard error

1.39 1.00 0.44 1.15

18, p 5

2

0.34* 0.03* 0.24* 0.12*

0.00

21 766 (740)

0.10 0.01 0.10 0.03 6 100.03 5 167.01 933.02, df 5

1.40 1.03 0.79 1.13

18, p 5

0.00

19 523 (708)

Control variable s: fam ily status, educatio n, years w orked, not em ployed, ISE I index, degree of urbanisa tion, cohort; ownership parents (for Germ any).

ownership subsidies with family situation in Germany might play a role in this difference. It is also possible that the shorter waiting tim es in the Netherlands are partly expressed in low transition rates for people with children: when hom e-ownership before childbirth is com mon, those who have not succeeded in becom ing a hom e-owner before childbirth might be a selective group. More often than Germans, the Dutch buy a hom e before they have children. This is true in particular for the younge r cohorts (Mulder, 1997). The com monness of ® rst hom e-ownership after ® rst childbirth was also found to be speci® c for Germany in com parison with the US (Clark et al., 1997). Educatio n and W ork The in¯ uences of factors associated with work and education on the transition to

hom e-ow nership are all in the expected direction. Higher education positively in¯ uences the likeliho od of becom ing a hom e-ow ner, as do years spent doing paid work, socioeconom ic status and self-employment. For Germany, all these in¯ uences are signi® cant; for the Netherlands, those of years spent working and self-employm ent are not (Table 4). T he differences between the countries are quite interesting. The num ber of years spent doing paid work has a stronger effect in Germany. T his is in line with the idea that Germans need more tim e to save before buying a hom e. Socioeconom ic status and current employm ent, however, have less im pact in Germany than in the Netherlands. It seems as if longer-term factors are more im portant in GermanyÐ where hom e-ownership is attained only after a long-term period of savingÐ and current econom ic positio n is more

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

FIRS T-T IM E HO ME -OW N ERSHIP

im portant in the NetherlandsÐ where the mortgage is of prim e im portanc e and banks use current status to judge future paym ent capability. Self-employm ent and education also seem to have som ewhat different effects in the two countries, but these differences are not signi® cant (Table 6). As expected, work does not have the same im pact for males and females (T able 7). Current non-em ploym ent keeps males from becom ing hom e-owners more strongly than it does females. For Germany, the effect of female employm ent is not even signi® cant. The total num ber of years spent working seems to have a greater im pact for females than for males. For the Netherlands, this difference is signi® cant (an interaction term for sex by years spent working contributed signi® cantly to a model for males and females together; not show n). Apparently, the female incom e in two-earner couples is more im portant for accumulating savings than for current paym ent capability. Differences betw een males and females in the effect of education and socioeconom ic status are signi® cant in neither of the two countries. Parental Resources No in¯ uence of socioeconom ic status of the father on the likeliho od of becom ing hom eowner can be detected. T he effect of father’ s education is not signi® cant (Table 4). Substituting the ISEI index for education did not result in a signi® cant effect. Not even when indicators for individ ual resources were left out of the model could any effect be found. Hom e-ow nership of the parents, however, appears to be of prim e im portanc e. The effect found is for both countries larger than the effect found by Henretta (1987) for the US. For Germans, the odds of making the transition to hom e-ownership in any given year are estimated to be almost twice as high when they lived in a hom e owned by their parents at age 15 than when they did not. Above we hypothe sised three mechanisms which might underlie this effect: a stronger preference of children of hom e-owners to becom e hom e-ow ners them selves; inherit-

