Source: TW
Death, sex and fertility: Female infanticide in rural Spain, 1750-1950
Francisco J. Beltrán Tapia1,
Norwegian University of Science and Technology
Francisco J. Marco-Gracia,
Universidad de Zaragoza
Abstract
Relying on longitudinal micro data from a Spanish rural region between 1750 and 1950, this article evidences that families mortally neglected a significant fraction of their female babies.+++(5)+++
On the one hand, baptism records exhibit exceptionally high sex ratios at birth, especially during the 19th century. On the other hand, our data shows that having no previous male siblings increased both the probability of male baptisms and female mortality during the first day of life. These findings seem to be concentrated at higher parities and among landless and semi-landless families which were subject to harsher economic conditions and therefore more likely to resort to extreme decisions under difficult circumstances.+++(5)+++
Crucially, the fact that the results are robust to employing data from birth and death registers rules out the possibility that under-registration explains this pattern.
Lastly, although these practices were definitely more important during the traditional demographic regime, discriminatory patterns affecting female mortality shortly after birth were still visible during the first decades of the 20th century, thus proving that son preference continued to be a strong cultural norm within these societies.
JEL Codes: I14, I15, J13, J16, N33
Keywords: Sex ratios, Infant mortality, Infanticide, Gender discrimination
* This research received financial support from the Research Council of Norway (Project 301527) and the Spanish Ministry of Science and Education (Project ECO2015-65582).
1 Corresponding author: francisco.beltran.tapia@ntnu.no
1. Introduction
Many pre-industrial societies regulated population size by resorting to infanticide and the mortal neglect of unwanted infants and children (Langer 1974; Harris and Ross 1987; Hrdy 1999; Hanlon 2016). These practices have traditionally targeted girls in India, China and Japan, among other countries characterised by strong patriarchal traditions that favour males (Sen 1990; Das Gupta 2003; Bhaskar and Gupta 2007; Drixler 2012; Gupta 2014). Although women’s status in historical Europe was definitely more advantageous than in other parts of the world, Europe was not a gender-equal paradise and women were discriminated in many dimensions (Szoltysek et al. 2017; Carmichael and Rijpma 2017; Dilli et al. 2019). Son preference, affecting the propensity to have additional children, was indeed present in many regions in pre-industrial Europe (Kolk 2011; Manfredini et al. 2013, Sandström and Vikström 2015, Reher and Sandström 2015)1.
There is, however, little evidence that European families neglected their female babies.
Derosas and Tsuya (2010), for instance, found no clear pattern of female infanticide in the three samples that they were analysing: Casalguidi in Italy (1819-1859), Sart in Belgium (1812-1874) and Scania in Sweden (1829-1867). Summarising this view, Lynch (2011) argues that the household formation patterns, as well as cultural and religious values, prevented female infanticide in historical Europe. Several studies have though challenged this view and suggested that families resorted to female infanticide as a means of controlling the size and sex composition of their offspring in different regions and periods (Bechtold 2000; Hynes 2011; Hanlon 2016; Beltrán Tapia and Raftakis 2019)2.
Relying on longitudinal micro data from a small rural region in North-eastern Spain between 1750 and 1950, this article evidences that discriminatory practices increased female at birth. Firstly, aggregate sex ratios obtained from baptismal records were exceptionally high, even during the 19th century when the quality of registration was very high. Secondly, the data shows that having no previous male siblings, a feature that could trigger discrimination in the presence of son preference, increased both the probability of male baptisms and female mortality around birth during this period. Taken together, these results strongly support the idea that these families were neglecting a significant fraction of their female babies in 19th-century Spain.
These findings seem to be concentrated at higher parities and among landless and semi-landless families which were subject to harsher economic conditions and therefore more likely to resort to extreme decisions under difficult circumstances. Crucially, the fact that the results are robust to employing data from birth and death registers rules out the possibility that under-registration explains this pattern: although female misreporting would bias sex ratios at birth upwards, it would have the opposite effect with mortality rates. Lastly, although these practices were definitely more important during the traditional demographic regime, discriminatory patterns affecting female mortality shortly after birth were still visible during the first decades of the 20th century, thus proving that son preference continued to be a strong cultural norm within these societies.
This article supports previous studies that challenged the idea that there were no missing girls in historical Europe (Beltrán Tapia and Gallego-Martínez 2017; Beltrán Tapia 2019). In 1 See also Bohnert et al. (2012) for the US.
2 Interestingly, Hynes (2011) and Hanlon (2016) consider that families could also target boys depending the circumstances. Another strand of the literature suggests that gender-discriminatory practices unduly increased female mortality during infancy and childhood via an unequal allocation of food, care and/or workload (Tabutin 1978; Johansson 1984; Pinnelli and Mancini 1997; Baten and Murray 2000; McNay et al. 2005; Horrell and Oxley 2016; Beltrán Tapia and Gallego-Martínez 2017; 2020; Beltrán Tapia 2019).
1
this regard, it shows that the high sex ratios in infancy and childhood found in previous studies are not driven by problems with the quality of the registers, but by female excess mortality. In addition, the micro-data used here sheds light on how these societies ended up showing unbalanced number of boys and girls by illustrating that an important fraction of these missing girls were neglected at birth. While the neglect of female babies was possibly a conscious family decision that affected a relatively small number of families under certain circumstances, our data cannot distinguish whether the death of infant girls was the result of direct female infanticide or more indirect forms of mortal neglect (starvartion, smothering, exposure, etc.)3.
Given that anecdotal evidence hardly exists, it seems that families succeeded in disguising female infanticide as natural deaths.
Likewise, the finding that families neglected female infants suggests that these societies regulated the size and sex composition of their offspring despite the religious norms against it.
The dominant view defends that the low number of children raised in European families was the result of different ways of reducing fertility. This could be achieved either indirectly via high age at marriage and a high singleness rate (Hajnal 1965; Wrigley and Schofield 1981), or directly by adjusting birth intervals or implementing fertility-stopping rules (Schofield 2000; Van Bavel 2004; Bengtsson and Dribe 2006; Reher and Sanz-Gimeno 2007; Marco-Gracia 2019; 2020)4. Our results imply that some sort of “death control” was also present (Langer 1974; Harris and Ross 1987; Hrdy 1999; Hanlon 2016; 2017). Resorting to infanticide could have been a drastic solution to regulate fertility, especially in already large families subject to economic stress. In this regard, girls entailed a larger burden to the household income because both the need to provide a dowry militated against girls and because there were less female waged labour opportunities (and worse paid)** **
The evidence provided here shows that families especially targeted girls, but boys could have undoubtedly fell victims of these practices as well. The fact that families somewhat increased the mortality of their children also has implications for our understanding of the demographic transition. The subsequent reduction of mortality rates would therefore partly respond to the relaxation of these practices. Likewise, many of these infanticides were never registered as births or deaths, thus indicating that fertility and mortality rates in pre-industrial Europe were higher than what it has been traditionally assumed.
Our findings are especially relevant because they take place in an area where nuclear households prevailed and inheritances were equally distributed among all children, regardless of their sex. Likewise, women maintained full control of the resources they brought to the marriage and were entitled to freely dispose of their patrimony through wills at their death (Jarque Martínez and Salas Ausens 2007, 126-127). Moreover, although it is true that women did not enjoy the same status in the labour market than men (lower salaries, less workdays, etc.), female waged labour was widespread and their contribution was crucial to sustain the household economy (Borderías et al. 2018; Germán Zubero 2009, Lana Berasain 2007). These features are not particularly related to patriarchal societies (Szoltysek et al. 2017), so the results open up the possibility of finding similar or even more extreme manifestations of son preference in other European regions.
