1THE IMPACT OF PUBLIC POLICIES ON FAMILIES AND [607372]

1THE IMPACT OF PUBLIC POLICIES ON FAMILIES AND
DEMOGRAPHIC BEHAVIOUR

Anne H. Gauthier1
August 2001

Paper presented at the ESF/EURESCO conference ‘The second demographic transition in
Europe’ (Bad Herrenalb, Germany 23-28 June 2001)

INTRODUCTION

Public policies have an undeniable effect on families. Among other things, they regulate
the conditions of employment, define eligibility to welfare benefits, provide education
and health services, and define the rights and responsibilities of parents. Public policies,
thus, shape family life in defining opportunities and constraints. Yet, public policies have
been claimed to have a much more pervasive effect on families. They have been claimed
to be encouraging some types of family structures over others, and to be providing
incentives or disincentives to cohabit, marry, divorce, and to have children in or outside
wedlock. According to a right-wing perspective, generous social and welfare policies
have ‘destroyed’ traditional family values and have encouraged non-traditional family
forms. Popenoe (1988), for instance, claims that generous support for families has
contributed to the ‘decline’ of families in Sweden.2

The links between public policies and demographic behavior are however much more
complex than those claimed by right-wingers. They involve the levels of benefits, the

1 Department of sociology, University of Calgary, 2500 University Dr. NW, Calgary (Alberta), Canada
T2N 1N4, fax: 403-282-9298, email: [anonimizat]

2 For an interesting reply to Popenoe’s attack on the Swedish family, see Sandqvist et al (1992).

2conditions of eligibility, the income and opportunity sets of individuals, as well as the
norms, stigma, and sanctions associated with the receipt of benefits and with non-
traditional forms of behavior, etc. Isolating the impact of social and welfare benefits on
demographic behavior is therefore a complex exercise, and not surprisingly, one that has
led to contradictory findings.

In this paper, I review the theoretical premises and empirical evidence linking public
policies and demographic behavior. As such, the paper expands and updates the reviews
by Chesnais (1996), Demeny (1987), Gauthier (1996a), Hecht and Leridon (1993), and
McNicholl (1998). Because of the vastness of this field of research, I confine the
discussion to the impact of policies on three main demographic and economic behavior:
fertility, family structure, and the labor force participation of mothers. I thus leave aside
the impact of policies on immigration and mortality. I also leave aside the impact of
policies on fertility in the context of high fertility.

The paper is divided into four main sections. Section 1 discusses the theoretical
framework at the basis of the analysis of the impact of public policies on families.
Section 2 reviews the ‘famous’ examples often cited in the literature as evidence of the
impact of public policies on families. This section also reviews some counter-examples
that are less often cited. Section 3 reviews results from multivariate analyses, while
Section 4 concludes the paper and identifies some future avenues of research.

From the onset, two clarifications are warranted. First, it is clear that although I have
made an effort at doing a thorough review of the literature, this review can by no means
be considered as exhaustive.3 Second, I also need to be clear about the types of policies
covered. As pointed out by Ermisch (1986), both population policies and policies without
a specific demographic target can affect demographic behavior. This includes labor
market policies, monetary and fiscal policies, education policies and subsidies, social
security policies, family law, etc. For reasons of space, I restrict the review to policies

3 I have relied on large databases of articles and books (including Sociological Abstracts and the searchable
database of Population Index), but I may still have missed contributions not covered by these databases. I
also restricted the review to studies published in French or English.

3directly targeted at families such as direct and indirect cash transfers for families with
children, means-tested welfare benefits, maternity and parental leave benefits, and
childcare facilities and related subsidy programs.

1. THEORETICAL PREMISES

Before reviewing the empirical studies on the impact of policies on demographic
behavior, I first examine the theoretical premises linking policies and demographic
behavior. Such a theoretical discussion is important as it shows light on the complexity of
the relationship between policies and demographic behavior. But more importantly, it
also makes explicit the assumptions behind the presumed causal relationship. To simplify
the discussion, I initially focus on the impact of policies on fertility. The theoretical
framework can however easily be expanded to cover other demographic behavior.

In their review of childrearing and fertility, Rindfuss and Brewster (1996) argue that:
‘insofar as labor force participation acts as a constraint on fertility, we would expect
fertility to rise in response to any easing of the worker-mother conflict’ (p. 263). By
extension, they furthermore argue that: ‘We would expect, other things being equal, that improvements in childcare availability, acceptability, and quality, and decreases in its
cost would have a positive impact on fertility’ (p. 271). At the core of these hypotheses,
is the assumption that childbearing is a rational decision, and that parents weigh the costs
and benefits of having children against their income, career expectations, own standards
concerning the quality of care for children, etc.

Such a rational choice framework has at its origin in the neoclassic economic framework
in which the decision to have a child, to marry, to divorce, or to take up employment is
assumed to depend on the respective cost and benefit of each alternative, subject to an
income constraint and to an individual’s preferences (Becker 1981; Cigno 1991). Thus,
according to the neoclassic economic theory of fertility, the decision to have a child is
subject to an economically rational decision (a utility maximization process), and is a

4function of the economic cost and benefits of children, subject to an income constraint.
According to this model, any reduction in the cost of children (as a result of public
subsidy) or any increase in income (as a results of transfer payment) is expected to
increase the demand for children (Gauthier and Hatzius 1997). In a revised version of this
model, Becker and Gregg (1973) introduced a quality component and argued that any
increase in income is expected to result in either a higher number of children, or in
children of higher quality (i.e. higher cost). This general model has also been applied to
the decision to marry and divorce (see Becker 1973, 1974, 1981).

This traditional economic model has been very influential in the literature and is at the
core of the assumed relationship between policies and demographic behavior.4 One
should however not lose sight of the fact that this neoclassic model lies on two major
assumptions: that the individual has full information on the cost and benefits of various
alternatives, and that having a child, marrying, or divorcing, is the result of an
economically rational decision. These assumptions have been questioned by numerous
scholars. First, it has been argued that it is doubtful that individuals have the full
information concerning the cost and benefits of various alternatives, for example, the cost
and benefits of children. Imperfect information is more likely to be the case.
Consequently, more recent variants of the rational choice theory have relaxed the full
information requirement, and have formulated a ‘milder’ requirement, namely that
individuals take their decisions based on the situational information available to them –
regardless whether or not this information is accurate or complete (Goldthorpe 2000).
For example, it could be argued that teenage girls take their decision to have, or not to
have, a child based on the information that is available to them – that is, information that
will vary with the teenagers’ own situation or circumstances. Consequently, teenage girls
may not necessarily have the full information about the cost of children, about the levels
and eligibility conditions of welfare benefits, and about their life opportunities, but they
take their decision to have a child based on the information that is available to them at
that particular moment.5

4 From the onset, it has also been heavily criticized. See for example, the paper by Judith Blake (1968).

5 See Raimondo (2001).

5
Second, scholars have also questioned the rationality requirement. And again, both strong
and weaker rationality requirements have been formulated (Goldthorpe 2000). In
particular, even though the decision to become a teenage mother may look as being
‘irrational’ to an economist (in view of the high cost of children and the low levels of
welfare benefits), the decision may be rational in view of the teenager’s own
circumstances, including her perceived prospects of other life alternatives, and the
perceived cost and benefit of children. As Goldthorpe (2000) puts it, an action may be
rational: ‘simply in the sense of being “appropriate” or “adequate” given actors’ goals
and given their situation of action which is taken to include their beliefs’ (p.120).
Furthermore, and here deviating substantially from the neoclassical theory, the costs and
benefits of various alternatives may be both economic and non-economic. For instance,
becoming a teenage mother may provide a sense of personal worth and responsibilities,
and may provide the teenager with a higher status in her immediate neighborhood.6

In short, it may be possible to view the link between public policies and demographic
behavior in a broader theoretical perspective, and to take into account the possibility of
‘imperfect’ information (available under the individual’s specific situation or
circumstances), non-economic costs and benefits, and the role of societal or community
norms and sanctions. This broader theoretical framework appears in Figure 1 below. The
classical formulation appears in rectangular boxes, while additional elements have been
added in ellipses. The word ‘perceived’ has also been added to reflect the possibility of
imperfect information, as discussed above.

