Policy-based Population Projections for the European Union: A Complementary Approach Policy-based Population Projections for the European Union: A Complementary Approach* Philip Cafaro, Patrícia Dérer Abstract: We present new population projections out to 2100 for the countries of the European Union and for the EU as a whole under a wide range of fertility and migration scenarios. As policy-based projections rather than forecasts, they aspire not to maximize predictive success regarding future demographic developments, but to accurately show the impact of different migration and socio-economic policy choices on population numbers. Our chief aim is to clarify those policy choices for European citizens and policymakers. We show that demographic policies have the potential to markedly increase or decrease future populations across the EU. Mi- gration policy offers greater scope for infl uencing future population numbers than policies aimed at infl uencing national fertility rates. In countries with particularly low fertility rates or high emigration levels, egalitarian economic and family support policies have the potential to limit future population decreases to a signifi cant ex- tent, especially in combination. In most cases, EU nations are well placed to stabilize or slowly reduce their populations by continuing status quo policies or with moder- ate changes. Thus they are well placed to achieve one of the necessary conditions for creating ecologically sustainable societies. Keywords: Population projections · European Union · Fertility · Migration · Population policy Comparative Population Studies Vol. 44 (2019): 171-200 (Date of release: 28.10.2019) Federal Institute for Population Research 2019 URL: www.comparativepopulationstudies.de DOI: 10.12765/CPoS-2019-14en URN: urn:nbn:de:bib-cpos-2019-14en1 * This article contains supplementary material in the form of three online appendices: Detailed Discussion of Projection Scenarios (appendix I) DOI 10.12765/CPoS-2019-15en, Projection Pa- rameters (appendix II) DOI 10.12765/CPoS-2019-16en, and Population Projections for all EU countries (appendix III) DOI 10.12765/CPoS-2019-17en: URL: http://www.comparativepopulationstudies.de/index.php/CPoS/article/view/310/285 URL: http://www.comparativepopulationstudies.de/index.php/CPoS/article/view/310/286 URL: http://www.comparativepopulationstudies.de/index.php/CPoS/article/view/310/287 • Philip Cafaro, Patrícia Dérer172 1 Introduction In this paper we present new, complementary population projections for the 28 member countries of the European Union and for the EU as a whole. These projec- tions, building on the work of Eurostat and numerous individual demographers, differ from previous national and region-wide projections primarily in projecting a wider range of fertility and migration scenarios farther out into the future than has typically been done and in linking these scenarios explicitly to policies. As policy-based projections rather than forecasts, they aspire not to maximize predic- tive success regarding what will happen, but to accurately show the long-range demographic impacts of different policy choices, so as to clarify those choices for Europe’s leaders and citizens. Our central claim is that these new projections clarify the population impact of a wide range of current policy choices better than recent projections from the UN, Eurostat, and national statistical bureaus. In addition, we ask these research ques- tions: 1) How much can feasible family-support policies increase or decrease Euro- pean populations? 2) How much can alternative migration policies do so, particularly over the long term? 3) Do changes to family-support policies or immigration numbers have the big- ger potential to infl uence future population size within the European Union? 4) What regional differences exist among EU countries in terms of population futures and policy choices? 5) What policies might best mitigate excessive population decreases in extra low fertility or high emigration countries? 6) How can migration and fertility policies affect the extent of aging in the EU? These new projections can also show which demographic policies lead toward stable or slowly decreasing populations in the densely populated EU and which toward higher ones – important information for citizens concerned to create eco- logically sustainable societies. In what follows, we review recent efforts to project European populations (Section 2), describe our own scenario choices (Section 3) and projection methods (Section 4), present results and discuss answers to our research questions (Sections 5 and 6), and offer our main conclusions (Section 7). 2 Background Successful countries need to make well-informed policy decisions. Today the na- tions of the EU face policy choices that will infl uence their long-range demographic development, leading to a wide range of possible population futures. But we live in a rapidly changing world, including sudden fl uctuations in migration levels, with Policy-based Population Projections for the European Union • 173 unexpected demographic surprises, such as persistently low fertility rates in some developed nations. Aware of this demographic uncertainty, national statistical bu- reaus tend to be leery of projecting more than a few decades into the future, and their alternative scenarios, if given at all, are usually narrowly constrained. This lim- its the usefulness of their projections for long-range policy analysis, or for informing their citizens about the full range of policy options. For example, the report presenting the most recent national population projec- tions from Germany to a general audience shows only two fertility scenarios (1.4 TFR and 1.6 TFR) and two migration scenarios (annual net migration of 100,000 and 200,000) with higher fertility and migration scenarios relegated to model assump- tions beside the main variants (Federal Statistical Offi ce of Germany 2015). Such a small variation in the total fertility rate arguably does not capture the fertility chang- es possible through changes to family support programs or other economic policies (see discussion below). Similarly, the two immigration scenarios hardly account for the range of policy choices facing a country where annual net immigration has aver- aged 259,000 over the past twenty years and varied widely (from – 56,000 in 2008 to 1.2 million in 2015) and where there is widespread support both for greatly increas- ing immigration (Social Democrats, Die Grünen) and greatly decreasing it (Christian Democratic Union, Alternative für Deutschland). The report makes no attempt to explicitly link different fertility rates and immigration numbers to particular policies. In addition, the projections only go out to 2060, with longer-term demographic im- pacts left unexplored. It is easy to lose sight of the goal of policy clarifi cation while pursuing predic- tive rigor. Probabilistic population forecasts have begun to supplant traditional co- hort-component projection methods, based on superior accuracy and their ability to quantify uncertainty (Wilson/Rees 2005; Azose et al. 2016). Recently the Italian national statistical bureau took this approach, using stochastic models to estimate future fertility, mortality and immigration trends, which they present with 90 per- cent confi dence intervals (Istat 2018). Stochastic projection models may provide superior short-term forecasts, but an unremarked consequence of using them is that median scenarios with confi dence intervals tend to displace presentation of a variety of alternative scenarios – thus breaking any explicit connection between fu- ture