Marriage squeeze and changes in family formation: historical comparative evidence in Spain, France, and the United States in the twentieth century

Full text

MARRIAGE SQUEEZE AND CHANGES IN
FAMILY FORMATION: HISTORICAL
COMPARATIVE EVIDENCE IN SPAIN, FRANCE,
AND UNITED STATES IN THE XXTH CENTURY
Albert Esteve i Anna Cabré
244

MARRIAGE SQUEEZE AND CHANGES IN
FAMILY FORMATION: HISTORICAL
COMPARATIVE EVIDENCE IN SPAIN, FRANCE,
AND UNITED STATES IN THE XXTH CENTURY
Albert Esteve i Anna Cabré
244
Aquesta comunicació es va presentar al Population Association of America
2004 Annual Meeting Program
Centre d’Estudis Demogràfics
2004

Marriage Squeeze and Changes in Family Formation: Historical Comparative
Evidence in Spain, France, and United States in the XXth Century
Resum: Els efectes de la grandària de les cohorts en la formació de la família han estat
meticulosament estudiats, seguint el treball d’Easterlin, el qual identifica el mercat
laboral com el factor explicatiu. El present text proposa hipòtesis diferents, però alhora,
convergents: Amb el casament universal femení, les dones de cohorts de naixement
escasses, podrien casar-se més joves i amb més grans proporcions, que si el mercat
matrimonial fos el factor explicatori. Aquest tipus de casament restringit tindria uns
ràpids efectes estimulants en la nupcialitat femenina i per contra, petits efectes allà on
hi hagués un excés femení.
En dos recents treballs, els autors han desenvolupat mecanismes d’ajustament
satisfactoris pel segle XX a Espanya, pronosticant resultats en el canvi en les tendències
de la fertilitat interpretades per les cohorts nascudes després de 1980. Utilitzant les
darreres microdades comparables dels censos, mitjançant IPUMS-International, l’estudi
s’aplica a França i als Estats Units, on es busca generalitzar la comprovació. Aquests
casos són diferents per la seva cronologia i pels desequilibris de sexes i de moments
específics, com ara la post Segona Guerra Mundial.
Paraules Clau: Matrimoni, Mercat matrimonial, Família, Microdades.
Resumen: Los efectos del tamaño de las cohortes en la formación de la familia han sido
meticulosamente estudiados, siguiendo el trabajo de Easterlin, el cual identifica el
mercado laboral como el factor explicativo. El presente texto propone hipótesis
distintas, pero a su vez, convergentes: Con el matrimonio universal femenino, las
mujeres de cohortes de nacimientos escasos, podrían casarse más jóvenes y con más
proporciones, que si el mercado matrimonial fuese el factor explicativo. Este tipo de
matrimonio restringido, tendría unos rápidos efectos estimulantes en la nupcialidad
femenina y por el contrario, pequeños efectos allí donde hubiese un exceso de mujeres.
En dos recientes trabajos, los autores han desarrollado mecanismos de ajuste
satisfactorios para el siglo XX en España, pronosticando resultados en los cambios en
las tendencias de la fertilidad interpretadas para las cohortes nacidas después de 1980.
Utilizando los últimos microdatos comparables de los censos, mediante IPUMS-
International, el estudio se extiende a Francia y Estados Unidos, donde se pretende
generalizar la comprobación. Estos casos son distintos por su cronología y por sus
desequilibrios de sexos y de momentos específicos, como la post Segunda Guerra
Mundial.
Palabras clave: Matrimonio, Mercado matrimonial, Familia, Microdatos

Abstract:The effects of cohort sizes on family formation have been thoroughly studied,
following Easterlin’s seminal work, which identifies the labor market as the explanatory
factor. The present paper proposes a different but converging hypothesis: with universal
female marriage, women in shrinking birth-cohorts would marry younger and in greater
proportions, that is, the marriage market would be the explanatory factor. This kind of
marriage squeeze should have rapid stimulating effects on female nuptiality, contrary to
small effects where there is an excess of females.
In two earlier works, the authors have developed the mechanisms of adjustment and
tested them successfully for 20th Century Spain, predicting from findings a reversal of
fertility trends performed by the cohorts born after 1980. Using recent comparable
census microdata, through IPUMS-International, the study is extended now to France
and United States, where we seek to generalize the proof. These cases differ by their
chronologies and by the imbalances of sexes at specific moments, such as post World
War II.
Key works: Marriage, Marriage market, Family, Microdata.

