Tuesday, January 10, 2017

Sex selective abortion is likely less common in Georgia than previously thought

[This blog post was co-published with EurasianetThe views presented in this article do not necessarily reflect the views of CRRC-Georgia.]

Sex-selective abortion in Georgia is a topic that has caught international attention. From an Economist article published in September 2013 to a 2015 UN report, Georgia tends to be portrayed as having one of the worst sex-selective abortion problems in the world. Closer inspection of the data, however, suggests the issue may be blown out of proportion.

The first study to draw attention to the sex-selective abortion issue in Georgia was published in 2013 in the journal International Perspectives on Sexual and Reproductive Health, and relied on statistics compiled by the World Health Organization. The authors found a sex-at-birth ratio of 121 boys for every 100 girls born in Georgia from 2005-2009. That number suggested there was a problem: one of the most common estimates of the natural sex-at-birth ratio is 105 boys for every 100 girls, or 95.2 girls for every 100 boys. Any difference between the natural and observed ratios in favor of boys is generally thought to be an proxy for sex selective abortion.

The study suggested that Georgia had one of the largest sex selective abortion problems in the world.

However, a missing data issue, a rounding error, and an anomalous sex at birth ratio in 2008, in the original study drove up the reported sex at birth ratio in Georgia. In the article, the sex at birth ratio between 2005 and 2009 is actually the average of the ratios in 2005 and 2008.  Martin McKee, one of the co-authors of the study stated, "The figure of 121 boys to 100 girls in 2005-2009 was calculated on the basis of the data submitted to the WHO at the time, from which several years were missing."

The missing data had a very large effect on the results of the study. In 2008, the ratio of boys to girls born in Georgia was exceptionally high at 128 boys born for every 100 girls. In 2005, 113 boys were born for every 100 girls, another high year for Georgia. Using these two years leads to an average of 120 boys born for every 100 girls between 2005 and 2009.

Notably, when asked about the discrepancy between the article reported 121 boys and the 120 boys to 100 girls ratio in the data, McKee acknowledged, “A very small rounding error crept in.”

With the full data between 2005 and 2009, however, the average sex at birth ratio drops to 113 boys for every 100 girls, rather than 120 – about half the reported deviation from the natural rate.


On top of the missing data, the fact that 2008’s sex at birth ratio is an outlier further exaggerates the reported magnitude of sex selective abortion in Georgia. If between 2005 and 2009, the average ratio was 113 boys for every 100 girls, the average ratio for the same period excluding 2008 is 110 boys for every 100 girls. That is to say, by excluding 2008, there were 5 excess boys born for every 100 girls rather than 8.

To flip the statistic around by looking at the ratio of girls born for every 100 boys, the average between 2005 and 2009 was 88 including 2008 and 91 when excluding it.  Translating this into the number of missing girls by subtracting the number of girls expected from the number born according to official data, suggests 6.74 missing girls for every 100 boys born when including the 2008 data. Without 2008, this drops to 4.20.


The exact causes of the situation recorded in 2008 are unknown. Although a higher than natural sex at birth ratio favoring boys is often explained by sex selective abortions and infanticide, comparing an estimate of the number of missing girls to the number of abortions over time suggests that some other factor may be at work.

Dividing the number of missing girls by the number of abortions in a year provides an estimate of the share of abortions that would need to be sex selective for it to explain the sex at birth imbalance. These calculations would suggest that sex selective abortion increased from 6% of all registered abortions in 2007 to 24% in 2008.

The calculations suggest one of three things: there was a dramatic increase in sex selective abortions in 2008, the number of unregistered abortions dramatically increased and they were also predominantly sex selective, or something else was driving the anomalous sex at birth ratio.

In the other category, many possible explanations exist. Notably, given the often poor state of data collection at the municipal level in Georgia, where births are recorded, recording error could explain the discrepancy.

The data alone cannot tell us whether 2008 saw a dramatic increase in the number of sex selective abortions or something else drove the anomalous sex at birth ratio. What is clear is that Georgia’s problem with sex-selective abortion is smaller than often portrayed.

That isn’t to say it is not a problem. In 2015, there were still about 4 missing girls for every 100 boys born.

Understanding the magnitude of the problem though is a first step towards addressing it.

Dustin Gilbreath is a Policy Analyst at CRRC-Georgia. He co-edits the organization’s blog Social Science in the Caucasus

To view the data used to calculate the figures used in this article, click here.

2 comments:

Unknown said...

Are high blood pressure (http://www.independent.co.uk/news/science/mother-blood-pressure-woman-before-conception-sex-child-baby-boy-girl-gender-study-endocrinologist-a7523861.html) and toxoplasmosis (https://blog.cerge-ei.cz/?p=2374#comment-65715) also taking into consideration when studying the discepancy between boys and girls being born?

Dustin Gilbreath said...

Not in this case, though more complex studies probably have considered it. Notably, Hepatitis B is another issue that should be taken into account with the issue: http://homes.chass.utoronto.ca/~siow/332/oster.pdf