709

ance of the hom e; and gifts in the form of ® nancial help or a hom e bought by the parents for the child. T he effect of parental hom e-ownership is signi® cantly less strong in the Netherlands than in Germany. As hypothe sised, parental resources are apparently either less needed or less easily transferred (or both) in the Netherlands. Degree of Urbanisa tion As expected, the likeliho od of becom ing a hom e-ow ner is smaller with higher degrees of urbanisation (Table 4). Even though degree of urbanisation was measured differently in the tw o countries, the size of the effect is about the same: in highly urbanised housing markets, the odds of becom ing a hom e-ow ner are about one-third of those in non-ur banised housing markets. Cohort Differences People belonging to successive birth cohorts make the transition to hom e-ow nership at increasingly younge r ages; the cohort differences are larger in the Netherlands than in Germany. T his is con® rmed by the multivariate analyses (Table 4). T he parameters for cohort are signi® cant for both countries, but the parameters for the Netherlands show a stronger growth over the cohorts than those for Germany. For Germany, the odds of becom ing a hom e-owner for cohort 1960 are estimated to be exp (0.75), or 2.21 tim es higher than that for cohort 1920. For the Netherlands, this ® gure amounts to exp (1.72), or 5.59 tim es (cohort 1955±64 com pared with cohort 1900±24). This difference between the countries is signi® cant (p , 0.01; Table 6). T he cohort differences re¯ ect the changes in real incom es and the relative costs of hom e-ow nership. But even with similar econom ic growth and growth of real incom es, the grow th of hom e-ownership has been different in the two countries. Although both the German and the Dutch government aim at stim ulating hom e-ownership, three

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

710

CL ARA H . M UL DE R AN D M ICH AE L W A GN ER

speci® c features of the German ® nancial system seem to prevent people from becom ing hom e-owners at younge r ages: the am ount of downpaym ent required to obtain a mortgage, the savings system (Bausparen), and the fact that landlords are allowed almost as much tax advantage as owner-occupiers. 7. Discussion This paper seeks to gain further insight into the transition to ® rst-tim e hom e-ownership and its connection with life-course and contextual factors. Hypothe ses about the in¯ uence of life-course differentiation and contextual factors are derived from a theory stating that hom e-ownership is in¯ uenced by preferences, resources and the relative costs of owning versus renting. The analyses are made using retrospective life-course data for W est Germany and the Netherlands. Previous analyses of the transition to hom e-ownership were made using panel data for W est Germany (Clark et al., 1997) and repeated cross-sectional surveys for the Netherlands (Dieleman and Everaers, 1994; Mulder and Hooim eijer, 1995). W e are the ® rst to use the retrospective life-course data available for Germany (GL HS data) and the Netherlands (SSCW and DFS data) for an analysis of the transition to hom e-ownership. Som e advantages of retrospective surveys over panel surveys are that they are cheaper and they allow for faster data collection and easier data handlin g, and the longer tim eframe. But retrospective surveys also have one im portant disadvantage: the retrospective measurement of incom e is hardly feasible. The only feasible alternative is using socioeconom ic status and education as indicators for incom e. But when the study of changes throug h tim e is of greater im portance than a reliable and exact measurement of incom e, the advantages of retrospective data by far outw eigh the disadvantages. Only when a panel has reached an `age’ of, say, 15 years or older can the study of longer-term trends be undertaken with panel data. The results show that hom e-ownership is strongly connected with status and events in

the family life course. The connection between the transition to hom e-ow nership and the events of marriage, ® rst and second childbirth are clearly dem onstrated in this article. Life-course differentiation in resources is also shown to be im portant. Parental hom e-ownership has a striking in¯ uence on the likelihood of becom ing a hom e-ow ner. If this remains true, the recent grow th of hom e-ow nership may lead to accelerated growth in future generations. A problem with the data is that the way of becom ing a hom e-owner is not know n. It would be interesting to distingu ish between buying a hom e and becom ing a hom e-owner through inheritance or gift, so as to gain a better insight into the role of family resources than can be obtained here. As argued earlierÐ by, for instance, Clark et al., 1994Ð the transition to hom e-ow nership is in¯ uenced by contextual factors besides individ ual life-course factors. Strong new evidence is found for this propos ition. As in the previous research, the local housing market turns out to be of great im portance. The same holds for national and temporal contexts. T hree major differences between W est Germany and the Netherlands are found. First, there is a difference in growth of transition rates to hom e-ownership. The growth was stronger and went on for a longer tim e in the Netherlands than in Germany. Secondly, there is a difference in the tim ing of ® rst hom e-ow nership in the life course. In Germany, considerably more so than in the Netherlands, hom e-ow nership is connected with marriage and particularly ® rst and second childbirth. Thirdly, there is a difference in the role of resources. Resources that are accumulated over a longer period or even over generations (savings and parental hom eownership) are more im portant in Germany, whereas current em ploym ent and socioeconom ic status are more im portant in the Netherlands. W hen considering these differences together, the follow ing picture seems to emerge. In Germany, there seem to be tw o mom ents in the family life course suitable for becom ing a hom e-owner: directly at

