www.asb.unsw.edu.

au

Last updated: 10/06/14 CRICOS Code: 00098G






Australian School of Business Research Paper No. 2014 ECON 29



It's Raining Men! Hallelujah?


Pauline Grosjean
Rose Khattar













This paper can be downloaded without charge from
The Social Science Research Network Electronic Paper Collection:
http://ssrn.com/abstract=2445285














Australian School of Business
Working Paper

1
Its Raining Men! Hallelujah?
Pauline Grosjean and Rose Khattar
*

3 June 2014
Abstract
We document the implications of missing women in the short and long run. We exploit a
natural historical experiment, which sent large numbers of male convicts and far fewer
female convicts to Australia in the 18
th
and 19
th
century. In areas with higher gender
imbalance, women historically married more, worked less, and were less likely to occupy
high-rank occupations. Today, people living in those areas have more conservative attitudes
towards women working and women are still less likely to have high-ranking occupations.
We document the role of vertical cultural transmission and of homogamy in the marriage
market in sustaining cultural persistence. Conservative gender norms may have been
beneficial historically, but are no longer necessarily so. Historical gender imbalance is
associated with an aggregate income loss estimated at $800 per year, per person. Our results
are robust to a wide array of geographic, historical and present-day controls, including
migration and state fixed effects, and to instrumenting the overall sex ratio by the sex ratio
among convicts.
Keywords: Culture, gender roles, sex ratio, natural experiment, Australia
JEL codes: I31, N37, J16, Z33


*
University of New South Wales. Corresponding author: Pauline Grosjean, School of Economics,
UNSW, Sydney NSW 2052, Australia. p.grosjean@unsw.edu.au. We are grateful to Chris Bidner,
Alison Booth, Monique Borgerhoff-Mulder, Sam Bowles, Rob Boyd, Rob Brooks, Gigi Foster,
Gabriele Gratton, Kim Hill, Hongyi Li, Leslie Martin, Suresh Naidu, Paul Seabright and audiences at
the Australian National University, the Institute for Advanced Studies in Toulouse, Hong Kong UST,
Monash University, University of Sydney, UNSW, the Santa Fe Institute and Sciences-Po for
comments and suggestions. We thank Renee Adams for financial help with this project. This paper
uses unit record data from the Household, Income and Labor Dynamics in Australia (HILDA) Survey.
The HILDA Project was initiated and is funded by the Australian Government Department of
Families, Housing, Community Services and Indigenous Affairs (FaHCSIA) and is managed by the
Melbourne Institute of Applied Economic and Social Research (Melbourne Institute). The findings
and views reported in this paper, however, are those of the authors and should not be attributed to
either FaHCSIA or the Melbourne Institute. All errors and omissions are ours.
2
1. Introduction
Traditional economic models link the division of labor between men and women to the
available technology, which determines the relative returns to male and female labor, and to
the conditions in the marriage market, which determine the outcome of bargaining between
men and women. However, recent work has highlighted the importance of cultural norms and
beliefs about appropriate gender roles (Fernandez 2009, 2014, Bertrand, Kamenica, and Pan
2013). This opens the question of how such beliefs emerge and persist. To answer this
question, Alesina, Giuliano and Nunn (2013) point to the technology that was available in
pre-industrial times. Past technology still matters because it shaped the division of labor
between men and women, which was then integrated into cultural norms that persisted over
time. Taking a natural step, we ask whether, and how, past shocks to the marriage market
have similarly enduring effects. Our focus on the marriage market also suggests specific
cultural persistence mechanisms, which rely on homogamy in marriage and on child
socialization within families, which we explore in details.
We exploit variation in marriage markets conditions that arose from an exogenous variation
in the sex ratio among an otherwise homogeneous population. Heavily male biased sex ratios
resulted from British policy to send convicts to Australia in the late 18
th
and 19
th
century.
Men far outnumbered women among convicts, by a ratio of 6 to 1 (Oxley 1996). This was at
a time when the physical skills required to build the country gave a large premium to male
labor. As can be seen in Figure 1, a male biased sex ratio endured in Australia for more than
a century due to male biased migration. It converged to parity after the end of the life span of
the last convicts, who were transported in 1868.
We rely on the variation across Australian counties in the historical sex ratio, measured
around the time at which convict transportation ceased. Historically, gender imbalance was
associated with women marrying more, participating less into the labor force, and being less
likely to work in high-ranking occupations. These results are robust in a panel of historical
counties, controlling for time and county fixed effects, which remove the influence of any
time invariant county characteristics that could be associated both with gender imbalance and
with female work outcomes. We then study the long-term implications by matching, for the
first time, 91 historical counties from the Australian Colonial Censuses to postal areas in the
3
1933 Census, the 2011 Census and in a nationally representative household survey. In areas
that had a larger imbalance historically, people today have more conservative attitudes
towards women working and women are less likely to have high-ranking occupations. This
glass ceiling effect was already present, and larger in magnitude, in 1933, before the onset
of large migration flows to Australia post-World War II.
A one unit increase in the historical sex ratio -one more man for a given number of women-
moves the average Australian today towards conservative attitudes by 6% at the mean. This
additional man is also associated with a 1 percentage point decrease in the share of women
employed as professionals, which represents nearly 5% of the population mean and 12% of
its standard deviation. Historical circumstances explain 5% of the variation in the share of
women employed as professionals that is left unexplained by traditional factors, even when
accounting for the share of men employed in similar professions.
We demonstrate an effect of gender imbalance: it is the local gender imbalance that has
shaped the patterns we document, not the convict settlement that generated the imbalance or
the overall gender imbalance in the economy. The results are unaffected by controlling for
the share of convicts in the overall population. Placebo specifications, in which sex ratios are
randomly allocated across counties while keeping the overall imbalance unchanged, give no
significant results. The results are also specific to views about women working; historical
gender imbalance does not explain sexism in general.
A concern for identification is that the gender imbalance was determined by characteristics
that still influence outcomes today. A careful study of historical circumstances enables us to
alleviate this concern. Convict transportation and the resulting gender imbalance reflected
British penal imperatives. Male convicts were in the majority because men caused more
crime than women in the United Kingdom. Even though prostitution was considered a crime,
it was not punished by transportation. Once in Australia, convicts would have to complete the
term of their sentence, generally 7 years, after which they were freed. However, they were
not confined to prisons. They either worked under the governments supervision or, later,
were assigned to employers. Potential employers were either free settlers or former convicts.
They would locate as a function of economic, mostly agricultural, opportunities. Convicts
were assigned to them as a function of their labor needs.
4
To account for the influence of economic opportunities on the sex ratio, we control for
geographical characteristics and for historical sectoral composition. We also control for state
fixed effects throughout, which remove any unobserved heterogeneity due to differences in
the legal environment or in the treatment of convicts. We also account for a wide range of
individual and contemporary controls, including the sectoral composition of the economy
today.
Our results are robust to an instrumental variable approach. We instrument the overall sex
ratio by the sex ratio among convicts, who had no choice where to locate. As before, we
remove the potential endogeneity in convict assignment by controlling for sectoral
composition. A legacy of a convict past independent of gender imbalance would violate the
exclusion restriction. We therefore control for the number of convicts together with historical
population size. Our IV estimates are consistent with our OLS estimates and similar in
magnitude.
Our historical results, contemporaneous to gender imbalance, are hardly surprising.
Economic and evolutionary biology models predict more conservative gender roles as a result
of male-biased sex ratios. Bargaining models namely predict that women will select a
wealthier partner, extract a better bargain and work less as a result (Grossbard-Schechtman
1984, Chiappori 1988, 1992, Chiappori, Fortin and Lacroix 2002). This is particularly
important when job opportunities for women are few. What is more surprising is that we find
that this effect has persisted to this day, when sex ratios have reverted back to parity.

We also find substantial implications for welfare. Women today are happier in their
relationship in areas that experienced more male biased sex ratios. Men, by contrast, are not.
However, women forego significant income. At the mean, historical gender imbalance is
associated with women foregoing $1,500 of yearly income. This is not fully compensated by
higher male income. A possible interpretation is that conservative gender norms are inherited
from a past in which they were adaptive because technology favored male labor, but are no
longer adaptive today.
Our interest is in whether and how past gender imbalance has shaped cultural beliefs about
gender roles. However, other mechanisms may explain persistence. For example, initial
5
gender imbalance may have influenced formal institutions, or distorted industrial
specialization into male-intensive economic activities. We rule out the first channel by
relying on within-country and even within-state variation, where formal legislation is
identical.
1
To account for the second channel, we control for industrial specialization today.
Another possibility is that past circumstances in the marriage market influenced respective
incentives of men and women to invest in education (Chiappori, Iyigun, and Weiss 2009).
We find very little evidence for this mechanism but still include education as a control.
Finally, strong sexual selection in the heavily male-biased situation that we describe may
have selected particular genotypes. While we cannot fully rule out this possibility, it seems
unlikely that genetic factors alone explain the large and systematic differences in attitudes,
labor force participation, and occupational choice that we document in this paper.
Even if we are confident that cultural beliefs explain our results, the question of how they
have persisted remains. Culture may persist because of the transmission of cultural traits
within families (vertical) or across unrelated individuals (horizontal) (Cavalli-Sforza 1981,
Bisin and Verdier 2001). We find, consistent with vertical transmission, that historical gender
imbalance is only associated with conservative views about gender roles among people born
of an Australian father. This is true both in high and low migration areas. Such differences
between people of different ancestry in the same areas also make it unlikely that the
relationship between historical gender imbalance and present-day outcomes is due to
unobservable local characteristics or to self-selection of people to localities on the basis of
taste.
Our focus on the marriage market suggests an additional persistence mechanism. Assortative
mating makes gender norms strategic complements among potential spouses. Strategic
complementarity implies that norms become evolutionarily stable (Young 1998) and that
even inferior conventions can persist over time, as shown theoretically by Tabellini (2008)
and Belloc and Bowles (2013). Accordingly, we find that historical gender imbalance is only
associated with conservative gender views in areas where homogamy, defined either along

1
The legal framework operating in Australia with respect to gender discrimination has been constant
across all states since the Sex Discrimination Act 1984 (Cth), which operates at a federal level. This is
a direct consequence of Australias Constitution, with any state laws inconsistent with this act invalid
to the extent of the inconsistency (Constitution s 109). The Family Law Act 1975 (Cth) unifies family
law in Australia at this federal level.
6
ethnic or cultural lines, is high. We also find evidence for a positive interaction between
homogamy and vertical transmission, which has a natural interpretation: parents want to
instill in their children norms that will make them attractive in the marriage market.
2

A potential criticism is that our results lack external validity because of Australias convict
history. We identify that our results are a legacy of gender imbalance, not convict heritage.
Besides, Australia only has a convict history because British convict transportation to the US
was no longer possible after US independence.
Our work is the first to shed light on the long-term effects of gender imbalance. We find that
a surplus of men reinforces conservative gender roles, keeps women under the glass ceiling,
and has substantial long-term welfare implications, with much lower independent incomes
for women. The effects persist for more than a hundred years, even though sex ratios have
long reverted to normal. This has important implications for the world today, where it has
been estimated that a hundred million women are missing (Hesketh and Xing 2006). This
problem is not limited to the most famous cases of China and India, but also concerns sub-
Saharan Africa, the US and Canada.
3
The sex ratio that our results are based upon (about 3
men for every woman) is higher than in China and India today
4
but comparable to that
prevailing in the West of the US at the end of the 19
th
century (2.13
5
). The study of the
determinants of such gender imbalances has attracted a large literature.
6
The study of its
consequences is more limited because of evident reverse causality issues. We discuss this
literature in more detail in Section 2 of this paper.
Most of the literature links variation in female labor force participation and gender roles
across regions and time in the available technology and conditions in the marriage market.
The rise in female labor force participation, the expansion of womens economic and political

2
Bisin, Topa and Verdier (2004) discuss different circumstances under which child socialization
effort and homogamy are complements or substitutes. However, we do not observe the child
socialization effort. We only observe the inheritability of cultural traits, which is the result of
socialization effort and environmental factors, so that we cannot comment on whether effort and
homogamy are substitutes or complements.
3
See Anderson and Ray (2010) for sub-Saharan Africa, Almond and Edlund (2008) for the US, and
Almond, Edlund, and Milligan (2013) for Canada.
4
The sex ratio prevailing in China and India today is estimated at 1.06 on average (Anderson and Ray
2010), but sex ratios at birth are much higher, at 1.20 and 1.09 respectively (Edlund and Lee 2009).
5
Authors calculation from US 1870 Census (Ruggles et al. 2010).
6
See Rao (1993), Edlund and Lee (2009), Chung and Das Gupta (2007), Hesketh and Xing (2006).
7
rights, as well as the reduction in fertility that has been observed in developed countries are
explained by technological change and the rise in returns to female labor (Goldin and Katz
2002, Greenwood, Seshadri and Vandenbroucke 2005, Doepke and Tertilt 2009, Doepke,
Tertilt and Voena 2012, Olivetti 2013). However, the stubbornness of the gender wage gap,
even in the most advanced economies, has pushed economists to consider the role of culture.
Fernandez (2008) and Fernandez and Fogli (2009) show that preferences for fertility and for
female labor force participation change slowly. This, our, and other work show that cultural
beliefs influence real work choices, family formation and welfare.
7
Alesina, Giuliano and
Nunn (2013) show that conservative gender norms stem from the introduction of plough
agricultural in pre-industrial societies. In contrast with this work, we illustrate a much more
rapid cultural change, we show that past conditions in the marriage market can shape cultural
beliefs and have long-term effects, and we discuss a particular mechanism that relies on the
specificities of the monogamous- marriage market.