numbers and current policy choices. Il Futuro Demografi co Del Paese provides readers with a single median projection (with 90 percent confi dence intervals) of Italy’s future population out to 2065, with no alternative scenarios (Istat 2018). But fertility, mortality and migration trends can develop in different ways due to policy choices, and it would have been helpful to clarify these. When we turn to region-wide projections, their value for policy clarifi cation also may be limited. Eurostat’s 2015 EU population projections run to 2080 and helpfully provide several alternative migration and fertility variants. Baseline scenarios are projected from recent trends by nowcast, convergence and trend models. TFRs for all countries are assumed to converge moderately to strongly around those of the highest fertility EU members, while net migration levels are assumed to converge relatively quickly around much lower levels (from ½ to ¼ of current levels by 2080) (Lanzieri 2017). Sensitivity tests are then provided for lower and higher migration • Philip Cafaro, Patrícia Dérer174 (+/- one third of the baseline net migration level) and for lower fertility (shrinkage of national fertility rates by 20 percent), but not for higher-than-baseline fertility. The projections provided in online tables (Eurostat 2019) are not clearly related to par- ticular policy choices, nor do they cover the range of plausible alternatives, nor are the results explained in accessible summary publications. The resulting projections provide readers with little sense of how increases or decreases in net migration or fertility could impact future population numbers. The most recent UN projections run to 2100 and use a probabilistic approach, resting on probabilistic fertility and mortality projections but deterministic migra- tion projections (Raftery et al. 2014; United Nations 2017a). They assume gener- ally constant levels of net migration until 2045-2050, which then gradually decline, reaching 50 percent of 2045-2050 levels by 2095-2100. Whatever the merits of this long-term forecast globally, it seems dubious for European countries, where migra- tion pressures are strong and are projected to grow stronger during this century (Lutz et al. 2019). Only one alternative migration scenario is presented, the zero net migration variant; this allows readers to estimate the impact of status quo migration levels to future population numbers through 2050, but not much more. Regarding fertility, the UN follows the tradition and adds projections for +/- 0.5 of base TFR. Combined with a “constant fertility” scenario, where fertility rates remain at 2010- 2015 levels, this provides some sense of how future fertility rates might infl uence future national population numbers. However, this information is not presented per- spicuously in summary publications (see United Nations 2017b). In any case, for low-fertility developed nations like those in the EU, a more fi ne-grained series of fertility and migration projections would be valuable – as well as some attempt to connect particular fertility and migration scenarios to particular policy choices. The public typically regard any population projection as a forecast designed to predict the future (Booth 2006). However, several recent efforts helpfully project alternative futures without emphasizing the superior likelihood of any of the alterna- tives projected. Such scenario-based projections emphasize that demography is not destiny and that policies can decisively infl uence demographic outcomes. The DEMIFER project (Demographic and Migratory Flows Affecting European Regions and Cities) is one effort to clearly link policy choices to demographic consequences using scenario designs (De Beer et al. 2010; Rees et al. 2012). Reasoning that fu- ture demographic changes would be infl uenced less by specifi c demographic policy choices than by overall economic trends (which in turn infl uence policies), research- ers developed four general scenarios. The complex scenarios depict more or less robust economies and more or less economic inequality within and between Euro- pean regions. Researchers then calculated expected changes to fertility and mortal- ity rates and immigration levels under each scenario. Their focus on NUTS2 regions provides a more fi ne-grained analysis than national projections and their holistic ap- proach arguably captures well potential changes in fertility rates, given the fact that societies’ overall support for family formation infl uences these rates more strongly than one-off policy initiatives (Frejka 2008). But the approach fails to model Euro- pean immigration policy choices, since future immigration levels, like current ones, will be set primarily by European governments rather than by economic trends. Policy-based Population Projections for the European Union • 175 The recently updated global projections from the International Institute for Ap- plied Systems Analysis provide the most ambitious approach we have seen linking policy choices to future population numbers. Building on the “shared socio-eco- nomic pathways” developed for climate change analysis, (Lutz et al. 2018) develop three scenarios representing slow, medium and rapid socioeconomic development, each with associated fertility, mortality and migration levels. They then split their medium scenario into three based on different migration levels: zero net migration; continuation of the average net migration level between 1960 and 2015; and double that average. All fi ve of the ensuing projections are run out to 2100. The result is the most explicit picture we have seen of the potential for different migration levels to infl uence future EU population numbers. Unfortunately, the potential impact of fam- ily support policies on future numbers is obscured, since their scenarios assume an implausibly large infl uence for increased education on future EU fertility levels, not taking into account already high European education levels and recent evidence for increased fertility among more educated women (Hazan/Zoabi 2015). They also mix changing fertility rates with changing migration and mortality levels in such a way as to render the impacts of family support policy choices opaque. 3 Scenarios In general contrast to the approaches discussed above, our EU projections disag- gregate the demographic impacts of fertility-focused policies and migration poli- cies, presenting separate projections with multiple policy scenarios for each. We set the parameters for fertility rates and annual immigration numbers with reference to recent rates and numbers in status quo scenarios. Then we build additional sce- narios by phasing in plausible changes to these rates and numbers, directly related to particular policy choices. Our scenarios are informed by reference to recent policy choices made or ad- vocated in the political sphere, and by analyses of the effectiveness of these poli- cies in the recent demographic literature. Given the wide range of policy proposals currently under discussion in the public sphere, we have opted to consider a wider rather than a narrower range of projections. 