Tomas Sobotka on Fertility Trends in Europe

Today I would like to recommend to you a power point presentation by Tomas Sobotka, entitled Fertility Trends in Europe, and I want to recommend it because it gives a really clear and accessible summary of the state of the art in the ongoing debate about the whys and wherefores of contemporary European fertility. In this sense this post is really a companion and follow up to my earlier review of European fertility (and here, and here) which was basically itself a review of an extensive article in the Economist on the topic.

Sobotka's interest in the presentation is primarily in what has come to be known as the Second Demographic Transition (SDT), and the principal question he wants to ask is whether below-replacement fertility an inevitable outcome of this transition? His conclusion is that it isn't. My feeling is, however, that at this point in time this is an almost impossible question to answer, since we are still a long way from knowing what will be the ultimate resting point of both the mothers age at first childbirth and long term completed cohort fertility. Sobotka gives a very extensive and exhaustive summary of many of the current arguments about the SDT, and some of his findings are even surprising, but in this post I would like to neatly sidetep such issues and instead focus on two major topics which also feature prominently in his exposition - a) "missing births" and "replacement migration" and b) "differential fertility" between the different regions of Europe and the uneven economic impacts of these.

Birth Postponement and Missing births

Now Tomas Sobotka is an extremly able and interesting young demographer, and his recent PhD Thesis "Postponement of childbearing and low fertility in Europe" (which is very readable for a piece of academic research and is online in its entirety here) is the best general introduction available to this important topic which lies at the heart of the whole low fertility problem. Postponement, we should recall, is simply the process whereby mothers start having children at ever later ages. During the years in which this process runs its course (and we still do not really know yet where the upper limit is here), a phenomenon known as the "birth dearth" occurs, since considerably fewer children are actually born and unusually low readings are registered on the "period" Total Fertility Rate indicator.

You could think of this process as being in some ways similar to the recent structural rise in energy prices (if thinking like this helps you at all), since annual inflation rates have registered very high readings as energy prices have become steadily more expensive when compared with other goods, but of course, once they reach their new higher levels (assuming they stay there) the sheer fact that they are high is not in itself inflationary (although, and as with the high first birth ages, there may be other consequences) so the monthly year on year CPI readings once more settle down again (in this case to their old lower level) after the structural transition is completed. And so it is with the period (which just means as measured during a certain period of time) TFRs.

As is illustrated in the chart below - which comes from Wolfgang Lutz - an increase in the mean age of childbearing results in a lasting loss of births, unless the childbearing age for some reason were to decrease again, and it is this lasting loss of births that causes so many structural problems in shape of the population pyramid. The resulting need to compensate in some form or another for these missing births has lead Sobotka to slightly redifine the idea of "replacement migration" as it was originally advanced by the UN population division (and here) to mean a short term injection of people into the 25 to 40 age group to both make up for earlier "missing births" and to compensate for the current deficit since people in this age group are liable to have chiuldren (more on all of this below). Since immigration is such a politically sensitive issue this may be a much more plausible and palatable way of presenting the problem of correcting structural population damage.




Now I have gone on at some length about all this "birth postponement" rigmorole, since if you don't "get" this bit (ie that women are having children later, and that this affects the shape of the population pyramid regardless of the level at which cohort fertility ultimately settles down) then you are going to struggle with the rest of the arguments in the Sobotka presentation and you are not really going to see why having reduced fertility over an extended period of time is any sort of big deal at all.

Three ‘stylised facts’ on European fertility


As is by now reasonably well known, at the present point in time the United States and France are the only two developed countries which have managed to recover something near the magic 2.1 Tfr replacement number. We have therefore few strong reasons to believe, and indeed we may well have a good many reasons to doubt, that a significant number of countries globally will finally prove capable of retaining replacement fertility. However, as I suggest above, any conslusions we reach here need to be extremely tentative, since in the longer term "path dependent" processes may well come into play, whereby what happens tomorrow and the day after may well exert significant influence on what gets to happen in 2030 or 2040, and given that none of this has ever happened before, and that we therefore have no precedents, all conclusions need to be very guarded and open to continuous modification.