CONTENTS
The marriage squeeze………………………………...………………………..…..... 1
Excess of men – scarcity of women: predictable effects……………….………...…. 4
The Spabish Experience: a prototypical case……………………………………….. 7
Generalizing the proof: historical evidence from France and the US
France ………………………………………………………………………….. 8
The United States……………………………………………………………….12
Discussion and further directions……………………………………………………15
Reference …………………………………………………………………………. 16
FIGURES CONTENTS
1. Live Births Spain 1930-1986……………………………………………………. 7
2. Proportion of never married by cohort (Spain 1986)……………………………. 8
3. Live Births France 1899-2002…………………………………………………… 8
4. Proportion of never married by cohort (French Census 1962)…………………... 9
5. Mean age difference between spouses by male
and female cohort (French Census 1962)……………………………………… 10
6. Age difference by Male Cohort 1917-1921 (French Census 1962)…………… 11
7. Age difference by Male Cohort 1911-1916 (French Census 1962)…………… 11
8. Age difference by Male Cohort 1906-1910 (French Census 1962)…………… 11
9. Live Births United States 1909-1998………………………………………….. 12
10. Proportion of never married by cohort (US Census 1980)……………………. 12

11. Mean age difference between spouses by male
and female cohort (US Census 1980)…………………………………………... 13
12. Age difference by Male Cohort 1935-1939 (US Census 1980)…………… 14
13. Age difference by Male Cohort 1929-1934 (US Census 1980)…………… 14
14. Age difference by Male Cohort 1924-1928 (US Census 1980)…………… 14

1
The Marriage Squeeze
An imbalance in the numbers of males and females at prime marriage ages
produces a marriage squeeze effect. Marriage patterns should reflect preferences in mate
selection when the sex-age composition of the marriage market is in perfect
equilibrium, which means the same number of males and females at all ages. The
optimum scenario, however, is not always achieved. For instance, sex differences in
mortality or migration patterns can easily create disequilibria in local marriage markets
by decreasing or increasing the population of males or females of certain ages. History
provides some illustrative examples on how war related mortality dramatically reduced
the number of males. In a classical work, Louis Henry (1966) explored for France how
the heavy lost of males during the First World War affected the marriage behavior of
both male and female cohorts at marriageable ages. Despite the fact that women were
trapped in a severe marriage squeeze, the proportion never marrying hardly changed.
This paradox encouraged Henry to further explore the mechanisms by which the
squeezed cohorts were able to overcome the situation without modifying excessively the
proportions never marrying.
The marriage market equilibrium can also be altered by in- and out-migration
movements. Thus, transoceanic migrations have changed the marriage markets of origin
and destination societies: in Europe the marriage market was favorable to men, whose
numbers were reduced by migration, in America and Oceania, final destinations of these
migrants, women were scarce. The massive male immigration to certain localities in
Australia during the XIX century provides an interesting example of the scarcity of
potential brides, which had important effects on marriage levels. On the other hand, in
rural-urban migrations, women have tended to migrate to urban areas in greater
proportions than men, increasing the level of bachelorhood in rural areas.
Along with mortality and migration, variations over time in the number of live
births, depending on strength and length, can also affect, the age-sex structure of the
marriage market a couple of decades later, as larger or smaller cohorts enter the
marriage market. From a theoretical point of view, isolating the effects of differential
mortality and migration by sex, if males and females marry within the same age, the sex
ratio at birth would explain one hundred per cent of the sex structure of the marriage