[Figure 1 about here]

In this expanded framework, family and welfare benefits may still affect the individuals’
demographic behavior, but their potential impact is no longer the result of a strict

6 This appears to be the case in some deprived communities. For example, the high teenage pregnancy rate
in remote communities of Northern Canada has been linked with the perceived elevated social status of being a mother, see: http://www. nunatsiaq.com/archives/nunavut 000531/nvt20519_01.html

6comparison between the cost of children, for instance, and the economic value of welfare
benefits. Their impact is instead subject to a wider range of economic, non-economic, and
normative considerations, processed within the realm of situation-based, or imperfect
information. The result is undoubtedly a much more fluid and complex model. In
particular, the model calls for the integration of information about the individuals’ own
circumstances in order to capture all the elements of the decision making process. The
increase in recent years of studies taking into account neighborhood conditions reflects
the attempt to move beyond the rigid boundaries of the neoclassical rational choice
framework.7

2. THE ‘FAMOUS’ EVIDENCE AND THE NEGLECTED COUNTER-
EXAMPLES

In contrast to the above complex model, several of the examples often cited in the
literature as evidence of the impact of welfare policies on demographic behavior rely on
simple univariate or bivariate analyses. For instance, they contrast the trends and levels of
fertility observed in countries in which different social policies are in force. I review
below some of these examples. For although they do not have the methodological and
statistical sophistication of studies that will be reviewed in the next section, they have
been presented, and perceived, as powerful evidence of the effect of policies on
demographic behavior. I review below three types of evidences: (1) differential fertility
trends; (2) discrepancies between ideal and actual number of children; and (3) perceived
causes of low fertility. For reasons of space, most of the examples are restricted to
fertility. In the next section, I expand the review to other types of demographic behavior.

7 The reader interested in a more thorough discussion of rational choice theory are referred to Goldthorpe
(2000), Hechter (1994), and Blossfeld and Prein (1998).

7
The differential demographic trends

Two cases have been widely cited as evidence of the impact of policy on fertility: the
cases of France and Germany. The respective trends in the countries’ total fertility rates
are illustrated in Figure 2. In the case of France, the evidence of a positive impact of
policies on fertility lies in the higher level of fertility observed in France as compared to
other Western European countries, and in France’s higher level of support for families
(especially in the immediate decades following World War II). According to Figure 2a,
fertility in France was higher than that observed in Belgium and Germany, especially in
the immediate post-World War II period. Between 1940 and 1999, France’s total fertility
has always been higher than that observed in Belgium, by an average of 0.2 children per
woman. Based on multivariate analyses, which I will come back to in the next section,
Ekert (1986) has also concluded that the higher family benefits provided in France have
resulted in a fertility level higher by about 0.2 child per woman.

The case of Germany has also often been cited as evidence of a positive effect of policies
on fertility. The evidence lies in the fact that until 1976 the fertility rate in East and West
Germany followed a similar trend. But starting in 1977, the difference, which was until then negligible, began to increase to reach 0.4 to 0.5 children per woman (figure 2b). It is
argued that the higher fertility observed in East Germany was the result of a series of
family policy measures introduced from 1976-77, including an extended maternity leave
and a paid childcare leave (Chesnais 1987; Vining 1984). More recent analyses carried
out by Monnier (1990) and Buttner and Lutz (1990) raise however some questions about
the long-term impact of such policy measures. Interestingly, since the end of the socialist
regime and reunification, fertility in East Germany has plummeted to unprecedented low
levels (Witte and Wagner 1995).

[Figure 2 about here]

8Other often cited ‘responses’ of aggregate demographic indices to the introduction of
policy measures include Romania’s sharp rise in fertility following the ban on abortion in
1966 (David 1993), Australia’s steep increase in the number of divorces following the
introduction of the no-fault divorce provision in 1975 (McDonald 1994), and Sweden’s
marriage boom following the reform of public widow's pensions in 1989 (Hoem 1991).8
Several of these demographic responses have however tended to be short-term,
suggesting that the impact was mainly on the timing of fertility rather than on the
completed cohort fertility. I will come back to this issue in the next section.

In contrast to the above example, some other examples, less often cited, cast serious
doubts as to the responsiveness of demographic indices to policies. Figures 2c and 2d
illustrate two such counter-examples: Britain and Quebec. In any typology of family
policy or welfare state regimes, Britain always appear among the least supportive
countries: a country of minimal support for families (Gauthier 1996b; Esping-Andersen
1990). Yet, fertility in Britain has been tracking remarkably that in France in recent
decades. Since 1965, the difference in fertility between the two countries has averaged
0.01 children per woman. I am obviously playing here the devil’s advocate through these
carefully selected examples, and it is clear that factors other than policies may be
contributing to the higher-than-average fertility level in Britain. It remains, that this
counter-example relies exactly on the same methodology than the example cited earlier as
evidence of the effect of the French pronatalist policy. This counter-example thus
suggests that a great dose of ‘carefulness’ is needed when using fertility differentials as
evidence of the effectiveness of policies.

The last example contrasts the fertility trends in Quebec and Canada (figure 2d), and
again casts serious doubts as to the impact of policies on fertility. Quebec’s historically
high fertility, started to decline rapidly from the early 1950s. In 1989, in response to the
very low levels of fertility, the Quebec government adopted a first series of pronatalist
measures, including a large third birth bonus. Yet, despite a short-term recovery, fertility

8 There have also been some studies on the impact of Swedish parental leave policies on fertility based on a
careful analysis of fertility trends (without any multivariate analysis). See for example, Hoem (1990) and
Sundstrom and Stafford (1992).

9in Quebec has remained either lower or equal to that of the rest of Canada where lower
support for families is in place. Of course, it is possible that Quebec fertility would have
been even lower in absence of the pronatalist measures adopted from the late 1980s.
However, it is clear that the impact of the measures introduced did not result in any major
increase in fertility. The impact of policies –if any– was probably small.

Discrepancies between ideal and actual number of children

The discrepancy between the ideal and actual number of children has often been used to
indicate the ‘window’ of opportunity of policies. People, it is argued, have fewer children
than what they considered as ideal because of barriers to fertility, including the high cost
of children and the incompatibility between family and work responsibilities.9 For
example, Chesnais (1996) states that: ‘the gap between the ideal and the reality [in terms
of number of children] demonstrates that public policies have failed to remove the
obstacles to the realization of fertility desires’ (p.736). Data appearing in Figure 3 shows
the difference between the ideal number of children in the countries of the European
Communities and the total period fertility rate in the late 1980s. The average gap is 0.55
children per woman. The gap is highest in Greece and Italy, and lowest in the United
Kingdom and France.

[Figure 3 about here]

Obviously, there are well-known problems associated with the use of the total period
fertility rate (Bongaarts and Feeney 1998) and the use of data on the ideal number of
children (Bongaarts 1998). Among other things, data on the ideal or expected number of
children tends to be highly volatile (Goldberg, Sharp, and Freedman 1959; Westoff and
Ryder 1977). Furthermore, when asked about the ideal number of children, people tend to

9 Examples of such argument to explain the gap between ideal and actual number of children, may be found
in Japan, see: http://www.mhlw.go.jp/english/wp/wp-hw/vol1/p2c5s1.html . The gap between ideal and
actual number of children is also noted in Switzerland, although with no reference to policies, see: http://www.statistik.admin.ch/news/archiv97/f p97005.htm

10refer to global norms and expectations rather than what they themselves consider as ideal.
For instance, responses to questions about the ideal number of children tend to usually
cluster around the 2-child norm, and very few people tend to report having zero or one
child as the ideal. Here again, caution is required in interpreting that type of data as
evidence of policy impact.

Despite these limitations and cautions, it is worth pointing to the relatively strong inverse
relationship between the fertility gap (defined as the difference between the ideal number
of children and the country’s total fertility rate) and the countries’ support for families in
the late 1980s.10 As seen in Figure 4, countries with low support for families tend also to
have the largest fertility gap. I should however stress that this correlation is highly
dependent on the type of demographic indicator used (here the total period fertility rate)
and the index of state support for families. As discussed later in this paper, the
measurement of state support for families is a difficult task.

[Figure 4 about here]

Perceived causes of low fertility

The third type of evidence that has been routinely used to point to the potential impact of
policies on demographic behavior is answers to surveys about the reasons for not wanting
more children. Data in Table 1 shows the percentage of respondents in a Eurobarometer
survey who agreed that reasons related to housing, childcare, or to the level of child
allowance can influence fertility (see Table 1 for the actual wordings of the question). All
three reasons are negatively correlated with fertility, but all display a weal correlation
(the largest one being the availability of childcare with a correlation of -.32 with fertility).

[Table 1 about here]

10 Note that there was a more recent Eurobarometer on the family in 1998. At the time of writing this paper,
the report had not been released.

11
Although this data suggests a potential impact of policies on fertility, it is not exempt
from possible biases. In particular, it is not clear whether the given reasons (e.g. lack of
childcare) had a determining impact on the respondents’ decision to have children, or if
such reasons were given by respondents as a post-hoc explanation or as a justification for
not wanting more children. For many people in Western societies, there is still some guilt
or feeling of selfishness associated with childlessness or not wanting a larger family.
Blaming the high cost of children, the lack of governmental support, or the lack of
adequate housing to justify one’s own decision not to have more children may be
perceived as being a socially more acceptable answer than simply saying that one does
not desire more children. Of course, I am not suggesting that the social or economic
reasons identified by respondents are not important. I am simply saying that even in
absence of the cited problem (e.g. too high cost of children), it is not clear that people
would necessarily have more children.

