Did the Promise of Baptism Really Boost Fertility in Georgia?

Yes! It Really Did!

Yesterday, I had an article at the Institute for Family Studies discussing a policy by Patriarch Ilia II of the Georgian Orthodox Church to personally baptize the 3rd or higher children of married Orthodox couples. I showed that immediately after his policy became effective in 2008, there were sharply higher births of higher-parity children and higher births to married women. For these and a variety of other reasons, there’s compelling evidence to suggest that a social or cultural figure substantively altered baseline fertility patterns for the country of Georgia.

But there was a bit of criticism. On Twitter, economic history blogger Pseudoerasmus suggested that I may be ignoring some important trends, namely, a shared U-shape in fertility across a large number of post-Soviet states, reflecting fertility losses in the chaos of independence, followed by fertility recovery. He also suggested I needed to control for income better.

These are totally fair criticisms. My article was limited in length, so I couldn’t address them as thoroughly as I would have liked. This post will address those criticisms.

Is there a common U-shape to post-Soviet fertility?

To start with, we need to discuss the data. Pseudoerasmus points to Total Fertility Rate data from the World Bank, as shown:

By that chart, Georgia’s trends look manifestly unimpressive.

But what’s going on here?

Well, in my article, I briefly discussed that there are major shortfalls in the data surrounding Georgia. For example, births tend to be slightly under-registered, so getting a true, complete fertility estimate requires an adjustment for under-registry. Luckily, somebody else has done this, a Georgian demographer named Giorgi Tsuladze. Conveniently, he also produced an independent estimate of Georgia’s population. Here’s a comparison of his population estimate vs. the World Bank’s vs. Georgia’s official figures:

So what’s going on here? Well, the 2002 Census adopted by Geostat and the World Bank was bonkers-wrong, because a set of ambiguously worded questions and low public trust in Census-takers led to a massive under-estimation of pre-2002 emigration from Georgia. The 2014 Census corrected for these problems.

Geostat used official birth, death, and improved migration data from 2002 onwards to estimate population from the 2002 baseline… but critically mis-estimated migration, natality, and mortality, and so massively missed the mark on 2014 population. Once the improved Census was taken, they trued up, with no backwards revisions.

World bank accepted the 2002 Census and the 2014 Census… and just smoothed gently between them. This method is a bit silly since those two Censuses are not comparable as they treated residency differently.

Tsuladze makes some plausible adjustments to the 2002 Census to harmonize it with prior and subsequent censuses. He then uses his own home-brew birth/death/migration estimates that he explains at length in the links I’ve posted above; spoiler, his methods are the best you can do for Georgian population, I think. And here’s the kicker: before the 2014 Census had been taken, Tsuladze had estimated 2014 population, and he was within 75,000 people of the total. That’s pretty darn good given the low information quality!

As such, I adopt Tsuladze’s estimates for Georgian births and for Georgian population. Tsuladze’s birth estimate is not that different from the World Bank’s, but his estimate for the population of given age groups of women, used to calculate the total fertility rate, is sometimes very different. As a result, here’s World Bank TFR vs. Tsuladze+Lyman for recent years, trying to mimic Tsuladze’s adjustments.

The world bank appropriately recognizes that Georgia’s official statistics woefully underestimate fertility. But they nonetheless miss a substantial amount of Georgia’s TFR thanks to incorrect population estimates, excessive smoothing, and in the later years probably a wrong estimate of total births as well. The World Bank understandably wants to avoid adopting unreliable sources for statistics, and biases towards official data rather than random academics, but this is a case where the academic in question did excellent work, and the bank should adopt the figures in question, not least because they predicted accurately.

What about other Soviet countries? Do they have similar problems?

Very few have as severe of data problems as Georgia, though some of the Central Asian countries give Georgia a run for its money. I’ve tried to compile the best estimates I can for each of the former constituent states of the USSR. The graph below shows them all.

I’ve highlighted Georgia. What you should notice is that the only other countries have a similar-scale change as Georgia had from 2007–2009 are Mongolia and Kazakhstan. In other words, there are some cases of similar-magnitude changes in fertility, but they are not common.

