Policies Addressing
the Tempo Effect in
Low-Fertility Countries
WOLFGANG LUTZ
VEGARD SKIRBEKK
GOVERNMENTS IN MANY industrialized countries are expressing concerns about
the long-term consequences of currently low fertility rates and the resulting rapid population aging for pension systems, health systems,
intergenerational equity issues, global economic competitiveness, and relative global political and cultural influence. A recent UN Population Division
survey of governments showed that without exception in countries where
period fertility was below 1.5, respondents considered this level be too low
(UN 2003; McDonald 2005). The European Commission recently initiated
a high-level discussion across the EU by publishing a Green Paper on demographic trends (European Commission 2005). Several national governments
have recently started to give significant public attention to the issues of low
fertility and population aging.
But the concern over demographic trends is not limited to governments and civil society organizations. Increasingly business leaders are concerned about the consequences for public finance and global competitiveness. A recent prominent example is a statement by the European Banking
Federation in which the chief economists of more than 4,500 commercial
banks in Europe highlight the economic dangers of population aging and
state rather bluntly as the first of their three main recommendations: “Increasing the birth rate is particularly important” (European Banking Federation 2004). Much like the governments, the bankers, however, do not
say how they hope to increase fertility.
In demographic terms, the future path of population aging is determined
by four forces: the current age structure of the population and the future
paths of fertility, mortality, and migration. Of these, only migration and fertility are at least theoretically candidates for government policies to counteract the widespread aging trend. The current age structure is a given, and as to
POPULATION AND DEVELOPMENT REVIEW 31(4): 703–723 (DECEMBER 2005)
703
704
POLICIES ADDRESSING THE TEMPO EFFECT
mortality, only policies aimed at increasing life expectancy and hence reinforcing population aging are politically feasible and ethically acceptable.
Continued population aging is largely programmed into the current European age structure, where low birth rates over the past two decades have
resulted in a small number of children and therefore decreasing numbers of
potential mothers in the future. Because of this, Europe has already developed a momentum toward population shrinking (Lutz, O’Neill, and Scherbov
2003). Possible policies aimed at increasing fertility and/or increasing immigration cannot be expected to reverse the aging trend, but they could moderate it and hence soften the expected negative consequences of population
aging. Calculations based on the 15 member countries of the European Union
(EU-15) in 2000 show that even in the case of a steep increase in immigration to produce a net inflow of 1.2 million persons per year, the old-age
dependency ratio (defined as the population above age 65 divided by the
population aged 15–64) is likely to almost double by 2050. Similarly, even a
steep hypothetical increase in fertility to a TFR of 2.2 could not stop significant population aging. Taken together, higher fertility and higher immigration can only moderate the aging trend. For the EU-15, it turns out that by
2050, 100,000 additional immigrants have the same effect on moderating
the increase in the old-age dependency ratio as a sustainable increase in the
total fertility rate by 0.1 children per woman (Lutz and Scherbov 2003).
In this article, we focus solely on the fertility component and particularly on possible policies that are aimed at affecting the tempo of fertility,
that is, the timing of births over women’s life cycle. The tempo effect depresses the level of period fertility and hence lowers the number of births
in a calendar year as long as the mean age of childbearing increases. Such
policies may be particularly relevant if some sort of “low-fertility trap” exists as we discuss below. We summarize the demographic rationale for policies aimed at influencing the tempo of fertility, review evidence about how
the timing of graduation from school affects the timing of fertility, and finally present some calculations about how hypothetical changes in the tempo
of fertility induced by school reforms would influence the future paths of
population aging in Austria, Bavaria, and Italy.
Entering or avoiding a “low-fertility trap”?
Peter McDonald recently observed that low-fertility countries tend to fall in
two distinct groups, those where the TFR has stayed above 1.5 and those
where it has fallen and remained below this critical level (McDonald 2005).
There are currently 28 countries with TFRs below 1.5, and a recent UN compendium on national population policies indicates that the government of
each of these countries considers this level of fertility too low (UN 2003).
McDonald hypothesizes that it is more difficult for a country to raise fertility to, say, 1.6 once it has fallen to levels of 1.3 or 1.4 than to keep fertility
WOLFGANG LUTZ / VEGARD SKIRBEKK
705
around 1.6. From this assumption he derives the policy recommendation
that countries should strive to keep fertility above the critical level of 1.5.
The assumption of a nonlinear dose–response relationship in the field
of possible policy impacts on fertility levels is a welcome addition to the
inconclusive literature on what level of fertility ought to be considered “too
low” and how governments might try to influence fertility levels. One can
elaborate on this hypothesis by seeking to identify nonlinear feedback
mechanisms that result in a bifurcation process that makes a level of period
TFR around 1.5 a watershed between different demographic regimes. It
would then follow that once this Rubicon is crossed, it will be difficult to
reverse this regime change. Recent work by Rindfuss et al. (2004) on social
transitions in Japan supports this assumption of nonlinear, self-reinforcing
processes in social change with thresholds and traps.
Is it justified to call this possible mechanism of irreversible (or hardly
reversible) regime change a “trap,” a term that neither McDonald nor
Rindfuss uses? If a trap is defined as an unpleasant situation (governments
would rather see higher fertility) into which one enters unintentionally and
can only escape with great difficulty, then indeed the described demographic
regime change may be called a trap. But in addition to postulating the possibility of such a low-fertility trap, it would be useful to identify the mechanisms at play in the self-reinforcing process toward lower and lower birth
rates and consequent accelerating population aging and shrinking population size. We describe three such mechanisms, one demographic, one economic, and one related to social norms.
The demographic process refers to the well-studied but in the public
discussion still not fully appreciated phenomenon of negative population
momentum. The dynamics of the age structure of a population imply that
as a result of low fertility in past years, fewer and fewer women (potential
mothers) will enter the reproductive ages in the future. The decline in this
population sub-group will then exert downward pressure on the absolute
number of births and the crude birth rate. It has been estimated that several countries and the EU as a whole have recently entered a period of
negative population momentum, that is, the age structure implies future
population shrinking even if fertility would instantly increase to replacement level (keeping mortality constant and assuming no migration) (Lutz,
O’Neill, and Scherbov 2003). With historically given age structures, this
negative momentum is an independent force implying fewer and fewer
births in the future. The lower the fertility rates in the near term, the stronger the force of negative momentum in the long term. While this demographic component of the “low-fertility trap” is purely an accounting effect at the aggregate level, the following two mechanisms relate to
behavioral effects.
The economic mechanisms that could lead to a downward spiral in
fertility from one generation to the next can be derived from the first part
706
POLICIES ADDRESSING THE TEMPO EFFECT
of Richard Easterlin’s relative income hypothesis, which postulates that family size results from the combination of consumption aspirations (which are
largely formed in the family of origin) and expected income (Easterlin 1980).