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

FIRS T-T IM E HO ME -OW N ERSHIP

marriage (for those who have the savings and/or parental resources available) or when children are com ing. In the Netherlands, hom e-ownership is increasingly pursued by childless couples, probably often in anticipation of having children. These differences can be understood in terms of differences in house prices, housing policy (subsidies and other regulations) and other differences in the ® nancial and legal systems. The high house prices in Germany, the high amount of dow npaym ent required to obtain a mortgage and the stim ulation of saving for the own hom e are supposedly the causes of the difference in both tim ing in the life course and the role of resources. The role of parental resources is probably further strengthened in Germany because of the favourable conditions for parent±child transfers. The high prices, together with the strong positio n of private landlords, may also account for the fact that hom e-ow nership rates did not increase as much in Germany as in other European countries. The strong grow th of hom e-ow nership in the Netherlands and its increasingly early occurrence in the life course is undoub tedly connected with changes in Dutch housing policy. Exam ples of these changes are the rise of allow ed annual rent increase, the increased availability of mortgage guarantees in com bination with full tax deductibility of mortgage interest, and the redirection of subsidies from the social rented sector tow ards the ow neroccupied sector. To conclude, we think our results suppor t BoÈ rsch-Supan’ s (1993) view that differences in housing consum ption betw een countries are to a large degree generated by housing policies. Both countries’ governments take the same positio nÐ nam ely, that families should be able to ow n their hom es. Both countries claim they stim ulate private hom e-ownership. The details of the policy measures by which they do so, however, are quite different and so is the im pact of the policy measures. These differences are partly uninten ded outc om es of a series of decisions in a long history of designing and redesigning tax systems, subsidies and regulations.

711

Ironically, this partly uninten ded difference in policy measures appears to have a substantial im pact on the hom e-ownership decisions of W est German and Dutch couples and families.