It is really for this sort of reason that we on this blog tend to focus on the here and now, and try to analyse and interpret the impact of low fertility on the present dynamics of those countries which have had below replacement fertility over an extended period of time, and especially on the economic and social dynamics of the countries in question. Whatever the impact of below replacement fertility in the longer term, our present experience shoud be able to offer a useful yardstick with which to guide and orient our deliberations.

In this context Sobotka offers us three "stylised facts" about European fertility. These are:

1/. Fertility rates in Europe are very low and further declining

2/. Current low fertility will necessarily lead to a rapid population ageing and to a decline in population size

3/. These trends are unsustainable in the long run and constitute serious threats to economy, labour market, welfare system, and thus also to the foundations of European societies

Truth be said, Sobotka is none too sure about any of these, even if, as he notes, ideas such as these have served as the justification for many of our present concerns about the longer term consequences of low fertility, leading to the use of expressions like “fertility implosion,” “baby deficit” and fears of a longer term ongoing population decline. Such concerns about lew fertility and its consequences are really nothing new, since similar concerns were already being expressed about low birth rates and their consequences in the 1920s and 1930s in a variety of countries ranging from France and Sweden to the United States (indeed to some extent it was this early arrival of below replacement fertility which lead France and Sweden to introduce the early versions of their pro-natalist policies). These early concerns were largely to disappear with the arrival of the post WWII baby booms.

Essentially Sobotka questions the usefulness of such stylised facts since he argues that, in the first place European fertility rates are not uniformly low - and even less is is it the case that lowest-low fertility is "spreading" across Europe (although it may be spreading out of Europe to other parts of the globe, and especially Asia). In particular there is strong regional variation in fertility patterns across Europe - and it is not at all clear that the longer term measure of completed cohort fertility is still declining in many European societies. Given this, the second stylised fact becomes questionable, since although ageing is a given for all developed economies as life expectancy rises, the rapidity of this ageing depends on the levels of achieved fertility and the levels of inward "replacement" migration, and neither of these are givens. So we don't necessarily simply have to sit back and accept that we are - as societies - ageing rapidly. Of course if we have ongoing low fertility, and we do simply sit back on our laurels and fail to react - as Germany, Italy and Japan are discovering to their cost - rapid ageing is the more or less inevitable consequence.

Lastly Sobotka would like to challenge the third idea - that such rapid ageing processes represent severe challenges to our economies, our labour markets and our welfare systems. Here his argument, such as he presents it, seems to be on much weaker ground, and looking at the three leading rapid agers it is hard to argue that their economies and institutions remain unaffected, even if a good of time and energy could be expended arguing about whether or not these countries are on sustainable paths in the long run.

In order to challenge the first two stylised facts Sobotka lists a number of hypotheses which he considers to be worth testing in order to examine the validity of the stylised facts under review. He suggests that:


- Extremely low period fertility rates are linked to fertility postponement and are likely to be temporary

- Pronounced regional differences are likely to prevail

- Second Demographic Transition is not necessarily linked to below-replacement fertility

- Immigration can substitute most or all of the births ‘missing’ due to below-replacement fertility

- Very low fertility and the prospects of population decline do not constitute an all-European problem.


- as a consequence fears of a European population implosion seem exaggerated

To illustrate his case Sobotka divides the old European Union 15 into four regions -North, South, German Speaking and West. As he shows in the following chart, we can discern very pronounced differences in fertility across the regions of Europe.

It is clear that treating the European Union as a single entity for demographic purposes is a rather dubious procedure to say the least. Southern and German-speaking Europe stand out as having a very different fertility pattern from Northern and Western Europe, and the principal difficulty arises in determining why this should be. The position becomes even more complicated if we add-in Central and Eastern Europe, where the ferility pattern, at least on the surface, resembles that of Southern and German speaking-Europe.