3 See Langer (1972) for the different ways that families could dispose of unwanted children.
4 See Van de Walle (1992) on the debate about whether an ideal family size existed before the demographic transition.
2
2. Data and historical background
This study focuses on a small rural area in Aragon, in North-Eastern Spain, that is located around 19-40 kilometres away from Zaragoza, the regional capital (see Map 1). This area, a combination of plains and foothills near the Huerva river, comprises 13 small municipalities5.
Their total population was approximately 5,525 inhabitants in 1750, 8,315 in 1857 and 9,556
in 1950. The statistical analysis relies on the complete Church registers of these villages, whose records provide high-quality information on all baptisms, marriages and deaths that occurred between 1575 and 1999 (although the starting date varies by location; more details about the
‘Alfamén and Middle Huerva Database’ in Marco-Gracia 2017; 2019). Document 1 in the appendix offers an example of these registers. In total, this dataset contains information on 88,989 individuals (84,686 up to 1950), including name, sex, place and date of birth, parents’
names and date of death, among others, thus permitting the reconstitution of the life history of these individuals and their families. This longitudinal dataset has also been complemented with information on occupation and literacy contained in population lists (1747-1830), population censuses (1857, 1860) and electoral rolls (1890-1955)6. Table A1 in the Appendix provides the number of observations by period, classified by sex, father’s occupation and father’s literacy.
Map 1. Area of study: Middle Huerva (Aragón, Spain)
Note: Dark dots refer to the localities studied here (except Zaragoza, the provincial capital) and the corresponding shaded areas to their municipal boundaries. Apart from rivers (in blue) and main roads (grey), the map also depicts neighbouring villages (white dots).
5 The localities are: Alfamén, Aylés, Botorrita, Cosuenda, Jaulín, Longares, Mezalocha, Mozota, Muel, Torrecilla de Valmadrid, Tosos, Valmadrid, and Villanueva de Huerva.
6 The information on father’s occupation has been grouped into six categories: (1) day-labourers and small owners who were unable to make a living exclusively from their properties, (2) farmers with enough land to secure their livelihood, (3) sheep and goat shepherds, (4) artisans, (5) elites, mostly doctors, teachers and council representatives, and (6) other occupations, a heterogenous group including mule drivers, soldiers, low-skilled occupations working for the municipality, among other occupation that do not fit within the other groups. Due to the lack of information, data on occupations and literacy is missing for around one-third of the individuals. This scarcity is more important in the earlier periods and very low in the final years analysed here.
3
We should bear in mind however that registration quality greatly improved from 1750
onwards. Infant and child mortality rates before that date are too low, so under-registration of deaths is likely to be an issue. Although registration quality had been improving throughout the 18th century, the year 1774 was especially important: the priests in these villages received a pastoral visit conveying the orders from the Archbishop of Zaragoza. These mandates emphasized the importance of keeping accurate and detailed parish records7. We will therefore restrict most of our analysis to the period 1750-1950.
The area of study, 13 villages covering around 500 kms2, hosted a population who mostly lived in nuclear households and was essentially devoted to agriculture (mostly wheat and some wine) and sheep grazing. Our records show that around 85 per cent of the male working population was engaged in the agricultural sector between 1800 and 1950. Average fertility was relatively stable around 6-7 children among complete families up to 1900 and declined rapidly thereafter following the demographic transition. Infant and child mortality rates were very high though and only around half of the children survived to their fifth birthday. ** **
Mortality rates began to decline in the last third of the 19th century due to increasing living standards. The decline firstly benefited children in their late childhood and spread later to younger cohorts. Infants were the last one to join this trend and their survival chances only significantly increased from 1900 onwards when hygienic conditions and their mothers’ health improved, an improvement that was especially visible during early childhood. Anthropometric evidence also indicates that standards of living were extremely low: the average male height was around 160 centimetres in mid-19th-century, well below their European counterparts or their fellow Spanish in other regions (Martínez-Carrión et al. 2016; Hatton and Bray 2020). In an area where most of the population enjoyed living standards close to subsistence levels, choices mattered and discriminatory practices could have had lethal consequences. ** **
Although the quantitative information provided by the sources is exceptional, qualitative evidence on how parents treated their children is scarce or non-existent. Although the inheritance system prescribed that all property was divided equally among all children, son preference seems to have formed part of the shared norms in this area during the period of study. In this regard, Spanish women did not enjoy the same status as men: legally subordinated to their fathers and husbands, women were expected to remain within the domestic realm and those who did work in paid jobs received significantly lower wages (Camps 1998; Sarasúa 2002; Borderías et al. 2010; Borderías and Muñoz 2018).
The custom of the dowry was also something that militated against girls as plenty of studies have documented for regions in South and East Asia (Das Gupta et al. 2003; Bhalotra et al. 2020)8. Many state regulations frequently attempted to limit this practice in order to make it less onerous (Martín Rodríguez 1984, 264). Coherent with the state’s interests in promoting fertility (so as to strengthen the nation with man and military power), the reason behind restricting the amount that parents had to pay their daughters to secure their marriages was mainly to facilitate weddings (and subsequently fertility). It however clearly implies that this practice was a burden for the families. The norm itself, consigned in the Pragmática of 1623
but alive during the period of study here, indicated that the costs associated with the dowry had detrimental consequences for families’ estates because they sometimes had to borrow the money. In this regard, Harding (1984: 103) mentions that a farmer lost his family state in the late 19th century because it had to provide dowries for his eight daughters.
7 While birth registers before that date only reported the name of the parents, they began to include that of grandparents. Death registers similarly now included the names of their relatives. Likewise, these records now reported not only where these people lived but also where they were born.
8 See also Beltrán Tapia and Raftakis (2019) on the discriminatory effect of the dowry in 19th-century Greece.
4
Fertility decisions seemed to have indeed been related to the sex composition of the surviving children (Reher and Sanz-Gimeno 2007, Reher and Sandström 2015; Marco-Gracia forthcoming). Those families who did not have previous sons not only were more likely to continue bearing more children but also had shorter birth intervals, so fertility control strategies where used to control fertility and ensure at least a male heir. Reher and González-Quiñones (2003) also found that the probability of dying of daughters increased more than that of boys after the death of the mother, thus suggesting that son preference was stronger for fathers. It also appears that women were discriminated through an unequal allocation of resources within the household, both in terms of nutrition and educational investments (Sarasúa 2002; Borderías et al. 2014). As well as to the relative backwardness of this area, literacy rates also testify to how unequally parents treated boys and girls: while around 40 per cent of men were literate in 1860, less than 5 per cent of women were able to read and write. Next sections address whether these underlying attitudes towards boys and girls translated into their respective survival chances at birth.
3. Aggregate sex ratios at birth
Baptism registers provide the name, sex, place and date of birth, parents’ and godparents’
names. Our records contain a total of 88,989 baptisms between 1575 and 1999. Although they are not birth registers, this information was registered quickly after birth because catholic beliefs stressed that if a child died without baptism, he or she could not enter heaven and ensure spiritual salvation (Hanlon 2016, 536; Minello et al. 2017). Our information confirms that new-borns were baptised within the same day that they were born from the late 18th century to the 1880s. The distance between birth and baptism increased in the following decades due to the decline in mortality rates and changing social and religious behaviour. Figure 1 plots sex ratios at baptism between 1575 and 1975. The number of boys per hundred girls remained exceptionally high between 1625 and 1750, averaging 113.5 (17,868 births). While the second half of the 18th century witnessed relatively balanced sex ratios at birth, these figures increased again during the 1790s and remained relatively high during the whole 19th century, and especially so in the 1800s and the 1860-1880s (reaching figures above 110 boys per hundred girls). The average sex ratio during the whole 19th century was 108.8, high enough to be statistically different from 105 at the 99 per cent confidence level (33,421 births; p-value=0.0006). Sex ratios became relatively normal from 1900 and remained so during the 20th century. As discussed in the previous section, registration quality greatly improved during the second half of the eighteenth century. Thus, our subsequent analysis will exclude the period before 1750. By doing so, we are adopting a conservative research strategy that disregards the most extreme sex ratio and therefore mitigates the possibility of finding spurious results.