References A L LIS ON , P. D. (1982) D iscrete-ti me m ethods for the analysis of event histories , in: S . L E INHAR DT (Ed.) Sociolog ical M ethodolo gy, pp. 61±98. San Francisco : Jossey-B ass. A L ONS O , W . (1964) Location and L and Use: Toward a G eneral Theory of L and Rent. Cam bridge, M A: M IT Press. B OÈ R SCH- S U PA N , A. (1985) Tenure choice and housing dem and, in: K. S TAH L and R. J. S TR UYK (Eds) U.S. and W est G erm an Housing M arkets: Comparative E conom ic A nalyses , pp. 55±113. W ashingto n: T he U rban Institute Press. B OÈ R SCH- S U PA N , A . (1993) H ousing market regulations and housing m arket perform ance in the U nited S tates, G erm any, and Japan, in: R. B L ANK (E d.) Social Protection versus E conomic Flexibilit y, pp. 119±156. Chicago: U niversity of Chicago Press. B OS VEL D , W . (1996) The Ageing of F ertility in E urope: A Comparative Dem ographic -analyti c Study . A msterdam : Thesis Publisher s/ PDO D B OUR NE , L . S. (1981) The G eography of Housing . L ondon: E dward Arnold. B RUÈ C KNE R , E . (1993) Lebensverl aÈ ufe und gesellsch aftlicher W andel: Konzeption, Design und M ethodik der E rhebung von L ebensverlaÈ ufen der Geburtsjah rgaÈ nge 1919± 1921 . Berlin: Max-P lanck-In stitut fuÈ r Bildungsforschu ng (M aterialien aus der Bildungsforschu ng N r 44). B RUÈ C KNE R , H. and M AY ER , K. U . (1995) Lebensverl aÈ ufe und gesellsch aftlicher W andel: Konzeptio n, Design und M ethodik der Erhebung von Lebensverl aÈ ufen der G eburtsjahrgaÈ nge 1954± 1956 und 1959± 1961 . Berlin: M ax-P lanck-I nstitut fuÈ r Bildungsforschu ng (Materialie n aus der Bildungsf orschung N r 48). È R FU B UNDE SMINIS TE RIUM R AUM OR DNUN G , B AU WE SE N UN D S T AÈ D TE BAU (1990) Haus und W ohnung im Spiegel der Statistik . Bonn. CBS (S TA TIS TICS N E THE RL ANDS ) (1996) Statistics on housing and construc tion in the Netherlan ds; increases and decreases in the dwelling stock since 1970, Netherlan ds Journal of Housing and the Built E nvironm ent, 11, pp. 69±77. C HEV AN , A. (1989) The grow th of hom e ownership: 1940±1980, D emography , 26, pp. 249±266. C LAR K , W . A. V . and D IEL EM AN , F. M . (1996) H ousehold s and H ousing: Choice and O ut-

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

712

CL ARA H . M UL DE R AN D M ICH AE L W A GN ER

com es in the H ousing M arket. New Brunsw ick: Centre for Urban Policy Research. C L ARK , W . A . V., D EUR LO O , M. C. and D IE LE MA N , F. M . (1994) T enure changes in the context of micro-lev el fam ily and m acro-lev el econom ic shifts, Urban Studies, 31, pp. 137±154. C L ARK , W . A . V., D EUR LO O , M. C. and D IE LE MA N , F. M. (1997) E ntry into hom e ownership in Germ any: com parisons with the U nited States, Urban Studies, 34, pp. 7±19. C OU RGE AU , D. (1985) Interacti on betw een spatial mobility, fam ily and career life-cycl e: a French survey, European Sociolo gical Review , 1, pp. 139±162. C OU RGE AU , D. and L EL IEÁV RE , E . (1992) Interrel ations betw een ® rst hom e-ow nership, constitution of the fam ily, and professi onal occupati on in France, in: J. T R USSE LL , R. H A NKINS ON and J. T ILT ON (E ds) D emographi c Applicati ons of Event H istory Analysis , pp. 120±140. O xford: Clarend on P ress. D EUR LOO , M. C., C LA RK , W . A. V. and D IE LE MA N , F. M. (1994) T he m ove to housing ow nership in tem poral and regional contexts , E nvironm ent and P lanning A, 26, pp. 1659±1670. D GVH (1993) Eigen woningbe zit in cijfers . The Hague: M inisterie VROM /DG VH. D IEL EM AN , F. M. and E V ERA ER S , P . C. J. (1994) From renting to owning: life course and housing m arket circum stances, Housing Studies , 9, pp. 11±25. D IET ER ICH , H., D RAN SFEL D , E. and V O SS , W . (1993) Urban Land and Property M arkets in Germ any. London: UCL P ress. D OLIN G , J. (1976) The fam ily life cycle and housing choice, Urban Studies , 13, pp. 55±58. D UNCA N , G . J. and H IL L , M. S. (1985) Conceptions of longitud inal househol ds: fertile or futile?, Journal of E conom ic and Social M easurem ent, 13, pp. 361±375. E AS TE RL IN , R. A. (1980) B irth and F ortune: The Im pact of Num bers on Personal W elfare. New York: Basic Books. F E DDE S , A . (1995) W oningm arkt, regulerin g en in¯ atie: het na-oorlo gse volkshui svestingsbeleid van tien Noordw est-E uropese landen vergeleke n (H ousing markets, regulati on and in¯ ation: post-w ar housing policy in north-w est Europe) . U trecht: N etherlan ds Geograph ical Studies 194. F O RREST , R. (1983) The m eaning of hom eow nership, Environm ent and P lanning D , 1, pp. 205± 216. G ANZ EBO OM , H . B. G., D E G R AAF , P . M. and T RE IMA N , D . J. (1992) A standard internati onal socio-eco nom ic index of occupati onal status, Social Science Research , 21, pp. 1±56. G OLL AND , A. (1996) Housing supply, pro® t and housing product ion: the case of the United Kingdom , the Netherlan ds and G erm any,