One of the problems we encouter at this point is that of getting one single theory to fit all the available known facts. One possible way forward, as Sobotka suggests, is to try to link the arrival of very low fertility (let's say here below Tfr 1.5) with the spread of the one-child family model. This seems to work in some cases, but runs into the important exception of the German-speaking countries and some parts of central Europe (and possibly outside Europe with Japan) where what also seems to be important is the number of women who never have any children.




Clearly the Chinese case would fit this model, but since there are very unusual factors at work in the Chinese case I would not be in any hurry to draw any important conclusions from this. On the other hand a comparison of societies in terms of the proportions of women having children in the higher birth orders can also prove to be revealing. In some countries the number of women who now have one child, and one child only is now quite high, while in others (Scandinavia, for example, as exemplified by Norway in the chart below) it has fallen rather than risen in recent decades.



If we look at the Spanish case a bit more closely, and compare it with the Swedish one, we can see some interesting details. One of the questions which confronts us when we come to look at birth postponement and longer term completed cohort fertility levels is the question of birth "recovery" - or how many children women in the higher age groups actually have. Here Sobotka presents some interesting data, since we find that what makes the big difference between the near replacement fertility countries and the lowest low fertility ones is very often the number of children women over 30 actually have. So if we look at a comparison of the lifetime probability of having (nother) child among women aged 30 as between Spain and Sweden we see that the evolution of this probability has not been very different between one country and the other.




But if we then go on to examine the same question in terms of the probability of having second and third children a very different picture emerges. In both the transition from the first to second child and the transition from the second to third one Spain dropped below Sweden in the mid 1980s and - even though there has been some slight recovery since the late 1990s - has remained significantly below ever since.

Is European Fertility Declining or Rebounding?

The next important question is to determine whether or not longer term fertility - the so called completed cohort fertility, as opposed to the shorter term, period, Tfr) - will continue to decline, or has actually stabilised. The answer to this difficult but necessary question is that it is still too early to say. We need to see data from at least one or two more cohorts to get a better picture, and since one cohort completes every five years, we are not going to get any definitive indication anytime soon. But we also know that - such concerns apart - longer term completed cohort patterns are not the same across countries, and once again Southern and German-speaking Europe seem to show a different, and more worrying, pattern to that of Western Europe and Scandinavia.

Sobotka enters into quite a bit of detail on the various techniques that have been developed to try and get a better measure of real period fertility (and this is hardly surprising since this is one of his specialist areas). There is no real consensus on techniques at the present time, but we do have some sort of reasonable idea of the order of magnitude of the tempo effect.

(click on image for better viewing)

The upshot really is that what is important is the depth and duration of the postponement transitition. As Sobotka himself points out the gap between the period TFR and the cohort CTFR may persist for a period of anything up to 30-40 years, and as such this is hardly a mere blip. We need to think seriously about the population structure effects of such a long drawn out process. And we still don't know how high average first birth ages can rise, so it is hard even to reach any form of meaningful conclusion about duration of the tempo effect as between countries, although we obviously do know that the effect is present and may well not be benign.


So much for the great unknowns. But what about what we already do know? Well we know that during the years of the postponement transition we have a phenomenon of "missing births" (which makes its presence felt at the level of maintaining a replacement-fertility-driven stable population pyramid), and we do know that one way of adressing this deficit is via increased immigration.

It is clear that immigration has an impact on childbearing trends and patterns in the receiving countries. Among other considerations it is obviously hardly coincidental that the two countries who have come closest to achieveing replacement fertility - France and the United States - have both experienced significant levels of inward migration (although, of course, in both cases migration alone is far from providing a complete explanation of the high levels of achieved fertility). Immigration may operate on fertility via two mechanisms: the migrants themselves may come from countries with higher fertility levels, and the new migrants may be strongly concentrated in child bearing ages. This latter factor clearly does not have a direct impact on the period Tfr in and of itself, but it does affect the annual flow of births, and in this sense may have a direct impact on the birth deficit.