2
market. Given, though, that almost universally men marry younger women, an increase
in the number of births will lead to a scenario where the size of a male cohort will be
smaller than subsequent female births cohorts. On the contrary, when the number of
births decreases, male birth cohorts will be larger than subsequent female cohorts.
Beyond the causes that reside behind a marriage squeeze, disequilibria in the
marriage market can be grouped on the basis of which sex squeezed. Therefore, two
different types can be created: (1) excess of women - scarcity of men; (2) excess of men
-scarcity of women. To the best of our knowledge, little has been done to specifically
compare both types of marriage squeeze. Assuming exactly the same length and
strength, does the scarcity of men produce the same effects on the marriage patterns as
the scarcity of women? To answer this question, let’s first briefly explore what are the
main consequences of a marriage squeeze.
Marriage squeeze effects can be tested using theoretical models. As shown by
Shoen (1983), in a fictitious marriage market, changes in age-sex composition influence
marriage behavior by changing the level and distribution of marriages. An interesting
conclusion derived from Shoen’s work, but not later developed by the author, is that a
decrease in the number of births, which catches men in a marriage squeeze, has greater
effects than an increase of births, which produces the inverse situation. Thus, this idea
implies that there are somehow asymmetric effects on the marriage market depending
on the kind of marriage squeeze.
Despite conclusions arising from models, empirical analyses of the marriage
squeeze are less conclusive. A marriage squeeze does not alter the level (proportion of
ever-married) or the distribution (difference between male and female mean ages at
marriage) of marriages in the way that would be expected from a theoretical model.
Squeezes in the marriage market have been solved through different mechanisms. For
instance, it is commonly stated that the rapid demographic growth occurred in Sub-
Saharan countries has contributed to polygamy. But in those societies where polygamy
does not exist, marriage squeeze has to be solved in a different way. It is generally
observed that the level of marriages is rarely affected by a marriage squeeze, even in
severe situations. Henry (1966) demonstrated that a strong difference in age-sex
composition, produced by large losses of men during First World War, had a
surprisingly slight effect on the level of marriage. Societies, concludes Henry, adapt

3
themselves easily to the sex-age disequilibria by modifying mainly the distribution of
marriages. Brides and grooms appear to adapt to, rather than to be constrained by, the
age distribution of partners available. The marriage market is flexible; age preferences
seem to be far from rigid. The conclusions reached by Bhrolchain (2000), in her study
of English and Welch females born between 1917 and 1922 caught in a marriage
squeeze, and Schoen (1983), in his studies on the US, point in the same direction. The
twentieth century US experiences indicate that the marriage squeeze had also little
effect in the level of marriage, that is the proportion of ever married, but a considerable
effect on its distribution. As some of the examples cited above suggest, the main
mechanism by which a specific population overcomes a marriage squeeze is by
changing the distribution of marriages. Age preferences flexibly accommodate
substantial changes in the number of partners available. Age preferences in the marriage
market are shaped by opportunity. Moreover, historical differences between male and
females mean age at marriage must not be seen as a sign of esthetic preferences, but as
the current balance of infinite age adjustments from the past (Cabré 1993). The almost
universal fact that men are older than their spouses seems to be the consequence of a
universal past too, based on excess female mortality caused by childbearing and for the
greater trend to remarry among widowers. This argument differs from that of Guttentag
and Secord (1983), who postulate that the observed mean age difference between
spouses is due to a balance of power.
In the case of Brazil, with increasing cohorts and decreasing mortality, younger
cohorts are larger than older cohorts, creating a constant ‘excess of women - scarcity of
men’ marriage squeeze type, which according to Green and Rao (1992) is considered to
be one of the causes behind the rise of consensual unions in that country.
Finally, the degree of change to be expected from a marriage squeeze depends
on its tightness and context where it takes place. Measuring the tightness of the
marriage squeeze has been a difficult task. No method offers a perfect solution that
takes into account all the complexities of the marriage market. The simplest measure is
the sex ratio that compares men at certain ages with female x ages younger, x being the
observed mean age difference between spouses. The main advantage of the sex ratio is
its easy and straightforward interpretation, which is always to provide an illustrative
picture of what is the strength of the imbalance. But comparing men to women x year