According to Easterlin the post–World War II baby boom in the United States
resulted from the combination of low aspirations (parents of the baby
boomers were relatively poor) and high economic growth during the early
postwar period leading to high expected income and general optimism. This
line of reasoning can also explain the subsequent fertility decline as a combination of increased material aspirations in the next generation with a less
optimistic economic outlook. Disregarding the second part of Easterlin’s hypothesis in which he assumes that expected income is a function of cohort
size, which implied the unsubstantiated expectation of a second baby boom,
one can directly apply the first part of the relative income argument to current and future fertility in low-fertility countries. Material aspirations of
young people have been rising over recent decades as a consequence of
increasing parental wealth, high consumption standards communicated by
the media, and possibly even smaller family size (youngsters have to share
with fewer siblings). At the same time the long-term economic outlook has
been darkening in part because of the prospect of population aging. As documented in youth surveys around Europe, the expectations of people entering the labor market today are not optimistic. Youth unemployment is high
in many countries (despite smaller cohorts entering the labor market); there
are fewer secure jobs; and recent steps toward reforming various systems of
social security (even if rather minor steps so far) have given rise to a somber view of the future among the younger generation. In contradiction to
the second part of Easterlin’s hypothesis, which postulates positive effects
resulting from small cohort sizes, recent findings from global surveys suggest that fewer younger people mean fewer start-ups of new enterprises
and fewer jobs, given that peak entrepreneurial activity takes place in ages
25–44 years (Global Entrepreneurship Monitor 2004). Moreover, firms may
move away from areas with smaller young cohorts, since fewer potential
workers will be available (Shimer 2001). Hence the economic story of the
“fertility trap” argument would go as follows: lower fertility leads to faster
population aging and thus to deeper cuts in the welfare state, less job creation, and an expectation of lower economic growth in the future; at the
same time, aspirations for personal consumption are still on the rise owing
to parental wealth and fewer siblings; and the match of high aspirations
and pessimism about the economic future will result in even lower fertility.
This assumed economic mechanism has the potential to create a continuing
downward spiral toward lower fertility.
Lastly, a plausible mechanism operates in the realm of normative
change with respect to ideal family size. If one assumes that the norms and
expectations of the younger generation are formed by what they see around
WOLFGANG LUTZ / VEGARD SKIRBEKK
707
themselves during the process of socialization, then this constitutes a direct
feedback mechanism from the family size of the previous generation to the
ideal family size of the next generation. Goldstein, Lutz, and Testa (2003)
have proposed this hypothesis in the context of the appearance of belowreplacement fertility ideals among the younger generation in the Germanspeaking countries. Those countries were among the first to experience the
decline to very low fertility levels in the late 1970s and early 1980s, which
now with a generational lag could influence the norms of today’s young
potential parents. This hypothesis has found empirical support in a multilevel analysis by Testa and Grilli (2004), who showed with regional European data—after controlling for a large number of social and economic factors—that fertility ideals among the young are significantly lower in areas
where the fertility of the parents’ generation has already fallen to low levels. Assuming that the controls adequately cover regional peculiarities other
than level of fertility a generation ago, this finding supports the possibility
of a downward spiral. This follows the same logic as described by Rindfuss
et al. (2004: 855): “Changes in attitudes likely create a feedback mechanism, influencing behavior; and changes in behavior likely create a feedback mechanism influencing attitudes.” Once the number of children (siblings, friends, children seen in other families) experienced during the process
of socialization falls below a certain level, one’s own ideal family size becomes lower, which may result in further declining actual family size and
still lower ideals in the subsequent generation.
If true, these three possible mechanisms of a self-reinforcing process
toward lower and lower fertility have all the characteristics of a trap. Since
these kinds of low-fertility conditions have never existed before in human
history, it is impossible to test empirically whether such “low-fertility trap”
mechanisms are indeed relevant in determining fertility trends. One can
only refer to informed reasoning with an element of speculation. But if the
existence of a low-fertility trap is considered a real danger (and we currently see no reason to rule it out), then the best and safest strategy is to
avoid falling into the trap by introducing policy interventions to prevent
fertility from falling below a certain critical level for an extended period.
We stated above that McDonald’s recommendation for governments
is to prevent the TFR from falling below 1.5. But what is the recommendation for governments in countries where the TFR has already fallen below
this level? The logic of the argument would suggest that in those cases, as a
matter of urgency, fertility should be brought back above 1.5 before the
regime change is irreversible. But is there a magic trick to raise the TFR by
some 0.3 births speedily, a new policy that has not yet been tried? This
magic bullet may well exist in the form of tempo policies that give period
fertility a short-term upward kick. Policies that address the tempo of fertility and stop the further increase of or even lower the mean age at child-
708
POLICIES ADDRESSING THE TEMPO EFFECT
bearing without necessarily affecting completed cohort fertility could be the
right policy tool enabling a population to escape a possible low-fertility trap
before it closes.
The rationale for tempo policies
Methodologically, this article follows up on two recent contributions and
aims at operationalizing the policy-relevant aspects of those articles. In 2003
in an article in Science entitled “Europe’s population at a turning point,”
Lutz, O’Neill, and Scherbov emphasized that the current fertility-depressing effect of an ongoing increase in the mean age at childbearing will have
a significant and lasting effect on population dynamics in Europe, played
out in population decline and accelerated population aging. This so-called
tempo effect on fertility has recently received much attention in the demographic literature (see, e.g., Bongaarts and Feeney 1998; Kohler and Philipov
2001). This work is based on the analytical insight that period fertility is
currently low in Europe for two reasons: 1) a tempo effect: women are delaying births to later ages, resulting in fewer births in the calendar years
during which this delay happens; 2) a quantum effect: women are not having enough births to achieve replacement level. If women do not forgo postponed births altogether, delayed childbearing does not affect the total number
of births women have over the course of their lives, but it still lowers period birth rates as long as postponement continues and hence the delay
contributes to further population aging and decline. In fact not all postponed births will be recuperated, and increases in the mean age at childbearing tend to reduce the quantum of the fertility of the cohorts experiencing such increases (tempo–quantum interactions).
Much of the demographic work in this context has focused on estimating fertility rates that adjust for the tempo effect, seeing it as a disturbance that should be eliminated in order to come up with a “purer” fertility
measure, the tempo-adjusted TFR. Lutz, O’Neill, and Scherbov (2003) turn
this approach upside down and focus on the tempo effect not as something
that should be ironed out, but rather as something that could provide a
point of leverage for attempts to influence the level of period birth rates.