Netherlan ds Journal of Housing and the B uilt E nvironm ent, 11, pp. 5±30. H AU RIN , D. R. and G IL L , H . L . (1987) E ffects of incom e variabil ity on the dem and for owneroccupied housing, Journal of Urban E conom ics, 22, pp. 136±150. H E NRE TT A , J. C. (1984) P arental status and child’ s hom e ow nership, A merican Sociolog ical R eview , 49, pp. 131±140. H E NRE TT A , J. C. (1987) Fam ily transitio ns, housing m arket context, and ® rst hom e purchase by young m arried househo lds, Social Forces, 66, pp. 520±536. K E NDIG , H. L. (1984) Housing careers, life cycle and residenti al m obility: implicatio ns for the housing m arket, Urban Studies , 21, pp. 271± 283. K E NDIG , H. L . (1990) A life course perspect ive on housing attainm ent, in: D . M Y ERS (E d.) H ousing D emography : Linking Dem ographic Structures and Housing M arkets , pp. 133±156. M adison, W I: The U niversity of W isconsin Press. L A SSAR RE , D . (1986) Moving into hom e ownership, Journal of E conom ic Psycholog y, 7, pp. 161±178. L A UTE RBACH , W . and L UÈ SCHE R , K. (1996) E rben und die Verbunde nheit der Lebensver laÈ ufe von Familienm itgliede rn, KoÈ lner Zeitschrift fuÈ r Soziolog ie und Sozialpsy chologi e, 48, pp. 66± 95. L O IKKA NEN , H. A. (1992) Housing dem and and tenure choice: evidence from Finland, Netherlands Journal of Housing and the B uilt Environm ent, 7, pp. 9±30. L U NDQV IS T , L. J. (1992) Dislodgin g the W elfare State? Housing and P rivatizati on in Four E uropean Nations . D elft: Delft Universit y Press. M A RTE NS , M. (1985) Owner-occ upied housing in E urope: postw ar developm ents and current dilem m as, E nvironm ent and P lanning A, 17, pp. 605±624. M A YER , K. U . (1973) D imensions of m obility space: som e subjectiv e aspects of career m obility, in: W . M UÈ L LE R and K . U. M AY ER (Eds) Social Strati® cation and Career M obility , pp. 88±115. Paris: M outon. M A YER , K . U. and B R UÈ CKN ER , E . (1989) Lebensverl aÈ ufe und W ohlfahrt sentw icklung: Konzeptio n, Design und M ethodik der Erhebung von Lebensverl aÈ ufen der G eburtsjahrgaÈ nge 1954± 1956 und 1959± 1961 . Berlin: M ax-P lanck-I nstitut fuÈ r Bildungsforschu ng (Materialie n aus der Bildungsf orschung N r 35). M E GBO LUG BE , I. F . and L INNE M AN , P. D. (1993) H om e ownership, Urban Studies , 30, pp. 659± 682. M IC HE LS ON , W . (1977) E nvironm ental Choice, H um an B ehavior, and Residentia l Satisfact ion. N ew Y ork: Oxford U niversity Press.