The key thing to point out here is that there’s not actually that much of a shared trend among these countries. A lot of them do show a period of decline and then recovery, but many, including Georgia, don’t evince that trend during this time period. Some show decline and then stability, like Moldova or Turkmenistan. Some show rough stability across the period, like Georgia’s closest neighbors Azerbaijan and Armenia, and Georgia itself, except its single big leap. Some show declines throughout the period, like Uzbekistan.

Now, it is true that, Russia, Belarus, Ukraine, and the Baltics show a similar trend. But that trend isn’t one that Georgia shared before or after the big jump. In other words, there’s no “there” there. There just isn’t a common post-Soviet fertility experience. In 10 of these countries, 2016 fertility is higher than 1995 fertility! The years of lowest-ebb for fertility show some significant variation as well.

So, Georgia’s experience isn’t just some standard post-Soviet U-shape. It exhibits a fairly unique shape, breaking with trends we observe in other countries.


Does Georgia’s fertility match nearby regions?

But maybe countries are too big of units. Maybe we need a more comparable unit for Georgia. We can also look at fertility in Russia’s Caucasus Federal subjects, specifically the ones in close proximity to Georgia. This gives us a bigger sample size of discrete region-specific fertility trends that may be nearby or comparable to Georgia. Here’s a graph of fertility rates in all of the Russian Federal subjects and the independent countries in the Caucasus:

Some of the Russian Federal subjects have big changes in similar years as Georgia. Check out Ingushetia’s rapid rise and fall!

But the trick is that many of these regions are very small. Georgia is the 3rd most populous of these regions in every year after Krasnodar Krai and Azerbaijan. The average population of the remainder of the regions is 1.5 million people, less than half the size of Georgia, and the median size is 860,000. The big-change region of Ingushetia has well under half a million residents.

Let’s simplify a bit and group these regions. First, let’s group by Russian regions vs. other regions, getting a population-weighted TFR.

Georgia had fertility that was about typical for the Caucasus region: higher than typical for its Russian neighbors, but lower than its Armenian or Azeri neighbors. Armenia and Azerbaijan saw a spike around 2003 or 2004, which Georgia also experienced. Then when Russian fertility fell in 2005, so did Georgian. But when Russian fertility rocketed north from 2006 to 2007, Georgian fertility barely inched up, more similar to the South Caucasus. Then when Georgian fertility did vault upwards in 2007 to 2008, so did Russian, while the South Caucasus fell… and from 2008–2009 when Georgia rose again, both Russian and South Caucasus fertility fell. Since then, the three regions have evinced different trends; Russia has shown a gradual rise, South Caucasus a rise then fall, while Georgia has basically been flat.

What happened in Russia in 2006–2008? I’ll return to that. But for now, let’s cut the data another way: by religion! Let’s compare regions that are 30%+ Muslim vs. other Caucasus regions.

Muslim regions have tended to have higher fertility, while Orthodox regions have lower fertility. Georgia’s fertility has been anomolously high for an Orthodox region throughout the period, but by 2009 was higher than that experienced by even Muslim parts of the Caucasus. There is a 2006–2008 rise in Muslim and Orthodox fertility, and they do have some similar pre-trends with Georgia. But again, Georgia’s rise is exceptional in size, and really remarkable for an Orthodox country.

So what happened in Russia in 2006–2008?

Well, they offered enormous incentives. In a previous article for IFS, I argued that financial incentives have a limited impact on fertility. But in 2006, Russia offered an incentive for kids nearly equal to a year’s income. It would be equivalent to the United States offering $30,000 to $40,000 for having a second kid. Russia is spending $53 billion a year in a much poorer country to boost its fertility: and it has worked! Fertility has risen! From 1.29 to 1.78 in 2015 (it fell in 2016)! That’s a very big cost for still way-below-replacement-rate fertility! To get a similar percentage of GDP, the US would need to shell out $750 billion on childbearing incentives. If the US dropped $750 billion per year on fertility incentives, I feel very confident saying that fertility would rise sharply. When we say fertility incentives have a small effect, what we’re saying is that they bang for your buck is low, not that no amount of money could ever change fertility. Russia also began giving special awards, honors, and public plaudits to parents with many families, and state media promoted childbearing as well. A coordinated financial and social campaign in a country with abnormally-low fertility (well below desired fertility!) can almost certainly boost fertility.