2. Conceptual Background
Economic and evolutionary biology alike predict that a male biased sex ratio will result in
conservative gender roles. Economic marriage models predict that the bargaining position of
one gender is proportional to its scarcity (Becker 1973, 1974). Accordingly, Pollet and Nettle
(2008) find that the importance of mens wealth for marriage in the US at the beginning of
the 20
th
century is positively correlated with local sex ratios. Addressing the possible
endogeneity between local marriage conditions and local sex ratios, Abramitzky, Delavande
and Vasconcelos (2011) exploit variation in World War I related deaths in France. They find
that a shortage of men is associated with men marrying more and marrying up. Our results, as
well as Francis (2011) in the case of Taiwan after the influx of the Chinese Nationalist Army
in 1949, show that, conversely, a shortage of women leads to women marrying more.
8
An

7
Fortin (2005) shows how gender role attitudes influence labor market outcomes. Alesina, Giuliano
and Nunn (2013) establish a relationship between beliefs and participation of women in the economy
and in politics. Bertrand, Pan and Kamenica (2013) find that households in which women earn more
than men are less likely to form and, once formed, are more likely to lead to divorce.
8
Pollet and Nettle (2008) find, based on the 1910 US Census, that the effect of socioeconomic status
on mens marital success was stronger in states in which males outnumbered women. However, they
do not address the potential endogeneity of sex ratios.
8
improvement in womens bargaining position resulting from higher sex ratios is also
predicted to reduce female labor force participation (Grossbard-Schechtman 1984, Chiappori
1988, 1992, Chiappori, Fortin and Lacroix 2002). This is supported by empirical evidence
(Angrist 2002, Chiappori, Fortin and Lacroix 2002).
While economists put forward bargaining as the main mechanism through which the sex ratio
affects gender roles, evolutionary biologists discuss others. Males may respond to
competitive pressures by making themselves more attractive to females, which involves
resource transfers as in bargaining models, but not only (Kokko and Jennions 2008). Any
signal correlated with the future possible reproductive success of offspring, such as a
beautiful tail or the tidiness of a nest, will be valued. In humans, signals of wealth or of ones
commitment to provide economic resources will be attractive to women. Male-biased sex
ratios may also have negative consequences for females. These ensue when males divert
resources from the female and offspring towards competing with other males, or when males
engage in mate guarding, which restricts female freedom. However, in all these
circumstances the prediction applied to humans is that male-biased sex ratios will lead to
more conservative gender roles with females working less outside the home.
We study a particular society, in which the mating system consists of monogamous marriage.
Marriage markets are plagued with asymmetric information, which is a particularly severe
problem for women. As shown by the recent dating literature, men are principally attracted to
looks, an essentially observable characteristic. Fisman et al. (2006) show, in the context of
speed dating experiments, that women are attracted to intelligence, a proxy for income
generation capacity that is not as easily observable. Men have to signal their qualities to
potential spouses. While everybody easily imagines the behavior that the speed-dating
Americans in Fisman et al. (2006)s study adopt as a response, the circumstances of 19
th

century Australia may have led to another kind of behavior.
The opportunity cost of a bad match for a woman depends on the technology. It is
particularly high if womens economic opportunities outside of marriage are limited or
unattractive, as it was in 19
th
century agricultural Australia. In this context, convincing a
woman of ones commitment to work enhances marriage prospects. Adhering to the view that
a man should earn the money and a woman should stay at home, and providing enough so
9
that she does not have to work, may achieve this. Bargaining also predicts that women will
work less, particularly in an environment in which female labor is arduous and poorly
rewarded.
As argued by Alesina, Giuliano and Nunn (2013), conservative gender roles and low female
labor force participation imprint onto cultural norms about the appropriate role of women in
society. Gender roles of the past then persist in the long run because culture changes slowly.
Following Richersen and Boyd (2008), culture is defined as rules of thumb that affect
behavior in complex and uncertain environments and that people acquire from other people
through teaching, imitation and other forms of social transmission (Richersen and Boyd
2008, p. 5). Cultural traits that are successful, which in our context means getting a wife, will
spread.
The economic literature discusses two main channels of cultural transmission: horizontal and
vertical (Cavalli-Sforza 1981, Bisin and Verdier 2001). Culture spreads horizontally across
peers, mainly through imitation. Culture spreads vertically from parents to children, through
imitation and active parental socialization (Bisin and Verdier 2001, Doepke and Zilibotti
2008). Vertical transmission is inherently sticky. The implied hysteresis explains why
historical sex ratios may have persistent effects, even long after sex ratios have reverted back
to normal.
Our focus on the marriage market suggests another persistence mechanism. Assortative
mating on the marriage market implies that views about gender roles are strategic
complements among potential spouses. More similar individuals are more likely to be
married to each other, more likely to stay married (Becker et al. 1977, Heaton 1984, Lehrer-
Chiswick 1993) and are happier in their relationship, as we will show in the empirical part of
this paper. In the marriage market, if certain norms make matching more likely, and a match
more successful and happier, they will prevail and persist over time. The marriage market is
also one in which experimentation is costly, because time is often of the essence and
uncertainty is high. The conservative gender roles that were reinforced by the 19
th
century
Australian environment of gender imbalance and high premium on male labor, may then
persist in the long run solely because they are mutual best responses in a homogamous
marriage market. Young (1998) shows theoretically that norms that are mutual best responses
10
are evolutionarily stable. Positive feedbacks of this kind are at the core of the persistence of
even Pareto-dominated conventions in Belloc and Bowles (2013). Thus, conservative cultural
traits may persist, even when they are no longer adaptive once the sex ratio has reverted back
to normal and economic opportunities for women have increased.
Belloc and Bowles (2013) discuss conditions that make the transition from one convention to
another more likely. Cultural change has the characteristics of a collective action problem.
The greater the cost of deviating from a given set of cultural traits and the bigger the
population size, the less likely any cultural change will occur. Deviation and experimentation
may be particularly costly in the marriage market, where time is of the essence, uncertainty
substantial, and search costs relatively high. If holding modern views leads to long delays in
finding a spouse, people will conserve traditional views. However, immigration should make
experimentation easier and may accelerate transition towards modern gender views. This
does not necessarily require immigrants to hold modern views. Uncertainty should be enough
to lead more people to experiment with different views. Conversely, homogamy in the
marriage market, a proxy for the strength of strategic complementarity of gender views,
should be associated with stronger persistence of conservative views. We test both
predictions in the empirical analysis.
Homogamy will interact with vertical transmission in a very natural way, as discussed in
Bisin, Topa and Verdier (2004). Parents want to instill in their children norms that will make
them attractive in the marriage market. If they anticipate that the prevalence of conservative
gender views is high among potential spouses, parents will try harder to transmit such views
to their children.
It is also possible that genetic selection explains persistence. Male biased sex ratios
exacerbate sexual selection. The behaviors and cultural traits that were successful at securing
mates in 19
th
century Australia may have spread through natural genetic-selection, precisely
because they led to marital success. This is certainly a possibility but given the complex
nature of the interactions between genes, traits and behaviors, it is unlikely that this process
alone drives the specific and systematic differences in attitudes, labor force participation, and
occupational choice that we document in this paper.
11
There are other reasons, apart from its influence on cultural norms, why initial gender
imbalance may also have persistent effects. For example, initial gender imbalance may have
shaped formal institutions. We rule out this channel by relying on within-country and even
within-state variation, where formal legislation is identical.
9
Initial abundance of males and
low participation of the few females around in the labor force possibly distorted industrial
specialization into male intensive economic activities. Persistent effects until today may have
nothing to do with culture, but with the kind of jobs available. To account for this, we control
extensively for industrial specialization, historically and today. Bargaining conditions in the
marriage market may also influence educational choices by men and women (Chiappori,
Iyigun, and Weiss 2009). We find very little empirical evidence for this but still control for
education to account for this possible channel.

3. Historical Background: Gender Imbalance, Female Work and Marriage in 19
th

Century Australia
After independence of the United States, Australia became the new destination of choice for
the United Kingdoms overflowing jail population. Between 1787 and 1868, 132,308 and
24,960 convict men and women were transported to Australia, mostly to Tasmania and New
South Wales, which initially also included Queensland, the Australian Capital Territory, and
Victoria (Oxley 1996, p. 3). The majority of convicts were transported for property offences,
such as petty theft (Oxley 1996).
Even though natural births and free migration acted to reduce the extent of the gender
imbalance, the extent of free migration to Australia was rather limited until the 1830s and
the imbalance was sustained by ongoing convict transportation over nearly a century. Male
convicts consisted more than 80% of the adult male population of New South Wales in 1833.
The sex ratio stood at 3.54 men for every woman in the general population, and at more than

9
The legal framework operating in Australia with respect to gender discrimination has been constant
across all states since the Sex Discrimination Act 1984 (Cth), which operates at a federal level. This is
a direct consequence of Australias Constitution, with any state laws inconsistent with this act invalid
to the extent of the inconsistency (Constitution s 109). The Family Law Act 1975 (Cth) unifies family
law in Australia, at this federal level.
12
8 among the convict population in the state at the time.
10
This imbalance made the authorities
worry that the disproportion of the sexes would have evil effects as men experienced
difficulty in getting wives (Select Committee on Transportation 1837-1838, p. xxvii).
Free migration was also biased towards men, given that economic opportunities consisted
mainly of pastoralism, and mining after the discovery of gold in the beginning of the 1850s.
We collect data on the historical gender ratio and on the structure of the colonial economy
from the Censuses taken in the 19
th
century in each of the six Australian states.
11
Our main
measure of the sex ratio in the rest of the paper will be taken from the first Census in each
state (see Figure 1 in Appendix), because we need to rely on the earliest possible measure of
the gender imbalance and of its exogenous component, which came from convict
transportation. More than 60% of the current population of Australia now lives in areas that
were already covered by the first Censuses in each state. For historical regressions, we will
consider the full panel of 19
th
century Censuses, roughly from 1836 to 1891, as listed in
Table A1. The panel is unbalanced across states because of their status as independent
colonies at the time and because for some of the years in some of the colonies, the maps of
counties are not available.
12

The unit of observation in the Census is the historical administrative district: a county or
police district. Other data sources, such as colonial musters that counted transported people,
have high reporting error and are not representative of the entire population since
participation was not compulsory (Camm 1978, p. 112). We therefore rely on the measure of
the gender ratio from the Census. The overall sex ratio reported in the first Census, at 3 men
for every woman, underestimates the sex ratio that prevailed in early Australia. As illustrated
in Figure 2, the first Censuses were taken after convict transportation ceased in Queensland
and Victoria. Still, the Censuses for Tasmania, New South Wales, and the Australian Capital
Territory include separate categories for convicts, so that we can recover the convict sex ratio,