Such attempts are called tempo policies. Lutz et al. calculate that at the TFR
level of the 15 member states of the European Union in 2000, a hypothetical end to postponement would instantly raise the period TFR from 1.5 to
1.8, which, cumulated over several decades, would have very significant
effects in moderating future population decline and aging. Their analysis
demonstrates that the changing age of childbearing represents an important facet of population dynamics in Europe that requires special attention.
It is worth noting that tempo policies in the opposite direction, that is, aiming to increase the mean age at childbearing in order to speed up fertility
WOLFGANG LUTZ / VEGARD SKIRBEKK
709
decline, hence lower the rate of population growth in developing countries, are already part of the demographic literature (see Bongaarts 1994)
and indeed of China’s “later, longer, fewer” population policy.
Goldstein, Lutz, and Scherbov (2003) explicitly address another more
methodological point in this context. Stable population theory states that
under conditions of below-replacement fertility, a longer mean age of a generation implies a slower shrinking of the population. This force operates in
the opposite direction to the tempo effect described above. What is the balance of these two opposing effects? The authors show both analytically and
through a set of alternative simulations that the tempo effect far outweighs
the effect of the mean length of generations for the coming centuries. The
second effect would catch up only about 250 to 300 years in the future, if it
is assumed that all postponed births are later recuperated. If one also includes estimates of tempo–quantum interactions, then the relative strength
of the tempo effect becomes so preponderant that the effect of mean length
of generations can safely be disregarded. In other words, this feature of stable
population theory does not call into question the assertion that in actual
European populations, a near-term end to postponement of childbearing
or even a decrease in the age at childbearing would have significant demographic consequences for reducing shrinkage of the total population and
reducing population aging.
In contemporary Europe (with the exception of France), explicitly pronatalist policies have met pronounced public resistance. Family policies in
Europe today tend to be based on an equal-opportunity rationale and aim to
help women combine childrearing with employment. So far such policies,
which come in a variety of forms under differing national conditions, seem
to have had little or no effect on period fertility in the countries with lowest
fertility (McDonald 2002; OECD 2003; Gauthier 2002). In this context, policies explicitly addressing the timing of births may be a useful addition to the
toolbox of fertility-relevant policies available to policymakers. Such tempo
policies could also have other important rationales such as increasing the efficiency of the educational system (see discussion below) and improving health.
In terms of health, further postponement of childbearing not only raises the
risk that women will remain involuntarily childless, it also leads to burgeoning numbers of often cumbersome infertility treatments and increases the
health risks associated with late pregnancies for both mothers and children.
Hence, policies aimed at creating the conditions that allow women to have
their children at an earlier age, or that at least do not encourage further delay, could turn out to be win-win strategies, responding to individual health
concerns as well as public demographic and economic concerns.
What public policies could help stop the increase in the mean age at
childbearing or even lead to a decrease in the near term? Strong social and
economic forces work in the direction of ever-increasing mean ages at child-
710
POLICIES ADDRESSING THE TEMPO EFFECT
birth. The prevailing view is that women want to finish their education,
become established in a job, and find a reliable long-term partner before
they enter the demanding life phase of raising children, which commits them
for at least 15 to 20 years. And the standards of what is considered satisfactory establishment in a professional career as well as what partnership is
reliable enough to have children seem to be rising. How could such forces
favoring later childbearing be reversed?
In theory, there are two ways in which childbearing could take place
earlier in the life cycle: a reordering of life events (such as having children
before finishing education); or maintaining the usual sequence of events
but shortening the phases that precede the birth of children (such as compressing the period of education). The first strategy would require a fundamental change in the current incentive structures that young adults face
when they plan for their parallel careers in education and work on the one
hand, and partnership and children on the other. Changing economic incentives in combination with providing new models for successful combinations of education/work and family could well influence family formation strategies (Andersson 1999; Esping-Andersen 1999; OECD 2003), but
would require significant changes in well-established norms and institutions. The second approach, trying to shorten the life phases that typically
precede childbearing, seems to be a better candidate for short-term interventions with near-term effects on period fertility rates. This second strategy is the focus of rest of this article.
We will discuss the possible role of education reforms as a strategy to
introduce a downward force on the mean age at childbearing. For reasons
entirely unrelated to the timing of births, efforts in many European countries are underway to lower the age at which young men and women finish
their secondary or tertiary education without lowering educational attainment or schooling quality.
Age at graduation and the timing of
demographic events
Studies on the timing of life events suggest that individuals tend to sequence
events in adulthood according to rigid schemes: leaving school precedes entering the labor market, having a child, and other events in adulthood (Billari,
Manfredi, and Valentina 2000; Blossfeld and De Rose 1992; Rindfuss,
Bumpass, and St. John 1980). An increase in the school-leaving age, particularly at the tertiary level, is likely to raise the age of entering parenthood, because women usually postpone having children until they have
completed their education (Black, Devereaux, and Salvanes 2004; Blossfeld
and Huinink 1991). In effect, a change in the timing of one event is likely
to affect the timing of subsequent events.1
WOLFGANG LUTZ / VEGARD SKIRBEKK
711
This effective incompatibility between education and childbearing
seems to have become stronger over time, at least in the United States
(Rindfuss, Bumpass, and St. John 1996). Hence the age of leaving school
(at whatever level) and the timing of fertility seem to have become more
closely linked over time. Furthermore, even in countries like Norway, where
parental benefits make it easier to combine having children with being a
student, enrollment in education strongly reduces the probability of childbearing (Kravdal 2001), indicating that fertility choices during education
are not strongly influenced by public policies.
As educational attainment has increased during recent decades, the
mean age at childbirth in most European countries has increased considerably, and total fertility rates have dropped below replacement levels (Council
of Europe 1999). For example, in Italy the mean age at first birth increased
from 28.9 (1990) to 29.8 (1996), while the TFR fell from 1.33 to 1.19. When
discussing changes in the length of education and in the average age at graduation, one must distinguish between the changing age at which a certain
level of education is completed and the change in the average level of educational attainment.
Educational reforms that lower the age at leaving school are underway, or are being planned, in several European countries. School regulations can affect the graduation age by 1) compressing or extending the duration of schooling required for a specific educational degree, or 2) changing
the age of entering school, which for a given schooling period would alter
the graduation age.
Changes in the required duration of schooling are underway across
Europe. The Bologna declaration (1999) harmonizes, and often shortens,
European tertiary education lengths. In the Italian case it means a change
from a system where students on average spend seven years in the university system (Cnvsu 2001) to a rigid system in which a bachelor’s degree is
earned in three years and a Master’s degree is earned in two years, thus
discouraging tendencies to spend a longer time studying than necessary.2
At the secondary level, the German Federal state of Saarland shortened the
duration of the academic track from nine years to eight years in 2001, and
several other Federal states, including Bavaria, are implementing similar
changes. School shortening reforms are implemented in order to increase
the flow of students through the system, to increase the supply of labor for
the economy, and to improve the cost-effectiveness of the educational system. The possible effects of these reforms on the level of period fertility
have so far not been considered.
