https://orcid.org/0000-0002-7098-686X
![Sample our Global Development journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5929/TOC-Subject-Sample-2021-Global-Development.jpg"})
Abstract
The collapse of communism led to highly skewed sex ratios in Albania, which had a long patriarchal tradition before the advent of communism. While the use of sex-selective abortions in the region is well-known, little is known about other forms taken by revealed son preference, such as differential stopping behaviour and birth spacing. Pooling data from the Demographic Health Surveys in 2008–2009 and 2017–2018, we find evidence of a higher proportion of boys being born at the last birth, indicating that parents practice differential stopping behaviour. Using a logit model, we also show that in sonless households parents shorten the birth intervals significantly, endangering mothers’ and children’s health. We conclude that differential stopping behaviour and short birth spacing are prevalent in all regions and across the socio-economic spectrum.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 The ultrasound technology was first available in 1987 in the main maternity hospital in Tirana, but after 1993 it started to become available in most private clinics across the country, and ultrasound scans were performed at low cost for the patients (Guilmoto et al., Citation2012).
2 Edwards (Citation2003) notes that the very existence of a word to describe a groom living with the bride’s parents (kollovar or dhëndër brënda) is an indication that the practice is uncommon. Similar words exist in Greek and Turkish.
3 Internal and international migration was prohibited by law during communism.
4 Albania is one of the main sources of prostitutes in Europe (Arsovska, Citation2013). Indeed, the number of Albanian prostitutes in the European Union is about one per cent of the Albanian population (ABC News, Citation2006). To lure victims, promises of a false marriage or employment are often made (UN Women, Citation2016; Van Hook et al., Citation2006).
5 The ‘bride room’ is a bedroom within the groom’s family household designated for use by the new couple. It may also be a separate dwelling depending on the family’s wealth. The groom and his family are responsible for furnishing the room or buying and furnishing the dwelling (de Rapper, Citation2012).
6 The bride’s trousseau includes pieces of furniture, garments, some household linen and embellishments. It also may include small gifts (for example, cloth napkins, socks, etc.) for the groom and his family. The trousseau was traditionally displayed at the wedding. While this tradition has fallen in desuetude, the trousseau remains a common expense for the bride’s family in contemporary Albania.
7 Based on the 2011 Census, 59 per cent of Albanians declared to be Muslim, 17 per cent to be Christian, and the remaining 24 per cent declared other religious denominations (INSTAT, Citation2011).
8 Another possible cut-off would be for birth-to-conception intervals of <6 months which are considered as very high-risk by the World Health Organization (Citation2005). However, as a mere 5 per cent of birth intervals are of <6 months in our data, the logit results showed no clear pattern and are thus not presented here but are available on request.
9 As common in the literature, we do not have data on household’s wealth at the time of conception, but only data on household’s wealth at the time of the survey. As wealth can fluctuate sharply over a lifetime, we use education, which is more stable over time, as a proxy for wealth.
10 Son preference affects women’s birth spacing in two opposite ways. On the one hand, a couple desiring a son may shorten birth intervals in the quest for a son. On the other hand, the same couple may use abortion to eliminate female fetuses and, thus, prolong the observed birth interval (as our birth intervals are calculated using birth but not terminated pregnancy). We have thus calculated the birth intervals (in months) and the proportion of birth intervals 24 months or less by sex of the previous children and the sex of the birth (tables available in the Supplementary Materials). The only time we reach statistically significant/marginally insignificant differences is for the transition to birth order 4 in both sonless and son-only families who have longer birth intervals before the birth of a son. From our data, it thus seems that the two opposite effects of son preference on birth intervals roughly cancel each other.
11 We, however, lose statistical significance for households with four children in the sub-period 1976–1990 and 2001–2018, possibly due to the small sample size.
12 The only exception is for the period 1976–1990 for families with three children at first birth, for which the 95 per cent confidence interval includes the natural sex ratio at birth of 107.
13 Empirical evidence in many countries shows that sex-selective abortions are not commonly practised at first birth, which is used to justify the use of sex of firstborn as an instrumental variable (see, for example, Kugler & Kumar, Citation2017 and Lee, Citation2008).
14 The sex ratios by birth order presented by INSTAT are different from the sex ratio at a specific birth order for a given parity that we present in this article. While the sex ratio at birth for a given birth order is for all children born at this level, the sex ratio at a specific birth order in this article is for a specific parity size and thus, depends on the willingness of parents to stop childbearing based on the sex composition of previous children. The two can thus diverge substantially.
15 11 per cent of women in DHS claim to have had at least one terminated pregnancy.
16 Some countries, such as Canada, consider birth-to-conception intervals too short if of <18 months. As a robustness check, we have estimated the model using 18 months as the cut-off. The results are very similar and are available in the Supplementary Materials.
17 As common in the literature, some birth intervals are truncated, which may bias our results. As a robustness check, we have re-estimated the model using women aged 35 or above at the time of the survey (95 per cent of Albanian women give birth before age 35). The results are in line with the original results and are presented in the Supplementary Materials.
18 Another way to divide the country is to contrast the Northern part, the Geg region, with the Southern part, the Tosk region. There are a few noteworthy historical differences between the Geg and the Tosk. First, while people from both regions can understand each other, the dialect is different. Second, before the Ottoman occupation, in 1430, the Geg were mostly Catholics, while the Tosk were mostly Orthodox. Finally, and potentially more importantly, the customary law of Kanun was followed only in the Geg region. Today, Kanun is followed almost exclusively in the Mountain region. While Kanun was followed only in the Geg region, the Tosk was also a strongly patriarchal society in which the culture of revenge was strong. The differences between the two regions are thus modest, and we do not expect to see strong differences in son preferences between the Geg and the Tosk. We, nevertheless, divided the sample between Tosk and Geg and re-run the estimate for differential stopping behaviour and birth spacing. The results available in the Supplementary Materials do not point toward any major differences in revealed son preference across the two regions.