10
Authors calculation from the 1833 NSW Census (Historical Census and Colonial Data Archive).
11
The online data from the Historical Census and Colonial Data Archive was supplemented by the
actual Census report due to errors for the 1881 Tasmanian Census. Only the Census reports are
available consistently across the period, some of the individual records having burned.
12
Table A1 only indicates one set of observations in Queensland in 1861. Until 1859, Queensland
was part of New South Wales. After 1861, we loose observations because substantial redistricting
took place and the maps indicating the new districts are not available for this study. All the results we
describe below are unchanged if we exclude the 1861 Queensland observations.
13
which was not influenced by free settlers. As expected, the sex ratio is much larger in the
convict population than in the general population by over 5 times. As an extreme example, in
the county of Bligh, New South Wales, there were 219 men for every woman.
Figures 3 and 4 map the sex ratio among the general and convict populations. All convicts
and most of the population included in the historical Census - Aboriginal people were not
accurately counted in the Colonial Census -- came to Australia by sea.
13
Yet, by the time we
measure them, people of both sexes had made their way in the hinterland and along the coasts.
The concentration of sexes has no definite pattern: high and low sex ratios were found in the
hinterland as well as along the coast.
Essential to our identification strategy is to understand what determined the variation in
population and sex ratios across space. Upon arrival, convicts were not confined to prison
cells. Initially, they were assigned to work under government supervision. Later, as the cost
of caring for large numbers of convicts became too high, convicts were assigned to private
individuals for employment. Employers were government officials, free settlers, and ex-
convicts. The placement of convicts was dictated purely by labor requirements and decided
centrally. Governor Bligh of New South Wales described the system in 1812:
They (the convicts) were arranged in our book () in order to enable me to
distribute them according (cited in Nicholas 1988, p. 15, emphasis added).
Tables 1 and 2 provide descriptive statistics on the historical gender imbalance and the
structure of the economy. Table 1 presents statistics for the historical panel from 1836 to
1881. We stop in 1881 because substantial redistricting took place after 1881, and we do not
yet have access to the subsequent maps. Panel A of Table 2 presents statistics from the first
Census, from which variables are taken for most of the subsequent analysis. This Table also
documents how different counties with high or low sex ratios were from one another (above
or below the median historical sex ration, which is 2.05). Agriculture was the largest
employer in Australia at the time, with 25% of the employed labor force. However, areas
with high or low sex ratio were not statistically different from one another in the share of
people employed in agriculture,. Next were domestic services with 15% and manufacturing

13
There were three main ports of entry: Port Jackson (New South Wales), Port Phillip (Victoria), and
Hobart (Tasmania).
14
and mining, with a combined total of 12%. While the share of people in domestic services
was lower in high sex ratio areas, the share of people employed in manufacturing and mining
was no different. Table A2 in Appendix includes the rest of historical employment categories,
11 in total.
Cheap convict labor seem to have been particularly valued in men. One explanation is that
only men had the physical strength required for agricultural work and building the country
(Nicholas et al 1988). Despite accounting for 19% of the convict population, convict women
accounted for only 6% of the convict labor force (Nicholas et al 1988, p. 70). In 1810, of the
190 jobs advertised in the Sydney Gazette, only seven were for women (Bloodworth 1990).
Yet, convict women possessed analogous skills to British women (Oxley 1996, p. 171) and
their domestic service skills were greatly needed in the settlement (Oxley 1996, p. 189).
Moreover, female labor was even cheaper than males. Wages were set by the government. In
1816, Governor Macquarie of New South Wales announced that male and female convicts
must be paid 10 and 7 per annum, respectively (Nicholas et al 1988 p. 131).

Meredith and
Oxley (2005, p. 56) document an even larger, 46%, gender pay gap. Hence, traditional
economic determinants alone cannot explain that convict women were undervalued and
underemployed in the labor market (Nicholas et al 1988, p. 15). Alford (1984, p. 243) infers
that the notion that remaining within the home was a womans proper place played, already,
a large role.
Some convict women were confined to female factories.
14
Female factories were a
combination of textile factory and female prison (Salt 1984, p. 142) for women who had
borne a child out of wedlock, displeased their assigned master, or committed a crime.
Women worked in female factories for very low or no wage.
15

Overall, Governor Macquarie of New South Wales put it best when he stated that convict
women had 3 choices: become a domestic servant, live in a female factory, or marry (Alford

14
No analogous male factory existed. New South Wales had 3 female factories in the counties of
Cumberland, Northumberland and Macquarie. Queenslands county of Brisbane had 1 and Tasmania
5; 2 located in Hobart and the rest in Launceston, George Town and Campbell Town.
15
Third class women, those who committed a crime in the colony or misbehaved in the factory,
received no wage (Salt 1984, pp. 86, 105)
15
1984, p. 29). In the circumstances described above, together with a high demand for wives,
marriage seemed like the most attractive option.
The authorities concern that the scarcity of women would lead to a difficulty in getting
wives was well founded. Men were more than half as likely to be married than women (see
Table 1), who were under great pressure to be married. According to the historical Census,
and Table 1, more than 70% of women in Australia were married in the 19
th
century, against
only 60% in Britain at the same time period (Alford 1984, p. 26). In areas with sex ratios
above the mean, women were more likely to be married but they had slightly less children
(see Table 2).
OLS panel estimates examining the historical relationship between gender imbalance and
marriage rates, female labor force and the quality of female occupations are displayed in
Table 3. All specifications control for county fixed effects, which remove the influence of
time invariant county characteristics that could be associated both with gender imbalance and
with marriage or female work outcomes. For each dependent variable, specifications in the
first column only includes county fixed effects, and add time fixed effects in the second
column. Because the panel is unbalanced, we have grouped years together and consider half
decades as time fixed effects. We check that the results are unchanged when we model time
from the 1836 start date.
In areas with higher gender imbalance, women were more likely to be married but men were
less likely to be married. The effects are significant at the 1% level and large in magnitude.
An increase by one unit in the sex ratio, that is to say an additional man for a constant
number of women, was associated with a roughly 2.5 points increase in the female marriage
rate. More male biased sex ratios are also associated with lower female labor force
participation and with a lower proportion of women employed in high-ranking occupations,
such as commerce and finance. These effects are statistically significant at the 1% level,
robust to the inclusion of county fixed effects and to controlling for the countys male labor
force participation or males employed in similarly high ranking occupations, and they are
large in magnitude. An increase by one standard deviation in the sex ratio, which is 1.6 (for a
mean of 1.88), is associated with a reduction in female labor force participation and in the
share of women in high raking occupations by 13 and 14 standard deviations, respectively,
16
even when controlling by male labor force participation and by the share of men in high
raking occupations.
OLS panel estimates examining these relationships in the historical cross-section provided by
the first Census available in each state are displayed in Table A3 in Appendix. The results are
qualitatively and quantitatively similar. For each dependent variable, the specification in the
first column only includes state fixed effects. The specification in the second column adds
controls for geographical characteristics (latitude and longitude) and for the main
employment categories -agriculture, domestic services and manufacturing- in order to
account for differences in sex ratios that were dictated by the imperative of the labor
assignment scheme.
To sum up, panel and cross section estimates indicate that in areas with higher gender
imbalance, women married more, had more children and were less likely to work in high-
ranking occupations. In the circumstances of 19
th
century Australia that we have described,
marriage was the most appealing option for women, and they were probably better off as a
result, even though we cannot directly measure this from available data. In the next sections,
we show how 19
th
century economic and marriage conditions have shaped cultural traits and
how they still influence outcomes in the present day. Section 4 explores the implications for
cultural attitudes and for women in the workplace. Section 5 explores the implications for
satisfaction and welfare.

4. The legacy of gender imbalance on culture and women in the workplace
In this section, we explore the long-term consequences of gender imbalance for female labor
force participation and occupational choices and how it has shaped the cultural values of
Australians. We also investigate the channels of transmission that underlie the persistence of
attitudes towards gender roles over more than a century. First, we discuss how we link
historical gender imbalance to present-day opinion surveys and Census data.


17
4.1.Data
We explore the legacy of male biased sex ratios on female labor force participation and
occupational choices with data from the Australian Census. To explore modern-day outcomes,
we retain the most recent, 2011, Census. The unit of observation is the postal area. There are
a total of 2,516 postal areas. To capture the legacy of gender imbalance on female work
choices, we focus on women in high-ranking occupations: women employed as professionals
as a proportion of the employed female labor force.
We also retain information from the Census on 18 categories of industrial specialization,
migration, and sex ratio today. Some areas of Australia still have a gender imbalance (see
Panel C of Table 1), particularly those areas heavily involved in the mining industry, but
these represent only a small fraction of postal areas.
16

Australia experienced its first significant influx of free migrants after the discovery of gold in
New South Wales and Victoria in the 1850s. However, deteriorating economic conditions in
the late 19
th
century and the White Australia Policy in the early 20
th
century restricted
migratory flows during those periods (McLean 2013). The second period of mass
immigration into Australia occurred after the Second World War and the relaxation of the
White Australia Policy in the 1970s, which also changed the composition of migrants to
Australia. In order to capture outcomes before these changes, we rely on data on female work
and occupations from the 1933 Census.
Data on cultural attitudes today are from the Household, Income and Labor Dynamics in
Australia Survey (HILDA), a nationally representative survey available since 2001.
Questions on attitudes towards gender roles were included in the 2001, 2005, 2008 and 2011
waves. HILDA also includes questions about satisfaction with ones spouse and links
partners attitudes and socio-demographic characteristics. We exploit the linked husband and
wife data set in Section 5.
The main question that captures views about gender roles asks to what extent respondents
agree that: it is better for everyone involved if the man earns the money and the woman
takes care of the home and children. Response categories range from 1 (strongly disagree)

16
5% of postal areas have sex ratios above 1.2 and 1% above 1.5.
18
to 7 (strongly agree). We recoded this so that a higher value indicates stronger disagreement
with this statement, which we interpret as the respondent holding more progressive attitudes.
The location of respondents is identified by the postal area. After matching to historical data,
dropping multiple responses and non-responses, we are left with a sample of 42,917
individuals in 1,518 postcodes.
We retain several individual characteristics from HILDA as potential controls, notably
gender, age, education, marital status, state of residency and parents country of birth.
Descriptive statistics are provided in Panel B of Table 2. The balance of these covariates
across areas below or above the mean historical sex ratio is also presented in Table 2. There
is no statistically significant difference across high and low historical sex ratio areas in terms
of age, gender, or ancestry composition, income, education, or in gender balance today.
However, areas that were more imbalanced historically are more likely to be rural areas today,
although the difference is only barely statistically significant at the 10% level. Still, we will
include whether a postal area is rural as a control in all specifications.
Postal areas in modern-day datasets and in the 1933 Census are not equivalent to historical
counties. Prior to this study, shape files on Australian historical census boundaries did not
exist. We digitized hard copies of maps from the National Library of Australia and from State
Libraries in order to construct these boundaries. The Appendix describes the list of maps
used. 12 counties from the colonial Censuses had to be dropped because of incomplete maps.
We were then able to match historical districts to postal areas.
The matching process is illustrated through an example in Figure A1 in the Appendix. When
a postal area was found in multiple counties, we assigned it to the county in which it was
mostly located. For the shaded area in Figure A1, this was Hunter. Figure A2 illustrates the
final matched output.

4.2. OLS results
Having matched historical gender imbalance to postal areas, we are able to examine its
legacy on attitudes, female labor force participation and occupational choices. Figure A3 in
19
the Appendix shows that progressive attitudes towards gender roles, female labor force
participation, and the proportion of women employed in high-ranking occupations are all
negatively correlated with the historical sex ratio. The unconditional relationships between all
three outcomes and historic sex ratios are all significant at the 1% level and robust to the
removal of outliers, such as areas with more than 10 men for every one woman.
We explore the robustness of the legacy of historical gender imbalance on present-day
individual attitudes and on female work by estimating the following equations:
(1)
!
y
ipc
= "
1
+ #
1
SexRatio
c
+ X
pc
G
$
1
+ X
c
H
%
1
+ X
ipc
C
&
1
+'
s
+'
t
+(
ipc