The age at entering school also varies by country; it can be as young as
four or as old as seven years in Europe (UNESCO 2003). In recent years,
the school entrance age has been changing to younger ages in several countries. Only seven American states required enrollment in school below age
[ok?]
712
POLICIES ADDRESSING THE TEMPO EFFECT
seven years in 1965; in 1992, 25 states did so (US Department of Health,
Education and Welfare 1965; Education Commission of the States 1994). A
recent, industry-sponsored German proposal (Lenzen 2003) suggested lowering the school entrance age to four years and reducing the typical duration of primary and secondary education to ten years, in order to lower the
age at exiting the educational system.
Several factors influence the time elapsed between graduation and first
birth and the timing of subsequent births. These include the role of social
norms related to expected entry to the labor market, financial support given
to young individuals, and the degree of wage flexibility, which affects the
likelihood of employment (Bentolila and Ichino 2000; Planas 1999; OECD
1999, 2000). Since the net effect of these different influences is unclear
from a theoretical perspective, we will examine them using empirical data
from Sweden.
The fertility effects of school-leaving age: The case
of Sweden
Women who attain higher education tend to have fewer children and to
have them at a later age, but nonrandom educational sorting makes drawing conclusions about causality in this relationship difficult. Women who
leave school at different ages tend to have different levels of education as
well; therefore, they differ by other characteristics—such as preferences, abilities, opportunities, and family background—that affect both educational attainment and fertility decisions. To identify the causal effects of a change in
the school-leaving age, we present a study based on a natural experiment3
that produces a variation in the explanatory variable (graduation age) that
is uncorrelated with other influences on fertility.
Skirbekk, Kohler, and Prskawetz (2004) analyzed a dataset of 863,304
nonimmigrant Swedish women born between 1946 and 1962 and made
use of the fact that Swedish children are enrolled in school in the calendar
year in which they become seven years old. Therefore, children who are
born during two consecutive months, December and January, differ by 11
months in the age at graduating from school.
This policy of the Swedish school system results in an exogenous variation in the age at completing compulsory and higher education, as parents
presumably do not time the births of their children with age at school entry
in mind. Birth months thus affect age at graduation in the same way as a
random sorting of individuals into higher and lower school-leaving ages.
The Swedish setting is, therefore, well suited to investigating the causal link
between the timing of births and the variation in school-leaving age.
Skirbekk et al. found that variation in the school-leaving age has a
strong and consistent effect on the timing of demographic events in adult-
713
WOLFGANG LUTZ / VEGARD SKIRBEKK
hood. Women who are born in January leave school at an 11-month higher
age, and this results in a 4.9-month later age at first birth relative to those
born in December (the previous month). Specifically, women born in January had their first child on average at age 25.3 while those born one month
earlier had their first child at the average age of 24.9. Figure 1 shows this
pattern for the four quarters of the mother’s birth date in the year and her
age at first child; there is a clear downward slope, with women born in the
first quarter being 3.1 months older when they have their first child relative to women born in the fourth quarter within a given year.
The timing of the second birth is also affected by the school-leaving
age. The birth interval between the first and the second child remains virtually unaffected by the variation in the age at first birth caused by different years of graduation. Women born in the first quarter of the year have
only a 0.05-months shorter birth interval relative to women born in the
fourth quarter of the year, indicating that there is no conscious compensation for birth-month-induced variation in the age at first birth. These findings suggest that the timing of fertility is strongly connected to the time of
leaving school.
A school reform that lowers the school-leaving age will, in addition to
affecting the individual’s biological age at the time of school exit, also affect
the social age of the social reference group (the class peers). The social age of
the group would be raised because its members would assume the life cycle
roles of worker and parent at an earlier biological age than would their
counterparts who left school at later ages.
By increasing the social age of those affected, a lower age at graduation would increase the effect of a younger school-leaving age on the timFIGURE 1 Age of mother at birth of first child for
Swedish women born 1946–62
25.2
Age
25.1
25
24.9
1
2
3
Quarter of birth in calendar year
4
714
POLICIES ADDRESSING THE TEMPO EFFECT
ing as well as the quantum of childbearing through tempo–quantum interactions. The impact of a more “mature” social influence at a younger
biological age would heighten the probability of having children at a
younger age. This could in turn imply an increase in completed cohort fertility. A change in the school-leaving age is likely to have a stronger influence on the timing of fertility than the one identified in the Swedish birth
month experiment, since both biological and social age at the time of graduation influence fertility decisions. Taken together, it is plausible that a oneyear change in the school graduation age would be reflected in a unit shift
in the age of entering parenthood. But we will also experiment with other
assumptions.
Hypothetical forecasts for Austria, Bavaria,
and Italy
To investigate the demographic impacts of a change in the timing of fertility resulting from a younger age at graduation, we project the consequences
of fertility changes for the young cohorts in Austria, Italy, and the German
Federal state of Bavaria, which with more than 12 million inhabitants is
bigger than many European countries.4 To study the unconfounded fertility effects, we assume zero net migration for all three regions. Life expectancy is assumed to increase by 1.5 years per decade, for a total of 7.5 years
for the 50-year duration, from 2000 to 2050. In Austria, life expectancy
increases from 81.5 to 89 years for women and from 75.7 to 83.2 years for
men; in Bavaria from 79.0 to 86.5 for women and 73.0 to 80.5 for men;
and in Italy from 81.6 to 89.1 for women and 75.8 to 83.3 for men.
We discuss five scenarios based on different assumptions about the
future course of period TFR. In all cases where the effect of a school reform
(resulting in an assumed lower mean age at childbearing) is being simulated, we assume that the reform first affects the female birth cohort of 1995.
Hence, the effect on the period TFR will be gradual and will increase as
more and more women born after 1995 enter the main reproductive ages.
For the cohorts born before 1995, the rates are assumed to stay constant at
the level described in the specific scenario assumptions. The assumed TFRs
for the three populations and summary results of population projections
based on five scenarios are given in Table 1.
First, we consider a constant period TFR scenario (Scenario 1). It presents a reference case for comparison, in which all age-specific period fertility rates remain constant throughout the projection period. This scenario
assumes that the current tempo effect continues. Since the mean age at
childbearing cannot rise forever, it also implies that the weakening of the
tempo effect will be compensated by a declining quantum in order to produce a constant period TFR.