3.1 Fertility scenarios All European countries have completed the demographic transition to small families with long-lived members (Frejka 2008). While there is considerable variation (from a current TFR of 1.34 in Italy and Spain to 1.92 in France), no European country is at or above replacement fertility (Table 1). Although a few analysts see low fertility and smaller populations as part of the natural evolution of successful societies (Matanle 2017; Götmark et al. 2018), most policy-makers see them as problems to be fought (European Commission 2014). Low fertility rates have induced EU nations to make a variety of attempts to raise them, including bonuses and tax breaks for having more children, and more gener- • Philip Cafaro, Patrícia Dérer176 ous provision of child care services. Such efforts have been extensively studied. Despite mixed evidence in the literature (Gauthier et al. 2013), by now policy-makers have a fairly clear sense of what works best for permanently raising fertility levels in developed nations: comprehensive Nordic-style policies that make work-life bal- ance easier for women and couples, combined with strong economic safety nets and more egalitarian societies (Thévenon/Gauthier 2011; Balbo et al. 2013; Poll- mann-Schult 2018). We have also learned which policies do not work, or which only provide a short-term boost to fertility rates by infl uencing the timing of child-bear- ing, rather than completed fertility. These include isolated payments to encourage larger families within unchanged contexts of economic or career insecurity (Kalwij 2010; Kim 2014, although see Hong/Sullivan 2016 for some contrary evidence for the effectiveness of fertility subsidies.) Given this general understanding, we have developed the fi ve family support policy scenarios summarized in table 2 below. In the status quo policy scenario, we imagine nations continuing their existing economic and family support policies. Under this scenario, we hold TFR steady for higher fertility countries (see Table 1) and phase in small fertility increases for lower and medium fertility countries, in line with prevailing trends and other recent projections. Then in our two egalitar- ian policy scenarios, moderate and strong, we visualize countries enacting policies that make it easier for couples to form and sustain families, or creating more gener- ous economic safety nets and doing more to redistribute wealth, so as to sustain a relatively egalitarian economic structure, or both. In contrast, in our two neo-liberal policy scenarios, moderate and strong, we imagine countries retreating from their Tab. 1: Fertility rates in 2016 for EU countries according to Eurostat, divided into three groups Lower fertility countries Medium fertility countries Higher fertility countries TFR < 1.5 TFR 1.5 < TFR < 1.7 TFR > 1.7 Italy 1.34 Austria 1.53 Latvia 1.74 Spain 1.34 Hungary 1.53 Denmark 1.79 Portugal 1.36 Bulgaria 1.54 United Kingdom 1.79 Cyprus 1.37 Finland 1.57 Ireland 1.81 Malta 1.37 Slovenia 1.58 Sweden 1.85 Greece 1.38 Estonia 1.60 France 1.92 Poland 1.39 Germany 1.60 Luxembourg 1.41 European Union 1.60 Croatia 1.42 Czech Republic 1.63 Slovakia 1.48 Romania 1.64 The Netherlands 1.66 Belgium 1.68 Lithuania 1.69 Source: Eurostat database, “Fertility indicators” table Policy-based Population Projections for the European Union • 177 current levels of family support, cutting their economic safety nets, or encouraging greater economic inequality – or some combination of all three A key question regarding these scenarios is the extent to which policy changes in one direction or the other will infl uence national fertility rates. There is much disagreement among demographers about the effectiveness of policy in infl uenc- ing fertility (Luci-Greulich/Thévenon 2013). Yet that they have some infl uence seems undeniable. Hilgeman/Butts (2009: 113) found that improved access to childcare increased European fertility rates, predicting that if “Italy were to increase its enroll- ment from 6 percent to 64 percent, to match that in Denmark, the realized fertility Tab. 2: Projection assumptions made under different family support scenarios. The impact on fertility is always phased in by 2036 Scenario Policy changes Impact on fertility rates (1) status quo economic and family support policies Continue existing level of family support, existing economic safety net, existing levels of economic equality and equality between the sexes Lower fertility countries: + 0.2 TFR Medium fertility countries: + 0.1 TFR Higher fertility countries: no change in TFR (2016 value) (2) moderate egalitarian policy shift Improve economic safety net, decrease economic inequality within society, increase policies that support family formation, commit to equality between the sexes All three categories: status quo TFR + 0.15 (3) strong egalitarian policy shift Greatly improve economic safety net, greatly decrease economic inequality within society, greatly increase policies that support family formation, strongly commit to equality between the sexes All three categories: status quo TFR + 0.3 (4) moderate neo- liberal policy shift Reduce economic safety net, accept growing economic inequality within society, decrease support for family formation, ignore inequality between the sexes All three categories: status quo TFR - 0.15 (5) strong neo-liberal policy shift Greatly reduce economic safety net, accept greatly increased economic inequality within society, greatly decrease support for family formation, ignore inequality between the sexes All three categories: status quo TFR - 0.3 Source: own design • Philip Cafaro, Patrícia Dérer178 per woman would be predicted to increase by an average of 0.97 children.” (Adsera 2004) concluded that the gap between desired and achieved fertility was lower for Spanish women with more secure public sector jobs than for those with less secure private sector jobs. Although Gauthier et al. 2013 argue that this “gap” is often over- estimated and should not be used as evidence for the potential impact of family- friendly policies, it is often cited as evidence at least for their possible impact (Eu- ropean Commission 2005). One summary found small yet signifi cant increases in developed nations’ fertility rates under a wide range of policy interventions (OECD 2011). Acknowledging the uncertainty around this issue, we assume a relatively small but not insignifi cant space for family support and general economic policies to increase or decrease future fertility rates in the EU: plus or minus 0.15 TFR for our “moderate” policy change scenarios and plus or minus 0.30 TFR for our “strong” scenarios. This range is broadly in line both with studies that have looked at the im- pacts of particular policies in particular EU countries and with the idea that the range of fertility rates exhibited among EU countries is partly but not wholly a function of different national policy choices. See appendix I for additional discussion. 3.2 Migration scenarios Regarding migration, we develop a similarly wide range of policy-dependent sce- narios, recognizing the greater uncertainty around future migration numbers com- pared to future fertility rates (Azose et al. 2016). Migration numbers vary widely across the EU, as illustrated by table 3, showing individual countries’ average an- nual net migration over the past twenty years. In recent decades, many EU countries have evolved from relatively low to rela- tively high immigration regimes. More recently, high immigration levels have pro- duced a strong populist reaction and some attempts to reduce these numbers, but whether this leads to permanently reduced immigration in the future remains to be seen. On the one hand, many EU citizens would like to see immigration reduced for social, economic, or political reasons (Connor/Krogstad 2018). They perceive cur- rent immigration levels offering few benefi ts and considerable harms. On the other hand, many citizens (sometimes the same ones) feel a humanitarian obligation to help their poorer neighbors to the south and east by allowing more immigration. Many business leaders see increased immigration as the solution to