Moreover, although we follow the literature in using 105, the “natural” sex ratio at birth, as our benchmark for comparison (Chao et al. 2019), the historical figure in absence of human manipulation might be slightly lower, thus making our findings even more significant. The probability of miscarriages is higher in high-mortality environments (Woods 2009). Given that the female survival advantage also exists in utero (Di Renzo et al. 2007; Dipietro and Voegtline 2017), more boys are expected to die before birth, thus pushing down the “natural” sex ratio at birth9. There is very little research on how this figure should look like in historical populations (Visaria 1967; Chahnazarian 1988), but modelling the evolution of sex ratios at birth over time 9 The mechanisms behind the higher vulnerability of male foetuses are still largely unknown (Dipietro and Voegtline 2017). As well as in perinatal and neonatal mortality, the female biological advantage continues through infancy and childhood (Waldron 1998; Drevenstedt et al. 2008; United Nations 2011; Peacock et al. 2012; Peelen et al. 2017; Zarulli et al. 2018).
5
in Europe (1750-2015) using the Human Mortality Database yields an estimated figure close to 104 in high-mortality environments as those existing in 18th- and 19th-century Spain (see figure A1 in the Appendix)10, thus making the figures observed here even more extreme. It is important to note that this historical estimate should be considered as a maximum threshold because it is based on observed sex ratios at birth, thus potentially including information from countries where female infants were being neglected11.
Fig. 1. Sex ratios at baptism, 1575-1975
Note: 25-year moving average. Dotted lines represent the 95 per cent confidence interval. Source: AMHDB.
Although female under-registration could have been an issue before 1750, it is very unlikely that it explains the extreme figures observed during the 19th century. In this regard, baptism was free of charge, so there were no incentives to avoid registration12. It is possible though that families did not register a child if he or she died shortly after birth in order to avoid the funeral fee (or if these families thought that the child was going to die shortly). If this was the case, sex ratios at birth can be considered net of neonatal deaths. This potential issue however would not prevent detecting female infanticide and/or neglect. Due to the female biological advance, more boys would be expected to die during the first hours and therefore being under-reported. However, as mentioned before, religious beliefs powerfully dictated that children should be baptised as soon as possible, so it is very unlikely that new-borns went 10 If under-registration of births was higher for girls in the past, this figure would be even lower because the sex ratios observed in the 18th and 19th centuries would be biased upwards.
11 In this regard, sex ratios at birth in Mediterranean countries tend to be relatively high, especially in earlier periods. Unfortunately, the series for Eastern European countries do not usually allow looking at the period before 1950.
12 In 1697, the synodal constitutions of the Archbishopric of Zaragoza established which ceremonies celebrated in the province were free and which ones involved paying a fee (and its amount). These constitutions were valid until 1943. ** **
6
unregistered. Moreover, the distance between birth and baptism is very similar for boys and girls in all periods analysed here, thus suggesting that there were no sex differences baptismal patterns.
Importantly, during the Epiphany Mass, the first festivity of each year (January 6), the local priest read aloud all birth, marriage and death registers for the previous year to make sure every event was recorded13.The priest consigned in the source that this reading was made during the Misa Mayor. Claims happened rarely, thus evidencing the high-quality of the records. When they were made, a register was added to the previous year’s book. This procedure was also in the interest of the families because the existence of the Church register was required for many subsequent formalities (such as getting married, etc.). If under-registration is not likely, such extreme sex ratios can only occur if families were disposing of some of their female babies before getting baptised14.
3. Individual characteristics and the probability of being born male Analysing the individual-level information associated to each of these births allows shedding light on the behavioural nature behind these high sex ratios. Given that twins not only suffered extremely high mortality rates, but also generated an unexpected shock to the household resources that can distort our analysis, we have excluded from the analysis those children born in families who raised twins. Apart from mitigating unobserved heterogeneity, this approach allows us to unveil discriminatory practices taking place within ordinary families.
The restricted sample contains 45,325 individuals born between 1750 and 1950. Table A2 in the appendix shows that this restriction hardly affects the socio-economic composition of our sample. If anything, the fraction of males is now lower, so excluding these children implies adopting a conservative research strategy because it reduces the likelihood of detecting practices that resulted in a higher number of male baptisms.
Table 1 reports the results of estimating a logit model assessing whether probability of being male at birth is related to different individual characteristics: birth parity, the number of children alive at the moment of birth and the sex composition of those previous children.
Theoretically, the probability of being male or female at birth is independent of the sex of previous children. In this regard, panels A and B compare whether what matters is having no males or no females siblings alive at the moment of birth. While column (1) presents the baseline specification, column (2) adds a set of control variables: mother’s age, father’s occupation and literacy, and village and period fixed effects. Lastly, columns (3) to (6) repeat the analysis focusing on different parities. Table A3 in the appendix reports summary statistics of the variables employed here.
13 The Epiphany celebrates the revelation that God is incarnated as Jesus Christ and is probably one of the most important Catholic celebrations, so all village gathered for the Misa Mayor (Major Mass). Our records indicate that this mechanism to correct potential errors in the records was at a place at least from the late 18th century.
14 Differential-stopping rules do not affect societal sex ratios at birth (Basu and De Jong 2010, 523).
7
Table 1. Probability of being male at baptism, 1750-1900
Dep. Variable: Probability of being male at birth
PANEL A
All births
Parity 2
Parity 3
Parity 4
Parity 5+
(1)
(2)
(3)
(4)
(5)
(6)
-0.005
-0.009
-0.059
0.020
0.146**
-0.044
Children alive at birth
(0.010)
(0.026)
(0.114)
(0.118)
(0.064)
(0.031)
-0.003
0.055
-0.125
0.184
0.457***
-0.116
No males at birth
(0.018)
(0.113)
(0.172)
(0.161)
(0.143)
(0.119)
Parity:
0.022
0.078
Child order = 2
(0.028)
(0.072)
0.014
0.233***
Child order = 3
(0.033)
(0.084)
0.041
0.235**
Child order = 4
(0.057)
(0.104)
0.013
0.188
Child order = 5+
(0.046)
(0.127)
Controls
NO
YES
YES
YES
YES
YES
Observations
33,101
6,510
1,102
973
830
2,384
Pseudo R2
0.000
0.003
0.009
0.009
0.024
0.005
PANEL B
All births
Parity 2
Parity 3
Parity 4
Parity 5+
(1)
(2)
(3)
(4)
(5)
(6)
0.001
-0.024
0.066
-0.073
0.001
-0.016
Children alive at birth
(0.010)
(0.023)
(0.218)
(0.080)
(0.068)
(0.030)
0.033
-0.032
0.125
-0.111
-0.189
0.058
No females at birth
(0.028)
(0.071)
(0.172)
(0.121)
(0.167)
(0.162)
Parity:
0.031
0.057
Child order = 2
(0.029)
(0.065)
0.026
0.203***
Child order = 3
(0.038)
(0.066)
0.054
0.200***
Child order = 4
(0.060)
(0.048)
0.021
0.164
Child order = 5+
(0.047)
(0.102)
Controls
NO
YES
YES
YES
YES
YES
Observations
33,101
6,510
1,102
973
830
2,384
Pseudo R2
0.000
0.003
0.009
0.008
0.019
0.004
Coefficients estimated using a logit regression model. Robust standard errors in parentheses (clustered at the village level); *** p<0.01, ** p<0.05, * p<0.1; For simplicity, the intercept is not reported. Controls include mother’s age, father’s occupation, father’s literacy, village and period fixed effects.