So if the financial incentives to having a kid in Russia changed from 2006 to 2008 by the equivalent in the US of $30,000 per kid, well then, that’s pretty gigantic.

And unsurprisingly, the Russian fertility response has been very robust! Because of course it has been. Their fertility jumps a year before Georgia’s because Russia’s fertility campaign began a year before Georgia’s.

It turns out, numerous post-Soviet countries have expanded benefits for childbearing since the early 2000s… so it’s no surprise that short-run fertility has risen! Now, we do not yet know if these increases will last, or if they’ll boost completed fertility. And the most effective financial incentives have positively gargantuan price tags. So this isn’t a way of saying, “fertility policy is highly effective.” It’s just a way of noting that the comparable-country time series is itself contaminated by some extraordinary exogenous factors.

Taken together, I think there’s very little reason to think that we are observing some kind of macro-regionally-correlated fertility trend that explain’s Georgia’s jump. I would suggest that it’s reasonable to think that at least a third or half of Georgia’s fertility spike is causally associated with Patriarch Ilia’s campaign, one way or another. There may be some “natural” pressure to go up… but it’s just not enough to explain the jump.


Do income changes explain fertility?

It is widely demonstrated in the academic literature that changes in economic prosperity alter fertility. Generally speaking, higher levels of economic development reduce fertility, but within-development-level economic shocks (so, economic cycle booms and bust) tend to boost fertility for income increases, reduce it for declines. So, we can look at the PPP per capita for our countries and see if there’s a time series correlation.

For some countries, it’s very clear. Here’s Latvia:

Rising prosperity in the 2000s boosted fertility, the recession slashed it, and the recovery boosted it again. Textbook case. So what does Georgia look like?

The fertility boom occurred during a period of particularly weak economic growth. There’s just no way that in 2008 and 2009 Georgia’s reproductive-age females were suddenly flush with cash; think about what was happening in the global economy at that time. Rather, this fertility boom happened in spite of weak income performance. And as Georgia’s income has risen since, TFR has been basically unchanged, just as it was basically unresponsive to income before 2007.

In other words, while income certainly has some effect, including an income variable in a regression for Georgia would mostly serve to make the implied effect of Patriarch Ilia’s campaign even bigger.

Again, I’m not saying 100% of increased fertility is Patriarch Ilia’s doing. I’m just saying a lot of it is.


Conclusion

The economics blogosphere and twitterverse is a wonderful place because I can post and article and get thoughtful, useful feedback and commentary within minutes. And the wonders of the internet mean I can find all the data to check those comments and verify my work within hours, and publish a follow-up in days. I am as convinced as I was a few days ago that Patriarch Ilia caused a demographically-significant boom in Georgia’s fertility. He’s not solely responsible, Georgia’s conditions are unique, and there may have been favorable winds at his back from regional trends. But Georgia’s experience is too extreme, too unique, and too uncorrelated with possible confounds to not grant that Patriarch Ilia had some meaningful effect. I look forward to Pseudoerasmus telling me what I’ve missed in this post!

Check out my Podcast about the history of American migration.

If you like this post and want to see more research like it, I’d love for you to share it on Twitter or Facebook. Or, just as valuable for me, you can click the recommend button at the bottom of the page. Thanks!

Follow me on Twitter to keep up with what I’m writing and reading. Follow my Medium Collection at In a State of Migration if you want updates when I write new posts. And if you’re writing about migration too, feel free to submit a post to the collection!

I’m a native of Wilmore, Kentucky, a graduate of Transylvania University, and also the George Washington University’s Elliott School. My real job is as an economist at USDA’s Foreign Agricultural Service, where I analyze and forecast cotton market conditions. I’m married to a kickass Kentucky woman named Ruth.

My posts are not endorsed by and do not in any way represent the opinions of the United States government or any branch, department, agency, or division of it. My writing represents exclusively my own opinions. I did not receive any financial support or remuneration from any party for this research.