Page 61 - IJPS-2-2

Mark Lyons-Amos

predominately marital, with little evidence of cohabitation, but with fertility behaviour somewhat
delayed. The final class captures generous European welfare states, which afford considerable wel-
fare provision as well as a high degree of support for childcare and legal protection for cohabitation.
These legal frameworks coincide with more common cohabitation and fertility behaviour relatively
late in the life course. Countries in this class belong to either Liberal or Social Democratic welfare
regimes (Blossfeld, 2006).
These findings demonstrate the major advantages of the latent class approach. By relaxing the as-
sumption of a Normal/continuous distribution among higher order units, I am able to identify clus-
ters of similar higher level units, which is informative in itself due to the neat division along the
Hajnal line. Another major advantage is that the higher level groups are formed using contextual
information, which can provide an additional point of interpretation. It should be noted that this is
also a limitation of the study: the requirement for contextual information is now greater than for ei-
ther random or fixed effects multilevel models. Additionally, a more cautious approach should be
applied for a researcher wishing to make causal statements: the fact that much country level infor-
mation is relatively recent means that some of these indicators are likely to be endogenous with fer-
tility and partnerships behaviour (Perelli-Harris and Sánchez-Gassen, 2012; McDonald, 2000). The
final advantage is that this approach allows straightforward comparison of individual level beha-
viour between clusters: the lateness of fertility behaviour in Western European higher support set-
tings compared to Eastern Europe is intuitive and straightforward and links with existing under-
standing of different fertility norms in Western and post-Socialist countries (Sobotka, 2003; Bloss-
feld and Drobnic, 2001). The divisions also link to previous typologies of welfare states, reinforcing
the advantage of this approach for the validation of theoretical models. In terms of the groupings
revealed, the contrast between the class containing the Netherlands and the other Western class is
similar in nature to the cluster analysis of Kammer et al. (2012) who empirically derive a ‘hybr-
id’ class of welfare provision. This reinforces the idea of a more limited welfare class (Class 2),
which is characterised by a Western looking structure but with restrictions compared to the most ge-
nerous welfare regime. One source of departure is the lack of variation in terms of Western patterns
within Class 3: indeed this class includes a number of countries with differing welfare patterns such
as the United Kingdom (Liberal), Norway (Sociodemocratic) and France (Christian Democratic)
within the same grouping (Esping-Andersen, 1999). That said, the range of countries included in this
analysis differs compared to both the original and revised Esping-Andersen analytic samples: indeed
the reason for excluding Eastern European and post-Soviet countries was due to the fact they were
deemed to be transitional and thus qualitatively different from the capitalist welfare systems, which
was reinforced by Blossfeld (2006). This analysis reinforces this point: Eastern European demo-
graphic and welfare trends interact to produce a unique pattern distinct from Western Europe.
Conflict of Interest and Funding

No conflict of interest reported by the author.

Acknowledgements
All analyses were run using the iridis-3 high performance computing facilities at the University of
Southampton. Thanks go to the anonymous reviewers for their comments which have improved the
article, and to Trevor, my proofreader.
References

Adserà A. (2004). Changing fertility rates in developed countries: the impact of labour market institutions. Journal of
Population Economics, 17(1): 17–43.
http://dx.doi.org/10.1007/s00148-003-0166-x.
Ahn N and Mira P. (2002). A note on the changing relationship between fertility and female employment rates in

International Journal of Population Studies | 2016, Volume 2, Issue 2 55