As shown in column (2), the probability of being male clearly increases with parity, especially for those children born third or fourth. This might be expected providing that perinatal and in utero mortality is mitigated in subsequent pregnancies and therefore especially 8
benefits boys that tend to die more when subject to tougher conditions15. Interestingly, when each parity is analysed independently, the probability of being a boy increases in parity 4 when there are no previous males alive, an effect that is not visible if there are no previous females alive (see how column 5 compares in panels A and B). It appears that the pressure to have a boy increased at high parities in those families that did not have any male descendant to the point to significantly alter the natural sex variability. In particular, at parity 4 and controlling for the other variables in the model, not having any boy alive increased the estimated probability of being male at birth from 50.9 to 61.8 per cent.
In order to further explore how these issues may have evolved over time, table 2 replicates the analysis for different periods (1750-1850, 1850-1900 and 1900-1950) and only focusing on those births happening at parity 4. The results show that the likelihood of having a son baptised significantly increased when there are no males in the family up to 1900. Having only girls however did not alter the sex ratio at birth during this period. This contrast is no longer visible after 1900 when having no male or female siblings does not affect the sex of the new baby. This implies that son preference was especially marked during the traditional demographic regime, a result that was also apparent when looking at crude sex ratios at birth (figure 1). The fact that this finding disappears in the early 20th century is compatible with a relaxation of cultural norms about the need of having male descendants. In this regard, improved living standards is likely to have mitigated economic constraints and the demographic transition towards lower fertility and mortality levels may have also altered parents’ perceptions about their children’s mortality and therefore their behaviour16. Likewise, the increasing availability of female labour opportunities, especially in the neighbouring Zaragoza, may have also affected the perceived relative value of girls.
Table 2. Probability of being male at baptism at parity 4 (by period), 1750-1950
Dep. Variable: Probability of being male at birth (at parity=4) PANEL A
1750-1850
1850-1900
1900-1950
(1)
(2)
(3)
(4)
(7)
(8)
0.163*
0.759*
No males at birth
0.267***
0.371**
-0.282**
0.083
(0.084)
(0.455)
(0.088)
(0.167)
(0.120)
(0.113)
Controls
YES
YES
YES
YES
YES
YES
Additional controls
NO
YES
NO
YES
NO
YES
Observations
2,402
163
1,618
677
1,407
749
Pseudo R2
0.007
0.081
0.008
0.025
0.009
0.013
Dep. Variable: Probability of being male at birth (at parity=4) PANEL B
1750-1850
1850-1900
1900-1950
(1)
(2)
(3)
(4)
(7)
(8)
-0.051
-0.822
No females at birth
-0.090
-0.044
0.167
0.015
(0.076)
(0.521)
(0.075)
(0.170)
(0.185)
(0.184)
Controls
YES
YES
YES
YES
YES
YES
Additional controls
NO
YES
NO
YES
NO
YES
Observations
2,402
163
1,618
677
1,407
749
Pseudo R2
0.006
0.089
0.006
0.021
0.008
0.013
15 Notice that column (2) controls for mother’s age, a dimension that would have the opposite effect. When mother’s age is not included, in column (1), this effect is not visible, probably due to having two countervailing effects: while mother’s age increases perinatal and in utero mortality, having several pregnancies decreases it.
16 Although it has been argued that parental indifference to their infants was very common in high-mortality contexts (Ariès 1979; Zelizer 1985), Woods (2003) was highly sceptical about this hypothesis.
9
Coefficients estimated using a logit regression model. Robust standard errors in parentheses (clustered at the village level); *** p<0.01, ** p<0.05, * p<0.1; For simplicity, the intercept is not reported. The first set of controls include the number of children alive at birth, village and period fixed effects. The additional controls include mother’s age and father’s literacy.
Although splitting the analysis by occupation makes the results noisier due to reduced samples, it seems that female neglect around birth was concentrated among landless and semi-landless families. In this regard, table 3 shows that having no male offspring at parity 4 increases the probability of a male birth for the latter group but not for farmers17. These families were subject to harsher economic conditions and therefore more likely to resort to extreme decisions under difficult circumstances. Not only the expectation of having to provide a dowry militated against girls, but also labourers relied on the demand for waged agricultural labour and this clearly favoured males. In this regard, while the farmers’ daughters could work on the family farm, it was difficult for landless families to find paid jobs for their daughters. This circumstance also made boys more attractive because they could better complement the family income.
Table 3. Probability of being male at baptism (at parity 4) by occupation, 1750-1900
Dep. Variable: Probability of being male at birth (at parity=4) Panel A
Farmers
Labourers
(1)
(2)
(3)
(4)
0.121
0.027
No males at birth
0.481***
0.673***
(0.263)
(0.306)
(0.134)
(0.222)
Controls
YES
YES
YES
YES
Additional controls
NO
YES
NO
YES
Observations
447
279
661
380
Pseudo R2
0.0131
0.0234
0.0269
0.0286
Panel A
Farmers
Labourers
(1)
(2)
(3)
(4)
-0.340
-0.190
No females at birth
-0.224
-0.239
(0.306)
(0.372)
(0.152)
(0.295)
Controls
YES
YES
YES
YES
Additional controls
NO
YES
NO
YES
Observations
447
279
661
380
Pseudo R2
0.0162
0.0245
0.0218
0.0182
Coefficients estimated using a logit regression model. Robust standard errors in parentheses (clustered at the village level); *** p<0.01, ** p<0.05, * p<0.1; For simplicity, the intercept is not reported. The first set of controls includes the number of children alive at birth, village and period fixed effects. The additional controls include mother’s age and father’s literacy.
4. Mortal neglect of female babies from death records** **
The previous results linking the probability of being born male to individual characteristics support a behavioural explanation for the high sex ratios at birth found at the aggregate level.
17 The size of the samples does not allow replicating the analysis on other occupational categories (shepherds, artisans, elites and/or other occupations).
10
Although anecdotal evidence on female infanticide is almost non-existent, the asymmetry between the number of male and female baptisms, especially at higher parities and in families of low socio-economic status with no previous male descendants, suggest that some families disguised female infanticide as natural deaths. As discussed before, registering a birth was free, so there were no incentives to under-report girls. Death registers, however, provide further evidence that baby girls were neglected, especially under certain circumstances. If under-registration systematically targeted girls, it would have affected both births and deaths around birth. The latter is actually more plausible because, unlike births, registering deaths imply paying a fee. In any case, if under-reporting would bias sex ratios at birth upwards, it would cause the opposite effect on deaths. Female deaths during the first hours and days would then be under-reported, thus providing more evidence of female infanticide or neglect.
In this regard, table 4 reports the results of estimating a logit model assessing whether the probability of dying during the first day of life was affected by the number of previous children alive and the sex composition of those children (controlling for other characteristics). Having no surviving brothers at birth had no effect on male mortality but significantly increased female mortality during the first day of life. This is extremely telling because they support the evidence provided previously regarding unbalanced sex ratios at birth. The number of girls dying shortly after birth is significantly higher when no other sons are alive, thus suggesting that some families were neglecting their new-born girls because they only wanted a boy. In particular, and according to the coefficient estimated in column (4), the probability of a girl dying during the first day shifts from 0.8 per cent when there was already at least one son in the family to 2.5
per cent when there is none. The distinct effect of the variables in the model, children alive at birth and parity, on male and female mortality around birth also suggests that the chance of survival was lower for baby girls, relative to boys, in already large families.