Thus while the volume of migration at the levels postulated in the original UN proposal may be so large as to be politically and culturally unthinkable (and is in no case a substitute for addressing the underlying phenomenon of low fertility) it may well be that ‘replacement migration‘ in this more restricted sense could be very valuable, especially since the impact of immigration acts not only on fertility but also on population size, slowing aging and reducing the pace of any decline.

One big problem in determining the size of the impact of migration on fertility is that there is a lack of comparative cross-country data, since comparatively few studies have been carried out. Also there is a problem of definition. Who should be included, (all) immigrant women (or men), foreigners only, first generation, or second and third generation, legal and illegal etc? The use of data equating immigrants with foreigners is also problematic, especially in countries with high rates of naturalisations .

Measurement of Tfrs and age-specific fertility rates may also be problematic, especially for recent immigrants (and hence also for the 'foreign'nationality women). Also the event of migration is frequently interrelated with family formation and childbearing: immigrant fertility is also strongly dependant on timein the receiving country since immigration. Migrants period fertility rates may often be particularly high when short-duration stays are involved.

The net impact of immigrant women on Tfrs is hard to determine, but key factors are clearly the proportion of total births which are to immigrant women and the extent of fertility differences between immigrant and ‘native born' women.

So advancing the discussion somewhat since the times of the UN report - Replacement migration: Is it a solution to declining and ageing poplation (2000) - we could say that a new consensus may be emerging that migration cannot stop population ageing (since there is only a modest impact in slowing-down the process ) but that immigration may well be able to substitute for most of the birth-deficit even in countries with very low fertility (Spain, and even Italy?). Clearly the role which immigration can play depends crucially on the fertility rate in the migrant sending country. Thus migrants arriving in Spain from relatively high fertility Latin American countries like Ecuador or Brazil may have much more impact on the Tfr than those coming from very low fertility countries in Eastern Europe like Bulgaria and Romania. A French-German comparison might well prove useful here, since from the mid 1980s onwards Germany may well have received more migrants than France, but given that the origin countires of most of these migrants have themselves had low fertility, this has had only a limited impact on the fertility readings.

This factor needs to be thought about carefully in those Eastern European societies who themselves are looking presently for replacement inward migration to bolster up their limited local labour forces, since often it is assumed that migrants from countries further east like Russia, Ukraine or Moldova might plug the gap, but all these societies are themselves very low fertility ones.

What about European global demographic and economic marginalisation?


Surprisingly, having made such a strong case that what is most important about the European fertility picture is the extent of the regional varaince, Sobotka goes on to suggest that in terms of population and global influence an EU-US comparison is not only useful, it is also meaningful. Especially is this surprising when he goes on to argue that territorial expansion rather than fertility has enabled the EU to surpass the US in terms of population growth and (and I quote) "to keep pace with its economic power". Personally I think this kind of comparison is virtually useless, and doubly so in the context of a fertility analysis.



And the comparison becomes doubly problematic when it is applied to economic prowess. If I reproduce his arguments here, then it is only because they are sufficiently widespread to at least merit being refuted. If we look at the chart below, which in some sort of spurious way makes a PPP EU-US comparison, I am quite happy to say that this type of comparison has no worthwhile interpretive reading whatsoever, even if some of the people based in Brussels are fond of hearing it.


The whole process that Sobotka gets into here, and as I emphasise I respect him greatly as a demographer, and as an analyst of the postponement process, seems to me to be at best well off the point (since it seems to open up some of that good old EU-US football rivalry, when what we should be trying to do is understand what is going on in all our interests) and at worse totally illegitimate, since he had just spent the best part of his presentation stressing the strong regional differences across Europe, and how, in this sense, and idea of Europe-wide anything is likely to be pretty much a useless piece of information. And so it is in this case, as I will now argue on the basis of the charts below.

As Sobotka has stressed we have very significant differences in fertility levels between advanced countries at the present time, and these fertility differences tend to translate themselves into very different dynamics when it comes to population growth and ageing rates. So it is here that the inter-country (and possibly inter-regional within-country) differences appear to matter. If we look at a comparison of population growth rates as between Germany, Japan and the US the reason I am saying this should begin to become clearer.