the potential problem of shrinking numbers of workers and consumers (Legrain 2014; d’Albis et al. 2018), a view echoed by EU policy-makers (European Commission 2005, 2014). The situation is fl uid and the sheer range of immigration policies advocated by European political parties is impressive. We seek to capture this range in our projec- tions, with the following fi ve migration scenarios for those countries (20 out of 28) that have averaged net positive migration over the past twenty years: We build the status quo net migration scenario by averaging annual net migra- tion into a country for the past twenty years and projecting it out to 2100 (note that our status quo net migration scenario is identical to our status quo fertility policy scenario). Such a lookback period roughly mirrors contemporary policies and con- Policy-based Population Projections for the European Union • 179 ditions, while smoothing out yearly fl uctuations. We then build four further scenari- os around the status quo scenario: 2X and 4X status quo scenarios, which increase immigration, and a zero net migration scenario and half status quo scenario, which decrease it. See appendix I for further details. We treat the eight EU countries where emigration has exceeded immigration on average over the last two decades somewhat differently. For these countries, which are all in Eastern Europe, we have built a status quo net migration scenario around continued negative net migration, again taking the average annual net migration lev- el for the past twenty years as a starting reference but linearly increasing it to zero in 2100 (in this way, gradually increasing net migration compensates for a decreasing total population that could not support large constant negative net migration levels). We then project a zero net migration scenario, reaching equal numbers of emigrants and immigrants by 2026, and three positive net migration scenarios built around applying a multiple of the average annual net migration rate for the EU as a whole during the past twenty years to the countries in question, out to 2100. These higher net migration scenarios thus assume convergence in migration patterns across the EU, as Eastern countries catch up to Western ones by being more open to immigra- tion from outside the EU while fewer of their own citizens emigrate. As with our fertility scenarios, we believe our migration scenarios well capture the range of policy choices facing EU countries today. These two sets of fi ve alterna- tives present the full gamut of policy choices on immigration: from drastically cur- tailing it to drastically expanding it; from the Sweden Democrats to Sweden’s Green Party. Several readers of this study in manuscript have questioned the feasibility Tab. 3: Average annual net migration (1998-2017) of EU countries Source: calculated from Eurostat database,”Population change – Demographic balance and crude rates at national level” table Countries with positive average annual net migration Countries with negative average annual net migration European Union 1,188,235 Spain 270,112 Ireland 21,645 Slovakia -305 Germany 259,316 Czech Republic 18,747 Estonia -942 United Kingdom 230,107 Denmark 16,778 Croatia -4,259 Italy 229,093 Hungary 13,652 Poland -12,552 France 100,525 Portugal 12,262 Latvia -14,362 Sweden 50,024 Finland 11,104 Bulgaria -21,052 Belgium 42,575 Greece 8,390 Lithuania -27,212 Austria 40,547 Luxembourg 6,967 Romania -103,807 The Netherlands 26,427 Cyprus 5,855 Slovenia 4,128 Malta 3,642 • Philip Cafaro, Patrícia Dérer180 of the zero net migration and 4X net migration scenarios for positive net migration countries. We agree that these more extreme scenarios are less likely to occur than the more modest policy changes characterized within the range of our half status quo to 2X status quo scenarios; that is why we graph them with dotted lines rather than straight lines in our projections. Yet given the overriding goal of these projec- tions – to clarify the potential demographic impact of current policy choices – we believe these policy options should be included in our analysis. A long-term zero net migration scenario for high-immigration countries in west- ern and northern Europe is unlikely. Yet “no international migration” scenarios are included in the population projections of many national statistical bureaus and aca- demic demographers, serving as reference scenarios to decompose the impacts of internal population changes and international migration. Similarly, migration levels equal to 4X these countries’ recent net migration are highly unlikely to be sustained for many decades. The 4X net migration scenario thus is best seen as a “what if” scenario, allowing us to ask: “what might the demographic results be if a country persisted in allowing very high levels of immigration?” Several EU governments allowed unprecedented immigration for a time in 2015 and 2016; in the case of Germany, this resulted in even higher net migration in 2015 than would occur under our 4X status quo migration scenario. Exploring the pos- sible demographic impacts of persisting in such a course is valuable, even if such Tab. 4: Projection assumptions under different migration scenarios for countries with net positive migration Scenario Annual net migration level Fertility rate assumptions (1) status quo net migration Continuation of the country’s average annual net migration level for the past 20 years for the rest of the century TFR the same as under status quo fertility scenarios (2) 2X status quo net migration 2X average annual net migration level by 2026, held stable for the rest of the century Status quo TFR + 0.05 by 2036 (3) 4X status quo net migration 4X average annual net migration level by 2026, held stable for the rest of the century Status quo TFR + 0.2 by 2036 (4) half status quo net migration ½ average annual net migration level by 2026, held stable for the rest of the century Status quo TFR - 0.025 by 2036 (5) zero net migration Zero net migration by 2026, held stable for the rest of the century Status quo TFR - 0.05 by 2036 Source: own design Policy-based Population Projections for the European Union • 181 persistence is unlikely. Similarly, for other countries, the 4X status quo net migration scenario allows us to ask: “what might the demographic results be if a country al- lowed open borders?” Europe’s Green parties explicitly advocated an open borders immigration policy in the 2019 European parliamentary elections (European Greens 2019a/b), and many EU social democratic parties have joined them in resisting ef- forts to set numerical limits to refugee admissions into the EU while simultaneously pushing for a much more expansive defi nition of refugees (Žižek 2016). Truly open borders are hard to model, but the 4X status quo scenario seems to us a good proxy for such an unlikely but possible course of action, showing its possible demograph- ic impact. See appendix I for further discussion. Finally, we note that family support and migration policies can and will change si- multaneously. For that reason, we also project several combination scenarios in an attempt to better understand the full range of potential demographic futures facing EU countries and the impacts that policy choices may have on them. We describe these combination scenarios in section 5 of appendix I. We present them, along with all our scenario projections, in appendix III. Tab. 5: Projection assumptions under different migration scenarios for countries with net negative migration Scenario Annual net migration level Fertility rate assumptions (1) status quo net migration Linear increase from average annual net