Table 4. Probability of dying during the first day of life (by sex), 1750-1900
Dep. Variable: Probability of dying during the first day of life Males
Females
(1)
(2)
(3)
(4)
0.009
0.089
Children alive at birth
-0.028
0.332***
(0.065)
(0.151)
(0.101)
(0.129)
-0.004
0.115
No males at Birth
0.339***
1.064**
(0.136)
(0.376)
(0.106)
(0.465)
Parity:
-0.515***
-1.068**
Child order = 2
-0.733***
-0.816
(0.160)
(0.438)
(0.118)
(0.541)
-0.980***
-1.361**
Child order = 3
-0.481**
-1.345*
(0.207)
(0.573)
(0.228)
(0.744)
-0.840***
-0.707
Child order = 4
-0.882***
-0.397
(0.133)
(0.584)
(0.289)
(0.648)
-0.726***
-1.365*
Child order = 5+
-0.462
-0.096
(0.163)
(0.750)
(0.342)
(0.648)
Controls
YES
YES
YES
YES
Additional controls
NO
YES
NO
YES
Observations
13,659
3,008
13,109
2,681
Pseudo R2
0.026
0.061
0.026
0.068
Coefficients estimated using a logit regression model. Robust standard errors in parentheses (clustered at the village level); *** p<0.01, ** p<0.05, * p<0.1; For simplicity, the intercept is not reported. While the first set of controls include village and period fixed-effects, the additional controls include mother’s age, father’s occupation and father’s literacy.
11
Likewise, table 5 in evidences that the probability that a baby girl died during the first day of life was higher in landless or semi-landless families with more children alive at birth, an effect that is not visible for boys despite that they should be theoretically weaker. The distinct effect of parity on male and female mortality around birth in these families also point in the same direction. Moreover, although the effect of having no male siblings on female mortality is not statistically significant, this is probably due to the reduced number of observations. The size of the coefficient is actually very similar to the one shown in Table A3, thus suggesting that this feature is likely to have increased female mortality rates (notice also the negative sign in the male columns). These results confirm those reported in the previous section on the probability of being born male. Families from low socio-economic status that relied on the demand for waged labour were subject to more strenuous circumstances and therefore more likely to be involved in this kind of practices. As well as the burden imposed by the dowry, the fact that boys were more easily employable in these agricultural markets also contributed to reducing the relative value of girls.
Table 5. Probability of dying around birth (by sex and father’s occupation), 1750-1900
Dep. Variable: Probability of dying during the first day of life Farmers
Labourers
Males
Females
Males
Females
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Children alive
-0.064
0.069
0.024
0.116
-0.007
-0.122
0.449** 0.768***
at birth
(0.169)
(0.244)
(0.210)
(0.211)
(0.114)
(0.204)
(0.177)
(0.272)
No males
-0.250
-0.211
-0.799
-0.177
-0.607*
-0.422
0.227
0.954
at birth
(0.550)
(0.744)
(1.008)
(0.805)
(0.358)
(0.419)
(0.260)
(0.642)
Parity:
-0.753**
-0.987*
= 2
-2.439*** -2.204*
-0.674**
-1.149*
-0.934
0.537
(0.345)
(0.577)
(0.845)
(1.311)
(0.310)
(0.688)
(0.718)
(0.930)
-0.800*
-1.315**
= 3
-2.122**
-0.991**
-2.429*
-0.891
-0.230
(0.435)
(0.571)
(1.032)
(0.478)
(1.247)
(0.690)
(0.755)
-0.836
-0.810
= 4
-2.435**
-1.102
-0.452
-0.812
-1.416
0.590
(0.689)
(1.056)
(1.192)
(1.337)
(0.412)
(0.581)
(0.923)
(1.064)
-0.942**
-1.151
= 5+
-2.364***
-1.294
-0.841*** -1.956***
-1.437
0.292
(0.464)
(1.147)
(0.810)
(1.452)
(0.211)
(0.588)
(1.072)
(1.478)
Controls
YES
YES
YES
YES
YES
YES
YES
YES
Add. controls
NO
YES
NO
YES
NO
YES
NO
YES
Observations
1,624
1,036
1,553
772
2,210
1,307
2,041
1,209
Pseudo R2
0.0348
0.0720
0.1016
0.1213
0.0389
0.1079
0.0583
0.1186
Coefficients estimated using a logit regression model. Robust standard errors in parentheses (clustered at the village level); *** p<0.01, ** p<0.05, * p<0.1; For simplicity, the intercept is not reported. While the first set of controls include village and period fixed-effects, the additional controls include mothers age, father’s occupation and father’s literacy.
Lastly, table 6 explores whether these patterns are also visible during the first half of the 20th century. Although individual characteristics did no longer influenced the probability of being male during this period (as shown in table 3 in the previous section), this additional exercise shows that they continued influencing female mortality during the first day of life. On the one hand, having no male siblings clearly increased the probability that baby girls died shortly after birth. On the other hand, although the number of children alive no longer seems to increase female mortality, the effect of parity still shows a distinct discriminatory pattern. In this regard, while being born at higher parities reduce male mortality rates during the first day of life, that is not the case for females. Although less important than before the demographic 12
transition as aggregate sex ratios indicate, the neglect of female babies persisted during the first decades of the 20th century.
Table 6. Probability of dying around birth (by sex), 1900-1950
Dep. Variable: Probability of dying during the first day of life Males
Females
(1)
(2)
(3)
(4)
0.127
0.063
Children alive at birth
0.042
0.020
(0.121)
(0.104)
(0.098)
(0.158)
0.092
-0.180
No males at birth
0.136
0.961**
(0.201)
(0.290)
(0.376)
(0.451)
Parity:
-1.071***
-0.874
Child order = 2
-0.150
-0.718
(0.370)
(0.591)
(0.351)
(0.819)
-1.618***
-3.102**
Child order = 3
-0.868*
0.043
(0.424)
(1.266)
(0.495)
(0.650)
-1.194***
-1.561**
Child order = 4
-0.673
0.338
(0.383)
(0.660)
(0.633)
(0.724)
-1.527***
-1.927***
Child order = 5+
-0.638
-0.160
(0.396)
(0.483)
(0.504)
(0.774)
Controls
YES
YES
YES
YES
Additional controls
NO
YES
NO
YES
Observations
3,724
1,952
3,344
1,547
Pseudo R2
0.0453
0.0824
0.0390
0.0764
Coefficients estimated using a logit regression model. Robust standard errors in parentheses (clustered at the village level); *** p<0.01, ** p<0.05, * p<0.1; For simplicity, the intercept is not reported. While the first set of controls include village and period fixed-effects, the additonal controls include mothers age, father’s occupation and father’s literacy.
5. Conclusion
Although documenting female infanticide in historical Europe has proved difficult, this article shows that the number of male baptisms in the study area was exceptionally high up to the early 20th century. In addition, individual-level info shows that the probability of male baptisms was significantly higher when there were no previous male siblings alive. Likewise, this same feature also increased female mortality rates around birth. Taken together, these findings strongly support the idea that a significant fraction of female babies were neglected or disposed away. Although there is hardly any anecdotal evidence on female infanticide, the highly unbalanced number of male and female baptisms and deaths during the first day of life, especially at certain parities and among landless and semi-landless families, suggest that some families disguised female infanticide as natural deaths. The mortal neglect of girls around birth was possibly a conscious family decision that affected a small number of families under certain circumstances. Apart from the need to provide dowries for their daughters, the fact that there existed less female waged labour opportunities also militated against girls. Although these practices were more visible during the period previous to the demographic transition, they persisted into the first decades of the 20th century, thus evidencing that son preference continued to be a strong cultural norm within these societies.