Unfortunately, very few surveys provide information as to whether or not respondents
would have more children if more governmental support were provided. The nine-
country Population Policy Acceptance Survey carried out in the early 1990s is
exceptional. In addition to being asked about what should be the governments’ policy
priorities, respondents were also asked whether or not they would have an additional
child if their preferred measures were introduced. Note however that the actual increase
in governmental support was not quantified for respondents. Nonetheless, results are
highly informative as they suggest that if the respondents’ preferred policy measures
were introduced, fertility would increase by about 0.1 to 0.2 children per woman
(Kamaras, Kocourkova, Moors 1998).11 In other words, about 1 or 2 respondents out of
10 would be influenced by the policies that they themselves identified as high priorities,
and would have an additional child. Of course, since this data is based on opinion surveys
it remains at the level of a hypothetical child, and various circumstances could lead the
respondents not to have this hypothetical child.

11 For a discussion of policy acceptance and their potential impact on fertility, see also Palomba, Bonifazi,
and Menniti (1989).

12
The above discussion and examples are bound to be partly controversial as they challenge
some well-accepted ‘facts’. However, what I wanted to point out is that those ‘well-
accepted facts’ often rely on relatively weak evidence, for which counter-examples may
easily be found. Too many factors that may have affected fertility trends are moreover
left uncontrolled, and too much is put on counterfactuals (e.g. fertility would have been
lower if it had not been of the strong family policy). In the next section I review studies
based on multivariate analyses of aggregate- or individual-level data. These studies, I
believe, provide a much more solid evidence as to the potential impact of policies on
demographic behavior. And in general they tend to suggest that policies have no, or a
very small, effects on demographic behavior.

3. EVIDENCE FROM MULTIVARIATE ANALYSES

I review below studies that include some policy-related indicators as explanatory
variables in multivariate analyses. I consider the effects of policies on fertility, family
structure, and mothers’ labor force participation. An overview of the studies reviewed in
this section appears in Appendix. The overview summarizes each study in terms of the
dataset used for the multivariate analysis, the type of statistical technique used, the
dependent variable, the policy variables, and the main findings. Before commenting on
the results of these studies, it is important to understand the type of analytical strategy
used by scholars to isolate the impact of policies from other determinants. In particular, it
is important to point out that because of the nature of the topic, very few studies are based
on ‘real’ experiments involving a ‘treatment’ and ‘control’ groups. The United States has
carried out over the years a series of demonstration projects using that type of research
design. In most cases, however, studies on the impact of policies on demographic
behavior are based on ‘naturally occurring’ experiments that exploit variations over time
in the level of benefits (for example a sudden increase in benefits) or variations across
countries or regions.

13
The impact of policies on fertility

Studies that examine the impact of policies on global measures of fertility appear in
Tables A.1, while studies that examine the impact of policies on the probability of out-of-
wedlock birth and on fertility of young mothers appear in Table A.2.

With regard to the first set of studies, all of them suggest a positive impact of policies on
fertility. Higher family or child benefits are associated with higher levels of fertility. In
most cases, however, the impact of policies is estimated to be small. On the basis of a
cross-national analysis, Blanchet and Ekert-Jaffe (1994), for instance, estimate the impact
of family policies at 0.2 children per woman. Using a similar research design, Gauthier
and Hatzius (1997) estimate that a 25 percent increase in family allowances would result
in an increase of the total fertility rate of 0.07 children per woman. Studies that used data
on age- and parity-specific fertility rates furthermore conclude that the impact of policies
on fertility is most likely to be on the timing of fertility rather than on the total number of
children. For example, Ermisch (1988) found that more generous child allowances in
Britain increase the likelihood of higher-parity births – but also encourage young
motherhood.

The above examples all involve measures of cash benefits, and one may wonder the
extent to which fertility is more, or less, responsive to benefits that instead directly
address the issue of compatibility (or incompatibility) between work and family
responsibilities. Studies based on Canadian, German, Norwegian, and Swedish data all
conclude that benefits, such as maternity or parental leave and childcare subsidies, have a
positive impact on fertility. The effect is however also estimated to be small. On the basis
of Norwegian data, Kravdal (1996) estimated that a 20-percentage point increase in the
provision of childcare would result in an increase of no more than 0.05 children per
women in completed cohort fertility. And on the basis of Canadian data, Hyatt and Milne
(1991) estimated that a one percent increase in the real value of maternity benefit would
result in an increase in the total fertility rate between 0.09 and 0.26 percent.

14
The second set of studies examines the impact of policies on the probability of births
outside wedlock and on the fertility of young women. The large majority of these studies
are based on American or British data, and most of them examine the impact of means-
tested benefits on fertility. Findings are mixed, ranging from no significant effect of
policies to small positive effects. For example, Duncan and Hoffman (1990) concluded
that receipt of the means-tested Aid to Families with Dependent Children (AFDC) has no
statistically significant impact of the probably of teenage out-of-wedlock birth. Similarly,
Fairlie and London (1997) concluded that AFDC has no significant impact on higher-
parity births for welfare recipients. On the other hand, Plotnik (1990) concludes that
welfare benefits have some impact on the probability of teenage out-of-wedlock birth for
Black and White teenagers, but not for Hispanics. Thus, contrary to right-wing claims
that teenage girls have babies to take advantage of welfare benefits (Murray 1984), the
evidence is not strong to support such claims.

The impact of policies on family structure and family dynamics

As mentioned in the theoretical section of the paper, policies may also affect family
formation and dissolution, as well as the living arrangements of families. Relevant studies
are summarized in Table A.3. Again, most of the studies use American data and examine
the effect of means-tested benefits. Results are mixed, suggesting that policies have either
no effect, or a small one, on the probability of becoming a lone-mother via divorce and
on female headship. For example, on the basis of data from the American Panel Study of
Income Dynamics, Hoffman and Duncan (1995) concluded that AFDC benefits slightly
increase the rates of marital dissolution. On the other hand, on the basis of data from the
Census Population Survey (CPS), Moffitt (1990) concludes that welfare benefits have no
statistically significant effect on marital status.

Results suggest however that AFDC may provide some incentive to cohabit rather than
marry (Moffitt, Reville, and Winkler 1998), and that welfare benefits may have an effect

15on the living arrangements of lone-mothers, by allowing them to live independently
rather than being part of a larger household. Again, however, results are contradictory.
Ellwood and Bane (1985) concluded that welfare benefits strongly increase the likelihood
that young mothers lived independently, while Hutchens, Jakubson and Schwartz (1988)
concluded that welfare benefits have no effect on the living arrangements of single
mothers.

The impact of policies on mothers’ labor force participation

Finally, the last set of studies that I reviewed examines the impact of policies on mothers’
labor force participation. There is a large body of literature on the potential disincentive
effect of welfare benefits on work (for example see the review by Bishop 1980). For the
purpose of this paper, I however narrowed down the analysis to the labor force
participation of mothers. I moreover distinguish cash benefits from other employment-
related benefits such as maternity leave and childcare subsidies. With regard to cash
benefits, results overwhelmingly suggest that means-tested benefits have a potentially
disincentive effect on the probability of taking up work, or staying in employment. For
example, on the basis of British data, Ermisch and Wright (1991) concluded that higher
welfare benefits increase the exits from, and reduce entries to, full-time employment for
single-mothers. Similarly, on the basis of American data, Blank (1985) concluded that
higher welfare benefits reduce the labor force participation of mothers who are head of
household.

The results with regard to other types of benefits are however mixed. While some studies
suggest that higher maternity leave benefits and childcare subsidies encourage mothers to
take up employment, others find no significant effect. For example, McRae (1993)
concluded that receipt of contractual maternity pay (i.e. benefits provide by the employer
and which top up extant state-wide provisions) has a positive impact on the return to
work after childbirth in Britain. On the other hand, Klerman and Leibowitz (1999)
concluded that maternity leave legislation in the United States does not influence

16mothers’ job continuity. Results are also mixed with regard to the impact of childcare
provision and subsidies. On the basis of Swedish data, Gustafsson and Stafford (1992)
estimated that the provision of public childcare encourages the labor force participation
of mothers with preschoolers. On the other hand, on the basis of German data,
Kreyfendeld and Hank (2000) found no evidence of an impact of the provision of
childcare on mothers’ labor force participation.

Limitations of multivariate studies

Overall, thus, the multivariate studies provide mixed conclusions as to the effect of
policies on demographic and economic behavior, once other factors such as education,
income, etc. are ‘controlled’ for. The effect – if any– tends moreover to be small.
Methodological issues may be at the basis of these inconclusive findings, especially since
the above studies rely on very different datasets and use very different statistical models.
Some of the datasets contain individual-level data while others contain aggregate-level
data (especially among studies on the effect of policies on fertility), and some of the
datasets are longitudinal (based on panel data or retrospective data), while others are
cross-sectional. Some analyses include only one type of welfare benefits, for instance
AFDC, while others include multiple types of welfare benefits, and some analyses
measure the value of welfare benefits at the individual level, while others measure the
value of welfare benefits at the aggregate level (at the level of state or country). Needless
to say, these methodological differences may partly account for some of the contradictory
findings.