migration for the past 20 years to zero by 2100 Status quo TFR (2) zero net migration Zero net migration by 2026, held stable for the rest of the century Status quo TFR (3) EU status quo net migration rate Net migration level equal to the recent average EU net migration rate by 2026, held constant for the rest of the century Status quo TFR + 0.05 by 2036 (4) 2X EU status quo net migration rate Net migration level equal to 2X the recent average EU net migration rate by 2026, held constant for the rest of the century Status quo TFR + 0.1 by 2036 (5) 4X EU status quo net migration rate Net migration level equal to 4X the recent average EU net migration rate by 2026, held constant for the rest of the century Status quo TFR + 0.2 by 2036 Source: own design • Philip Cafaro, Patrícia Dérer182 4 Methods Our population projections employ the deterministic cohort-component method. We used DemProj program version 5.71, part of the Spectrum policy development and planning tool developed by Avenir Health (Stover/Kirmeyer 2007; Stover et al. 2010). The program requires information on population numbers by age and sex in the base year (2016 in all scenarios), as well as current year data and future assump- tions about total fertility rate (TFR), age distribution of fertility, life expectancy at birth by sex, the most appropriate model life table, and the magnitude and pattern of international migration. In order to calculate dependency ratios, median age and proportion of 65+ population throughout the projections, we also used the DemProj module of Spectrum. Our source for baseline 2016 data on population numbers by sex and single age (0-80+), national TFRs, life expectancy at birth and net migration fi gures, is the European Statistical Offi ce (Eurostat). We use 2016 as our base year because at the time we made the projections, that was the most recent year for which actual TFR fi gures were available. Age specifi c fertility rates follow the United Nations’ most recent world population projections’ country specifi c fertility distribution data (United Nations 2017a). For changes to life expectancy, we use the sex specifi c as- sumptions of the baseline scenario of the Eurostat 2015 population projections until 2080, then hold these constant through the end of the century. In order to work with age-specifi c mortality rates, we use the UN’s country specifi c model life tables. To model the impact of changes in migration levels, we constructed status quo scenarios by averaging Eurostat’s annual national net migration fi gures for 1998 to 2017, then built alternative scenarios on multiples of that base. As we were not forecasting but building what-if policy scenarios, a model using readily available net migration level data was satisfactory for our purposes, although we are aware that multiregional models such as MULTIPOLES (Kupiszewski/Kupiszewska 2011) based on immigration levels and emigration rates are more accurate and methodologically elegant (Rogers 1990). Similarly, we did not attempt to model regional, sub-national migration, nor to classify different migration fl ows by motive, such as employment versus asylum seeking, nor did we distinguish between immigrants coming from EU countries or outside the European Union. Such approaches may provide a more fi ne-grained picture. But our focus in this study was squarely on national net migra- tion levels, since a primary goal was to model the future demographic impacts of migration policy, which is made primarily at the national level. Concerning “net mi- grants,” we assume equal proportions between the sexes and the DemProj program uses the UN’s country-specifi c migration age distributions. All changes to TFR in our scenarios assume a linear change phased in between 2016 and 2036, while all changes to net migration rates are phased in between 2016 and 2026. Appendix II provides tables showing the projection parameters (life ex- pectancy, fertility rates and net migration levels) set for all scenarios across all coun- tries in the study. Appendix III provides detailed population projections for all EU member states and the EU as a whole, in chart and table form. Policy-based Population Projections for the European Union • 183 5 Results To begin summarizing our results, consider our projections for the European Union as a whole. The current 28 countries in the EU had a combined population in 1950 of 379.8 million and 510.3 million in 2016. The region’s current (2016) total fertility rate is 1.60 and its annual net migration level over the past 20 years (1998-2017) averaged 1,188,235. Under a continuation of status quo family support and economic policies and an extension of current immigration levels in every country, we can expect a 3.6 per- cent decrease in the EU population (Fig. 1). This would hardly be overwhelming, spread out over eight decades. A relatively modest turn toward less generous sup- port for family formation, under the moderate neoliberal scenario, would acceler- ate population decrease, leading instead to a 14.2 percent population decline. On the other hand, more generous support for workers and their families, under the moderate egalitarian scenario, would increase fertility rates suffi ciently to avoid any population decline, even adding 27 million people to the total EU population. Given wide disparities among EU countries in economic support systems, commitments to egalitarian relations between the sexes and fi nancial support for having children, there seems suffi cient scope for enacting such policies and thus avoiding a decline if that is desired. The full range of fertility-related policy scenarios (dashed lines on fi gure 1), from strong neoliberal to strong egalitarian policies, leads to populations ranging in 2100 from 396.8 million (a 22 percent decline) to 592.3 million (a 16 per- cent increase): almost a 200 million spread. Fig. 1: European Union population projections under fi ve fertility scenarios. All scenarios assume the continuation of the past 20 years average annual net migration on the EU level Source: own calculations 300 400 500 600 700 800 900 1,000 1950 1975 2000 2025 2050 2075 2100 Strong Egalitarian Moderate egalitarian Status quo economic and social policies Moderate neoliberal Strong neoliberal 379.8 million 510.3 million 592.3 million 537.2 million 491.9 million 439.8 million 396.8 million Population in millions • Philip Cafaro, Patrícia Dérer184 Altering immigration levels provides much greater scope for increasing or de- creasing the total EU population, compared to changing fertility-related policies on the country level – an important result that also holds for most individual countries (Fig. 2). Cutting net average migration in half would reduce the EU population by an additional 70 million people, or an extra 14 percent compared to the population loss under the status quo scenario, for a total drop of 24 percent by 2100. Doubling net migration, conversely, would switch the EU’s population from declining by 52.6 mil- lion (10 percent) under the status quo to growing by 92.0 million (18 percent). That’s a swing of 214.6 million people across our most likely range of immigration policy changes. The spread across the full range of policy choices is much greater: over 600 million people, from swelling to 933.3 million in 2100 in the case of quadrupling status quo net migration numbers (an 83 percent increase), to contracting to only 318.9 million by reducing net migration to zero (a 38 percent decline). Our combination scenarios add further detail to the picture (see appendices I and III for details). Doubling recent average net migration levels across the EU while