The sources are though silent regarding how these girls were missing. There were of course many ways of disposing of unwanted children and these crimes were easy to commit and difficult to prove (starvation, dehydration, strangulation, drugged to death; smothering, exposure to elements, etc.; Langer 1972, 96; Harris and Ross 1987, 5-6). In his * Elementa* 13
medicinae et chirurgiae forensis, written in Latin in 1781 (and translated into Spanish in 1796), Joseph Jakob von Plenk, a prestigious Austrian physician, devoted 18 pages to describe different ways of committing infanticide, many of them practically indistinguishable from natural deaths. Given the high rates of stillbirth and neonatal mortality, infant deaths within married couples did not raise the suspicion of civil or religious officials (Hynes 2011, 509; Hanlon 2016, 537). Anecdotal evidence from religious trials indicates that killing unwanted new-borns before baptising them was a “well-established” custom in Early Modern Aragon (Tausiet 2001)18. The same source also stresses that many infants died due to neglect in the form of inadequate feeding or lack of attention. The Catholic Church strongly condemned infanticide, so increasing the mortality of new-born girls through indirect methods, rather than directly killing them, sounds indeed plausible19. As well as the “quiet disposal” of babies, Wrigley (1966, 105) argues that families could have not striven enough to keep them alive especially during the crucial first hours of life. Exposing them at the entrance of the church or in other less visible places hoping that someone would take care of the baby was also a possibility and it is quite likely that many of these abandoned children did not survive and die of hunger or cold. Boys could undoubtedly fell victims of these practices as well, but the evidence provided here shows that families especially targeted girls.
This article challenges the notion that there were no missing girls in historical Europe, and therefore suggests that these families regulated the size and sex composition of their offspring despite the religious norms against it. These findings have important implications for our understanding of the traditional demographic regime and the subsequent transition to lower fertility and mortality rates. The low number of children raised in European families cannot be longer explained just resorting to different methods of reducing fertility either indirectly (age at marriage, fraction of singles) or directly (spacing, stopping). By increasing the mortality of their unwanted children, these families also adopted a “death control” (Harris and Ross 1987; Hanlon 2017). Resorting to infanticide or infant neglect was a complementary strategy to control fertility that allowed families to limit family size and adjust the sex-composition of their offspring according to their preferences in a context where son preference prevailed. The gradual disappearance of these practices would therefore partly contribute to explaining the decline in fertility and mortality that took place during the demographic transition would therefore partly respond to the relaxation of these practices. Likewise, the fact that many of these infanticides escaped from birth and death registers also indicates that fertility and mortality rates in pre-industrial Europe were higher than what it has been routinely assumed.
This analysis, however, focuses on a small region in North-eastern Spain, so more research is needed to assess whether this behaviour was also shared in other European regions.
References
Alesina, A., Giuliano, P., and Nunn, N. (2013), On the origins of gender roles: Women and the plough, The Quarterly Journal of Economics 128, 2: 469-530.
Ariès, P. (1979), Centuries of childhood: A social history of family life (Penguin).
Arteta de Monteseguro, A. (1802), Disertación sobre la muchedumbre de niños que mueren en la infancia y modo de remediarla y procurar en sus cuerpos la conformidad de sus miembros, su robustez, agilidad y fuerza competentes (Zaragoza).
18 Mentioned in Hanlon (unpublished manuscript, 9-10). We are grateful to the author for sharing his working paper.
19 It should be stress that the attitude of the Church towards infanticide was much more permissive in Early Modern Aragon and considered these crimes as involuntary homicides due to lack of attention providing it took place within the private sphere of the family (Tausiet 1998, 83).
14
Basu, D., and De Jong, R. (2010). Son targeting fertility behaviour: Some consequences and determinants, Demography 47, 2: 521-536.
Baten, J., and Murray, J.E. (2000), Heights of men and women in 19th-century Bavaria: Economic, nutritional and disease influences, Explorations in Economic History 37 (4): 351-369.
Bechtold, B.H. (2001). Infanticide in nineteenth-century France: a quantitative interpretation, Review of Radical Political Economy 33: 165–187.
Beltrán Tapia, F.J. (2019), Sex ratios and missing girls in late-19th-century Europe, *EHES *
Working Paper 160.
Beltrán Tapia, F.J. and Raftakis, M. (2019), ‘All little girls, the bad luck!’ Sex ratios and gender discrimination in 19th-century Greece, EHES Working Paper 172.
Beltrán Tapia, F.J. and Gallego-Martínez, D. (2017), Where are the missing girls? Gender discrimination in 19th-century Spain, Explorations in Economic History 66: 117-126.
Beltrán Tapia, F.J. and Gallego-Martínez, D. (2020), What explains the missing girls in 19th-century Spain?, Economic History Review 73 (1): 59-77.
Bengtsson, T. and Dribe, M. (2006), Deliberate Control in a Natural Fertility Population: Southern Sweden, 1766-1864, *Demography * 43: 727-746.
Bhalotra, S., Chakravarty, A., and Gulesci, S. (2020), The price of gold: Dowry and death in India, *Journal of Development Economics * 143.
Bhaskar, V., and Gupta, B. (2007), India’s missing girls: biology, customs, and economic development, *Oxford Review of Economic Policy * 23, 2: 221-238.
Bohnert, N., Jåstad, H. L., Vechbanyongratana, J., and Walhout, E. (2012), Offspring sex preference in Frontier America, Journal of Interdisciplinary History 42, 4: 519-541.
Borderías, C. and Muñoz, L. (2018), ¿Quién lleva el pan a casa en la España de 1924? Trabajo y economías familiares de jornaleros y pescadores en Cataluña y Galicia, Revista de historia industrial 74: 77-106.
Borderías, C., Pérez-Fuentes P., and Sarasúa, C. (2010), Gender inequalities in family consumption: Spain 1850–1930. In: Addabbo T et al. (Eds.). Gender Inequalities, Households and the Production of Well-Being in Modern Europe (Ashgate Farnham): 179–196.
Borderías, C., Pérez-Fuentes, P., and Sarasúa, C. (2014), La desigualdad en el consumo familiar. Diferencias de género en la España contemporánea (1850–1930). *Areas * 33: 105–120.
Boserup, E. (1989), Woman’s role in economic development (Earthscan, London).
Camps, E. (1998), Transitions in women’s and children’s work patterns and implications for the study of family income and household structure: A case study from the Catalan textile sector (1850-1925). History of the Family 3 (2): 137–153.
Carmichael, S.G., and Rijpma, A. (2017), Measuring Agency, in J.L. van Zanden, A. Rijpma and J. Kok (eds.), Agency, gender and economic development in the world economy, 1850-2000. Testing the Sen hypothesis (Routledge, London): 51-72.
Chahnazarian, A. (1988), Determinants of the sex ratio at birth: Review of recent literature, Social Biology 35: 214-235.
Chao, F., Gerland, P., Cook, A.R., and Alkema, L. (2019), Systematic assessment of the sex ratio at birth for all countries and estimation of national imbalances and regional reference levels, Proceedings of the National Academy of Sciences 116 (19): 9303-9311.
Coale, A.J., Demeny, P., and Vaughan, B. (1968), Regional model life tables and stable populations.
Das Gupta, M., Zhenghua, J., Bohua, L., Zhenming, Z., Chung, X, and Hwa-Ok, B. (2003), Why is son preference so persistent in East and South Asia? A cross-country study of China, India and the Republic of Korea, *Journal of Development Studies * 40, 2: 153–187.
15
Derosas, R., and Tsuya, N.O. (2010), Child control as a reproductive strategy, in Tsuya, N.O., et al. (eds.), Prudence and Pressure: Reproduction and Human Agency in Europe and Asia, 1700–1900 (Cambridge University Press, Cambridge): 129–155.
Dilli, S., Carmichael, S.G., and Rijpma, A. (2019), Introducing the Historical Gender Equality Index, Feminist Economics 25 (1): 31-57.