(2)
!
y
pc
= "
2
+ #
2
SexRatio
c
+ X
pc
G
$
2
+ X
c
H
%
2
+T
pc
C
&
2
+'
s
+(
pc

where
!
y
ipc
is the survey-based measure of attitudes of individual i in postal area p, part of
historical county c.
!
y
pc
are the 2011 or the 1933 Census-based measures of female labor
force participation or females employed as professionals in postal area p, part of historical
county c.
!
SexRatio
c
is the historical sex ratio: the number of males over the number of
females in historical county c.
!
X
pc
G
is a vector of time-invariant geographic county
characteristics and
!
X
c
H
is a vector of historical controls.
!
X
ipc
C
and
!
T
pc
C
are vectors of
individual-level and county-level contemporary controls, respectively.
!
"
s
is a vector of state
dummies and
!
"
t
is a vector of HILDA wave dummies. All standard errors are clustered at the
county level, with 80 to 90 clusters observed in each regression.
!
X
pc
G
and
!
X
c
H
are intended to capture geographic and historic characteristics that may have
been correlated with the sex ratio in the past and may still influence present-day outcomes. In
particular, as discussed in the preceding Section, Australia in the 19
th
century was an
agricultural country. Agricultural opportunities influenced where the government assigned
convicts initially and then where free settlers, and their assigned convicts, located. In order to
flexibly account for geographic differences across counties that may be correlated with
agricultural potential, we control for latitude and longitude in all specifications. We also
control directly for the county historical economic specialization, by including in the
historical shares of the population employed in 11 sectors, from mining to carrying to legal
!
X
c
H
20
and medical to government. Table A2 in Appendix contains the full list of these sectors, and
Table 2 the list of main occupations. Due to data limitations outlined in Figure 2, some of
these measures are unfortunately not contemporaneous to our first measures of the sex ratio
in all states. This may result in lowering the coefficient on the sex ratio for these observations
if industrial specialization was influenced by the sex ratio. We do not include these measures
in the most basic specifications but check that all our results are robust to their inclusion.
In the models of individual attitudes, contemporary individual controls include gender,
marital status, age, income, education, and whether the respondent was born in Australia.
County-level controls include the sex ratio today, and 18 categories of industrial
specialization from the standard classification. All county-level controls are from the Census.
In the models of female labor force participation and occupational choices from the Census,
contemporary controls at the postal area include the sex ratio today, average education, and
whether the area is rural. Controlling for the proportion of married people or for the full
range of industrial specialization is problematic, as these are endogenous outcomes. However,
to account for sectoral differences across counties that influence the share of women
employed as professionals, we control for the share of men employed in a similar occupation.
Considering that we are exploiting within-country and even within-state variation, keeping
the formal legislation constant, controlling for the share of men employed in a similar
occupation should leave us with the variation that is due to culture, as opposed to formal
institutions, technology, or employment opportunities.
The estimates displayed in Table 4 show that where the gender imbalance was most severe in
the early days of white settlement in Australia, people are less likely to hold progressive
views about gender roles, and women are less likely to participate in the labor force; when
women do work, they are less likely to occupy high-ranking occupations. The relationship
between attitudes towards gender roles and historical gender imbalance remains statistically
significant at the 1% level even when controlling for the full set of geographic, historic, and
contemporary controls. At the mean, one more man historically for a given number of women
moves the average Australian today towards conservative attitudes by nearly 6 percentage
points (3.102*0.018).
21
The relationship between the historical sex ratio and attitudes today is specific to views about
women working. Gender imbalance does not explain sexist attitudes in general. Table A4
includes the estimation results of (1) in which the dependent variable captures respondents
views about the quality of female leaders. There is no significant effect of historical gender
imbalance on these attitudes.
The relationship between female labor force participation and the historical sex ratio is no
longer significant when we include all historical and contemporary controls. Female labor
force participation may be too gross a measure: it pools together female executives and check
out chicks. When we focus on the quality of female work instead, the relationship with
historical gender imbalance remains statistically significant when controlling for the full set
of controls, including the share of men in similar occupations. At the mean, one more man
historically is associated with slightly under 1 percentage point decrease in the share of
women employed as professionals, which represents nearly 5% of the population mean and
12% of its standard deviation.
The share of men employed in the same occupations explains a very large part of the
variation in the share of female professionals. By controlling for this, state fixed effects and
other controls, we are eliminating any influence of the legal environment, technology, and
employment opportunities for professionals. Still, we find a significant effect of historical
gender imbalance. In terms of the share of the variation explained, adding historical
characteristics to the full set of controls increases the R-squared by 2 percentage points. This
is equivalent to 5% of the remaining unexplained variation in the share of women employed
as professionals.
17

We observe that the coefficient associated with the historical sex ratio sharply drops in
magnitude when the historical shares of employment are added as controls. This suggests,
consistent with our interpretation above, that industrial specialization and the gender ratio
were simultaneously determined at the time.
All the results pertaining to attitudes and to the share in high ranking occupations are robust
to excluding metropolitan areas, and to excluding counties that had very few women

17
This is calculated as (0.567-0.546) / (1-0.546). 0.546 is the R-squared of a regression with all but
historical controls.
22
historically (less than 100). They are also robust to controlling for population density today,
or for the shares of different religions in the population. Some of these additional robustness
tests are presented in Table A5. Table A5 also displays the results of placebo specifications in
which historical sex ratios are randomly allocated to historical counties, while keeping the
overall shares of men relative to women constant. The results are not significant.

4.3. Additional robustness and instrumental variable results
Once concern is that due to the penal origins of the gender imbalance in Australia, we may be
identifying a legacy of convict history rather than gender imbalance history. One may further
argue that convicts were not representative of the rest of the population and were more prone
to hold conservative attitudes. Yet, historical accounts suggest otherwise. Convict men and
women were not hardened and professional criminals (Nicholas et al 1988, p. 3) but
ordinary working class men and women (Nicholas et al 1988, p. 7). We still check that our
results are robust to controlling for Australias convict past.
Convicts were only present in 36 of our 90 historical counties, only in Tasmania and the
colony New South Wales, which also included the Australian Capital Territory at the time.
We estimate equations (1) and (2) on this reduced population, controlling for the convict
population, together with total population, or for the proportion of convicts in the population.
However, because the proportion of convicts is highly correlated with the sex ratio
(correlation coefficient of 0.88), the results with the proportion of convicts should be
interpreted with caution. Due to the reduction in population size when considering convict
settlements only, we drop the full industrial specialization of the county today (18 categories)
but we keep the share of people in agriculture and the share of people in domestic services.
We only report results for attitudes towards female work and for the share of women in high-
ranking occupations, since results for female labor force participation are no longer
significant once all controls are included. For each dependent variable, the first 2 columns
present results with the sex ratio in the general population as the main independent variable
of interest, and the last two with the sex ratio among the convict population instead. We only
present results with the full set of controls.
23
Results are presented in Appendix Table 5. The relationships between historical sex ratio and
conservative gender norms and women employed as professionals today remain statistically
significant. Due the possible colinearity issue mentioned above, the statistical significance of
the coefficients drops, but remains around the 10% mark. The magnitude of the coefficient is
comparable to that in the whole population. An increase by one unit in the historical sex ratio
is associated with moving the average Australian now in former convict settlements by 3
percentage points, and with a reduction in the share of women employed as professional by
9%. The coefficients associated with sex ratios among convicts only are lower in magnitude,
bearing in mind that the average sex ratio is much larger among convicts than among the
general population (see the bottom panel of Table 5 for summary statistics in the
subpopulation of convict settlements).
The proportion or the number of convicts does not have a consistent and robust effect on
either of our outcomes. Conservative gender views and the glass ceiling today are thus due to
the gender imbalance that was caused by Australias convict past, not to convict character per
se.
The fact that we obtain robust results among convicts only, in New South Wales and
Tasmania, where we measure the sex ratio before the discovery of gold, rules out that our
results are not an artifact of Australian gold rushes.
Our results are robust to a battery of observable geographic, historical, and contemporary
controls. Yet, where men and women chose to locate historically may have been driven in
part by unobservable characteristics, for example on the basis of a taste for discrimination. To
address this concern, we adopt an instrumental variable approach. We instrument the overall
sex ratio by the sex ratio among the convict population. Convicts, by definition, were not free
to choose where to go. Where they ended up was not purely random, as we have discussed,
but it was influenced by agricultural opportunities and other employment prospects, which
we control for in both stages. A legacy of convict past independent of gender imbalance due
for example to convicts holding different views about gender roles than the rest of the
24
population, would violate the exclusion restriction. We therefore control in both stages for
the number of convicts.
18

Table 6 presents the results of instrumental variable specifications, where were regress
outcomes on the general sex ratio instrumented by the convict sex ratio. The first stage,
displayed in Panel B of the Table, is extremely strong. All the results discussed in Section 4.2
are robust to this instrumental variable strategy. The IV estimates with the full set of controls
in Columns 2 and 5 in Table 6 should be compared to the OLS estimates in Columns 1 and 5
in Table 5. They are similar: -0.056 for the OLS and -0.059 for the IV for attitudes, and -
0.820 and -0.813 for the OLS and IV, respectively, for women employed as professionals.

4.4. 1933 Results
We have so far documented the short-run implications of a male biased sex ratio, and its
implications in the long-run, about 150 years later. It is important for the validity of our
analysis to document medium-term implications, especially before the onset of massive
migration to Australia after the Second World War and the relaxation of the White Australia
Policy in the 1970s. We do so in the 1933 Census. We digitize and match the 1933 Census
to the Colonial Censuses. We obtain a match for 86 historical counties. The total population
of Australia in 1933 was around 4.5 million people. Due to male biased migratory flows, the
sex ratio stood at 1.16 men for every one woman (see Figure 1).
We estimate specification (2) and regress the female labor force participation and the share of
women employed in high ranking occupations in the 1933 Census on the historical sex ratio,
controlling for the historical shares of employments in 11 sectors, the historical population
and the usual geographic controls. There is no urban/rural indicator in the 1933 Census, so
we control instead for the share of people employed in agriculture in 1933, in addition to
tertiary education. For female labor force participation, we also control for male female labor

18
The presence of female factories, which hosted some female convicts, may have influenced the
convict sex ratio as well as attitudes towards these women, who were considered outcasts. Controlling
for the location of these factories does not alter our results. Besides, the endogeneity bias would run
against the direction of our main result: we would find more conservative attitudes where there were
more women. The effect of female factories is never significant.
25
force participation, and for the share or women in high-ranking occupations, for the share of
men in similar occupations.
Regression results are reported in Table 7. We present results with the full set of controls in
the first column for each dependent variable. We then add the contemporaneous sex ratio in
in 1933.
The relationship between female labor force participation and the share of women in high-
ranking occupations is negatively associated with the historical sex ratio. However, for
female labor force participation, once the 1933 sex ratio is controlled for, the coefficient
drops about 4 fold in magnitude and is no longer statistically significant. The relationship
between women in high ranking occupations and the historical sex ratio is statistically
significant around the 10% mark. In contrast with female labor force participation, it is
unaffected, and actually becomes more precisely estimated, when 1933 sex ratio is included
as a control. We also check that this result is robust to instrumenting the sex ratio by the sex
ratio among convicts. The reduced form estimate is presented in Column 5 and the second
stage IV estimate in Column 6.
These results are entirely consistent with the results documented so far. The relationship
between female labor force participation and the sex ratio is only present in the short-run.
Consistent with a bargaining model, in areas where females are more numerous, they can
marry up, extract a better bargain from their spouses, and work less. However, when the sex
ratio comes back to parity, as it is the case today, their bargaining power vanishes and women
have to go back to the workplace. However, the quality of their occupations is deeply
imprinted by the legacy of past sex ratios. This could be due to investment decisions by
women, who do not foresee the loss of their bargaining power in the future, or by attitudes
that have developed about appropriate gender roles and appropriate occupations for women.
In Appendix Table A6, we regress the share of men and women with a tertiary education in
2011 and in 1933 on the historical sex ratio, controlling for historical and contemporaneous
employment in different sectors of the economy, the usual geographic controls and the
contemporaneous sex ratio. Even though, as expected, the relationship between the historical
sex ratio and the share of females with tertiary education is negative both in 1933 and now, it
is not statistically significant. The relationship between the historical sex ratio and the share
26
of males with tertiary education is positive, but also not statistically significant. The evidence
hence argues more in favor of the second, cultural channel.
We have already documented large and persistent differences in attitudes towards women
working. The next section is devoted to studying the mechanisms of persistence of these
attitudes.

5. Cultural Transmission: The roles of fathers and of marriage homogamy
We have documented in this paper a relatively rapid adaptation of cultural norms as a
response to gender imbalance, even among a population that was homogenous. We then find
that these cultural norms have persisted over time, even after sex ratios have reverted back to
parity. In this section, we explore the traditional transmission channels that have been
discussed in the literature and we document that vertical transmission has played a strong role.
We also explore transmission channels that are specific to the marriage market and have been
relatively less explored so far by the literature on cultural persistence.