4
younger is not representative of the real distribution. In fact, a difference of 2-3 years is
not the dominant factor.
A more refined measure is to compute a ratio of men to women, but the total
number at each age is weighted by the probability of marriage (Akers 1969). By taking
into account probabilities for men and women separately the results change, because
somehow both probabilities are shaped by opportunity: the weighted sex ratio is
partially influenced by what it pretends to measure. Further developments have been
made to consider exclusively the single population or to break down the marriage
market in groups, based on sex, age, and educational characteristics. However, all these
measures, as Mc Donald points out (1995), suffer as well from the circularity problem.
Excess of men - scarcity of women: predictable effects
In the literature, marriage squeeze has been often analyzed in particular
conjunctures, such as heavy losses of men by exceptional mortality or in cases of local
or specific populations affected by imbalances due to migration, usually as a first step
for studying ethnic intermarriage (McCaa 1993). To our knowledge, little has been done
to study the effects of marriage squeeze caused by sharp and lasting decreases in
fertility, and occasional findings concerning these episodes have remained unexplained
under this particular angle (such as the puzzling continuous reduction of age at marriage
for only females in France during the fist third of the 19th century, stated by Henry et
Houdaille). However, this kind of marriage squeeze, characterized by the relative
scarcity of women, would be of particular interest because of its rapid stimulating
effects on nuptiality trends, contrary to its opposite, the female excess, which has been
proven to have effects much lower than expected, as shown in some of the works
mentioned above.
In recent decades, many developed countries have experienced dramatic
decreases in births. Some developing countries have followed them in this way and
many others will do so in the current or following decade. The case of decreasing births
is going to be experienced at one moment or another, at one speed or another, by most
countries of the World and by the ensemble of the World population, probably in the

5
first third of the 21st century. The relative scarcity of women caused by decreasing
fertility will predictably be aggravated by better survival rates, maintaining the
advantage in numbers of males at birth until ages progressively higher. Considering,
furthermore, that large areas of the world, mainly in Asia, are practicing extensive sex-
selection against the girl-child, the relative scarcity of women points as a common
sensible point in future marriage markets across different continents, excluding
substantial demographics-based transnational compensations.
Therefore, if the relationship between relative scarcity of women and increasing
trends in nuptiality and fertility could be properly established, this knowledge could
help to predict futures changes in these variables, or at least to signal in advance
elements influencing their trends. The present contribution is only one step in this
direction, but not the first one. In two former works, Cabré (1993 and 1994) has
developed the mechanisms of the adjustments, triggered by female relative scarcity and
has tested them successfully in 2Oth Century Spain, focusing on the nuptiality and
fertility comparative records of the cohorts 1930-1945, which were prototypical of the
theoretical case. The findings on these cohorts, parents of the Spanish “baby-boom”
(1955-1975), have predict a reversal of nuptiality and fertility trends for the cohorts
born after the dramatic drop of births occurred in late seventies when their members
attain the age of reproduction. In some of these cohorts men outnumber women three
years their junior by as much as 4 to 3.
The predicted changes proposed by Cabré are the following:
a) Not finding mates in the traditional three-years-younger age range (on average),
men will seek younger women (or the few closest-in-age older women still available).
b) In doing so, the first excess male cohorts will invade the traditional marrying
grounds of the younger cohorts, themselves in a situation of relative gender excess,
increasing thus the competition amongst men and causing the anticipation of options
and the general preference of men for marrying at younger ages.
c) Women will receive marriage proposals in greater number and at younger ages. It is
likely that they will accept marrying younger, but their preferences as to the older or
younger candidates are unclear.
d) Should female preferences go to older candidates, the difference in ages between
spouses will increase, favoring traditional marriage structures where gender status and
age status act in the same direction. This is likely to happen in the less educated
segments of the population.