WOLFGANG LUTZ / VEGARD SKIRBEKK
715
Scenario 2 refers to the hypothetical case of constant tempo-adjusted
period TFR at the current level. Here we assume that period fertility immediately jumps to the level of tempo-adjusted fertility, assuming an instant
end to the increase in the mean age at childbearing. These rates are then
held constant over time. To estimate the tempo effect for adjusting the TFR,
we apply the estimate of Lutz, Philipov, and Scherbov (2005) for a general
relationship between an increase in the mean age at childbearing and the
tempo effect. It shows that on average in contemporary Europe, an increase
in the mean age at childbearing (over all birth orders) of 0.1 years depresses
period TFR levels by 0.19 children. Lutz et al. base this estimate on data
from European countries from 1980 to 2000. If we apply this estimate and
take into account the changes in the mean age at childbearing,5 Austria’s
tempo-adjusted fertility is 1.69 and the unadjusted TFR in 2002 is 1.41. In
Bavaria, the tempo-adjusted fertility is 1.62 and the unadjusted TFR in 2002
is 1.36. In Italy, the tempo-adjusted fertility is 1.51 and the unadjusted TFR
in 2000 is 1.24.
Scenario 3 considers the case of a school reform that has the net effect of reducing the mean age at childbearing by two years for the cohorts
born in 1995 or later. This assumption could be interpreted as a 1:1 relationship between age at graduation and age at childbirth. It is implemented
by a simple downward shift in the age-specific fertility profile for the cohorts concerned. Since we do not have sufficient evidence for assuming an
exact quantitative relationship between a lower age at leaving school and
a lower mean age at childbearing (see discussion above), we found it safer
to make the assumptions in terms of a certain assumed shift in the age
pattern of fertility. There may be reason to assume that a lower schoolleaving age will not fully translate into a lower childbearing age. Hence
this scenario may be viewed as a high case. The effect of a one-year decline in the mean age at childbearing would be half of the effect calculated
under this scenario.
Scenario 4 studies the case of a school reform that also affects the
quantum of fertility. Here we investigate the likely possibility of a tempo–
quantum interaction on fertility, where the cohorts affected by the school
reform are not only two years younger at the time of childbearing, but
also have higher cohort fertility.6 We base our estimates on Kohler, Skytthe,
and Christensen (2001), who find that a one-year earlier initiation of childbearing increases cohort fertility by 3 percent. Thus our assumption of a
school reform that causes a two-year drop in the age at first birth would
lead to a 6 percent increase in fertility in the “reform with quantum effect” scenario.
Finally, Scenario 5 looks at the case of a school reform with a quantum effect in addition to the assumptions in Scenario 2. Here we assume
(as in Scenario 2) that the postponement of childbearing will end immedi-
716
POLICIES ADDRESSING THE TEMPO EFFECT
TABLE 1 Assumptions about the future course of period TFR and
about the mean age at childbearing in five scenarios and results for
the number of births, the old-age dependency ratio, and population
size, Austria, Bavaria, and Italy, 2005–2050
Scenario
2005
2020
2035
2050
Assumptions
Total fertility rate and
mean age at childbearing
Austria
S1
S2
S3
S4
S5
1.41
1.69
1.41
1.41
1.69
(28.3)
(28.3)
(28.3)
(28.3)
(28.3)
1.41
1.69
1.55
1.61
1.93
(28.3)
(28.3)
(27.4)
(27.4)
(27.4)
1.41
1.69
1.41
1.49
1.78
(28.3)
(28.3)
(26.3)
(26.3)
(26.3)
1.41
1.69
1.41
1.49
1.78
(28.3)
(28.3)
(26.3)
(26.3)
(26.3)
Bavaria
S1
S2
S3
S4
S5
1.36
1.62
1.36
1.36
1.36
(30.2)
(30.2)
(30.2)
(30.2)
(30.2)
1.36
1.62
1.51
1.54
1.86
(30.2)
(30.2)
(29.3)
(29.3)
(29.3)
1.36
1.62
1.37
1.45
1.73
(30.2)
(30.2)
(28.2)
(28.2)
(28.2)
1.36
1.62
1.36
1.44
1.71
(30.2)
(30.2)
(28.2)
(28.2)
(28.2)
1.24
1.51
1.24
1.24
1.51
(30.3)
(30.3)
(30.3)
(30.3)
(30.3)
1.24
1.51
1.41
1.46
1.77
(30.3)
(30.3)
(29.4)
(29.4)
(29.4)
1.24
1.51
1.24
1.32
1.61
(30.3)
(30.3)
(28.3)
(28.3)
(28.3)
1.24
1.51
1.24
1.32
1.60
(30.3)
(30.3)
(28.3)
(28.3)
(28.3)
Italy
S1
S2
S3
S4
S5
Results
Births
Austria
S1
S2
S3
S4
S5
72,038
86,609
72,041
72,041
86,609
62,695
76,009
67,715
70,398
86,004
49,350
69,809
49,930
53,188
75,048
41,359
60,904
42,884
47,231
70,150
Bavaria
S1
S2
S3
S4
S5
105,283
119,303
105,283
105,283
119,303
89,300
104,943
99,527
101,476
120,892
69,395
88,278
67,787
71,865
92,993
56,496
77,811
60,039
64,513
87,652
Italy
S1
S2
S3
S4
S5
489,260
595,665
489,260
489,260
595,665
344,013
421,158
387,157
399,019
492,987
292,399
416,919
286,038
303,656
438,400
212,067
316,832
223,279
243,012
368,651
/continued...
717
WOLFGANG LUTZ / VEGARD SKIRBEKK
TABLE 1 (continued)
Scenario
2005
2020
2035
2050
Old-age dependency ratio (65+/15–64)
Austria
S1
0.24
S2
0.23
S3
0.23
S4
0.23
S5
0.23
0.31
0.30
0.30
0.30
0.30
0.52
0.48
0.51
0.51
0.47
0.56
0.49
0.54
0.54
0.47
Bavaria
S1
S2
S3
S4
S5
0.24
0.24
0.24
0.24
0.24
0.28
0.28
0.28
0.28
0.28
0.48
0.46
0.48
0.48
0.45
0.49
0.45
0.48
0.48
0.43
Italy
S1
S2
S3
S4
S5
0.28
0.28
0.28
0.28
0.28
0.34
0.33
0.34
0.34
0.33
0.52
0.49
0.52
0.52
0.48
0.69
0.60
0.67
0.66
0.57
Population size
Austria
S1
S2
S3
S4
S5
7,975,335
8,036,270
7,954,387
7,954,387
8,036,270
7,718,934
7,941,119
7,728,071
7,738,721
8,040,808
7,197,142
7,708,453
7,283,157
7,338,575
7,905,116
6,202,044
7,017,948
6,326,473
6,435,905
7,329,567
Bavaria
S1
S2
S3
S4
S5
12,036,100
12,094,979
12,036,100
12,023,897
12,096,228
11,258,753
11,586,957
11,315,099
11,323,467
11,667,621
10,125,841
10,719,132
10,254,331
10,317,856
10,920,529
8,423,244
9,334,737
8,573,865
8,640,079
9,663,439
Italy
S1
S2
S3
S4
S5
56,670,032
57,254,477
56,670,032
56,670,032
57,253,746
52,126,321
54,100,986
52,417,109
52,453,609
54,496,983
46,390,530
49,680,383
46,966,837
47,235,320
50,847,809
39,860,844
44,917,031
40,428,714
40,961,171
46,504,488
718
POLICIES ADDRESSING THE TEMPO EFFECT
ately; to this assumption we add the effect of the school reform, including
the effect of associated tempo–quantum interactions. Clearly, this scenario
has the highest assumed future fertility rates. Table 2 presents a summary
of the assumptions of the five scenarios.