5.1. Vertical Cultural Transmission
If gender norms are transmitted within families, and if Australias past shaped a specific
norm in the way we describe, people whose parents are born in Australia should be more
likely to display this norm. The average age of HILDA respondents is 44 years old. This
makes their parents roughly born in the mid 1900s. If these parents were born in Australia,
the White Australia Policy makes it likely that they too were born of Australian parents. The
White Australia Policy, as its name indicates, strongly restricted migration to Australia from
1901 until the early 1970s.
We can test vertical transmission in more details by distinguishing between the mother and
the father. To do this, we add interaction terms between historical gender imbalance and a
27
dummy indicating an Australian father as well as a dummy indicating an Australian mother.
The excluded category consists of persons born to two non-Australian born parents.
19

Regression results are in Columns 1 and 2 of Table 6. The coefficient associated with the
historical sex ratio alone is no longer significant. This means that historical gender imbalance
has no influence on people who are not born of Australian parents. The main effect of having
an Australian father is positive and significant, but its interaction with the historical sex ratio
is negative, statistically significant and large in magnitude: more than twice as large as in the
sample as a whole. In other words, an Australian father transmits more progressive norms,
but not where the gender imbalance was high. With controls included, mothers do not
influence attitudes. This provides supportive evidence of vertical transmission from the father
only.
The crucial role played by fathers is consistent with our interpretation that a conservative
gender attitude is an investment by the male, who needs to transmit it to his son, so that he
can find a wife, and to his daughter, so that she can find a good husband.
We test for the presence of horizontal transmission in Columns 3 to 6 of Table 6. We
distinguish areas with high and low migration, based on the median proportion of migrants at
the postal area level in the 2011 Census. We find no evidence of migrants adopting
conservative gender norms through horizontal transmission. People with no Australian
parents are not influenced by the historical gender imbalance, even in low-migration areas
that had a distorted sex ratio, where they are surrounded by Australians who are more likely
to hold conservative views. Indeed, the coefficient associated with the historic sex ratio is
never significant, even when we restrict the sample to low-migration areas and we include
the usual controls in Column 4.
Vertical transmission is strong and overcomes the attenuating influence of other norms
brought by migrants. Even in areas with a lot of migration, Australian fathers still have a
fundamental influence on their childrens view. The coefficient associated with the dummy
for Australian father is significant both in high and low migration areas and similar in

19
Unfortunately, HILDA dataset does not contain information on respondents mothers occupational
status. We are thus unable to test for the mechanisms described in Olivetti, Patacchini and Zenou
(2013).
28
magnitude. Even though the magnitude is slightly higher in low migration areas, it is not
significantly so.
20

The differences among people of different ancestry within the same areas rule out that the
relationship between historical imbalance and present-day outcomes is due to unobservable
local characteristics or to self-selection of people to localities on the basis of taste.

5.2. Marriage Homogamy and Strategic Complementarities of Norms Among Spouses
We have discussed in Section 2 other mechanisms that might contribute to persistence, and in
particular the issue of coordination on the marriage market and homogamy. We document the
relationship between homogamy in the marriage market and the persistence of conservative
gender norms in Columns 7 to 10 of Table 6. We can define homogamy either along ethnic or
cultural lines. Ethnic homogamy is measured by the probability that ones partner is Australia,
when one is born in Australia. This probability averages 86%. Table A7 in Appendix shows
that homogamy brings direct benefits: people are happier in their relationship when married
to someone ethically or culturally similar. We define high and low homogamy areas if the
average probability in the postal area lies above or below the overall average. Historical
gender imbalance is only associated with conservative views today in areas where homogamy
is high (Column 9). By contrast, no legacy is found in areas with low homogamy (Column 7).
We can also measure homogamy directly along cultural lines, by the probability that ones
partner has the same views, when one has certain views, and the results are similar. A
potential concern is that homogamy itself is endogenous, so that the results cannot be given a
causal interpretation. However, the correlation between the historic sex ratio and homogamy
is less than 0.05. In addition, if we can rely instead on the level of homogamy predicted by
exogenous factors, such as urbanization, industrial sectors, and the share of Australians in the
population, the results are unchanged.
21


20
The coefficient on the interaction between historic sex ratio, Australian father, and low migration
area is 0.01 with a t-stat of 0.37. The results of this triple interaction are not included here.
21
These results are not displayed for economy of space but are available upon request.
29
We also find evidence of a positive interaction between homogamy and vertical transmission
in Columns 8 and 10 of Table 6. This has a natural interpretation: parents want to instill in
their children norms that will make them attractive in the marriage market.
Strategic complementarities in the marriage market are compatible with the apparent paradox
of rapid adaptation of cultural norms and yet cultural persistence. The situation we study is
that of a drastic disequilibrium in the marriage market, one able to lead to rapid adaptation of
norms towards norms that guarantee success in wooing a wife. Our interpretation of
persistence is that of these norms locked in, even after sex ratios have reverted back to
normal, because of strategic complementarity of gender views in the marriage market,
complemented with vertical transmission of norms within families.

6. Welfare Implications
We argue that conservative gender attitudes developed as an attractive male feature for a rare,
and hence powerful, potential spouse. In Section 3, we documented that women were more
likely to marry in in more male-biased areas. We also discussed how, given the options
available to women in 19
th
century Australia, this outcome was beneficial for women, too.
However, we had no direct measure of their welfare. We do for women today.
We analyze self-reported satisfaction. HILDA asks: how satisfied are you with your
relationship with your partner? with answers ranging from 0 (completely dissatisfied) to 10
(completely satisfied). The sample average is 8.30, with women slightly unhappier than men
(difference of -0.24, t-stat of 10.58).
We estimate equation (1) with answers to this marital satisfaction question as the dependent
variable, and we contrast the results for men and women. Regression results are in Table 7.
We first present specifications with only geographic controls and then with geographic,
historic and contemporary controls. The first two columns include an interaction term
between the gender of the respondent and the historic sex ratio. Women in areas where they
were fewer in the past are happier today, but only when their husband works. This is shown
in Columns 3 and 4 when we differentiate by partners employment status. This is consistent
30
with our interpretation. In an environment with a shortage of women, the highest quality men
get the women. In an environment with an equal number of women, as now, they get the
better women (the dating literature would argue: the prettier). Nowadays, there is no
shortage of women, but a shortage of jobs. The lower-quality men hence may have a
(possibly ugly) wife, but no job. But still, all women are taught by their father to value a
working husband in those areas. The ones with working husbands are the happy ones, not
those left with unemployed husbands.
While men are, in general, happier in their relationship than women, they are not more so in
areas with high historic gender imbalance. If anything, they are less happy, with a negative
coefficient associated with the interaction term in Column 1 and 2, although it is insignificant.
A particularly interesting result is that in high gender imbalance areas, those men who were
successful at getting married and who end up with the housewife that the preceding section
has shown corresponds to their wishes, are not happier (see Column 5). In fact it is the man
who ends up with a working woman who is happier (see Column 6). This makes intuitive
sense as the working woman shares the economic burden of the household. But the question
is then: why did men adopt the norm that men should work and women stay at home? Our
answer is that they inherited this norm from a time at which it was beneficial. Conservative
gender attitudes are like the peacocks tail. Like the male peacock with the big tail, which is
so attractive to the female but makes him unable to run away from predators, the man who
defends conservative gender roles may still be attractive to the woman but he is not happier
as a result.
The evidence on the positive effect on the happiness of women is consistent with a
bargaining model, in which women are able to extract a better bargain from their scarcity.
However, these women are foregoing $1,500 dollars every year (more than $2,000 pre tax
and more than $60,000 post-tax over a lifetime of employment). To see this, we estimate a
regression in which the dependent variable is the respondents personal annual income. We
include the full set of geographic, historical and contemporary controls, including the 18
categories of industrial specialization. We include an interaction term between a male
dummy and the historical gender imbalance. Results are in Column 1 of Table 8. Men make
much more money than women, $15,354 a year, on average. Women in high gender
imbalance areas make, at the mean, $1,500 less ($434*3.10). Meanwhile, the interaction term
31
between the male dummy and the historic sex ratio is small and insignificant. This means that
the shortfall in womens earnings is not compensated by higher earnings by men in those
areas, as the economists models would predict. Bargaining models predict a zero sum game.
It is not so. Analysis with Census data in Column 2 confirms that high historic gender
imbalance is associated with a sizeable aggregate welfare loss today. On average, every year,
every person in these areas loses out on nearly $800 of income.

7. Conclusion
This paper documents how cultural norms emerged as a response to a specific scarcity
situation: the lack of women. We show how sexual selection interacted with technology to
shape cultural beliefs. Our results illustrate how cultural norms can emerge as an adaptive
evolutionary response and, like the peacocks tail, serve reproductive fitness. Cultural
evolution has a decisive advantage over genetic evolution. Adopting cultural norms is faster
than growing a tail, easier, and cheaper. However, while certain cultural norms may increase
reproductive fitness, like the peacocks tail, they may harm overall fitness. Certain cultural
norms may be adaptive in the short run, but not necessarily in the long run. This is especially
so since technology changes much faster than culture.
Our results indicate that the presence of strategic complementarities between cultural norms,
here in the marriage market, implies that such norms will persist over time. We believe that
the presence of strategic complementarity between cultural norms solves the apparent
paradox of rapid adaptation of cultural norms and yet cultural persistence over long periods
of time. One implication is that persistence will be stronger and longer for norms that exhibit
stronger strategic complementarity and in situations, like the marriage market, where it may
be costly to experiment.
Our results suggest that industrial specialization may have been influenced by the initial
gender imbalance. In Columns 3 and 4 of Table 8, the effects on lost income are twice as
large as when we control for industrial specialization. To a large extent, one could argue that
industrial specialization today is endogenous to the occupational choices of all the
generations that followed the first settlers and therefore to the initial gender imbalance.
32
Developing a full-fledged structural framework to investigate to what extent initial gender
imbalance has shaped industrial specialization is beyond the scope of this paper and left for
future research.