Dipietro and Voegtline (2017), The gestational foundation of sex differences in development and vulnerabililty, Neuroscience 342: 4-20.
Di Renzo, G.C., Rosati, A., Sarti, R.D., Cruciani, L., and Cutuli (2007), Does fetal sex affect pregnancy outcome?, Gender Medicine 4, 1: 19-30.
Drevenstedt, G.L., Crimmins, E.M., and Vasunilashorn, S. (2008), The rise and fall of excess male infant mortality, Proceedings of the US National Academy of Sciences 105, 5016-21.
Drixler, F. (2013), Mabiki: Infanticide and population growth in Eastern Japan (Berkeley).
Germán Zubero, L. (2009), Coste de la vida y poder adquisitivo de los trabajadores en Zaragoza durante el primer tercio del siglo XX. In C. Forcadell Alvárez (ed.), Razones de historiador. Magisterio y presencia de Juan José Carreras. Zaragoza: Institución Fernando el Católico, 373-390.
Giuliano, P. (2015), The role of women in society: From preindustrial to modern times, CESifo Economic Studies 61 (1): 33-52.
Giuliano, P. (2018), Gender: A historical perspective, in S.L. Averett L.M. Argys and S.D.
Hoffman (eds.), Oxford Handbookd on the Economics of Women (Oxford, Oxford University Press).
Guinnane, T. W., and Ogilvie, T. W. (2014), A two-tiered demographic system: ‘insiders’ and
‘outsiders’ in three Swabian communities, 1558-1914 , The History of the Family 19, 1: 77-119.
Gupta, B. (2014), Where have all the brides gone? Son preference and marriage in India over the twentieth century, *Economic History Review * 67, 1: 1–24.
Hajnal, J. (1965), European Marriage Patterns in Perspective, in Glass, D. V. and Eversley D.E.C. (eds.), *Population in History, Essays in Historical Demography *(London, Edward Arnols): 101-143.
Hanlon, G. (2016), Routine infanticide in the West, 1500-1800, *History Compass * 14, 11: 535-548.
Hanlon, G. (2017), Death control in the West: New research on routine infanticide in Northern Italy from the 16th to the 18th century (unpublished manuscript; version: July 24, 2017).
Harding, S.F. (1984). Remaking Ibieca: Rural Life in Aragon under Franco (The University of North Carolina Press).
Harris, M, and Ross, E.B. (1987), Death, Sex, and Fertility: Population Regulation in Preindustrial and Developing Societies (Columbia University Press, New York).
Hatton, T.J., and Bray, B.E. (2020), Long run trends in the heights of European men, 19th-20th centuries, Economics & Human Biology 8, 3: 405-413.
Horrell, S., and Oxley, D. (2016), Gender bias in 19th century England: Evidence from factory children, *Economics & Human Biology * 133: 47-64.
*Human Mortality Database. * University of California, Berkeley (USA), and Max Planck Institute for Demographic Research (Germany). Available at www.mortality.org (data downloaded on 09.09.2018).
Hrdy, S.B. (1999), Mother nature: A history of mothers, infants and natural selection (New York).
Hynes, L. (2011), Routine infanticide by married couples? An assessment of baptismal records from 17th century Parma, *Journal of Early Modern History * 15: 507-530.
16
Jarque Martínez, E., and Salas Ausens, J.A. (2007), La vida en la historia de Cosuenda. Siglos XVI-XIX (Zaragoza: Institución Fernando el Católico).
Johansson, S.R. (1984), Deferred infanticide: excess female mortality during childhood, in Hausfater, G., Blaffer, S. (eds.), Infanticide: Comparative and Evolutionary Perspectives (Aldine, New York): 463–485.
Knodel, J. and De Vos, S. (1980), Preferences for the sex of offspring and demographic behavior in 18th- and 19th-century Germany: an examination of evidence from village genealogies, Journal of Family History 5, 2: 145-166.
Kolk, M. (2011), Deliberate birth spacing in nineteenth century northern Sweden, European *Journal of Population/Revue européenne de Démographie * 27, 3: 337-359.
Lana Berasain, J. M. (2007), El poder de compra de jornaleros y criados. Salarios reales y mercados de trabajo en la Navarra rural, 1781-1936, Investigaciones de Historia Económica 7: 37-68.
Langer, William L. (1972), Checks on Population Growth: 1750-1850, Scientific American 226, 2: 92-99.
Lynch, K.A. (2011), Why weren’t (many) European women “missing”?, History of the Family 16: 250-266.
Manfredini, M., Breschi, M., and Fornasin, A. (2016), Son preference in a sharecropping society. Gender composition of children and reproduction in a pre-transitional Italian community, Population 71 (4): 641-658.
Marco-Gracia, F. J. (2017), Influence of the Social and Economic Changes in the Fertility Control: an analysis of rural Spain based on microdata (16th-21st centuries), PhD
dissertation
(University
of
Zaragoza).
Available
at
https://zaguan.unizar.es/record/61544/files/TESIS-2017-044.pdf (downloaded on 18.05.2020).
Marco-Gracia, F. J. (2019). Fertility control due to short-term economic stress in rural Aragón (Spain), 1801-1909, Revista de Historia Económica – Iberian and Latin American Journal of Economic History, on-line. Doi: 10-1017/S0212610919000144.
Marco-Gracia, F.J. (forthcoming). Adapting family size and composition: childhood mortality and human agency in fertility decisions in rural Spain, 1750-1949, Journal of Interdisciplinary History.
Martín Rodríguez, M. (1984), Pensamiento económico español sobre la población (Piramide).
Martínez-Carrión, J.M., Cámara, A.D., and Pérez-Castroviejo, P.M. (2016), Parámetros antropométricos de los reclutas españoles antes de la transición nutricional. Análisis de las desigualdades territoriales (1858-1913), Nutrición Hospitalaria 33, 6: 1477-1486.
McNay, K., Humphries, J., and Klasen, S. (2005), Excess female mortality in Nineteenth-century England and Wales, *Social Science History * 29, 4: 649-681.
Minello, A., Dalla-Zuanna, G., and Alfani, G. (2017), First signs of transition: The parallel decline of early baptism and early mortality in the province of Padua (northeast Italy), 1816‒1870, Demographic Research 36, 27: 759-802.
Peacock, J.L., Marston, L., Marlow, N., Carvert, S.A., and Greenough, A. (2012), Neonatal and infant outcome in boys born very prematurely, Pediatric Research 71, 3: 305-310.
Peelen, M., Kazemier, B., Ravelli, A., De Groot, C., Van der Post, J., Mol, B., Hajenius, P., and Kok, M. (2017), Impact of fetal gender on the risk of preterm birth, a national cohort study, Acta Obstet Gynecol Scand 95: 1034-1041.
Pérez Moreda, V., Reher, D., and Sanz-Gimeno, A. (2015), *La conquista de la salud. *
Mortalidad y modernización en la España contemporánea (Madrid: Marcial Pons Historia).
Pinnelli, A., and Mancini, P. (1997), Gender mortality differences from birth to puberty in Italy 1887-1940, in C.A. Corsini and P.P. Viazzo (eds.), The decline of infant and child 17
mortality. The European experience: 1750-1990 (The Hague: Martinus Nijhoff Publishers): 73-93.
Reher, D., and González-Quiñones, F. (2003), Do parents really matter? Child health and development in Spain during the demographic transition, Population Studies 57, 1: 63-75.
Reher, D., and Sandström, G. (2015), Dimensions of Rational Decision-Making during the Demographic Transition; Aranjuez (Spain) Revisited, Historical Life Course Studies 2: 20-36.
Reher, D., and Sanz-Gimeno, A. (2007), Rethinking Historical Reproductive Change: Insights from Longitudinal Data for a Spanish Town, *Population and Development Review * 33, 4: 703-727.