As the projection results in Table 1 show, the absolute number of births
is declining in all scenarios in each of these countries. This is because fertility, even under the highest scenarios, will still be below replacement level
and because smaller cohorts of women will enter the reproductive ages as a
consequence of the low fertility in the past (the negative momentum of
population growth).
The old-age dependency ratio shown in the table is defined as the population above age 65 years divided by the population aged 15–64. Since this
ratio takes into account only adults and older persons, one would expect a
delayed effect of changes in the birth scenarios. And indeed our calculations show a clear variation among scenarios beginning around 2030. After
that the differences turn out to be quite sizable. In Bavaria and Italy, the
lowest fertility scenarios (Scenarios 1 and 3) result in old-age dependency
ratios that increase from initial levels of 0.24 and 0.28, respectively, to about
0.5 and 0.7 by 2050, whereas in the cases of a declining mean age at childbearing, assumed to be a consequence of school reform combined with an
end of the tempo effect (Scenario 5), these ratios would increase less rapidly, to about 0.4 and 0.6.
Table 1 also shows the changes in total population size attributable to
the posited fertility changes under otherwise identical mortality and migration assumptions. For Austria, Bavaria, and Italy, none of the scenarios can
stop the significant population decline that is implied by the declining number of births as described above. But the extent of decline is still surprising,
even when one considers that it is occurring in a closed population (one
without migration). Under all scenarios, the population of Bavaria would
decline from a current level around 12 million to between 8.4 and 9.7 mil-
TABLE 2
Summary description of scenarios
Scenario
Tempo
adjustment
S1
S2
S3
No
Yes
No
S4
No
S5
Yes
Educational reform—
2 years younger
school-leaving age
No
No
Yes, childbearing shifts
2 years toward younger ages
Yes, childbearing shifts 2 years
toward younger ages
Yes, childbearing shifts 2 years
toward younger ages
Tempo–quantum
interaction
No
No
No
Yes, cohort fertility
increases by 6%
Yes, cohort fertility
increases by 6%
WOLFGANG LUTZ / VEGARD SKIRBEKK
719
lion in 2050. For Italy, the extent of decline is even more pronounced, with
the population shrinking from 57 million to between 40 and 47 million by
2050. But the point of this exercise is not to look at absolute changes (for
this we need realistic migration assumptions), but rather, by comparing the
scenarios, to assess the relative impacts of possible school reform effects.
In summary, the five scenarios presented here suggest that changes in
the age at childbearing that might result from lowering the school-leaving
age by two years might have significant long-term effects on population
dynamics. If we compare the constant period TFR scenario (Scenario 1) with
the school reform plus quantum effect scenario (Scenario 4), we see that
the absolute number of births by 2020 could be 12 to 16 percent higher in
the case of school reform. In terms of the old-age dependency ratio, the
difference attributable to an assumed education reform is on the order of 1
to 3 percentage points in 2050. Considering what a single percentage point
means in terms of expenses for social security, these are very significant
long-term impacts that make a closer analysis of the effects of school reform on the mean age at childbearing a worthwhile effort.
Such tempo policies will have an even more significant long-term demographic effect if the mechanisms of the hypothesized “low-fertility trap”
are at work. In this case, pushing period fertility above a certain critical
level can help to avoid the feedback mechanisms that would tend to pull
the quantum of fertility to lower and lower levels, with additionally severe
consequences for the speed and extent of population aging.
Discussion
Population policies aimed at affecting the tempo of fertility are a new concept, and possibly a powerful and socially acceptable way to increase period
fertility rates where these rates are considered to be too low. As discussed
in Lutz, O’Neill, and Scherbov (2003) and Goldstein, Lutz, and Scherbov
(2003), this concept is based on a sound demographic rationale, but is far
from mature in terms of its possible social implementation. In this article
we focused on education reforms that would lead to a more efficient school
system, with a younger mean graduation age as one possible form of tempo
policy. Such reforms are currently being discussed as a possible means to
improve the supply of skilled young labor and to reduce the social cost of
education. If these reforms would also help stop the trend toward rising
mean ages at childbearing, which is desirable for individual health reasons,
they would also represent a potentially significant aggregate demographic
gain. The purpose of this article was to propose the approach and initiate a
discussion that we hope will result in many more attempts to substantiate
our claim and will also address factors other than length of education that
might affect the tempo of childbearing.
720
POLICIES ADDRESSING THE TEMPO EFFECT
In discussing school reforms aimed at lowering the age at graduation,
one often hears the argument that students would be less mature when
leaving school and their human capital would be lower. While maturity is
difficult to measure, the human capital argument can be evaluated and seems
to be misguided. Surveys show that countries with a one- to two-year earlier exit from secondary school often have the same or higher human capital levels than countries with higher graduation ages (Mullis et al. 1998). A
recent study revealed no differences in earnings or higher education attainment in a cohort of German students who, owing to a school reform, obtained the same lower secondary school degree with almost a year less
schooling (Pischke 2003). Likewise, school length does not affect student
performance in Switzerland, where students obtain the same degree after
12, 12.5, or 13 years depending on their canton of residence (Skirbekk 2006).
Instead, school inefficiencies that in some countries lead to lower human
capital levels are likely to be caused by such factors as the organization of
schooling, the number of hours taught per year, the type of teaching strategies applied, and the age at which students are separated into academic or
vocational training (Braathe and Ongstad 2001; Eurydice 2000; Weiss 1995).
Low period fertility is not only of concern in Europe. Several Asian
countries have TFRs below 1.5 and are confronted with the prospect of significant population aging as a consequence. The concern about low fertility
has been particularly pronounced in Japan, South Korea, and Singapore.