There are four related issues. First and foremost, there is the issue of the measurement
itself of policies. As pointed out, some studies include in their statistical model only one
form of welfare benefits while in others several forms of welfare benefits are considered.
The absence of consistent series on other forms of welfare benefits, or on other forms of
support for families, most often explains the omission of such policies from multivariate
analyses. For example, governmental support for housing is often excluded, as is

17governmental support for health and education. Similarly, the large majority of studies
focus on the impact of cash benefits on demographic behavior, while fewer consider the
impact of policies related to maternity and parental leave or to childcare.

Second, there is the issue of imperfect information. As discussed earlier, the potential
effect of welfare benefits, say on teenage birth, is empirically measured in multivariate
models by including the value of welfare benefits to which a teenager would be eligible
to if she were to have a child. This value of benefits (measured at the state level in
American studies) is then contrasted to the potential income of a potential spouse, and to
the teenager’s own estimated future income. While these estimations are done with great
statistical sophistication, one may wonder if these are ‘inputs’ that are well known to the
teenage girl when she is about to have sex with a partner. In fact, one can suspect that her
knowledge of benefits and her estimation of her potential income are rather limited.
Interestingly, when the mothers of teenage girls are themselves welfare recipients, and
thus, when the teenagers are more likely to have a better knowledge of the value of
welfare benefits, teenage girls are found to have a greater likelihood of giving birth out-
of-wedlock and not a lower one as may be expected (in view of the low level of welfare
benefits) (An, Haveman, Wolfe 1993). But of course, other factors may also be
responsible for this finding including the teenager’ perceived limited number of other life
alternatives.

Thirdly, there is the issue that is often not systematically discussed in the literature,
namely that a large proportion of pregnancies, even today, are still unplanned – at least in
the United States. For example, in the United States, about 60 percent of births out of
wedlock to never-married women are said to be unintended (Terry-Humen, Manlove,
Moore 2001). Of course, one may argue that welfare benefits may encourage women to
have a child rather than have an abortion when they find themselves unintentionally
pregnant. However, abortion still being a socially highly divided issue in the United
States, the decision to have, or not to have, an abortion is likely not to be an easy one, and
one that may potentially have nothing to do with the value of welfare benefits.

18Finally, there is the issue of the exact mechanisms by which policies may affect
demographic and economic behavior. Going back to the theoretical framework presented
earlier, it is possible that policies affect demographic behavior according to a strict
economic rational choice theory. For instance, it is possible that the reduction in the cost
of children provided by governmental transfers and subsidies results in an increase in the
demand for children. Other mechanisms are however possible. In particular, a higher
governmental support for families may alter the norms and preferences for children, and
may indirectly result in higher fertility. Alternatively, the causal mechanism linking
policies and demographic behavior may be due to a third factor. For instance, in the case
of teenage girls, it is possible that the level itself of welfare benefits has no direct effect
on the probability of becoming a teenage mother. However, a high proportion of welfare
recipients in the teenager’s neighborhood, and a high proportion of teenage mothers, may
all contribute to shaping the teenager’s perception of her own life opportunities, and may
influence her probability of becoming a teenage mother. In this case, welfare benefits,
thus, would only indirectly influence teenage pregnancy through neighborhood
characteristics. As pointed out repeatedly in this paper, the relationship between policies
and demographic behavior is undoubtedly complex.

4. CONCLUSION

I started this paper by referring to right-wing scholars and politicians who believe in the
undeniably negative impact of policies on families, in encouraging lone-parenthood,
births outside wedlock, and in discouraging employment. The analysis presented in this
paper calls for much caution. In particular, the mechanisms that theoretically link policies
and demographic outcomes are complex involving imperfect information and decisions
that are rationally bound by very specific circumstances. These complex mechanisms are
usually not part of empirical analyses. As reviewed in this paper, the most ‘famous’
evidence on the impact of policies on demographic behavior is based on simple
information about fertility trends, the gap between ideal and actual fertility, or the
perceived reasons for not wanting more children. But even in the case of multivariate

19analysis, data limitations often prevent researchers from taking into account all the
policies or welfare benefits that may affect the probability of having a child, getting
married, getting divorced, etc. There are consequently numerous methodological issues
that make the analysis of the impact of policies on demographic behavior particularly
difficult, and that most likely explain some of the inconclusive findings of empirical
analyses. This point has been raised by several authors. For instance, in his analysis of the
impact of public policies on fertility in Sweden, Walker (1995) concludes that: ‘Its
[parental benefit] strong connection to the female wage, combined with the large
movement in income tax rates and other factors connected to wages, makes it impossible
to estimate the separate effects of parental benefits’ (p. 246).

Does this suggest that we should give up our attempts to measure the impact of policies
on demographic behavior? Probably not. But we should also not lose sight of the fact that
most of the policies reviewed in this paper do not aim at influencing demographic
behavior, but instead aim at increasing the well-being of families. In addition to assessing
the impact of these policies on demographic behavior, it is therefore imperative to also
assess the impact of these policies on the well-being of families: Do policies manage to
lift families out of poverty? Do they reduce the seemingly incompatibility between work
and family responsibilities? Do they successfully support families in stressful situation?
Do they contribute to the successful development of children and to the successful
transition of teenagers to adulthood and parenthood? These questions have already been
addressed in the literature, but there is still scope for more work, especially from a cross-
national perspective. And furthermore, from a social justice perspective, they may be
much more important to address than the question of the impact of policies on
demographic behavior.

Acknowledgements
I am grateful to my research assistant, Monetta Bailey, for her help in searching and
reviewing the literature. I am also grateful to the participants of the ESF/Euresco conference on the second demographic transition for their comments on an earlier version of this paper.

20
BIBLIOGRAPHY

Acs, G. (1996) “The Impact of Welfare on Young Mothers’ Subsequent Childbearing Decisions” The Journal of Human Resources . volume 31 (4):898-915.
An, Chong-Bum; Haveman, Robert; and Wolfe, Barbara. (1993) “Teen Out-of Wedlock Births and Welfare Receipt: The Role of Childhood Events and Economic Circumstances” The Review of Economics and Statistics . volume 75 (2): 195-208.
Becker, G. S. (1981) A Treatise on the Family. Cambridge, MA: Harvard University
Press. Becker, Gary S. (1973) “A Theory of Marriage: Part I” Journal of Political Economy .
volume 81 (4): S13-S46. Becker, Gary S. (1974) “A Theory of Marriage: Part II” Journal of Political Economy .
volume 82 (2): S11-S26. Becker, Gary S. and Lewis, H. Gregg. (1973) “On the Interaction between the Quantity and Quality of Children” Journal of Political Economy . volume 81 (2) Part II: S279-
S288. Bell, Stephen H. and Orr, Larry L. (1994) “Is Subsidized Employment Cost Effective for Welfare Recipients? Experimental Evidence form Seven State Demonstrations” The
Journal of Human Resources. volume 29 (1): 42-61.
Bingley, Paul; Lanot, Gauthier; Symons, Elizabeth and Walker, Ian (1995) “Child Support and the Labor Supply of Lone Mothers in the United Kingdom” The Journal of
Human Resources . volume 30 (2): 256-279.
Bishop, John H. (1980) “Jobs, Cash Transfers and Marital Instability: A Review and Synthesis of the Evidence” The Journal of Human Resources . volume 15 (3): 301-334.
Blanchet, D. (1987) “Les effets démographiques de différentes measures de politique familiale: un essai d’évaluation” Population . volume 42 (1): 99-127.
Blanchet, Dider and Ekert-Jaffe, Olivia. (1994) “The Demographic Impact of Fertility Benefits: Evidence from a Micro-Model and from Macro-Data”. In: The family, the
market and the State in ageing societies. John Ermisch, and Naohiro Ogawa (eds).
Oxford : Clarendon Press: 79-104 Blank, Rebecca M. (1985) “The Impact of State Economic Differentials on Household Welfare and Labor Force Behavior” Journal of Public Economics. volume 28: 25-58.