6
e) If female preferences go the younger closer-aged candidates, older candidates
trapped by the change in trend will seriously risk becoming permanent bachelors.
Nevertheless, the option for same-age partners seems more probable among the more
highly educated segment. Older single educated man could chose mates among less
educated younger women, passing all the burden on younger men with less education,
who would then face very high chances of not marrying or marrying very late.
f) In the fifties and sixties, many man in the situation described above migrated to
other countries, entering more favorable marriage markets and improving their
competitive value in their own country. Nowadays, it is more likely that the
immigration of foreign women will help balance the sexes.
g) Cohabitation out of marriage will loose its attraction for men, being considered a
more precarious form of union as compared to marriage.
h) Male-caused divorce should slow its increase or even decrease, while female-
initiated divorce could rise quite sharply. The final result is unclear.
i) Widow or divorced women will experience a substantial increase in their chances of
remarrying, now quite low, while widowers and divorced men will encounter added
difficulties in remarrying, particularly if they desire to mate with young, never-
married women.
j) All the above should generate a more family-friendly set of values and a consistent
rise in fertility.
What we now intend to do is to extend the study to France and the United States,
seeking a wider generalization of the proof in cases which differ by their chronologies
and by the imbalances of their marriage markets at sensible moments such as post
World War II. Contrary to Spain, USA and France experienced the post-war baby-
boom, which certainly cannot be simply explained by one single factor but which might
keep some relation with marriage squeeze, especially in the second part of it, when the
decreasing cohorts born during the thirties entered their reproductive life. Also, France
and the US are different as to the situation of the marriage markets when their baby-
boom started: France had a very unbalanced market, marked by the deep scars in birth
statistics caused by World War one, and a “tail” of marriage squeeze could have easily
fueled the baby-boom there, while the US had a more regular and well-balanced market.
Given these differences among the three countries, comparisons are expected to
enlighten us with respect to the existence and possible effects of marriage squeeze.

7
The Spanish Experience: a prototypical case
A structural excess of women, for different reasons, has historically
characterized the Spanish marriage market. The increasing tendency in the number of
births during the century, the higher male mortality rates, and male out migration rates
are behind the historical excess of women. However, the consistent deficit of women
has appeared three times during the century, but we will focus on the effects produced
by the decrease of births during the late thirties (Figure 1).
As it has been stated before, some of the mechanisms explained before were
already verified in the “experimental” Spanish cohorts 1939-1945: proportions of ever-
married went significantly up for women and somewhat down for men (Figure 2), age
differences increased, age at marriage decreased very substantially for both sexes, with
the decline initiated among females, several years before males; fertility increased; and
men out-migrated more than women. Study of differentials in adjustments by education
or social class was not done but can be now undertaken using census microdata that
were previously unavailable. Other variables mentioned above, such as divorce or
cohabitation, were not relevant given the legal framework existing in Spain prior to
1978.
Figure 1. Live Births Spain 1930 - 1986
Source: Movimiento Natural de la Población Española (INE) (Vital Statistics)
The good results of the test on Spain led to another verification (Cabré 1994): in
the case of the Spanish region of Catalonia, where fluctuations in the numbers of births
had been sharper, the changes undergone by all the tested variables were also higher
than in the whole Spain and always in the expected direction.
350,000
450,000
550,000
650,000
750,000
1925 1935 1945 1955 1965 1975 1985

8
Figure 2. Proportion of never married by cohort ( Spain 1986)
Source: INE Padrón municipal de habitantes 1986.
Generalizing the proof: historical evidence from France and the US
France
In the course of the twentieth century, France has experienced two major
contractions in the number of births. The first, and most striking, occurred during the
First World War. This paper focuses exclusively on the effects of this first birth dearth.
As shown in Figure 3, in 1916 the number of births had fallen to half of the number
observed in 1913. Although the reduction was severe, it only affected a couple of years.
By 1920, the number of births had returned to the level prior to the war. This
contraction in birth rate and its subsequent effects on the marriage market have not been
examined in detail. It is not necessary to compute highly sophisticated measures to be
persuaded that men born between 1912 and 1915 were caught in a severe marriage
squeeze.
Figure 3. Live Births, France 1899-2002
Source: INSEE, par Vallin, J. et Mesle, F. "Tables de mortalité françaises pour les XIXè et XXè siècle et projections
pour le XXIè siècle", INED
5.00%
6.00%
7.00%
8.00%
9.00%
10.00%
11.00%
12.00%
13.00%
14.00%
15.00%
1899
1902
1905
1908
1911
1914
1917
1920
1923
1926
1929
1932
1935
1938
1941
1944
1947
1950
Men
Women
200,000
300,000
400,000
500,000
600,000
700,000
800,000
900,000
1,000,000
1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