REFERENCES
Abramitizky, Ran, Adeline Delavande, and Luis Vasconcelos. 2011. Marrying up: the role
of the sex ratio in assortative matching, American Economic Journal: Applied
Economics, Vol. 3: 124-157.
Alesina, Alberto, Nathan Nunn, and Paola Giuliano. 2013. On the origins of gender roles:
women and the plough, The Quarterly Journal of Economics, Vol. 128(2): 469-530.
Alford, Katrina. 1984, Production or reproduction? An economic history of women in
Australia, 1788-1850, Oxford University Press, Melbourne.
Almond, Douglas and Lena Edlund. 2008. Son biased sex Ratios in the US 2000 Census,
Proceedings of the National Academy of Sciences of the U.S.A. Vol. 105 (15): 5681-
5682.
Almond, Douglas, Lena Edlund and Kevin Milligan. 2013. Son Preference and the
Persistence of Culture: Evidence from Asian Immigrants to Canada. The Population and
Development Review, Vol. 39 (1): 75-95.
Anderson, Siwan and Debraj Ray, 2010. Missing Women: Age and Disease, Review of
Economic Studies, Oxford University Press, vol. 77(4), pages 1262-1300.
Angrist, Joshua. How Do Sex Ratios Affect Marriage and Labor Markets? Evidence from
America's Second Generation, Quarterly Journal of Economics Vol. 117 (3): 997-1038.
Bateson, Charles. 1974. The convict ships: 1787-1868, A. H. & A. W. Reed, Sydney.
Basu, Kaushik. 2006. Gender and Say: A model of household behaviour with endogenously
determined balance of power, The Economic Journal, Vol. 116: 558-580.
Becker, Gary S. 1973. A Theory of Marriage: Part I, Journal of Political Economy, Vol.
81(4): 81346.
Becker, Gary S. 1974. A Theory of Marriage: Part II, Journal of Political Economy, Vol.
82(2): S1126.
Bertrand, Marianne, Jessica Pan and Emir Kamenica. 2013. Gender identity and relative
income within households, NBER working paper 19023.
Bisin, Alberto and Thierry Verdier. 2001. The economics of cultural transmission and the
dynamics of preferences, Journal of Economic Theory, vol. 97, no. 2, pp. 298-319
Bisin, Alberto, Giorgio Topa and Thierry Verdier. 2004. Religious Intermarriage and
Socialization in the United States. Journal of Political Economy, 112(3), p. 615-664
Bloodworth, Sandra 1990, The poverty of patriarchy theory, Socialist Review (Australian),
vol. 2, pp. 5-33
Camm, Jack. 1978, Historical geography and past populations: 19th century New South
Wales, Australian Geographical Studies, vol. 16, no.2, pp. 111-125.
Cavalli-Sforza, L. and Feldman, M. 1981. Cultural Transmission and Evolution: A
Quantitative Approach. Princeton, NJ: Princeton University Press
Chiappori, Pierre-Andre. 1988. Rational household labor supply. Econometrica 56, 63-90
Chiappori, Pierre-Andre, 1992. Collective Labor Supply and Welfare, Journal of Political
Economy, vol. 100(3): 437-67.
33
Chiappori, Pierre-Andr, Bernard Fortin, Guy Lacroix. 2002 Marriage Market, Divorce
Legislation, and Household Labor Supply, The Journal of Political Economy, Vol. 110
(1): 37-72.
Chiappori, Pierre-Andre, Iyigun, Murat, and Weiss, Yoram. 2009. Investment in Schooling
and the Marriage Market. American Economic Review, 99(5), p. 1689-1713
Chung and Das Gupta 2007 Why is Son Preference Declining in South Korea? The Role of
Development and Public Policy, and the Implications for China and India. Policy
Research Working Paper 2007.
Commonwealth of Australia Constitution Act 1900 (Imp).
Doepke M, Tertilt M. 2009. Womens liberation: Whats in it for men? Quarterly Journal
of Economics, Vol 124 (4):154191
Doepke M, Tertilt M. and Alessandra Voena 2012. The Economics and Politics of Womens
Rights, Annual Review of Economics, Vol. 4: 6.1-6.34.
Doepke, Matthias and Fabrizio Zilibotti, 2008. Occupational Choice and the Spirit of
Capitalism, The Quarterly Journal of Economics, vol. 123(2), pages 747-793, 05.
Edlund, Lena. 2006. Marriage: Past, Present, and Future? CESifo Economic Studies, Vol.
52 (4): 621-639.
Fernandez, Raquel, Alessandra Fogli, and Claudia Olivetti. 2004. Mothers and sons:
preference formation and female labor force dynamics, The Quarterly Journal of
Economics, Vol. 119(4): 1249-1299.
Fernandez, Raquel. 2008. Culture and Economics, in SN Durlauf & Blume LE (eds), The
new Palgrave dictionary of economics, Palgrave Macmillan.
Fernandez, Raquel and Alessandra Fogli 2009, Culture: an empirical investigation of beliefs,
work and fertility, American Economic Journal: Macroeconomics, Vol. 1(1): 146-177.
Fisman, Raymon, Sheena Iyengar, Emir Kamenica, Itamar Simonson. 2006. Gender
Differences in Mate Selection: Evidence from a Speed Dating Experiment, Quarterly
Journal of Economics, Vol. 121 (2): 673-697
Fletcher, B 1976, Colonial Australia before 1850, Thomson Nelson (Australia) Limited,
Australia.
Fortin, Nicole. 2005. Gender role attitudes and the labor-market outcomes of women across
OECD countries, Oxford Review of Economic Policy, Vol. 21(3): 416-433
Francis, Andrew M. 2011. Sex ratios and the red dragon: using the Chinese Communist
Revolution to explore the effects of the sex ratio on women and children in Taiwan,
Journal of Population Economics, Vol. 24: 813-837.
Goldin, Claudia, and Lawrence Katz. 2002. The power of the pill: oral contraceptives and
womens career and marriage decisions, Journal of Political Economy, Vol. 110: 730-
770.
Grafen, Allen. 1990. Biological Signals as Handicaps, Journal of Theoretical Biology Vol.
144: 517-546.
Greenwood, Jeremy, Ananth Seshadri, and Guillaume Vandenbroucke, 2005. The Baby
Boom and Baby Bust, American Economic Review, vol. 95(1), pages 183-207.
Grosjean, Pauline. Forthcoming, A history of violence: the culture of honor and homicide in
the US South., Journal of the European Economic Association.
Grossbard-Shechtman, A. (1984). A theory of allocation of time in markets for labour and
marriage, The Economic Journal, vol. 94: 863-882
Hauk, Esther and Maria Saez-Marti. 2002. On the Cultural Transmission of Corruption,
Journal of Economic Theory, Elsevier, vol. 107(2): 311-335.
34
Hesketh Therese and Zhu Wei Xing. 2006. Abnormal sex ratios in human populations:
causes and consequences. Proceedings of the National Academy of Sciences of the
U.S.A. Vol. 103 (36): 13271-13275.
National Wage and Equal Pay Cases (1972) 147 CAR 172.
Kokko, Hanna and Michael D. Jennions. 2008. Parental Investment, sexual selection and sex
ratios, Journal of Evolutionary Biology, Vol. 21: 919-948.
Meredith, David and Deborah Oxley. 2005. Contracting convicts: the convict labor market
in Van Diemens Land 1840-1857, Australian Economic History Review, Vol. 45(1):
45-72.
Nicholas, S (eds) 1988, Convict workers: reinterpreting Australias past, Cambridge
University Press, Australia.
Nunn, Nathan and Leonard Wantchekon. 2011. The slave trade and the origins of mistrust in
Africa, American Economic Review, Vol. 101(7): 3221-3252.
Olivetti, Claudia. 2013. The female labor force and long-run development: the American
experience in comparative perspective, NBER-Spencer Conference, Cambridge,
Massachusetts.
Olivetti, Claudia, Eleanora Patacchini and Yves Zenou. 2013. Mothers, Friends and Gender
Identity. NBER working paper No. 19610.
Oxley, Deborah. 1996. Convict maids. Brown Prior Anderson, Australia.
Rao, Vijayendra, The Rising Price of Husbands: A Hedonic Analysis of Dowry Increases in
Rural India, Journal of Political Economy, Vol. 101 (4): 666677.
Ruggles, Steven, J. Trent Alexander, Katie Genadek, Ronald Goeken, Matthew Schroeder,
and Matthew Sobek. Integrated Public Use Microdata Series: v. 5.0 Uni. Of Minnesota,
2010.
Salt, Anette 1984, These outcast women: The Parramatta female factory 1821-1848, Hale
and Iremonger Pty Ltd, Sydney
Select Committee on Transportation 18371838, Report, Minutes of Evidence, Appendix,
Index, House of Commons Parliamentary Papers
Summers, Anne 1976, Damned whores and Gods police, Penguin Books, Australia
Trivers, Robert. 1972. Parental investment and sexual selection. In Sexual Selection and the
Descent of Man 1871-1971. (B. Campbell, ed.), 139:179. Aldine Press, Chicago.
Voigtlander, Nico and Hans-Joachim Voth. 2012. Persecution perpetuated: the medieval
origins of anti-Semitic violence in Nazi Germany, The Quarterly Journal of Economics,
Vol. 127(3): 1339-1392.
Young, Peyton. 1998. Individual Strategies and Social Structure: An Evolutionary Theory of
Institutions. Princeton: Princeton University Press.
Zahavi, Amotz. 1975. Mate selectiona selection for a handicap, Journal of Theoretical
Biology, Vol. 53 (1): 205214.



35
TABLES
Table 1: Summary statistics in Historical Panel (1836 to 1891)
Variable Obs. Mean Std. Dev. Min Max

Sex ratio 446 1.88 1.57 0.93 18.83
Sex ratio among convicts 115 22.36 35.37 0.63 261
Proportion of convicts in population 450 6.34 13.71 0 71.84
Proportion of convicts in population in convict settlements 124 23.00 17.31 0.44 71.84
Female labor force participation 295 32.09 26.36 0 99.85
Female labor force participation among married women 205 43.93 28.03 0 100
Women in high-ranking occupations 247 24.45 23.17 0 77.97
Proportion of adult males married 406 39.53 12.93 5.34 138.58
Proportion of adult females married 406 72.37 26.60 6.26 284.36
Fertility (number of children per adult woman) 410 1.90 0.36 0.45 3.09
Notes: for list of available data by state and by year, see Table A1 in Appendix.

36
Table 2: Summary statistics for Present-Day Regressions and Balance of Covariates
Variables Obs. Mean s.d. Min Max
Above - below
median historic
sex ratio (ii)
t-stat of
diff. (iii)
Panel A: Historical Data
Sex ratio (Historical sex ratio henceforth) 92 3.08 2.95 1.01 18.83 3.19 3.66***
Sex ratio among convicts (Convict sex ratio henceforth) 35 27.65 42.01 1.27 219 18.1 2.55**
Historical Population 91 2764 12194 36 101080 -234 -1.02
Industrial sectors % of county pop. employed in: (i)
! Agriculture, pastoral, horticulture or winegrowing 90 23.43 12.41 1.87 74.62 4.38 1.45
! Domestic and personal service 90 14.37 13.58 2.89 80.81 -8.17 -2.47**
! Manufacturing, Mining 90 11.79 17.08 0 100 -3.31 -0.68
Male married 83 28.85 11.60 6.25 59.61 -15.3 -7.77***
Female married 83 74.64 9.10 57.33 100 8.80 5.40***
Fertility (number of children per adult woman) 83 1.69 0.30 1 2.47 -0.11 -1.47
Female labor force participation 92 54.14 4.74 42.87 67.96 -2.18 -2.39**
Women in high-ranking occupations 75 5.34 5.47 0.00 22.86 -3.73 -3.42***
Panel B: HILDA matched with the historical Censuses
Progressive Attitude Gender Roles 42931 4.47 1.98 1 7 -0.04 -0.3
Individual Controls:
Above 40 years old 42931 0.51 0.5 0 1 0.03 0.53
Log of income plus 1 42931 6.62 4.99 0 14.51 -0.51 -1.23
Australia born 42931 0.77 0.42 0 1 -0.01 -0.24
Beyond year 12 education 42931 0.33 0.47 0 1 0.02 0.37
Legally married dummy 42931 0.51 0.5 0 1 -0.02 -0.39
Postal area controls:
Contemporary gender ratio 42930 0.97 0.07 0.64 3.57 0.02 0.82
Rural 42930 0.72 0.45 0 1 -0.19 -1.67*
Panel C: 2011 Census matched to the historical Censuses
Female labor force participation 1898 56.01 9.47 0 100 -2.54 -2.83***
Male labor force participation 1898 66.44 9.7 0 100 -1.62 1.57
Women in high-ranking occupations (professionals) 2066 20.94 8.2 0 46.5 -3.76 -3.23***
Additional Postal area controls:
Proportion with professional college education 2071 0.21 0.05 0 1 0 0.75
Panel D: 1933 Census matched to the historical Censuses
Female Labor Force Participation 559 27.59 8.30 12.24 57.61 -4.10 -1.76*
Male Labor Force Participation 559 103.34 5.15 47.28 152.83 0.81 1.89*
Women in High Rank Occupations 559 9.18 5.62 0 25.35 -2.62 -2.15**
Notes: (i) for full list of all historical industrial sectors, see Appendix. (ii) and (iii) Differences and t- are from regressions with state fixed effects and robust s.e.
37
Table 3: Historical Panel Results: Gender imbalance and historical female work, and marriage
1 2 3 4 5 6 7 8 9 10
Prop. married women Prop. married men Female labor force participation Women in high-ranking occupations


Sex ratio 2.396*** 2.481*** -2.135*** -1.515*** -21.136*** -12.961** -13.205** -19.663*** -13.832*** -14.100***
(0.714) (0.794) (0.495) (0.285) (6.482) (5.076) (5.579) (3.452) (3.551) (3.603)
Male labor force particip. 0.088***
(0.026)
Male in high-ranking occup. 0.110
(0.106)

Observations 412 412 412 412 205 205 205 247 247 247
Number of counties 94 94 94 94 70 70 70 77 77 77
R-squared 0.026 0.234 0.055 0.120 0.077 0.335 0.377 0.101 0.617 0.619
County FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Year FE No Yes No Yes No Yes Yes No Yes Yes
Notes: The table reports OLS estimates. The unit of observation is a historic county. Sex ratio is the number of men to the number of women in a county. Female (respectively,
Male) Labor Force Participation is the proportion of females (respectively, male) employed, as a proportion of married females (respectively, males). Women (respectively, Men)
in high-rank occupations is the proportion women (respectively, men) employed in commerce and finance, as a percentage of the employed females (respectively, males). See
Table 1 for summary statistics. All regressions are with a constant. Robust standard errors are reported in parenthesis. ***, ** and * indicate statistical significance at the 1%, 5%
and 10% level, respectively.
38
Table 4: Gender imbalance, gender roles attitudes and female work today: OLS results
1 2 3 4 5 6 7 8 9 10
Progressive Attitude Female labor force participation Women in high-rank occupations

Historical sex ratio -0.043*** -0.027** -0.018* -0.502*** -0.134 -0.023 -1.125*** -0.460*** -0.287*** -0.270**
(0.012) (0.011) (0.009) (0.156) (0.145) (0.177) (0.268) (0.138) (0.085) (0.117)

Geographic controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Historical controls No Yes Yes No Yes Yes No Yes Yes Yes
Individual controls No No Yes - - - - - - -
Contemporary controls No No Yes No No Yes No No Yes Yes
Men in high-rank occupations - - - - - - No No No Yes