improving family support through egalitarian social and economic policies leads to a 38 percent population increase by the end of the century. In contrast, cutting im- migration in half and cutting back on social and economic support for family forma- tion leads to a population decline of over two hundred million people (a 40 percent decline). Interestingly, halving recent net migration levels combined with strong improvements in family support leads to a relatively stable population (decreasing Fig. 2: European Union population projections under fi ve fertility scenarios. Status quo migration is the continuation of the past 20 years average annual net migration level (1,188,235). Migration scenarios use total fertility rates varying between 1.65 and 1.90, with higher immigration levels projected to drive higher TFRs Source: own calculations 300 400 500 600 700 800 900 1,000 1950 1975 2000 2025 2050 2075 2100 4X status quo migration 2X status quo Status quo net migration (1,188,235) Half status quo zero net migration 379.8 million 510.3 million 933.3 million 602.3 million 457.7 million 387.7 million 318.9 million Population in millions Policy-based Population Projections for the European Union • 185 only 3.6 percent by the end of the century). This might be a particularly appealing policy combination on a continent where support for mass immigration has waned while concerns about low fertility levels remain strong. 5.1 Regional differences Despite their shared experience passing through “the second demographic transi- tion,” EU nations all face somewhat different population futures and policy choices. In particular, it is worth highlighting the substantial differences among four regions of the EU. Central and Eastern Europe Under status quo policies, many low fertility and/or net emigration countries in this region are on track for steep population declines this century.1 These include Ro- mania, facing a 74 percent decline by 2100, Bulgaria (60 percent), the Baltic states (between 38 percent and 86 percent) and Hungary, facing a decline of 29 percent by 2100. However, egalitarian economic and social policies could mitigate population declines among this group by increasing fertility rates and (potentially) decreas- ing emigration numbers. In tandem with modestly increased immigration, much of the projected depopulation could be avoided. For example, under a combination scenario of zero net migration fused with strong family support policies, Romania’s population decline shrinks to 22 percent by 2100, while under the same scenario over the same time-frame, Bulgaria’s population only decreases 28 percent. Add positive net immigration at the current EU level and these populations would in- stead increase, Romania by 17 percent and Bulgaria by 6 percent. Hungary, which has had small positive net migration over the past twenty years, could limit its population decline to 11 percent by holding net migration steady and embracing strong family support policies. Instead, its current government has pro- posed the sort of pro-natalist policies that have proven ineffective in the past, such as subsidies to encourage families to have three or four children. More effective, both narrowly in terms of increasing fertility and broadly in terms of social well- being, would be policies to help young couples have the one or two children many of them want, without having to sacrifi ce their careers or go into poverty. Southern Europe Low fertility/high net migration countries in this region are propping up population numbers through high immigration levels under status quo policies. Without heavy immigration, Spain’s and Italy’s populations would already be declining, since their TFRs have been below replacement rate since the early 1980s and are now well be- low it. Under zero net migration scenarios, Spain’s population decreases by 46 per- 1 Sobotka (2003) argues that the countries of Central Europe and Eastern Europe are following somewhat different paths, the former moving toward western European norms of later child- rearing and the latter shifting toward larger proportions of one-child families. • Philip Cafaro, Patrícia Dérer186 cent and Italy’s by over 50 percent by 2100. These countries cannot cut back sig- nifi cantly on immigration without either embracing egalitarian economic and family support policies to increase fertility rates, or accepting signifi cant population de- clines this century. Greece and Portugal, averaging more modest positive net migration over the past twenty years, instead face steep population declines in coming decades un- der status quo scenarios and all projected fertility scenarios. Their choices going forward are somewhat similar to Italy’s and Spain’s, however: accept higher net mi- gration numbers, boost support for family-formation and create stronger economic safety nets for common citizens in order to maintain their populations, or accept rapid population decline. Interestingly, for Italy and Spain, the status quo scenario and a scenario combining half status quo migration and a strong egalitarian shift in family support policies lead to relatively similar results, coalescing around a fairly stable population, while for Greece and Portugal this combination scenario leads to signifi cantly less population loss. A combination of increased support for fam- ily formation and decreased immigration might be appealing to many residents of these countries, especially given relatively high unemployment rates throughout the region. Northern European countries With high fertility and high immigration levels these countries will see signifcant population growth this century under status quo policies, and some growth un- der most plausible policy scenarios. An example is Sweden, which is on track to increase its population 52 percent by 2100 under status quo fertility and immigra- tion policies. Increases in immigration levels could lead to much larger populations throughout northern Europe, as seen in the table below: With wealthy societies, low population densities (by European standards) and a post-war tradition in some countries of taking in large numbers of refugees, sig- nifi cant increases in immigration levels are a real possibility – although there is also considerable opposition to such increases, as demonstrated in recent elections. Tab. 6: Percentage change from current population by 2100, under different immigration scenarios in Northern Europe. Scenarios leading to > 50 percent population increase are shaded Zero net ½ status Status quo 2X status 4X status migration quo migration quo quo Denmark - 19 - 2 16 54 142 Finland - 35 - 22 - 11 15 71 Sweden - 9 23 52 124 276 Source: own calculations Policy-based Population Projections for the European Union • 187 Western European countries With medium to high fertility and high immigration levels, these countries are most- ly on track for relatively stable to moderately increasing populations under status quo policy scenarios (Table 7). While increases in immigration levels could lead to much larger populations, such as increases of 186 percent in Belgium and 167 per- cent in the UK under our highest immigration scenario, decreases could lead to signifi cantly smaller ones, such as a 38 percent decline in Germany and a 24 percent decrease in the Netherlands at zero net migration. As in the rest of Europe, policy- induced changes in fertility rates could also change future population numbers sub- stantially, albeit to a lesser extent. 