In recent years the governments of these countries have implemented policy
packages, including tax cuts, housing support, and cash benefits, that are
explicitly pronatalist in a way that would make them not easily acceptable
in contemporary Western Europe. But so far period fertility rates seem to
remain unaffected by such measures, or else the forces tending toward lower
and lower fertility overwhelm whatever positive effect such policies might
have. In South Korea, the total fertility rate recently fell below 1.2 despite
new pronatalist measures. In Singapore, the prime minister set up a special
task force under his personal supervision to deal with the issue of low fertility because past pronatalist policy packages did not result in fertility increases and period fertility continued to fall. With a high political priority
assigned to the issue and the evident failure of previous measures, those
Asian countries might consider the option of tempo-related policies. In those
countries, childbearing remains almost universally within wedlock and marriage is typically postponed until the couple can afford an apartment, which
means that they must first begin earning money in the labor market. In this
context providing student couples who are ready to marry and have children with subsidized campus housing and childcare to enable them to
achieve their childbearing desires might result in an increase in student fertility. During study time, young parents may have more flexibility for arranging their time and accommodating childcare than in the case of both
WOLFGANG LUTZ / VEGARD SKIRBEKK
721
partners trying to improve their status within highly competitive companies. Thinking of such possible policies in terms of the tempo effect in addition to the usual cohort perspective might lead to effective fertility-enhancing measures.
Finally, this article introduced the hypothesis of a “low-fertility trap.” If
the social and economic mechanisms assumed under this hypothesis are at
work and have the potential to exert a so-far-unexpected strong pull toward
lower and lower fertility, the consequences for the countries concerned would
be severe. Although more empirical work is needed to study the different
elements of this hypothesis, we see no reason to dismiss the hypothesis on
grounds of implausibility. Viewed in a long-term perspective, the record of
demographic transitions demonstrates that the balance of births and deaths
can be disturbed for many decades because fertility is strongly embedded in
the system of social norms, and demographic regimes can be very persistent
once they are well established. Many decades of birth rates only sluggishly
adjusting to declining mortality rates resulted in historically unprecedented
population growth. One cannot rule out the possibility that analogous forces
of social momentum, once a new low-fertility regime has been established,
will result in decades of “undershooting” birth rates, resulting in historically
unprecedented population aging and fertility-induced decline.
Notes
The authors thank Statistics Austria, Yvonne
Tollmann at the Bayerisches Landesamt für
Statistik und Datenverarbeitung, and the Italian National Institute of Statistics (ISTAT) for
providing the data for the population projections.
The authors acknowledge the financial support
provided by the European Commission’s Research Training Network entitled “Demographic
Sustainability and European Integration,” Contract No. HPRN-CT-2001-00234.
1 In the three regions we examine below,
almost everyone attends compulsory primary
schooling, over 90 percent attend at least some
secondary schooling, and an increasing share
of the population, currently about 20 to 30
percent, attend tertiary education (Eurostat
2005). Hence, education reforms that affect
schooling up to secondary levels will affect a
larger part of the population. On the other
hand, late childbearing is particularly pronounced among the highly educated, and
policy options to alter the university graduation age may be highly effective for the university educated.
2 Information available at «htttp://www.
bologna-berlin2003.de/pdf/bologna_
declaration.pdf».
3 For a discussion of the use of natural
experiments, see Rosenzweig and Wolpin
(2000).
4 Sources of data: Statistics Austria,
Bayerisches Landesamt für Statistik und
Datenverarbeitung, and the Italian National Institute of Statistics.
5 The annual rise in the mean age at childbearing in Austria is estimated to be 0.15 years,
which is calculated as the average change from
1997 to 2000. Bavaria’s annual rise in the
mean age at childbearing is 0.138 years, which
is calculated as the average annual change
from 2001 to 2002. Italy’s annual rise in the
mean age at childbearing is 0.142 years, which
is calculated as the average annual change
from 1991 to 2000.
6 This result may be compatible with the
results from Skirbekk, Kohler, and Prskawetz
(2004), who find no impact on the fertility out-
722
come of a lower biological age when leaving
school. This is due to the fact that a decrease
in the school-leaving age would increase the
POLICIES ADDRESSING THE TEMPO EFFECT
individual’s social age, which could lead to
higher cohort fertility levels.
References
Andersson, G. 1999. “Childbearing trends in Sweden, 1961–1997,” European Journal of Population 15(1): 1–24.
Bentolila, S. and A. Ichino. 2000. “Unemployment and consumption: Are job losses less painful
near the Mediterranean?,” Centre for Economic Policy Research Discussion Paper No.
2539.
Billari, F., P. Manfredi, and A. Valentini. 2000. “Macro-demographic effects of the transition to
adulthood: Multistate stable population theory and an application to Italy,” Mathematical
Population Studies 9(1): 33–63.
Black, S. E., P. J. Devereaux, and K. Salvanes. 2004. “Fast times at Ridgemont High? The effect
of compulsory schooling laws on teenage births,” NBER Working Papers 10911. Cambridge, MA: National Bureau of Economic Research. Can be downloaded at «http://www.
nber.org/papers/w10911».
Blossfeld, H.-P. and A. De Rose. 1992. “Educational expansion and the changes in entry into
marriage and motherhood: The experience of Italian women,” Genus 48: 73–89.
Blossfeld, H.-P. and J. Huinink. 1991. “Human capital investments or norms of role transition?
How women’s schooling and career affect the process of family formation,” American Journal
of Sociology 97: 143–168.
Bongaarts, J. 1994. “Population policy options in the developing world,” Science 263: 771–776.
Bongaarts, J. and G. Feeney. 1998. “On the quantum and tempo of fertility,” Population and
Development Review 24(2): 271–291.
Braathe, H. J. and S. Ongstad. 2001. “Egalitarianism meets ideologies of mathematical education—instances from Norwegian curricula and classrooms,” International Reviews on Mathematical Education 33(5): 1–11.
Cnvsu. 2001. Rapporto annuale sullo stato delle università 2001 (Annual Report on Universities
2001), available online at «www.cnvsu.it» (last access February 2005).
Council of Europe. 1999. Recent Demographic Developments in Europe. Strasbourg: Council of Europe Press.
Easterlin, R. A. 1980. Birth and Fortune. New York: Basic Books.
Education Commission of the States. 1994. Compulsory School Age Requirements. Washington,
DC: US Department of Education.
Esping-Andersen, G. 1999. Social Foundations of Postindustrial Economies. New York: Oxford University Press.
European Banking Federation. 2004. Letter, No. 15 (July). Brussels: The European Banking
Federation «http://www.fbe.be».
European Commission. 2005. “Confronting demographic change: A new solidarity between
the generations,” Green Paper. Brussels: Commission of the European Communities.
Eurostat. 2005. Key Data on Education in Europe 2005. European Commision. Education and Culture.
Eurydice. 2000. Key Data on Education in Europe 1999/2000. Brussels: European Commission.
Gauthier, A. 2002. “Family policies in industrialized countries: Is there convergence?,” Population 3: 447–474.
Global Entrepreneurship Monitor. 2004. “National Team Reports” «http://www.gemconsortium.
org/category_list.asp?cid=164» (Web site accessed 21 April 2005).