21Blake, J. (1968) “Are Babies Consumer Durables? A Critique of the Economic Theory of
Reproductive Motivation” Population Studies. volume 22: 5-25.
Blau, D.M. and Robins, P.K. (1986) “Labor Supply Response to Welfare Programs: A Dynamic Analysis” Journal of Labor Economics . volume 4 (1): 82-104.
Blau, D. M. and Robins, P. K. (1988) “Child-care costs and family labour supply” Review
of Economics and Statistics . XXL, (3): 374-381.
Blau, D. M. and Robins, P. K. (1989) “Fertility, employment, and child care costs” Demography . volume 26 (2): 287-299.
Blau, David M. and Robins, Philip K. (1991) “Turnover in Child Care Arrangements” The Review of Economic and Statistics . volume 73 (1): 152-157.
Blossfeld, H. P. and Prein, G. (1998) (eds.) Rational Choice Theory and Large-Scale
Data Analysis. Boulder, CO: Westview Press.
Bongaarts, J (1998) “Fertility and reproductive preferences in post-transitional societies” Population Council . On-line:
http://www. popcouncil.org/pdfs/wp/114.pdf
Bongaarts, J. and Feeney, G. (1998) “On the quantum and tempo of fertility”. Population
and Development Review . 24: 271-91.
Bradshaw, J.; Ditch, J.; Holmes, H. and Whiteford, P. (1993) “A Comparative Study of Child Support in Fifteen Countries” Journal of European Social Policy . volume 3 (4):
255-271. Brouillette, Liliane; Felteau, Claude and Lefebvre, Pierre. (1993) “The effects of financial factors on fertility behavior in Quebec” [Les effets de la fiscalite sur les comportements de fecondite au Quebec.] Canadian Public Policy/Analyse de Politiques .
volume 19 (3): 260-78 pp. Guelph, Canada. Bureau de la statistique du Québec (1983). Démographie québécoise : passé, présent,
perspective , Québec, B.S.Q.
Buttner, Thomas and Lutz, Wolfgang. (1990) “Estimating Fertility Responses to Policy Measures on the German Democratic Republic” Population and Development Review.
volume 16 (3): 539-555. Cain, G. and Wissoker, D. (1990) “A Reanalysis of Marital Stability in the Seattle-Denver Income Maintenance Experiment” American Journal of Sociology . volume 95:
1235-1269.

22Caudill, Steven B. and Mixon, Jr., Franklin G. (1993) “A Note of the Effects of AFDC
Payments on Birthrates” Rivista Internationale di Scienze Economuichie e Commerciali .
volume 40 (4): 379-384.
Chesnais Jean-Claude. (1996) “Fertility, Family, and Social Policy” Population and
Development Review . volume 22 (4): 729-739.
Chesnais, Jean-Claude. (1987) “When one people becomes two: one Germany and the other” [Quand un peuple en devient deux: une Allemagne et l'autre.] Population et
Societes, volume 209: 4 pp. Institut National d'Etudes Demographiques [INED]: Paris,
France. Cigno, Alessandro (1991) Economics of the Family . Oxford: Clarendon Press.
Cigno, Alessandro. (1986) “Fertility and the Tax-Benefit System: A Reconsideration of the Theory of Family Taxation” The Economic Journal . volume 96: 1035-1051.
Cigno, Alessandro and Ermisch, John. (1989) “A Microeconomic Analysis of the Timing of Births” European Economic Review . volume 33: 737-760.
Danziger, S et. al. (1982) “Work and Welfare as Determinants of Female Poverty and Household Headship” Quarterly Journal of Economics . volume 97 (3): 519-534.
David, Henry P. (1993) “Population Policy and Reproductive Behavior: Incentives and Disincentives” Advances in Population . volume 1: 1-33.
Demeny, Paul. (1987) “Pronatalist policies in low-fertility countries: patterns, performance, and prospects”. In: Below-replacement fertility in industrial societies:
causes, consequences, policies, edited by Kingsley Davis, Mikhail S. Bernstam, and Rita
Ricardo-Campbell. Population and Development Review . Vol. 12, Suppl., 1987. 335-58
pp. Population Council: New York, New York. Duncan, Greg J and Hoffman, Saul D. (1990) “Welfare Benefits, Economic Opportunities, and Out-of- Wedlock Births Among Black Teenage Girls” Demography .
volume 27 (4): 519-535. Ekert, Olivia. (1986) “Effets Et Limites Des Aides Financieres Aux Familles: Une experience et un modele” Population . volume 2 : 327-348.
Ellwood, D. and Bane, M. J. (1985) “The Impact of AFDC on Family Structure and Living Arrangements” Research in Labour Economics, volume 7. Ronal Ehrenberg (ed.)
Greenwich: JAI Press. Ermisch, John. (1986) “Impacts of Policy Actions on the Family and Household” Journal
of Public Policy . volume 6 (3): 297-318.

23Ermisch, John. (1988) “The Econometric Analysis of Birth Rate Dynamics in Britain”
The Journal of Human Resources . volume 23 (4): 563-576.
Ermisch, John. (1991) Lone Parenthood: An Economic Analysis . Cambridge: Cambridge
University Press. Ermisch, John F. and Wright, Robert E. (1991) “Employment Dynamics Among British Single Mothers” Oxford Bulletin of Economics and Statistics . volume 53 (2): 99-122.
Esping-Andersen, G. (1990). The Three Worlds of Welfare Capitalism. (Oxford: Polity
Press). European Commission (1990) European Public Opinion on the Family and the Desire for
Children. Eurobarometer 32. (Brussels: Commission of the European Communities).
Fairlie, Robert W. and London, Rebecca A. (1997) “The Effect of Incremental Benefit Levels on Births to AFDC Recipients” Journal of Policy Analysis and Management .
volume 16 (4): 575-597. Gauthier, Anne Helene and Hatzius, Jan. (1997) “Family Benefits and Fertility: An Econometric Analysis” Population Studies . volume 51: 295-306.
Gauthier, A.H. (1996a). “The measured and unmeasured effects of welfare benefits on families: Consequences for Europe's demographic trends” in: Europe's Population in the
1990s . D. Coleman (ed.) (Oxford: Oxford University Press): 297—331.
Gauthier, A.H. (1996b). The State and the Family; A Comparative Analysis of Family
Policies in Industrialized Countries . (Oxford: Clarendon Press).
Gensler, Howard J. and Walls, W. David. (1995) “The Effect of Welfare on Labor Force Participation: The Case of Low Income Female Household Heads” International Journal
of Sociology of the Family. volume 25 (1): 81-89.
Georgellis, Yannis and Wall, Howard J. (1992) “The fertility effect of dependent tax exemptions: estimates for the United States” Applied Economics . volume 24 (10): 139-
145. Goldberg, David; Sharp, H. and Freedman, Ronald. (1959) “The Stability and Reliability of Expected Family Size Data” The Milbank Memorial Fund Quarterly . volume 37: 368-
385. Goldthorpe, J. (2000). On Sociology; Numbers, Narratives, and the Integration of
Research and Theory. (Oxford: Oxford Univ Press).
Gottschalk, Peter. (1988) “The Impact of Taxes and Transfers on Job Search” Journal of
Labor Economics . volume 6 (3): 362-375.

24
de Guibert-Lantoine, C., Monnier, A. (1997). ‘The demographic situation of Europe and the developed countries overseas: an annual report’. Population , an English selection,
revue de l'Ined. vol. 9: 243-268. Gustafsson, S., Stafford, F. (1992) “Child care subsidies and labour supply in Sweden” Journal of Human Resources . volume 27 (1): 204-30.
Hausman, Jerry A. (1980) “The Effects of Wages, Taxes and Fixed Costs on Women’s Labor Force Participation” Journal of Public Economics . volume 14: 161-194.
Hecht, Jacqueline; Leridon, Henri. (1993) “Fertility policies: a limited influence?” In: The changing population of Europe, edited by Daniel Noin and Robert Woods.: 62-75
Cambridge, Massachusetts/Oxford, England: Blackwell. Hechter, M. (1994) “The Role of Values in Rational Choice Theory” Rationality and
Society . volume 6: 318-333.
Heckman, James J. (1974) “Effects of Child-Care Programs on Women’s Work Effort” Journal of Political Economy . volume 82 (2): S136-S163.
Heckman, James J. and Walker, James R. (1991) “Economic models of fertility dynamics: a study of Swedish fertility” Research in Population Economics . volume 7: 3-
91. Hoem, Jan M. (1990) “Social policy and recent fertility change in Sweden” Population
and Development Review . volume 16 (4): 735-48.
Hoem, J.M., (1991). “To marry, just in case …: the Swedish widow's-pension reform and the peak in marriages in December 1989” Acta Sociologica . volume 34: 127-135.
Hoem, Jan M. (1993) “Public Policy as the Fuel of Fertility: Effects of a Policy Reform on the Pace of Childbearing in Sweden in the 1980s” Acta-Sociologica . volume 36 (1):
19-31. Hofferth, Sandra and Collins, Nancy. (2000) “Child care and employment turnover” Population Research and Policy Review . volume 19: 357-395.
Hoffman, Saul D. and Duncan, Greg J. (1988) “A Comparison of Choice-Based Multinomial and Nested Logit Models: The Family Structure and Welfare Use Decisions of Divorced or Separated Women” The Journal of Human Resources . volume 23 (4):
552-562. Hoffman, Saul D. and Duncan, Greg J. (1995) “The Effect of Incomes, Wages, and AFDC Benefits on Marital Disruption” The Journal of Human Resources . volume 30 (1):
19-41