5.2 Impact on population aging Fertility and migration policies will have only a moderate infl uence on the pace of future population aging, with migration policy changes having a stronger potential impact. For the EU-28 population, the proportion of the population 65 years and older and the dependency ratio will be higher in 2100 than they are today under all scenarios (Table 8), so some aging is inevitable. Of all the scenarios, increasing long term average net migration 4 times would do the most to slow the pace of aging, with the dependency ratio increasing 10 percent instead of 17 percent as under the status quo scenario, while increasing the proportion of the population 65 years + by 2.2 percent instead of 7.2 percent. But in exchange for these modest improvements, these immigration policy changes would add an extra 475 million people to the total EU population (933 million instead of 458 million under the status quo scenario). Pursuing the strong egalitarian family policy scenario would improve the de- pendency ratio roughly as much as under the 2X status quo migration scenario. In the other direction, the strong neoliberal family policy scenario would increase the Tab. 7: Percentage increase from current population by 2100, under different immigration scenarios in Western Europe. Scenarios leading to > 50 percent population increase are shaded Zero net ½ status Status quo 2X status 4X status migration quo net migration quo quo Austria -36 -9 16 72 195 Belgium -23 >-1 26 74 186 Czech Republic -37 -25 -14 8 61 France -9 2 13 35 88 Germany -38 -19 -2 37 123 Ireland -6 23 51 111 245 Luxembourg -33 50 117 267 600 The Netherlands -24 -12 -2 25 82 United Kingdom -18 3 24 68 167 Source: own calculations • Philip Cafaro, Patrícia Dérer188 T a b . 8: T h e im p a c t o f fi v e m ig ra ti o n a n d fi v e f e rt ili ty p o lic y s ce n a ri o s o n t h re e m e a su re s o f so ci e ta l a g in g in t h e E U a s a w h o le M ig ra ti o n s ce n a ri o s F e rt ili ty s ce n a ri o s Y e a r S ta tu s q u o Z e ro n e t H a lf s ta tu s 2 X s ta tu s 4 X s ta tu s S tr o n g M o d e ra te M o d e ra te S tr o n g S ce n a ri o m ig ra ti o n q u o q u o q u o n e o lib e ra l n e o lib e ra l e g a lit a ri a n e g a lit a ri a n W o rk in g a g e p o p u la ti o n ( % ) 2 01 6 6 5 .3 6 5 .3 6 5 .3 6 5 .3 6 5 .3 6 5 .3 6 5 .3 6 5 .3 6 5 .3 2 0 5 0 5 7. 8 5 6 .4 5 7. 1 5 8 .9 6 0 .4 5 8 .7 5 8 .2 5 7. 4 5 7. 0 21 0 0 5 8 .8 5 6 .5 5 7. 9 6 0 .1 61 .3 5 7. 3 5 8 .2 5 9 .3 5 9 .6 D e p e n d e n cy r a ti o ( % ) 2 01 6 5 3 .2 5 3 .2 5 3 .2 5 3 .2 5 3 .2 5 3 .2 5 3 .2 5 3 .2 5 3 .2 2 0 5 0 7 3 .0 7 7. 3 7 5 .0 6 9 .7 6 5 .5 7 0 .5 71 .7 74 .3 7 5 .5 21 0 0 7 0 .0 7 6 .9 7 2 .8 6 6 .5 6 3 .2 74 .5 71 .9 6 8 .6 6 7. 7 P o p u la ti o n 6 5 + ( % ) 2 01 6 19 .2 19 .2 19 .2 19 .2 19 .2 19 .2 19 .2 19 .2 19 .2 2 0 5 0 2 7. 7 3 0 .0 2 8 .8 2 5 .7 2 2 .5 2 8 .9 2 8 .3 2 7. 1 2 6 .5 21 0 0 2 6 .4 2 9 .9 2 7. 9 2 4 .3 21 .4 3 0 .8 2 8 .5 2 4 .4 2 2 .7 S o u rc e : o w n c al cu la ti o n s Policy-based Population Projections for the European Union • 189 percentage of the population 65+ by an additional 4.4 percent in 2100, comparable to the zero net migration scenario, while it would increase the dependency ratio by an additional 4.5 percent, comparable to cutting immigration to one quarter of cur- rent levels. But once again, the impacts on total population numbers would dwarf the impacts on aging. 5.3 Comparison with other studies Our fertility and migration scenarios give projection results that are roughly compa- rable to the results from other recent European-wide studies, while linking those re- sults more clearly to available policy choices. Compare in this regard our migration scenario projections to recent projections from Eurostat (2015), the United Nations (2017b) and the Wittgenstein Centre for Demography and Global Human Capital (2018) in fi gure 3 below. Results from our zero net migration, status quo and 2X status quo scenarios are broadly comparable to the no migration, median or “most likely,” and high migration projections from these other studies. The clear outlier is our 4X status quo scenario which graphs a much higher sustained high immigration level than any major study to date. The justifi cation for providing this scenario has already been provided. Note that the supply is already in place, with hundreds of millions of young people in the developing world living in rapidly-growing societies that struggle to provide them with the jobs and material well-being they desire. Whether EU countries (or the EU as a whole, under a strengthened centralized government) will enact policies that increase immigration this much is doubtful, yet possible. 6 Discussion What conclusions can we draw from these projections regarding our main research questions? We would direct attention to fi ve key points. (1) Migration policy offers greater scope for infl uencing future population num- bers than changes to family support policies, or changes to other fertility- related economic policies in the EU. Changing fertility does infl uence population size. But in a context of relatively low-fertility regimes where policy changes’ impacts on fertility are often fairly small, changes to national immigration levels could have a much greater impact going forward. For Germany, Europe’s most populous country, the more likely scenarios for fer- tility variation range between 1.55 TFR and 1.85 TFR, leading to a potential differ- ence of 15.4 million people in 2100 (72.9 million versus 88.3 million) (fi gure 4). In contrast, the three more likely net migration scenarios generate a 3X greater spread, ranging across 46.6 million people (Fig. 5). Considering the full range of fertility change scenarios, including less likely but still possible low and high scenarios of 1.4 TFR to 2.0 TFR, doubles the potential range across Germany’s 2100 population to 31.1 million. But considering the full range of possible migration scenarios in- • Philip Cafaro, Patrícia Dérer190 creases the 2100 population spread by over 100 million more. While the full range of fertility scenarios all lead to populations in 2100 that are broadly commensurate with Germany’s present population of 82.2 million (80 percent to 118 percent of cur- rent numbers), the extreme migration scenarios would lead to much more drastic changes: from 62 percent of the current population under the lowest scenario to 223 percent under the highest. Similar results, with much higher ranges across mi- gration scenarios than fertility scenarios, obtain for most EU countries and for the EU as a whole, as we saw earlier. (2) Egalitarian economic and family support policies and increased net migration have signifi cant potential to mitigate excessive population decreases in the EU’s lower fertility countries. Fig. 3: Comparing the policy-based migration scenarios for 28 EU member states in the present study with the migration scenarios in other published region-wide or world-wide studies Sources: a – own calculation; b – Eurostat Database 2015; c – United Nations Popula- tion Prospects 2017; d – Wittgenstein Centre for Demography and Global Human Capital 2018. 