Goldstein, J., W. Lutz, and S. Scherbov. 2003. “Long-term population decline in Europe: The
relative importance of tempo effects and generational length,” Population and Development
Review 29(4): 699–707.
WOLFGANG LUTZ / VEGARD SKIRBEKK
723
Goldstein, J., W. Lutz, M. R. Testa. 2003. “The emergence of sub-replacement family size ideals
in Europe” Population Research and Policy Review 22: 479–496.
Kohler, H.-P. and D. Philipov. 2001. “Variance effects in the Bongaarts–Feeney formula,” Demography 38(1): 1–16.
Kohler, H.-P., A. Skytthe, and K. Christensen. 2001. “The age at first birth and completed fertility reconsidered: Findings from a sample of identical twins,” Working Paper WP-2001006. Rostock: Max Planck Institute for Demographic Research.
Kravdal, Ø. 2001. “The high fertility of college educated women in Norway: An artefact of the
separate modelling of each parity transition,” Demographic Research Vol. 5, Article 6.
Lenzen, D. 2003. Bildung Neu Denken! (Rethink Education!). Munich: Verlag Leske + Budrich for
the Vereinigung der Bayerischen Wirtschaft.
Lutz, W., B. C. O’Neill, and S. Scherbov. 2003. “Europe’s population at a turning point,” Science
299: 1991–1992.
Lutz, W. and S. Scherbov. 2003. “Can immigration compensate for Europe’s low fertility?,”
European Demographic Research Papers No. 1. Vienna: Vienna Institute of Demography
of the Austrian Academy of Sciences.
Lutz, W., D. Philipov, and S. Scherbov. 2005. “Components of natality trends in Europe,” manuscript. Vienna: Vienna Institute of Demography of the Austrian Academy of Sciences.
McDonald, P. 2002. “Sustaining fertility through public policy: The range of options,” Population 57(3): 417–446.
———. 2005. “Low fertility in Singapore: Causes, consequences and policies,” presentation at
the Forum on Population and Development in East Asia, Beijing, 16–17 May.
Mullis, I., M. Martin, A. Beaton, E. Gonzalez, D. Kelly, and T. Smith. 1998. Mathematics and
Science Achievement in the Final Year of Secondary School: IEA’s Third International Mathematics
and Science Study (TIMSS). Chestnut Hill, MA: TIMSS International Study Center, Boston
College.
OECD. 1999. Thematic Review of the Transition from Initial Education to Working Life: Final Comparative Report. DEELA/ED (99)11: 319–322. Paris: Organisation for Economic Co-operation
and Development, Directorate for Education, Employment, Labour and Social Affairs, Education Committee.
———. 2000. From Initial Education to Working Life: Making Transition Work. Paris: Organisation
for Economic Co-operation and Development.
———. 2003. “Low fertility rates in OECD countries: Facts and policy responses,” Social, Employment and Migration Working Papers 15. Paris: Organisation for Economic Co-operation and Development.
Pischke, J. S. 2003. “The impact of length of the school year on student performance and earnings: Evidence from the German short school year,” NBER Working Papers 9964. Cambridge, MA: National Bureau of Economic Research. Can be downloaded at «http://www.
nber.org/papers/w9964».
Planas, J. 1999. “School to work transition in Spain,” in D. Stern and D. Wagner (eds.), International Perspectives on the School-to-Work Transition. Cresskill, NJ: Hampton Press, pp. 413–422.
Rindfuss, R., L. Bumpass, and C. St. John. 1980. “Education and fertility: Implications for the
roles women occupy,” American Sociological Review 45: 431–447.
Rindfuss, R., S. P. Morgan, and K. Offutt. 1996. “Education and the changing age pattern of
American fertility,” Demography 33(3): 277–290.
Rindfuss, R., M. K. Choe, L. L. Bumpass, and N. O. Tsuya. 2004. “Social networks and family
change in Japan,” American Sociological Review 69(6): 838–861.
Rosenzweig, M. R. and K. I. Wolpin. 2000. “Natural ‘natural experiments’ in economics,” Journal of Economic Literature XXXVIII: 827–874.
Shimer, R. 2001. “The impact of young workers on the aggregate labor market,” Quarterly Journal of Economics 116(3): 969–1007.
Skirbekk, V. 2006. “Does school duration affect student performance? Findings from cantonbased variation in Swiss educational length,” Swiss Journal of Economics and Statistics 142(1).
724
POLICIES ADDRESSING THE TEMPO EFFECT
Skirbekk, V., H.-P. Kohler, and A. Prskawetz. 2004. “Birth month, school graduation and the
timing of births and marriages,” Demography 41(3): 547–568.
Testa, M. R. and L. Grilli. 2004. “The effects of childbearing regional contexts on ideal family
size in Europe: A multilevel analysis,” European Demographic Research Papers No. 4.
Vienna: Vienna Institute of Demography of the Austrian Academy of Sciences.
United Nations. 2003. World Population Policies. New York: United Nations, Department of Economic and Social Affairs.
UNESCO. 2003. Global Education Digest 2003: Comparing Education Statistics Across the World. New
York: United Nations Educational, Scientific and Cultural Organisation «http://portal.
unesco.org/uis/TEMPLATE/pdf/ged/GED_EN.pdf».
US Department of Health, Education and Welfare. 1965. State Law on Compulsory Attendance.
Washington, DC: US Department of Health, Education and Welfare, Office of Education.
Weiss, A. 1995. “Human capital vs. signalling explanations of wages,” Journal of Economic Perspectives 9(4): 133–154.
Policies Addressing the Tempo
Effect in Low-Fertility Countries
WOLFGANG LUTZ
VEGARD SKIRBEKK
The possible negative consequences of current low fertility levels are causing increasing concern, particularly in countries where
the total fertility rate is below 1.5. Social inertia and self-reinforcing processes may
make it difficult to return to higher levels
once fertility has been very low for some
time creating a possible “low-fertility trap.”
Policies explicitly addressing the fertility-depressing effect of increases in the mean age
at childbearing (the tempo effect) may be a
way to push up period fertility to somewhat
higher levels and help escape the “low-fertility trap” before it closes. Reforms in the
school system may affect the timing of childbearing by reducing the age at completion
of education. A more efficient school system,
which provides the same qualifications with
a younger school-leaving age, is potentially
capable of increasing period fertility and
hence exerting a rejuvenating effect on the
age composition, even if the levels of cohort
fertility remain unchanged. Such policies
may also have a positive effect on completed
cohort fertility.
WOLFGANG LUTZ is Leader, World Population Program, International Institute for
Applied Systems Analysis, Laxenburg, Austria.
VEGARD SKIRBEKK is a Research Scholar, World Population Program, International
Institute for Applied Systems Analysis, Laxenburg, Austria.