25
Hutchens, Robert; Jakubson, George and Schwartz, Saul. (1988) “AFDC and the Formation of Subfamilies” The Journal of Human Resources . volume XXIV (4): 599-
619. Hyatt, Douglas E. and Milne, William J. (1991) “Can public policy affect fertility?” Canadian Public Policy/Analyse de Politiques . volume 27 (1): 77-85.
Jenkins, Stephen P. (1992) “Lone Mothers’ Employment and Full-time Work Probabilities” The Economic Journal . volume 102: 310-320.
Kamaras, F., Kocourkova, J., Moors, H. (1998). ‘The impact of social policies on reproductive behavior’ in: Population, Family and Welfare; A Comparative Survey of
European Attitudes Volume 2. R. Palomba and H. Moors (eds) (Oxford: Clarendon
Press): 242—61. Klerman, Jacob Alex and Leibowitz, Arleen. (1999) “Job Continuity Among New Mothers” Demography . volume 36 (2): 145-155.
Kravdal, Oyestein. (1996) “How the Local Supply of Day-Care Centers Influences Fertility in Norway: A Parity-Specific Approach” Population Research and Policy
Review . volume 15 (3): 201-218.
Kreyenfeld, Michaela and Hank, Karsten (2000) “Does the Availability of Child Care Influence the Employment of Mothers? Findings from Western Germany” Population
Research and Policy Review . volume 19: 317-337.
Lefebvre, Pierre and Merrigan, Philip. (1998) “The Impact of Welfare Benefits on the Conjugal Status of Single Mothers in Canada: Estimates From a Hazard Model” Journal
of Human Resources . volume 33 (3) : 742-757.
Lefebvre, Pierre and Merrigan, P. (1997) “Welfare Benefits and Incomes of Canadian Families: A Dynamic Analysis of Marital-Cohabitation Dissolution” Canadian Journal
of Economics .
McDonald, P. (1994) Household and Family Trends in Australia. (Year Book Australia,
1994) McNicoll, Geoffrey. (1998) “Government and Fertility in Transitional and Post-transitional Societies” Population Council Working Paper . On-line:
http://www. popcouncil.org/pdfs/wp/113.pdf

McRae, S. (1993) “Returning to Work After Childbirth: Opportunity and Inequalities” European Sociological Review . volume 9 (7): 125-138.
Moffitt, Robert. (1983) “An Economic Model of Welfare Stigma” The American
Economic Review . volume 73 (5): 1023-1035.

26
Moffitt, Robert. (1990) “The Effect of the U.S. Welfare System in Marital Status” Journal of Public Economics . volume 41: 101-124.
Moffitt, Robert. (1994) “Welfare Effects on Female Headship with Area Effects” The
Journal of Human Resources . volume 29 (2): 621-636.
Moffitt, Robert A.; Reville, Robert and Winkler, Anne E. (1998) “Beyond Single Mothers: Cohabitation and Marriage in the AFDC Program” Demography. volume 35
(3): 259-278. Monnier, Alan. (1990) “The Effects of Family Policies in the German Democratic Republic: A Re-Evaluation” Population: An English Selection . volume 2: 127-140.
Murray, C. (1984). Losing Grounds; American Social Policy 1950—80. (US: Basic
Books). Olausson, Petra Otterblad , Bengt Haglund, Gunilla Ringbäck Weitoft and Sven Cnattingius (2001) “Teenage Childbearing and Long-Term Socioeconomic Consequences: A Case Study in Sweden”. Family Planning Perspectives . volume 33 (1)
On-line:
http://www.agi-usa.org/pubs/journals/ 3307001.html
Palomba, Rossella; Bonifazi, Corrado and Menniti, Adele (1989) “Demographic Trends, Population Policy and Public Opinion” Genus . volume 45 (3/4): 37-54.
Popenoe, D. (1988) Disturbing the Nest: Sweden and the Decline of Families in Modern
Society .
Plotnick, R. (1990) “Welfare and Out of Wedlock Childbearing: Evidence from the 1980s” Journal of Marriage and the Family . volume 52: 735-746.
Raimondo, G. (2001). Determinants of single motherhood in the United States.
Unpublished D.Phil thesis (Oxford University). Rindfuss, R. R. and Brewster, K. L. (1996). “Childbearing and fertility” Population and
Development Review . volume 22 (Suppl): 258-89.
Robins, Philip K. and West, Richard W. (1980) “Labor Supply Response Over Time” The Journal Of Human Resources . volume 15 (4): 524-544.
Robins, Philip K.; Tuma, Nancy Brandon and Yaeger, K. E. (1980) “Effects of SIME/DIME on Changes in Employment Status” The Journal of Human Resources.
volume 15 (4): 545-573. Sandqvist, Karin et. al. (1992) “Thriving Families in the Swedish Welfare State; Family Decline: A Rejoinder” Public Interest . volume 109: 114-122.

27
Schultz, T. Paul. (1994) “Marital Status and Fertility in the United States: Welfare and Labor Market Effects” The Journal of Human Resources . volume 29 (2): 635-669.
Sundstrom, Marianne; Stafford, Frank P. (1992) “Female labour force participation, fertility and public policy in Sweden” European Journal of Population . volume 8 (3):
199-215. Teitelbaum, M.S., Winter, J.M. (1985). The Fear of Population Decline. (London:
Academic Press). Terry-Humen, E; Manlove, J. and Moore, K. A. (2001) “Births Outside of Marriage: Perceptions vs. Reality”. Child Trends Research Brief . On-line:
http://www.childtrends.org/PDF/r b_032601.pdf

Vining, Daniel R., Jr. (1984). “Family salaries and the East German birth rate: A comment”. Population and Development Review. 10 no. 4 (Dec): 693-696.
Walker, Ian (1990) “The Effects of Income Support Measures on the Labour Market Behaviour of Lone Mothers” Fiscal Studies . volume 11 (2): 55-76.
Walker, Ian ; Bingley, Paul; Lanot, Gauthier and Symons, Elizabeth (1995) “Child Support and the Labor Supply of Lone Mothers in the United Kingdom” The Journal of
Human Resources . volume 30 (2): 256-279.
Walker, James R (1995) “The Effect of Public Policies on Recent Swedish Fertility Behavior” Journal of Population Economics . volume 8 (3): 223-51.
Westoff, Charles F. and Ryder, Norman B. (1977) “The Predictive Validity of Reproductive Intentions” Demography. volume 14 (4): 223-251.
Whittington, Leslie A; James, Alm and Peters, H. Elizabeth. (1990). “Fertility and the Personal Exemption: Implicit Pronatalist Policy in the United States” The American
Review . volume 80 (3): 545-556.

Winkler, Anne E. (1995) “Does AFDC-UP Encourage Two-Parent Families?” Journal of
Policy Analysis and Management. volume 14 (1): 4-24.
Witte, J. C. and Wagner, G.G. (1995) “Declining fertility in East Germany after unification: a demographic response to socioeconomic change” Population and
Development Review . volume 21 (2): 387-397.
Zhang, Junsen; Quan, Jason and Van Meerbergen, Peter. (1994) “The Effect of Tax-Transfer Policies on Fertility in Canada, 1921-88” The Journal of Human Resources.
volume 29 (1): 181-201.

28
Figure 1. Theoretical model of fertility decision

P e r c e i v e d
Societal & community
Norms and sanctions Preference for children vs.
Other goods

Perceived Non-economic cost and benefit Of children vs. other goods Decision to have a c h i l d Economic cost and benefit Of children vs. other goods
Perceived
Income constraint &
Life opportunities

29Figure 2. Total fertility rate, selected countries

Sources: Guibert-Lantoine and Monnier (1997); Teitelbaum and Winter (1985); Bureau de la statistique
du Quebec (1983); and on-line statistics from Statistics Canada, the Institut de la statistique du Quebec, and INED.
Figure 2a Figure 2b
Figure 2c Figure 2dTFR, East/West Germany
0.000.501.001.502.002.503.00
1920 1940 1960 1980 2000
YearNumber of children per woma n
Germany (West)
Germany (East)Total fertility rate, France, Belgium, and West
Germany
0.000.501.001.502.002.503.003.50
1900 1920 1940 1960 1980 2000
YearNumber of children pe r
womanGermany (West)
Belgium
France
Total fertility rate, France, England-Wales, and
Australia
0.00.51.01.52.02.53.03.54.0
1900 1920 1940 1960 1980 2000
YearNumber of children pe r
womanEngland and Wales
France
AustraliaTotal fertility rate, Quebec & Canada
00.511.522.533.544.5
1910 1930 1950 1970 1990 2010
YearNumber of children per woma n
Canada
Quebec