The same sort of comparison may be made between the UK, France and Italy.


Or between Germany, the UK and France.




What is clear is that while population in the UK, France and the US is still growing significantly, in Italy, Germany and Japan it has become almost completely stationary. And the reason why we are seeing such differences in population growth rates is not that hard understand, since the UK, France and the US are all ageing much less rapidly than Germany, Japan and Italy, and the principal explanation for this slower ageing is the relative fertility between these countires.

And this inter-country comparison is interesting, since, when we break per capita GDP growth rates down by country, we find a rather surprising result: those developed countries who have the highest population growth rates also have the highest per capita GDP growth rates. So in economic terms arguably differential fertility does matter, and much more than Sobotka seems to imagine. Lets look at the charts.



The position is pretty clear: since the mid 1990s the relative position of Germany, Italy and Japan in terms ofg PPS per capita has deteriorated considerably. But if we come to the UK, the US and France we find something very different, they all reveal a pretty similar pattern, barring the odd detail here and there, and that is that they have more or less maintained there relative position, and in any event they have all done much better than any of the previous three.


These charts are based on data prepared by Eurostat, and show the volume index of GDP per capita as expressed in Purchasing Power Standards (PPS) (with the European Union - EU-27 - average set at 100). If the index of a country is greater than 100, then this country's level of GDP per capita is higher than the EU average and, below 100 the position is vice-versa. The basic data is expressed in PPS which thus becomes a kind of common currency eliminating differences in price levels between countries in the process making possible meaningful volume comparisons of GDP between countries. Please note that the index, since it is calculated from PPS figures and expressed with respect to EU27 = 100, is valid for cross-country comparison purposes rather than for individual country inter-temporal comparisons. Nonetheless these charts are extraordinarily revealing.

As we can see the high-population-growth (near replacement fertility) group maintains its relative GDP per capita position reasonably intact, while in the the case of the low-to-declining-population growth (lowest-low fertility) group it steadily deteriorates. As we can see, in PER CAPITA income growth terms all three of the former hold their comparative position much better than all (or any) of the latter three.

The implications of all this are really quite profound. While - for catch up growth reasons - we should only expect that developed societies should lose their relative standing vis a vis emerging economies, we would never have imagined that one group among the more established economies should be losing impetus when measured against another group, and that this difference should be reasonably correlated with both population growth and rates of ageing. This should I would hope strike some at least among those of you reading this as a rather interesting result. It appears then population median age does seem to matter, and is going to be an important conditioner on the way in which relative living standards evolve. I clearly cannot go into this more here, since the post is already way too long, but I am chipping away at getting through to the underlying economic theory which is involved here in this post.

Conclusions

1/ Very low fertility is not an all-European problem, nor is it of course an exclusively European one.
2/ A new heterogeneity and new cleavages are emerging in Europe
3/ We have a comparatively high fertility belt‘ - the Nordic countries plus North-western Europe (Benelux, FR, UK, IRE) with Tfrs in the 1.7-2.0 region, and adjusted Tfrs and cohort CTFRs around 1.9, with replacement or above-replacement migration levels.

Then we have a low fertility-high, comparatively high recent migration group in Southern Europe (Italy, Spain, Greece).

Then there is a very long term low fertility, medium migration group, with the special characteristic that much of the migration has been from societies which themselves have low fertility (Switzerland, Austria, Western Germany).

And finally a very low fertility (adjusted TFR 1.5) - emigration group. This is a very dangerous mix, suggesting strong negative population momentum and rapid population decline (East Germany, Baltic countries, Bulgaria, Ukraine, Belarus, Moldova, Russia)

All countries face longer run ageing problems which will in no case be negligable. The first group should however be better equipped to handle the problems posed. The intermediate group will be faced with important welfare sustainability and economic growth issues which it is not clear at this point how they are going to resolve. In the case of the last of the above group of countries and region very low fertility may seriously undermine social dynamic and prosperity in even the relatively short term.