9
Data comes from the IPUMS microdata for the 1962 French 5 % census sample.
Census microdata are an invaluable source for social science research. However, some
precautions have to be taken in account when using a later census to explore an
unbalanced marriage market in a somewhat distant past. Mortality can change the
distribution of the age differences between spouses as measured in a subsequent census.
The influence of divorce, separation and remarriage also affect the observed distribution
of marriages. In this paper we do not address these potential biases, given that we do not
think that they affect significantly our results. In 1962, the 1912-1915 birth cohort was
46 to 49 year old. We do not expect at these ages major mortality effects nor high age
differences in their frequency of marriage.
The male marriage squeeze in France had a small but persistent effect on the
proportion of males and females ever married, as shown in Figure 4. By comparing
different cohorts at the same moment, we run the risk of mixing age and cohort effects,
but, in spite of this, some observations can be made. Differences between male and
female proportions of never married are bigger for cohorts born after 1914. With low
and constant female levels of never-married, male’s levels increase, which could be
some indication of the predicted effects.
Figure 4. Proportion of never married by cohort (French Census 1962)
8.00%
8.50%
9.00%
9.50%
10.00%
10.50%
11.00%
11.50%
12.00%
12.50%
13.00%
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
Source: Matthew Sobek, Steven Ruggles, Robert McCaa, et al., Integrated Public Use Microdata Series-
International: Preliminary Version 1.0. Minneapolis: Minnesota Population Center, University of Minnesota, 2002.
The distribution of marriage was notoriously changed by the imbalance in the
number of males and females available. The French Census of 1962 does not provide
data on the age of first marriage. However, by comparing age between spouses, it is
possible to compute the mean age difference for each male and female cohort. Male
cohorts caught in a marriage squeeze increased the mean age difference between
Men
Women

10
spouses (Figure 5). The mean age difference between spouses is greater for male
cohorts born between 1912 and 1917 than previous or later cohorts. This altered the
expected age-distribution of this variable in balanced situations, which usually shows an
increase by age. Female cohorts 1914-1919 also show an increase in the mean age
difference. A closer look on how male cohorts reacted to the marriage squeeze is
provided in Figures 6 to 8. In the absence of any major disturbance, the spouse’s age
distribution does not change significantly among consecutive cohorts.
Figure 5. Mean age difference between spouses by male and female cohort (French Census 1962)
2.50
2.70
2.90
3.10
3.30
3.50
3.70
3.90
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
Source: Matthew Sobek, Steven Ruggles, Robert McCaa, et al., Integrated Public Use Microdata Series-
International: Preliminary Version 1.0. Minneapolis: Minnesota Population Center, University of Minnesota, 2002.
The spouses’ age distribution of the male French cohorts affected by the First World
War birth decline, especially those born between 1911 and 1916, were convincingly
altered because the women that they had a priori to marry belonged to diminished
cohorts. Some of the distributions are bimodal. As the cohorts move away, they tend to
regain the spouses’ age distribution observed for the oldest cohorts. The imbalance in
the marriage market has been primarily solved by changes in the distributions of
marriages, although the level of marriages also was slightly altered in the expected
direction.
Finally, the French experience can be compared to the Italy situation. Bartiaux
(1994) obtained similar results in her study of the effects that the birth decline in Italy
during the First World War produced on the marriage market a couple of decades later.
Men
Women
(
-
)