30Figure 3. Ideal versus actual number of children, late 1980s

Sources: Graphed by the author based on data published in European Commission (1990).00.511.522.53
Ireland
United Kingdom
France
Denmark
Netherlands
Belgium
Luxembourg
Portugal
West Germany
GreeceSpain ItalyNumber of childrenIdeal number of children
Observed TFR

31
Figure 4: Relationship between fertility gap and state support for families, late 1980s

Notes:
1. The following country abbreviations were used: ; Bel: Belgium; Den: Denmark; Fra: France; Frg:
West Germany; Gre: Greece, Ita: Italy; Ire: Ireland; Lux: Luxembourg; Net: Netherlands; Por: Portugal; Spa: Spain; UK: United Kingdom.
2. Kid-Gap: difference between ideal number of children and observed total fertility rate; 3. Support: Index of cash benefits for families

Sources : Computed by the author from data published in European Commission (1990) and Bradshaw et al
(1993). SUPPORT12 10 8 6 4 2 0KID_GAP1.2
1.0
.8.6
.4
.2 Rsq = 0.5810
ukfralux beldenfrgporspanetireitagre

32
Table 1. Fertility and causes of low fertility, late 1980s Causes of low fertility
2
Country TFR1 Housing Childcare Child
allowances
Belgium 1.62 19.5 30.2 19.0
Denmar k 1.67 31.5 58.7 5.2
France 1.78 17.0 28.4 22.2 German
y – West 1.50 53.1 34.0 19.1
Greece 1.39 34.8 35.9 48.4 Irelan
d 2.15 43.5 22.8 25.6
Italy 1.33 37.5 38.8 10.1
Luxembour g 1.60 38.2 28.9 19.5
Netherlands 1.62 31.2 42 12.6 Portu
gal 1.51 41.5 41.2 25.1
Spain 1.36 46.9 27.9 26.2
United Kin gdom 1.83 51.3 27.4 15.1
Correlation with fertility – -0.06 -0.32 -0.15

Notes:
1- TFR: Total period fertility rate as of 1990.
2- Based on a survey carried out by the European Commission in 1989. The question asked was:
“Many things can influence the number of children parents decide to have. Here is a list of such factors. Could you please select the three you consider to be the most important nowadays in deciding the number of children parents are likely to have”.

Source : Computed by the author from data published in European Commission (1990).

33APPENDIX

Table A1. Overview of studies on the impact of policies on fertility

Country1 Authors (year) Data2 Methods of analysis Dependent variable Policy variables3 Findings

International –
OECD Gauthier &
Hatzius (1997) Official statistics
1970-96 Pooled cross-national
and time-series regression Total period
fertility rate Family cash benefits Small positive effect
of cash benefits on fertility.

International –
Western Europe Blanchet & Ekert-
Jaffe (1994) Official statistics, 11
countries 1969-83 Ordinary least squares
regression and two-stage least squares regression Total period
fertility rate Index of family
policy Positive and
significant effect of family policy on fertility.

International –
Western Europe Ekert (1986) Official statistics, 8
countries 1971-83 Ordinary least squares
regression Total period
fertility rate Index of family
policy Positive effect of
family policy on fertility.
Canada Brouillette,
Felteau, Lefebvre
(1993) Survey of consumer
finances 1985-88 Maximum likelihood
method Conditional fertility
probabilities Direct and indirect
cash transfers to
families Direct and indirect
cash transfers to
families have a
positive but small effect on fertility.
Canada Hyatt & Milne
(1991) Official statistics
1948-86 Ordinary least-squares
regression Total period
fertility rate (log) Maternity benefits Maternity benefits
have a significant but
small effect on
fertility. A 1% increase in maternity benefits would result in a 0.26% increase in
fertility.
Canada Zhang, Quan,
Meerbergen Official statistics
1971-83 Generalized least
squares Total period
fertility rate Tax exemption,
child tax credit, Tax exemption, child
tax credit and family

34Country1 Authors (year) Data2 Methods of analysis Dependent variable Policy variables3 Findings
(1994) family allowances,
maternity leave allowances have
significant positive effects on fertility.
Germany Buttner & Lutz
(1990) Official statistics
1964-87 Age-period-cohort
analysis Age specific
fertility rates Pronatalist policy
introduced on 1976 Statistically
significant positive effect of policy on
birth rate up to 5
years after implementation.
Norway Kravdal (1996) Family and
Occupation Survey
1988 Logistic regression Probability of first-
second-, and third-
birth Day care facilities The provision of day
care facilities has a
weak positive effect on fertility. A 20-percentage points increase in childcare enrolment rate would
result in an increase
in cohort fertility of .05 child per woman.
Sweden Hoem (1993) Official statistics
1961-90 Indirect
standardization Parity-specific birth
rate Parental leave policy Positive impact of
policies on the total
fertility rate.
Sweden Walker (1995) Official statistics
1955-90 Time-series analysis Total period
fertility rate Sweden’s social
insurance programs Parental benefits,
public child care availability, and child allowances have
reduced the price of
fertility since the early 1970s and thus, had a pronatalist effects. However,
these effects were

35Country1 Authors (year) Data2 Methods of analysis Dependent variable Policy variables3 Findings
small compared to the
larger and negative effects of trends in female wages and return to human
capital.
UK Cigno & Ermisch
(1989) 1980 Women and
Employment Survey Ordered probit model Completed fertility Tax and child
benefits Increases in women’s
hourly earnings net of tax reduce birth rates, higher child benefits raise completed
fertility.
UK Ermisch (1988) Official statistics
1971-86 Time series regression Parity- and age-
specific birth rates Child allowances More generous child
allowances increase the chance of third
and fourth births, and
also encourage early motherhood.
USA Georgellis &
Wall (1992) Official statistics
1913-84 Generalized least-
squares method Birth rate Real tax value of
dependent
exemption Tax exemption has a
positive impact, but
small, on fertility.
USA Whittington,
Alm, Peters (1990) Official statistics
1913-84 General least squares
regression General fertility
rate Real tax value of the
personal exemption Personal exemption
has a positive and significant effect on the birthrate.
Notes: 1- Country of study; 2- Data: the following acronyms are used: CPS: Current Population Survey; PSID: Panel Study of Inco me Dynamics; NLS: National
Longitudinal Survey; NLSY: National Longitudinal Survey of Youth; NSFH: national Survey of Families and Households; SIPP: Surve y of Income and Program
Participation. 3- Policy variables: the following acronyms are used: AFDC: Aid to Families with Dependent Children; AFDC-UP: Ai d to Families with
Dependent Children (Unemployed Parent).

36
Table A2. Overview of studies on the impact of policies on the probability of births outside wedlock and fertility of young
women
Country1 Authors (year) Data2 Methods of analysis Dependent variable Policy variables3 Findings

Sweden Olausson et al.
(2001) 1985 Swedish
Population Census & Population register Multiple logistic
regression Probability of teenage
birth Welfare benefits Receipt of welfare
benefits is positively associated with teenage birth.

UK Ermisch (1991) Women and
Employment Survey 1980 Proportional hazards
model Pre-marital birth Welfare benefits Higher welfare
benefits increase the likelihood that a yo ung
woman has a birth outside marriage.
However the effects
are not large.
USA Acs (1996) NLSY 1979-88 Discrete time hazard
models logit regression Probability of giving
birth to a second child before the age of 25 AFDC benefits &
Food stamps Welfare benefits have
no statistically significant impact on
subsequent
childbearing decisions.
USA An, Haveman,
Wolfe (1993) PSID 1968-87 Bivariate probit model Probability of teenage
birth out of wedlock Welfare benefits Teenage girls whose
mothers received welfare are more likely
to give birth out of
wedlock.
USA Caudill & Mixon
Jr (1993) Official statistics
1985-86 Ordinary least-square
regression Illegitimacy ratio
(ratio of births to
single mothers to the
total number of births per year) AFDC benefits Positive relationship
between welfare
payments and
illegitimacy rates.

37Country1 Authors (year) Data2 Methods of analysis Dependent variable Policy variables3 Findings

USA Duncan &
Hoffman (1990) PSID 1973-85 Logit model Probability of teenage
out of wedlock birth AFDC benefits Receipt of AFDC has
no statistically significant impact on the probability of teenage out of wedlock birth.

USA Fairlie & London
(1997) SIPP 1990 Logit model Probability of higher-
order birth for mothers who are AFDC recipients AFDC family cap The family cap policy
is not likely to result in a large reduction in the number of births to
AFDC recipients.
USA Plotnik (1990) NLSY 1979-84 Logit regression and
discrete time hazard models Teenage out-of-
wedlock childbearing Welfare benefits Some evidence (but
not strong) that welfare benefits have an effect on teenage
out-of-wedlock
childbearing for Blacks and whites, but not for Hispanics.
See notes Table A.1.