Sandström, G. and Vikström, L. (2015), Sex preference for children in German villages during the fertility transition, Population Studies 69 (1): 57-71.
Sarasúa, C. (2002), El acceso de niños y niñas a los recursos educativos en la España rural del siglo XIX’, in J. M. Martínez-Carrión (ed.), El nivel de vida en la España rural, siglos *XVIII–XX *(Salamanca, 2002): 549–612.
Schofield, R. (2000), ‘Short-run and secular demographic response to fluctuations in the standard of living in England’ in Bengtsson, T. and Saito O., Population and Economy: From Hunger to Modern Economic Growth (Oxford, Oxford University Press), 49-71.
Sen, A. (1990), More than 100 million women are missing, New York Rev. Books 37, Dec.
20: 61-66.
Szołtysek, M., Kluesener, S., Poniat, R., and Gruber, S. (2017), The patriarchy index: A new measure of gender and generational inequalities in the past, Cross-Cultural Research 5, 3: 228–262.
Tabutin, D. (1978), La surmortalité féminine en Europe avant 1940, Population 33 (1): 121-147.
Tausiet, M. (2001), Witchcraft as metaphor: Infanticide and its translations in Aragón in the 16th and 17th centuries, in S. Clark (ed.), Languages of Witchcraft (Palgrave: London): 179-195.
United Nations (2011), Sex differentials in childhood mortality (New York: United Nations).
Visaria, P.M. (1967), Sex ratio at birth in territories with a relatively complete registration, Eugenics Quarterly 14: 132-142.
Van Bavel, J. (2004), Deliberate Birth Spacing before the Fertility Transition in Europe: Evidence from Nineteenth-Century Belgium. Population Studies, 58 (1): 95-107.
Van de Walle, E. (1992), Fertility Transition, Conscious Choice, and Numeracy, Demography 29: 487-502.
Waldron, I. (1998), Sex differences in infant and early childhood mortality: Major causes of death and possible biological causes, in Waldron (ed.), Too young to die: Genes or gender (New York: United Nations): 64-83.
Williamson, N.E. (1976), Sons or daughters: A cross-cultural survey of parental preferences (Beverley Hill, Calif., Sage).
Woods, R. (2003), Did Montaigne love his children? Demography and the hypothesis of parental indifference, Journal of Interdisciplinary History 33, 3: 421-442.
Woods, R. (2009), Death before birth: Fetal health and mortality in historical perspective (Oxford: Oxford University Press).
World Health Organisation (2019), World Health Statistics 2019.
Wrigley, E.A. (1966), Family limitation in pre-industrial England, The Economic History Review 19, 1: 82-109.
Wrigley, E.A. and Schofield, R. (1981), The population history of England, 1541-1871: A *reconstruction *(Cambridge, Harvard University Press).
18
Zarulli, V., Barthold Jones, J.A., Oksuzyan, A., Lindahl-Jacobsen, R., Christensen, K., and Vaupel, J.W. (2018), Women live longer than men even during severe famine and epidemics, Proceeding of the National Academy of Sciences.
Zelizer, V.A. (1985), Pricing the priceless child: The changing social value of children (New York).
19
**APPENDIX **
Document 1. Register of baptism
Translation: At the parish church of Alfamén in May 20 (1787), I, the priest Joseph Garcia, vicar of this church, baptised a girl born the previous day. Daughter of Domingo Longares (born in Santa Cruz) and Isabel Morales (born in Ricla), legitimately married and parishioners of this church of Alfamén: granddaughter of Pedro Longares and María Ximeno on the father’s side; and Francisco Morales and Isabel Romeo on the mother’s. She was named Manuela. Her godmother was Antonia Pérez (born in Bardallur) who, in accordance to this spiritual relationship, has the obligation of teaching her the Christian doctrine in the absence of her parents.
Source: AMHDB.
20
Table A1. Frequency statistics (by year)
Sex
1550-99 1600-49 1650-99 1700-49 1750-99 1800-49 1850-99 1900-49
Total
Female
735
2,140
2,687
4,615
7,283
7,477
8,531
7,187
40,655
Male
744
2,329
3,047
5,214
7,688
8,104
9,309
7,596
44,031
Total
1,479
4,469
5,734
9,829
14,971
15,581
17,840
14,783
84,686
Father’s occupation
1550-99 1600-49 1650-99 1700-49 1750-99 1800-49 1850-99 1900-49
Total
Artisans
45
51
17
15
67
389
859
596
2,039
Elites
18
21
40
28
21
27
112
134
401
Farmers
10
17
40
52
248
1,680
3,353
3,456
8,856
Labourers
10
10
0
3
265
1,441
4,663
4,352
10,744
Other
11
35
53
37
47
119
318
424
1,044
Shepherds
5
13
7
5
33
264
565
586
1,478
Missing
1,380
4,322
5,577
9,689
14,290
11,661
7,970
5,235
60,124
Total
1,479
4,469
5,734
9,829
14,971
15,581
17,840
14,783
84,686
Father’s literacy
1550-99 1600-49 1650-99 1700-49 1750-99 1800-49 1850-99 1900-49
Total
Illiterate
0
10
0
9
50
1,996
6,166
2,800
11,031
Literate
5
3
9
9
32
606
3,879
6,431
10,974
Missing
1,474
4,456
5,725
9,811
14,889
12,979
7,795
5,552
62,681
Total
1,479
4,469
5,734
9,829
14,971
15,581
17,840
14,783
84,686
Source: AMHDB.
21
Fig. A1. Sex ratios at birth in Europe, 1750-2016
Source: Human Mortality Database. Coverage varies by country. Although the fitting line is estimated all the available data, the graph does not include the observations above 110 and below 100 so as to better visualize the trends. Excluding those observations from the estimation hardly alters the trend depicted here.
22
Table A2. Frequency statistics, 1750-1950
Whole sample
Restricted sample
Sex
Obs.
%
Obs.
%
Female
30,478
48,24
22,107
48,77
Male
32,697
51,76
23,218
51,23
Total
63,175
100,00
45,325
100,00
Period
Obs.
%
Obs.
%
1750-1799
14,971
23,70
9,691
21,38
1800-1849
15,581
24,66
10,009
22,08
1850-1899
17,840
28,24
13,104
28,91
1900-1949
14,783
23,40
12,521
27,62
Total
63,175
100,00
45,325
100,00
Father’s occupation
Obs.
%
Obs.
%
Farmers
8,737
36,38
6,521
33,48
Labourers
10,721
44,64
8,983
46,12
Shepherds
1,448
2,29
1,246
6,40
Artisans
1,911
7,96
1,658
8,51
Elites
294
1,22
285
1,39
Other
908
3,78
783
4,02
Total
24,019
100,00
19,476
100,00
Father’s literacy
Obs.
%
Obs.
%
Illiterate
11,012
50,15
8,301
47,91
Literate
10,948
49,85
9,027
52,09
Total
21,960
100,00
17,328
100,00
Note: The restricted sample excludes those individuals born in families who had twins.
Source: AMHDB.
23
Table A3. Summary statistics, 1750-1950
Obs.
Mean
St. Dev.
Min
Max
(1)
(2)
(3)
(4)
(5)
Sex (Male=1)
45,325
0.512
0.500
0
1
Dead during the first day of life
45,325
0.0152
0.122
0
1
Children alive at birth
45,325
1.560
1.715
0
14
No males at birth
45,325
0.526
0.499
0
1
No females at birth
45,325
0.538
0.499
0
1
Parity (birth order)
45,325
3.602
2.528
1
17
Mother’s age
30,313
29.99
6.399
15.5
50.0
Father’s literacy
17,328
0.521
0.500
0
1
Source: AMHDB.
24