11
Figures 6, 7, 8 Age difference by Male Cohort (French Census 1962)
0
2
4
6
8
10
12
14
131211109876543210-1-2-3-4
Husband's age - wife's age
%
1921
1920
1919
1918
1917
0
2
4
6
8
10
12
14
131211109876543210-1-2-3-4
Husband's age - wife's age
%
1916
1915
1914
1913
1912
1911
0
2
4
6
8
10
12
14
131211109876543210-1-2-3-4
Husband's age - wife's age
%
1910
1909
1908
1907
1906
Source: Matthew Sobek, Steven Ruggles, Robert McCaa, et al., Integrated Public Use Microdata Series-
International: Preliminary Version 1.0. Minneapolis: Minnesota Population Center, University of Minnesota, 2002.

12
The United Sates
The United States has experienced two major contractions in the number of
births during the twentieth century. While in France the number of births, with the two
exceptions mentioned above, was kept close between seven and eight hundred
thousands births per year, the number of births in the United States did not follow the
same pattern. During the first half of the century the number of births scarcely exceeded
three million, while in the second half of the century this figure has always been higher,
reaching its maximum level in the baby boom years 1960-1964.
Figure 9. Live Births, United States 1909-1998
U.S. Public Health Service, National Center for Health Statistics, Vital Statistics of the United States, 1993, vol. I,
"Natality", Table 12; National Vital Statistics Reports, Vol. 48, No. 3, "Births: Final Data for 1998," Table 1.
In this paper we will focus on the birth decline occurring between 1924 and
1933, which was less sudden but longer than the French contraction (Figure 9). The US
did not reach the level of 1924 again until twenty years later, 1944. Data comes from the
IPUMS 5% sample of census microdata for the US 1980 Population Census.
Figure 10. Proportions of never married by cohort (US Census 1980)
4.00%
5.00%
6.00%
7.00%
8.00%
191
4
191
5
191
6
191
7
191
8
191
9
192
0
192
1
192
2
192
3
192
4
192
5
192
6
192
7
192
8
192
9
193
0
193
1
193
2
193
3
193
4
193
5
193
6
193
7
193
8
193
9
Source: Matthew Sobek, Steven Ruggles, Robert McCaa, et al., Integrated Public Use Microdata Series-
International: Preliminary Version 1.0. Minneapolis: Minnesota Population Center, University of Minnesota, 2002.
2,000,000
2,500,000
3,000,000
3,500,000
4,000,000
4,500,000
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
Men
Women

13
Figure 11. Mean age difference between spouses by male and female cohort (US Census 1980)
2.00
2.20
2.40
2.60
2.80
3.00
3.20
3.40
3.60
3.80
4.00
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
Source: Matthew Sobek, Steven Ruggles, Robert McCaa, et al., Integrated Public Use Microdata Series-
International: Preliminary Version 1.0. Minneapolis: Minnesota Population Center, University of Minnesota, 2002.
Male cohorts theoretically caught in a marriage squeeze did not present any of
the predicted effects. Moreover, the proportion of never married seems even to
contradict our initial expectations. For the squeezed birth cohorts, the male proportion
of never married decreases and increases for females (Figure 10). The distribution of
marriages does not show any kind of disturbance nor in the mean age difference
between spouses (Figure 11), nor in the spouses’ age distribution by male cohorts
(Figures 12 to 14). The US results open a range of questions related to the suitability of
using census data, analyzing data at aggregate level, and, most important, not taking in
account other factors that, especially in the US, can be distorting our results, such as
immigration, divorce, remarriage and changes in nuptiality patterns. The intensity of
these factors eroded the expected effects of marriage squeeze. A more detailed analysis
is needed to understand how male cohorts resolved their marriage squeeze, to the extent
that it existed.
Women
(
-
)
Men

14
Figures 12, 13, 14 Age difference by Male Cohort (US Census 1980)
0
2
4
6
8
10
12
14
16
131211109876543210-1-2-3-4
Husband's age - wife's age
%
1939
1938
1937
1936
1935
0
2
4
6
8
10
12
14
16
131211109876543210-1-2-3-4
Husband's age - wife's age
%
1934
1933
1932
1931
1930
1929
0
2
4
6
8
10
12
14
16
131211109876543210-1-2-3-4
Husband's age - wife's age
%
1928
1927
1926
1925
1924
Source: Matthew Sobek, Steven Ruggles, Robert McCaa, et al., Integrated Public Use Microdata Series-
International: Preliminary Version 1.0. Minneapolis: Minnesota Population Center, University of Minnesota, 2002.