38
Table A3. Overview of studies on the impact of policies on family structure and family dynamics

Country1 Authors (year) Data2 Methods of analysis Dependent variable Policy variables3 Findings

CANADA Lefebvre &
Merrigan (1998) Family History Survey
1990 Proportional hazards
model Conjugal status of
single mothers Provincial welfare
benefits Welfare benefits
significantly affect the exit rates of single mothers towards marriage or
cohabitation.
USA Cain & Wissoker
(1990) Seattle-Denver Income
Maintenance Experiment 1970-4 Log-linear model Rate of marital
dissolution Negative income tax The negative income
tax has no effect on the rate of marital
dissolution.
USA Danziger et al.
(1982) CPS 1975 Logistic regression Female headship Welfare benefits Welfare benefits have
a small effect on female headship.

USA Ellwood & Bane
(1985) CPS 1975-6 Logistic regression Female head of
household AFDC benefits and
welfare benefits Welfare benefits have
a large impact on the living arrangements of young single mothers. A $100 increase in benefits would double
the likelihood that
young women live independently.
USA Hoffman &
Duncan (1988) PSID 1968-82 Multinomial logit and
nested logit Marital status &
welfare receipt AFDC benefits A reduction in AFDC
is associated with an
increase in the proportion of women who remarry according

39Country1 Authors (year) Data2 Methods of analysis Dependent variable Policy variables3 Findings
to the independent
logit model, but AFDC has no significant impact in the nested logit model.
USA Hoffman &
Duncan (1995) PSID 1967-83 Nested logit model Marital dissolution AFDC benefits AFDC benefits slightly
increase marital
dissolution rates.
USA Hutchens,
Jakubson,
Schwartz (1988) CPS 1984 Universal logit model Single mother
propensity to reside in
a subfamily or to head
her own household AFDC benefits Overall level of AFDC
benefits has no effect
on living arrangement.

USA Moffitt (1990) CPS 1969, 1977, 1985 Multiprobit model Probability of being
married and probability of being
female head of
household Welfare benefits The effects of welfare
payments on marital status and female
headship have
increased over time. The effects are negative but rarely statistically significant.

USA Moffitt (1994) CPS 1969-1989 Probit model Female headship State-specific welfare
benefits Welfare benefits have
positive effects on female headship.
USA Moffitt, Reville,
Winkler (1998) CPS, PSID, NLSY,
NSFH Multinomial logit Partner status
(cohabiting, married,
neither) AFDC, AFDC-UP Weak evidence that
AFDC provides
incentives to cohabit
rather than marry.
USA Schultz (1994) 1980 Census and state
information of welfare
and unemployment Ordinary least squares
regression Probability of woman
being married and
number of children AFDC, Medicaid ,
AFDC-UP Statistically significant
and negative effect of
AFDC and Medicaid

40Country1 Authors (year) Data2 Methods of analysis Dependent variable Policy variables3 Findings
benefits borne on currently being
married and on ever having borne children.
USA Winkler (1995) NSFH 1987 Probit model Probability of a
mother to be married or not AFDC-UP Program AFDC-UP does not
have a statistically significant effect on
the probability of
being married.
See notes Table A.1.

41
Table A4. Overview of studies on the impact of policies on mothers’ labor force participation
Country1 Authors (year) Data2 Methods of analysis Dependent variable Policy variables3 Findings

Germany Kreyenfeld &
Hank (2000) German Socio-
Economic Panel Study – 1996 Multinomial logit
model Labor force
participation of mothers Childcare provision
rate No significant effect of
the regional provision of public day care on female labor force participation.

Sweden Gustafsson &
Stafford (1992) Swedish Household
survey 1984 Logit and ordered
probit models Labor force
participation of mothers in two-parent households Childcare provision
and cost Higher provision and
lower price of childcare have a positive effect on mothers’ labor force
participation.
UK Bingley, Lanot,
Symmons & Walker (1995) Family Expenditure
Surveys 1979-1988 Multinational probit
random utility model Labor force
participation of lone mothers Child support
payments Lone mothers
receiving child support were more likely to work.

UK Ermisch &
Wright (1991) 1980 Women and
Employment Survey Proportional hazards
models Movement in and out
of full-time employment among single mothers Welfare benefits-
supplementary benefit, income support Higher welfare
benefits increase exits from, and reduce entries to, full-time employment.

UK Jenkins (1992) Lone Mothers Survey
1989 Probit Model Employment
probabilities of lone mothers Social assistance, wage
rate, child-care benefits, job availability, non-labor
income Higher wages and low
child-care costs result in higher employment probabilities, higher
social assistance is a
work disincentive, higher maintenance is

42Country1 Authors (year) Data2 Methods of analysis Dependent variable Policy variables3 Findings
associated with lower
employment probabilities.
UK McRae (1993) Survey of mothers
1987-88 Logit and probit
models Labor force
participation of mothers after
childbirth Maternity rights
legislation Contractual maternity
pay has a positive impact on return to
work after childbirth.
UK Walker (1990) Family Expenditure
Surveys 1979-1984 Probit model Labor force
participation of lone parents Welfare system –
supplementary benefits (SB) Welfare benefits create
a disincentive to labor force participation.

USA Bell & Orr (1994) AFDC Homemaker-
Home Aide Demonstrations program 1983-6 Ordinary least squares
regression Earnings and welfare
dependence of low-income mothers AFDC training and
subsidized employment programs Programs resulted in
significant increases in earnings and reduced dependence on welfare.

USA Blank (1985) CPS 1979 Ordinary least square
regression & maximum likelihood estimation Household labor force
and welfare participation of mothers who were head of household State-specific AFDC
benefits, tax rates, welfare benefits Differences in welfare
payments, wages and taxes across states create significant differences in labor
force and welfare
participation among low income households.
USA Blau & Robins
(1986) Employment
Opportunities Pilot
Projects household survey 1980 Maximum likelihood
estimation Labor force
participation of
married and single women Welfare benefits Welfare programs have
a significant work
disincentive effect.

USA Blau & Robins
(1988) Employment
Opportunity Pilot Maximum likelihood
logit model Family labor supply
and childcare demand Childcare costs Decision to enter labor
force and to purchase

43Country1 Authors (year) Data2 Methods of analysis Dependent variable Policy variables3 Findings
Project 1980 for 2-parent families
with children childcare are
significantly related to childcare costs.
USA Blau & Robins
(1988) Employment
Opportunity Pilot Project 1980 Event-history analysis Employment status of
married women and fertility Childcare costs Higher childcare costs
are associated with an increase in the rate of
leaving employment
and a reduction in the rate of entering employment. Higher childcare costs also
result in a lower
birthrate for non-employed women but not for employed women.

USA Blau & Robins
(1991) NLSY 1979–86 Maximum likelihood
Poisson model Employment turnover
Child care cost Higher childcare cost
is associated with lower employment turnover.
USA Gensler & Walls
(1995) CPS 1981 Probit model Labor force
participation of lone-
mothers Welfare benefits Welfare benefits
significantly influence
the decision to work.
USA Gottschalk (1988) Denver Income
Maintenance Experiment 1972 Maximum likelihood
estimation Labor force
participation of lone-mothers Taxes and transfers Only a small
proportion of ADFC recipients started
working in any month,
and an even smaller proportion leave welfare through work.

USA Haussman (1980) Gary (Indiana) Income Non-linear probit Probability of labor Negative income tax, A higher marginal tax

44Country1 Authors (year) Data2 Methods of analysis Dependent variable Policy variables3 Findings
Maintenance
Experiment 1971-4 model force participation for
black women head of household AFDC rate & higher transfer
payment lower the probability of labor force participation.
USA Heckman (1974) NLS 1966 Maximum likelihood
technique Labor force
participation of
married mothers Child care programs Child care costs and
quality influence the
decision to work.
USA Hofferth &
Collins (2000) National child care
survey 1990 Discrete-time logit
model Probability of exiting
paid work for mothers Cost, availability,
stability, and flexibility of childcare The components of
childcare affect the employment exits of
mothers, but the effect
varies with maternal wage.
USA Klerman &
Leibowitz (1999) NLSY 1978–90 Weighted logistic
regression Job continuity of
mothers Maternity leave
legislation Maternity leave
legislation does not
influence job
continuity.
USA Moffitt (1983) PSID 1976 Nonlinear maximum
likelihood Work hours of women
head of household AFDC benefits AFDC reduces work
by about 4 hours per week.

USA Robins, Tuma,
Yeager (1980) Seattle-Denver Income
Maintenance Experiments 1970-4 Maximum likelihood
regression Exit from and entry
into employment Negative income tax Negative income tax
significantly increases the length of spells out of employment.

USA Robins & West
(1980) Seattle-Denver Income
Maintenance Experiments 1970-4 Maximum likelihood
regression Desired hours of work Negative income tax Statistically significant
reductions in desired hours of work as a result of negative income tax treatments.
See notes Table A.1.

45