Observations 42,930 41,943 41,943 1,898 1,871 1,862 1,888 1,861 1,861 1,888
Number of counties 81 78 78 91 88 88 91 88 88 88
R-squared
0.017 0.019 0.174 0.033 0.064 0.195 0.130 0.218 0.308 0.547
State FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
HILDA wave FE Yes Yes Yes - - - - - - -
Notes: Columns 1 to 3 present results when the dependent variable is an individuals response to the statement it is better for everyone involved if the man earns the money and
the woman takes care of the home and children. Response categories range from 1 (strongly disagree) to 7 (strongly agree) (mean: 4.47), which we recoded so that a higher value
indicates more progressive attitudes. The unit of observation is an individual in a POA in a historic county. Geographic controls are POAs centroids latitude and longitude.
Historical controls are the proportion of residents working historically in 11 industries provided in Table A2 in Appendix (see Panel A of Table 1 for the main categories) and the
number of people living in the county, historically. Individual controls are gender, whether one is married, a dummy indicating whether a respondent is aged above 42 years, log
of income plus 1, whether one was born in Australia and whether one has education below year 12. These are all derived from HILDA. Contemporary controls in Columns 1 to 3:
number of men to women in a POA, whether POA is rural, and 18 variables accounting for the proportion of those employed in a POA employed in certain industries.
Contemporary controls in Columns 4 to 10: number of men to women in a POA, whether POA is rural, average vocational tertiary education of POA. Columns 4 to 6 present
results when the dependent variable is the 2011 FLFP as reported in the 2011 Census. (mean: 56.01). Columns 7 to 10 present results when the dependent variable is the proportion
of employed women employed as professionals from the 2011 Census (mean: 21.31). Men: high-rank occupations refers to the proportion of employed men employed as
managers or professionals. The unit of observation in Columns 4 to 10 is a POA matched to its historic county. Standard errors are reported in parenthesis and have been corrected
for heteroskedasticity and for clustering at the county level. Results are robust to using robust standard errors. ***, ** and * indicate statistical significance at the 1%, 5% and 10%
level, respectively.
39
Table 5: Attitudes and female work today and gender imbalance in former convict settlements and among convicts
1 2 3 4 5 6 7 8
Progressive attitudes Women in high-rank occupations

Historical sex ratio -0.056*** -0.084*** -0.820*** -0.433
(0.015) (0.023) (0.191) (0.282)
Convicts sex ratio -0.002** -0.002* -0.030*** -0.017
(0.001) (0.001) (0.010) (0.010)
Number of convicts (in 100s) 3.076 -2.460 -84.060 -198.395
(10.551) (10.847) (126.293) (135.362)
Proportion of convicts 0.805** -0.038 -3.204 -6.337**
(0.353) (0.219) (4.556) (3.038)

Geographic controls Yes Yes Yes Yes Yes Yes Yes Yes
Historical controls No Yes No Yes No Yes No Yes
Historical population Yes No Yes No Yes No Yes No
Individual controls No Yes - - No Yes - -
Contemporary controls No Yes No Yes No Yes No Yes
Men: high-rank occupations - - No Yes - - No Yes

Observations 13,850 13,850 13,850 13,850 661 661 661 661
Number of counties 32 32 32 32 35 35 35 35
R-squared 0.156 0.155 0.155 0.155 0.812 0.806 0.808 0.806
State FE Yes Yes Yes Yes Yes Yes Yes Yes
HILDA wave FE Yes Yes - - Yes Yes - -
Mean Dependent variable 4.47 22.06
Mean Historical Sex Ratio 2.60 2.30
Mean Convict Sex Ratio 16.61 12.90
Notes: See Table 4. This analysis is restricted to NSW and Tasmania. Geographic controls is the latitude of a POA. Historical controls is the share of the population employed
in domestic service and agriculture, historically. Contemporary controls for the attitudinal analysis are the number of men to women and the proportion of the population
employed in agriculture, as per the 2011 Census. The figures for the sex ratio in the population and among convicts only are not identical in Columns 1 4 and in Columns 5 - 8
because attitudes data from HILDA is only available in a sample of the population (32 counties against 35 counties, as indicated in the Table). Robust standard errors are reported
in parenthesis (small number of clusters). ***, ** and * indicate statistical significance at the 1%, 5% and 10%, respectively.
40
Table 6: Gender imbalance and attitudes and female work today: IV results
1 2 3 4
Panel A: Second Stage
Progressive attitude Women in high-rank occupations

Historical sex ratio -0.084*** -0.059** -1.094*** -0.813***
-0.015 -0.027 (0.219) (0.247)

Geographic controls Yes Yes Yes Yes
Historical controls No Yes No Yes
Individual controls No Yes - -
Contemporary controls No Yes No Yes
Male high-rank occupation - - No Yes

Observations 14436 13850 688 661
Number of counties 31 28 34 31
R-squared 0.023 0.156 0.302 0.812
State FE Yes Yes Yes Yes
HILDA wave FE Yes Yes - -
Panel B: First Stage: Historical Sex ratio
Progressive attitude Women in high-rank occupations

Convict sex ratio 0.055*** 0.039*** 0.049*** 0.037***
-0.001 -0.001 -0.003 -0.003

Observations 14436 13850 688 661
Number of counties 31 28 34 31
F statistic 2120 1518 212 249
Adjusted R-squared 0.694 0.795 0.704 0.85
Notes: See Table 4. The table reports 2SLS estimates. This analysis is restricted to NSW and Tasmania. Geographic controls is the latitude of a POA. Historical controls is the
number of people living in a county, the share of the population employed in domestic service and agriculture and the number of convicts living in a county, historically.
Contemporary controls for the attitudinal analysis are the number of men to women and the proportion of the population employed in agriculture, as per the 2011 Census. The
controls in each Column of Panel B are identical to those included in the same Column as Panel A. Robust standard errors are reported in parenthesis (small number of clusters).
***, ** and * indicate statistical significance at the 1%, 5% and 10% level, respectively.


41
Table 7: Gender imbalance and female work in 1933
1 2 3 4 5 6
Female labor force participation Women in high-rank Occupations
Estimation method OLS OLS IV

Historical sex ratio -0.183** -0.071 -0.071 -0.084* -0.289+
(0.091) (0.098) (0.045) (0.043) (0.185)
Sex ratio in 1933 -13.969*** 1.834 2.739 3.318
(2.644) (1.251) (2.894) (2.784)
Convicts sex ratio -0.012+
(0.008)

Geographic controls Yes Yes Yes Yes Yes Yes
Historic controls Yes Yes Yes Yes Yes Yes
Contemporary controls Yes Yes Yes Yes Yes Yes
Male labor force participation Yes Yes - - - -
Male in high-rank occupation - - Yes Yes Yes Yes
Convicts number - - - - Yes Yes

Observations 552 552 552 552 155 155
Number of counties 86 86 86 86 31 31
R-squared 0.697 0.741 0.731 0.733 0.717 0.724
State FE Yes Yes Yes Yes Yes Yes
Notes: The unit of observation is a POA in 1933. Geographic controls are POAs centroids latitude and longitude. In Columns 1 to 6, Historical controls are the proportion of
residents working historically in 11 industries provided in Table A2 in Appendix (see Panel A of Table 1 for the main categories) and the number of people living in the county,
historically. In Columns 7 and 8, Historical controls are reduced because of the drop in population size to: the proportion of residents working historically in the 2 main industries
(agriculture and domestic services, see Panel A of Table 1) and the number of people living in the county, historically. Contemporary controls are: the proportion of people
employed in agriculture in 1933, and average tertiary education in 1933, at the POA level. Female (respectively Male) labor force participation and Male labor force
participation are computed in the same way and represent proportion of female (respectively male) breadwinner as the percentage of the female (respectively male) population of
working age (15 to 70 years old). The population averages are: 27.59% for females and 103.34% for males. Women: High-rank Occupations and Men: High-rank Occupations
are computed in the same way and represent the proportion of employed females (respectively males) employed in commerce and finance in 1933. The population averages are:
9.18% for females and 9.15% for males. Standard errors are reported in parenthesis and have been corrected for heteroskedasticity and for clustering at the county level in Columns
1 to 6. Robust standard errors are in parenthesis in Columns 7 and 8 (small number of clusters). ***, **, *, and + indicate statistical significance at the 1%, 5%, 10% and 15% level,
respectively.
42
Table 8: Persistence: Vertical Cultural transmission by Fathers and Homogamy
1 2 3 4 5 6 7 8 9 10
Dependent variable: Progressive attitude
All areas Low migration High migration Low homogamy High homogamy

Historical sex ratio 0.004 0.016 -0.029*** 0.001 -0.001 0.028 -0.227 -0.215 -0.052*** -0.009
(0.018) (0.017) (0.010) (0.016) (0.152) (0.155) (0.425) (0.424) (0.010) (0.021)
Australian father 0.348*** 0.162** 0.067 0.193*** 0.076 0.133*
(0.064) (0.067) (0.075) (0.066) (0.114) (0.068)
Australian father * sex ratio -0.092*** -0.074*** -0.039* -0.045* -0.019 -0.055**
(0.034) (0.027) (0.020) (0.027) (0.019) (0.023)
Australian mother 0.121* -0.029
(0.069) (0.067)
Australian mother * sex ratio 0.044 0.029
(0.027) (0.022)

Geographic controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Historical controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Individual controls No Yes Yes Yes Yes Yes Yes Yes Yes Yes
Contemporary controls No Yes Yes Yes Yes Yes Yes Yes Yes Yes

Observations 41,943 41,943 20,982 20,982 20,961 20,961 21,295 21,295 20,651 20,651
R-squared 0.026 0.175 0.155 0.155 0.175 0.175 0.183 0.183 0.172 0.173
State FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
HILDA wave FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Notes: See Table 4. Low migration refers to POAs where the proportion of residents born in Australia is higher than the median proportion of residents born in Australia as per
the 2011 Census. High migration refers to POAs where the proportion of residents born in Australia is lower than the median proportion of residents born in Australia as per the
2011 Census. Low Homogamy refers to POAs below the median level of homogamy computed as the average proportion of people of Australian decent in the POA who married
someone also of Australian decent. High Homogamy refers to POAs above that median, which is 85%. Robust standard errors are reported in parenthesis for columns 5 and 6 as
the number of clusters is small.

43
Table 9: Marital satisfaction: women are happier; men are not
1 2 3 4 5 6
Satisfied with partner
All
Female with an
employed partner
Female with an
unemployed partner
Male with an
employed partner
Male with an
unemployed partner

Historical sex ratio 0.040*** 0.028** 0.046*** 0.036* 0.019 0.025**
(0.009) (0.011) (0.013) (0.020) (0.015) (0.012)
Male 0.245*** 0.267***
(0.029) (0.030)
Male * sex ratio -0.001 -0.001
(0.012) (0.011)

Geographic controls Yes Yes Yes Yes Yes Yes
Historical controls No Yes Yes Yes Yes Yes
Individual controls No Yes Yes Yes Yes Yes
Contemporary controls No Yes Yes Yes Yes Yes

Observations 31,282 30,526 9,480 3,160 4,931 7,607
Number of counties 80 77 74 71 69 70
R-squared 0.009 0.030 0.013 0.050 0.043 0.016
State FE Yes Yes Yes Yes Yes Yes
HILDA wave FE Yes Yes Yes Yes Yes Yes
Notes: See Table 4. The dependent variable is an individuals response to the question how satisfied are you with your relationship with your partner? Response categories range
from 0 (completely dissatisfied) to 10 (completely satisfied). Contemporary controls are the number of men to women and the proportion of the population employed in
agriculture, as per the 2011 Census. Standard errors are reported in parenthesis and have been corrected for heteroskedasticity and for clustering at the county level
44
Table 10: Income
1 2 3 4
Annual income Median income weekly

Historical sex ratio -898.410*** -434.217* -7.302** -5.055*
(318.586) (264.281) (3.550) (2.940)
Male 15,093.182*** 15,356.249***
(1,201.921) (1,209.566)
Male * sex ratio 65.199 33.762
(489.808) (487.969)

Geographic controls Yes Yes Yes Yes
Historical controls Yes Yes Yes Yes
Individual controls Yes Yes - -
Contemporary controls No Yes No Yes

Observations 47,975 47,975 1,871 1,861
Number of counties 78 78 88 88
R-squared 0.157 0.175 0.227 0.264
State FE Yes Yes Yes Yes
HILDA wave FE Yes Yes - -
Notes: See Table 4. Columns 1 and 2 present the results where the dependent variable is an individuals annual income,
as reported in HILDA. Columns 3 and 4 present the results where the dependent variable is the median weekly income
of a postal area, as reported in the 2011 Census. Contemporary controls in Columns 3 and 4 refers to the share of the
employed population employed in agriculture. Standard errors are reported in parenthesis and have been corrected for
heteroskedasticity and for clustering at the county level


45
FIGURES
Figure 1: Sex ratio in Australia: Number of men to every woman

Source: Australian Bureau of Statistics


1
1
.
5
2
2
.
5
3
1820 1840 1860 1880 1900 1920 1940 1960 1980 2000
year
46
Figure 2: Data sources in the Australian states and territories


47
Figure 3: Gender imbalance in mid 19
th
century Australia


48
Figure 4: Gender Imbalance among Convicts in News South Wales and Tasmania




APPENDIX: Its Raining Men! Hallelujah?
Pauline Grosjean and Rose Khattar



Table A1: Historical Observations

Colony
Year NSW TAS VIC SA WA QLD (i)

1836 18
1841 18
1842 16
1844 7
1846 18
1848 16 6
1851 18
1854 21 6
1856 18
1857 10 21
1859 6
1861 18 10 21 16 6 14
1866 16
1870 10 6
1871 18 21 16
1876 16
1881 10 21 16
Notes: (i) Queensland was part of New South Wales until 1859. We loose
observations after 1861 in Queensland because substantial redistricting took
place and the maps indicating the new districts are not available for this study.