15
Discussion and further directions
The results of this analysis do not point in the same direction. While the
prototypical case of Spain and to some extent the French experience, which differs
significantly in terms of strength and length of its marriage squeeze, verify the
adjustments predicted by Cabré (1993). The US experience does not, however, follow
the same pattern. Further research is necessary to explore the uniqueness of the US case.
Contextual aspects, such as the incidence of divorce, remarriage, and immigration, need
to be taken into account to explain differences among these cases. With regard to
Spain, an imbalance in the number of males and females produced by a birth dearth had
palpable effects on the marriage market. In other words, the Spanish marriage market
seems to be more sensible to unequal distributions of potential brides and grooms,
which probably correlates with the fact that it is a less flexible marriage market,
compared to the American or French. Certainly, the incidence of divorce and remarriage
is lower that in the other countries considered here.
On the other hand, migration also plays a key role in shaping the sex-age
structure of the marriage market by increasing or decreasing the size of birth cohorts.
Since the United States is a country of immigration and the flows are predominantly
male, we would expect that the marriage squeeze would be further aggravated rather
than attenuated by this circumstance. More research needs to be done in this direction.
The length and strength of the birth decline would affect the marriage market
differently when diminished cohorts reach marriageable age. It’s been observed, based
on the Spanish and French cases studied above and the cited case of Italy, that abrupt,
but short birth declines seem to have considerable effects on the marriage market. On
the contrary, gradual and less intense birth declines do not have the same effect, as
shown by the US experience.
To continues this line of research, we will search the historical record for
evidence of how societies have been solving the excess of men – scarcity of women
type of marriage squeeze, which we believe should occur in most of the developed and
less developed countries that have experienced dramatic decreases in births in recent
years.

16
References
AKERS, D.S., 1967. “On measuring the marriage squeeze”. Demography. 2: 907-924.
BARTIAUX, F., 1994. "Dénatalité et marché matrimonial. Le cas de l'Italie en 1930-
1950". Population. 1:91-118.
CABRÉ, A. 1993. "Volverán tórtolos y cigüeñas" a LUIS GARRIDO I ENRIQUE GIL
CALVO (eds.), Estrategias familiares. Madrid: Alianza Editorial, p. 113-131.
CABRÉ, A. 1994. "Tensiones inminentes en los mercados matrimoniales" a JORDI
NADAL (ed.) El mundo que viene. Madrid: Alianza Editorial, p. 37-62.
GREENE, M. E. & RAO, V. (1992) “A compressao do mercado matrimonial e o
aumento das unioes consensuais no Brasil”. Rev. Bras.Estudios Pop. Campinas,
9(2):168-182.
GUTTENTAG, M. & SECORD, P.F., 1983. Too many women?. Beverly Hills: Sage,
1983.
McCAA, R. 1993. "Gender in the Melting Pot: Marital Assimilation in New York City,
1900-1980", Journal of Interdisciplinary History 24(2):207-231.
McDONALD, P. 1995. "L'équilibre numérique entre hommes et femmes et le marché
matrimonial: le point sur la question". Population. 6:1579-1590.
NI BHROLCHAIN, M. 2000. “Flexibility in the marriage market”. Population. 55 (6):
899-939.
HENRY, L., 1966. “Perturbations de la nuptialité résultant de la guerre 1914-1918”.
Population. 2:163-166.
SHOEN, R., 1983. “Mesuring the Tightness of the Marriage Squeeze”. Demography
20:62-78.