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014

FIRS T-T IM E HO ME -OW N ERSHIP

M INIS TE RIE V ROM (1994) Volkshuisv esting in cijfers . The Hague: Ministerie VROM /DGV H. M ULD ER , C. H. (1997) E en kind of eerst een koophui s? Verander ingen in het tijdstip van het kopen van een huis ten opzichte van relatievorm ing en vruchtb aarheid (H aving a child or buying a hom e ® rst? Changes in the tim ing of buying a hom e in relation to partners hip form ation and fertility ), M ens & M aatschap pij, 72, pp. 4±20. M ULD ER , C. H. and H O OIM EIJER , P . (1995) M oving into ow ner-occu pation: com position al and contextual effects on the propens ity to becom e a hom e-ow ner, Netherland s Journal of H ousing and the B uilt E nvironm ent, 10, pp. 5±25. M ULD ER , C. H. and M ANT ING , D. (1994) Strategies of nest-leav ers: `settling dow n’ versus ¯ exibility , E uropean Sociolog ical R eview , 10, pp. 155±172. M ULD ER , C. H. and W AGNE R , M. (1993) M igration and m arriage in the life course: a m ethod for studying synchro nized events, E uropean Journal of P opulatio n, 9, pp. 55±76. M YER S , D . (1985) W ives’ earnings and rising costs of hom e ownership, Social Science Quarterly, 66, pp. 319±329. N EE DHAM , D. B., K RU IJT , B. and K OE NDE RS , P. (1993) Urban Land and Property M arkets in the Netherlan ds. London: UCL Press. N EUT ZE , M. and K EN DIG , H. L . (1991) Achievement of hom e ownership am ong post-w ar A ustralian cohorts, H ousing Studies , 6, pp. 3±14. P A PA , O. (1992) Housing System s in Europe: P art II, A Comparative Study of Housing F inance . Delft: D elft University Press. P ICKVA NCE , C. G. (1974) Life cycle, housing tenure and residenti al m obility: a path analytic approach , Urban Studies, 11, pp. 171±188.

713

P OW ER , A . (1993) Hovels to High Rise: State H ousing in E urope since 1850. L ondon: Routledge. S AUN DER S , P. (1990) A Nation of H om e Ow ners. L ondon: U nw in H ym an. T O M ANN , H . (1996) Private hom e-ow nership ® nance for low -incom e househol ds, Urban Studies , 33, pp. 1879±1889. U L TE E , W . C. and G AN ZEB OOM , H . B. G . (1995) Netherlan ds Fam ily Survey 1992± 93. (m achine-re adable data set). Nijmegen, N etherlands: D epartm ent of Sociology , Nijmegen U niversity . Codebook prepared by H. B. G . G anzeboom , S. Rijken, Septem ber 1993 edition. Changes and addition s m ade by H. B. G . G anzeboom and R. W eygold, January 1995 edition. V AN D ER S CHA AR , J. and H E RE IJGE RS , A. (1991) Volkshuisv esting: een zaak van beleid . U trecht: H et Spectrum . V AN W E ES EP , J. and V AN V EL ZE N , R. (1994) A com parative study of ® nancial housing instrum ents in selected industria lized countrie s: inventory and prospects . Utrecht: Faculty of G eograph ical Sciences (publish ed in: E xpertenko mm ision W ohnungs politik im A uftrag der Bundesreg ierung: M aterialba nd m it Sonderg utachten im A uftrag der Kom mission . Bundesm inisteriu m fuÈ r Raum ordnung , Bauw esen und StaÈ dtebau, 1995). V AN W OER KOM , B. (1994) Ontw ikkeling grondprijzen en kavels 1991. The H ague: Ministerie V ROM/D GVH . W AGN ER , M. (1989) R aÈ um liche M obilitaÈ t im Lebensverl auf: E ine empirische U ntersuchu ng sozialer B edingung en der M igration . Stuttgart : E nke. Y AM A GUCH I , K. (1991) Event History A nalysis . N ewbury Park, CA: Sage.

Downloaded from usj.sagepub.com at ZB MED on April 16, 2014