Table A2: Additional Descriptive statistics not included in Table 1

Variables Obs. Mean s.d. Min Max
Panel A: Historical Data
Industrial sectors % of county pop. employed in:
! Agriculture, pastoral, horticulture or winegrowing 90 23.43 12.41 1.87 74.62
! Domestic and personal service 90 14.37 13.58 2.89 80.81
! Manufacturing or Mining 90 11.79 17.08 0 100
! Manufacturing 90 6.76 8.78 0 56.16
! Mining 90 5.70 13.42 0 67.01
! Carrying 90 2.23 4.09 0 22.25
! Trade or commerce (such as accounting and retail) 90 1.98 2.80 0 17.64
! Literature, fine arts and sciences (such as teachers) 90 1.78 3.13 0 16.50
! Food 90 1.60 2.50 0 16.09
! Government 90 1.14 1.62 0 12.79
! Learned professions (such as legal and medical) 90 0.45 0.78 0 4.94
Panel B: HILDA and 2011 Census matched to historical Census
State of residency dummy:
! New South Wales 42,931 0.28 0.45 0 1
! Victoria 42,931 0.28 0.45 0 1
! Queensland 42,931 0.18 0.38 0 1
! SA 42,931 0.10 0.30 0 1
! WA 42,961 0.10 0.30 0 1
! Tasmania 42,961 0.04 0.19 0 1
! ACT 42,961 0.02 0.15 0 1

Table A3: Historical Cross Section Results: Gender imbalance and historical marriage rates, fertility and female work

1 2 3 4 5 6 7 8

Female marriage rate
(%, historic)
Female marriage rate
(%, historic, in convict
areas)
Male marriage rate (%,
historic)
Women in High-
Ranking
Occupations (%,
historic)

Sex ratio 2.159*** 2.160*** 2.857*** 3.127*** -1.614*** -1.468*** -0.327** -0.221**
(0.615) (0.638) (0.988) (1.012) (0.158) (0.149) (0.140) (0.111)

Geographic controls No Yes No Yes No Yes No Yes
Historical controls
sectors
No Yes No Yes No Yes No Yes

Observations 92 89 35 32 92 89 90 90
R-squared 0.776 0.785 0.798 0.858 0.641 0.717 0.602 0.669
State fixed effects Yes Yes Yes Yes Yes Yes Yes Yes

Notes: The table reports OLS estimates. Sex ratio is the number of men to the number of women in a county. Geographic controls is a POAs
latitude and longitude. Historical controls are the proportion of the population in agriculture, domestic service and manufacturing, historically. A
list of the industries is provided in Panel A of Table A1. The unit of observation is a historic county. All regressions are with a constant. Robust
standard errors are reported in parenthesis. ***, ** and * indicate statistical significance at the 1%, 5% and 10% level, respectively.

Table A4: Not machismo in general: Gender imbalance and perceptions of female political leadership

1 2 3
Attitudes towards women as political leaders

Historical sex ratio 0.026 0.001 -0.003
(0.047) (0.010) (0.010)

State fixed effects No Yes Yes
HILDA wave fixed effects No Yes Yes
Geographic controls No Yes Yes
Historic controls No Yes Yes
Individual controls No No Yes
Contemporary controls No No Yes

Observations 37,120 31,334 31,334
Number of counties 81 78 78
R-squared 0.000 0.005 0.100
Notes: See Table 4. The table presents results where the dependent variable is on the whole, men make better political leaders than women do. Response categories range from 1
(strongly disagree) to 7 (strongly agree), which we recoded so that a higher value indicates more progressive attitudes.
Table A5: Robustness

1 2 3 4 5 6 7 8
Progressive attitudes Women in high rank occupations

Historical sex ratio -0.024** -0.027* -0.018* -0.421** -1.736*** -0.183*
(0.011) (0.016) (0.009) (0.176) (0.483) (0.101)
Historical sex ratio (randomized) -0.005 -0.116
(0.006) (0.166)
Sample
Excluding
metropolitan
areas
Excluding
counties with
less than 100
women
Whole Whole
Excluding
metropolitan
areas
Excluding
counties with
less than 100
women
Whole Whole
Pop Density No No Yes No No No Yes No
Geographic controls Yes Yes Yes Yes Yes Yes Yes Yes
Historic controls Yes Yes Yes Yes Yes Yes Yes Yes
Individual controls Yes Yes Yes Yes - - - -
Contemporary controls Yes Yes Yes Yes Yes Yes Yes Yes
Male high-rank occupation - - - - Yes Yes Yes Yes

Observations 17521 40206 41943 41943 1058 1761 1888 1876
R-squared 0.171 0.173 0.174 0.174 0.278 0.461 0.576 0.417
State fixed effects Yes Yes Yes Yes Yes Yes Yes Yes
HILDA wave fixed effects Yes Yes Yes Yes - - - -
Notes: See Notes to Table 4 in main paper. Metropolitan areas are: Cumberland and Camden in NSW, Bourke, Evelyn, Grant and Mornington in VIC, Adelaide
in SA, Stanley in QLD. There are 13 historical counties with less than 100 women (4 in NSW, 2 in QLD, 5 in SA, 2 in WA). Pop. density is the population
density today. Robust standard errors clustered at the historical county level.


Table A6: Education as an outcome

1 2 3 4
2011 1933

Proportion
of males
with tertiary
education
Proportion
of females
with tertiary
education
Proportion
of males
with tertiary
education
Proportion
of females
with tertiary
education

Historical sex ratio 0.016 -0.008 0.002 -0.002
(0.061) (0.076) (0.003) (0.003)

Geographic controls Yes Yes Yes Yes
Historical controls Yes Yes Yes Yes
Contemporary controls Yes Yes Yes Yes

Observations 1,861 1,861 552 552
R-squared 0.741 0.610 0.983 0.867
Notes: The unit of observation is a POA. Geographic controls are POAs centroids latitude and
longitude. In Columns 1 to 6, Historical controls are the proportion of residents working historically in
11 industries provided in Table A2 in Appendix (see Panel A of Table 1 for the main categories) and the
number of people living in the county, historically. Contemporary controls: the sex ratio, whether POA
is rural, and 18 variables accounting for the proportion of those employed in a POA employed in certain
industries in 2011 and in Columns 3 and 4: the sex ratio and the proportion of people employed in
agriculture in 1933. Standard errors are reported in parenthesis and have been corrected for
heteroskedasticity and for clustering at the county level. ***, ** and * indicate statistical significance at
the 1%, 5% and 10% level, respectively.
Table A7: Homogamy makes people happy

1 2

Happy in your relationship
with your partner

Australian born -0.038 -0.224***
(0.034) (0.083)
Partner Australian born 0.066* -0.121
(0.036) (0.112)
Australian born * Partner Australian born 0.313**
(0.155)
Progressive attitudes

Partner Progressive attitudes

Progressive attitudes * Progressive attitudes

Geographic controls Yes Yes
Individual controls Yes Yes
Contemporary controls Yes Yes

Observations 25,976 25,976
R-squared 0.021 0.022
State FE Yes Yes
HILDA wave FE Yes Yes
Notes: The table reports OLS estimates. The dependent variable is the answer to
the question: how happy are you in your relationship with your partner?.
Answers are on a scale of 1 to 10. The average is 8.38. All the controls are the
usual controls (see Table 3). All regressions are with a constant. Robust standard
errors are reported in parenthesis. ***, ** and * indicate statistical significance at
the 1%, 5% and 10% level, respectively.

Map and Data Sources

Arrowsmith, J 1846, South Australia shewing the division into
counties of the settled portions of the province with situation of
mines of copper & lead, ca 1:935 000, National Library of
Australia.
Arrowsmith, J 1848, The colony of Western Australia, ca 1:975
000, National Library of Australia.
Australian Bureau of Statistics 1933, Census of the
Commonwealth of Australia, 1933, Cat. no. 2110.0, Australian
Bureau of Statistics, Canberra.
Australian Bureau of Statistics 2011, Australian Statistical
Geography Standard (ASGS), Cat. no. 1270.0, Australian
Bureau of Statistics, Canberra.
Australian Bureau of Statistics 2011, Socio-Economic Index of
Australia (SEIFA), Cat. no. 2033.0.55.001, Australian Bureau
of Statistics, Canberra.
Australian Bureau of Statistics 2013, Average Weekly Earnings,
Cat no. 6302.0, Australian Bureau of Statistics, Canberra.
Black, A & Black, C 1861, The British colonies of Australia,
New Zealand and Tasmania shewing the latest discoveries and
travellers' routes, ca. 1:4,200,000, National Library of Australia.
Buxton, JW 1863, The new map of Queensland, State Library
of Queensland.
Cross, J 1832, Chart of part of New South Wales, with plans of
the harbour, ca. 1:1 500 000, National Library of Australia.
Government of Queensland 1909, A review of Queensland
progress, Anthony J. Cumming, Queensland, p. 69.
Hall, S 1840, Van Diemens Land, ca: 1:800 000, National
Library of Australia.
Harris, WG 1862, Map of the province of South Australia, ca
1:1 850 000, National Library of Australia.
Mitchell, T 1834, To the Right Honorable Edward Geoffrey
Smith Stanley this map of the Colony of New South Wales, ca.
1:540 000, National Library of Australia.
Pearce, J 1885, Map of Western Australia shewing electoral
districts, 1885 compiled expressly for the Herald almanac and
directory, ca. 1:5 068 800, State Library of Western Australia.
Rapkin, J 1851, Victoria, or Port Phillip, ca. 1:3 000 000,
National Library of Australia.
Robertson, A 1858, Victoria, census districts and distribution of
the population, March 29th 1857, ca. 1:510 000, National
Library of Australia.
Waterlow & Sons 1859, Map of South Australia including the
recent discoveries, ca 1 inch to 20 miles, State Library of South
Australia.
Figure A1: Postal Are Shape File Overlaid by 1834 NSW Shape
File


Notes: The highlighted polygons each represent a county. Underlying this historic
map are polygons comprising of NSW POAs. Each POA polygon was matched to
its associated county polygon.

Sources: ABS 2011 Cat no. 1270.0 and 1834 NSW map digitised based on
Mitchell 1834.

Figure A2: Postal Area Shape File Overlaid by Historic County
Shape Files




















Sources: ABS 2011 Cat no. 1270.0 and historic county maps digitised based upon Cross
1832, Mitchell 1834, Hall 1840, Arrowsmith 1846 and 1848, Rapkin 1851, Robertson
1858, Waterlow & Sons 1859, Buxton 1861, Black & Black 1861,
Harris 1862 and Pearce 1885.






Figure A3: Bivariate correlations of outcomes today by historic sex ratio



Figure A4: Bivariate correlations of outcomes in 1933 by historic sex ratio



Notes: Each dot is the average value, by historical county. All relationships are statistically
significant at the 1% level. Estimated coefficients (with robust standard errors) are: A) -
0.06 (t-stat: 2.66) and -0.08 (t-stat: 2.21), B) -0.38 (t-stat: 3.56) and -0.53 (t-stat: 2.08), C)
-0.22 (t-stat: 2.27) and -0.30 (t-stat: 1.59)
Figure A4: Estimated coefficients (with robust standard errors) are: A) -0.35 (t-stat: 5.12)
and -0.48 (t-stat: 2.60), B) -0.30 (t-stat: 2.16 and , C) -0.78 (t-stat: 3.30).

Sources: HILDA and Australian Census