0 100 200 300 400 500 600 700 800 900 1,000 1950 1975 2000 2025 2050 2075 2100 4X status quo migration in this studya SSP2 double migrationd 2X status quo migration in this studya Eurostat higher migration sensitivity testb Eurostat Baseline projectionsb CEPAM Medium (SSP2)d UN medium varian (EU 28)c Status quo scenario in this studya Half status quo migrationa SSP2 zero migrationd Eurostat no migration sensitivity test UN zero migration variant (EU28)c Zero net migration in this studya Millions Policy-based Population Projections for the European Union • 191 Several eastern European countries with low fertility levels and high emigration rates are on track for rapid depopulation this century. While the dangers of small population decreases are often overblown, too large or too quick decreases could Fig. 4: Germany population projections under fi ve fertility scenarios: All scenarios assume the continuation of the past 20 years average annual net migration level (259,316) Source: own calculations Fig. 5: Germany population projections under fi ve migration scenarios: Status quo migration is the continuation of the past 20 years average annual net migration level (259,316). Migration scenarios use total fertility rates varying between 1.65 and 1.90, with higher immigration levels projected to drive higher TFRs Source: own calculations 40 80 120 160 200 1950 1975 2000 2025 2050 2075 2100 Strong egalitarian (2.0 TFR) Moderate egalitarian (1.85 TFR) Status quo economic and social policies (1.7 TFR) Moderate neoliberal (1.55 TFR) Strong neoliberal (1.4 TFR) Population in millions 69.3 million 82.2 million 97.1 million 88.3 million 80.3 million 72.9 million 66.0 million 40 80 120 160 200 1950 1975 2000 2025 2050 2075 2100 4X status qou 2X status quo Status quo net migration (259,316) Half status quo Zero net migration Population in millions 69.3 million 82.2 million 183.6 million 112.8 million 80.3 million 66.2 million 51.2 million • Philip Cafaro, Patrícia Dérer192 cause social problems in these societies (Rees et al. 2012; Jakovljevic/Laaser 2015). So it is good to know that part of this potential decrease could be avoided by adopt- ing the economic and family support policies that have proven so successful at boosting fertility rates and improving lives in northern Europe. Consider Poland, the most populous EU country facing signifi cant population decline (see projections in appendix III). Under status quo policies that provide little support for women who want to combine careers with creating a family (Mishtal 2009), Poland’s population is set to decline 44 percent by the end of the century. Combining increased net migration up to the current EU status quo net migration rate with a strong increase in sup- port for family formation would instead increase Poland’s total population 4 percent by 2100. Similar refl ections apply to Romania, Bulgaria, Italy, Portugal, Greece and other countries in eastern and southern Europe that could see signifi cant population losses over the next 80 years. It is particularly noteworthy that some of the same economic policies that could raise national fertility rates would also increase eco- nomic security, possibly decreasing emigration by younger workers. While these policies, such as improved child-care programs and paid family leave for workers, are expensive, they hold out hope for a slow, manageable population decrease rath- er than a precipitous decline in these countries (Frejka/Gietel-Basten 2016). (3) In most cases, EU nations are well placed to stabilize or slowly reduce their populations–thus achieving one of the necessary conditions for creating eco- logically sustainable societies. We can see this in many of the status quo scenarios, which tend to stay relatively fl at throughout the 21st century: often within 10 percent of current population fi g- ures (Holland, Italy, Germany) and if not, then within 20 percent of current numbers (France, Spain, Denmark, Czech Republic). In replenishing their populations, these countries have room to shift the balance somewhat between immigration and births to citizens, either in one direction or the other, while still keeping their populations stable or moderately declining. This can be seen by comparing the status quo fertil- ity and migration scenarios with the ½ net migration/strong family support policy combination scenarios in Table 9. A few countries, like Belgium and the United Kingdom, are set for larger popula- tion increases this century under a continuation of status quo fertility and migra- tion policies. Such increases could be mitigated by decreasing immigration levels, which would likely be more popular than decreasing fertility rates by cutting eco- nomic safety nets or support for family formation. On the other hand, higher net migration levels could lead to much high population growth across the continent. (4) An egalitarian shift in family support policies or increasing immigration lev- els are not effective remedies for population aging in the EU. In accordance with other recent studies (Craveiro et al. 2019; Lutz et al. 2019) and contrary to many political discussions, encouraging more immigration from external countries into the EU or increasing fertility rates by means of successful government initiatives will not signifi cantly slow population aging in the EU. Even increasing net migration levels 4 times would do little in the long-term to reduce the dependency ratio or the proportion of the population 65 years or older (see Policy-based Population Projections for the European Union • 193 Table 8). This is due to the fact that immigrants, like non-immigrants, age and retire, requiring social benefi ts and medical care. The fertility scenarios discussed in the present paper show an even smaller impact on aging measures than our migration scenarios. Meanwhile fertility and especially immigration increases show a much greater capacity for increasing total population numbers. In exchange for a small improvement in dependency ratios in 2100 (66.5 instead of 70.0 under the status quo), doubling net immigration into the EU over that period would add 145 million more residents to the total population (602 million rather than 458 million). Increasing immigration 4X would improve the dependency ratio slightly more, to 63.2. But that would still be a full ten points higher than today’s 53.2 and in exchange the EU would have to support 475 million more people than it would under a status quo migration scenario – with all the increased environmen- tal demands (Marques et al. 2019) and social contention (Harmon 2018) that such increased numbers imply. The moral is that to the extent aging societies create economic problems, the most effective remedies will involve increasing labor-force participation and other systemic changes, rather than increasing either net migra- tion or fertility rates (Götmark et al. 2018). Tab. 9: Percentage change from current population by 2100, under four policy scenarios, for the EU’s ten most populous countries. Scenarios resulting in relatively stable populations (+/- 15 percent) highlighted Status quo ½ status quo 2X status quo 2X status quo migration & migration & migration & migration & status quo strong egalitarian status quo strong egalitarian family support policy shift family support policy shift Germany -2 0 37 58 France 13 28 35 59 UK 24 29 68 90 Italy -8 -13 34 54 Spain 19 5 83 107 Poland -26a --- 12b 36c Romania -5a --- 31b 59c The Netherlands -2 11 25 46 Belgium 26 23 74 105 Greece -36 -25 -25 -11 EU total -10 -4 18 38 a Shows the EU status quo migration rate & status quo family support scenario b Shows the 2X EU status quo migration rate & status quo family support scenario c Shows the 2X EU status quo migration rate & strong egalitarian policy shift scenario Source: own calculations • Philip Cafaro, Patrícia Dérer194 (5) Demographic policies have the potential to signifi cantly raise or lower future populations, and hence to make it harder or easier for EU nations to create ecologically sustainable societies. In a world where the population growth rate is slowing down and most future population growth is projected for the developing world, it is easy to overlook the importance of demographic policies to developed countries’ futures. Arguably this is misguided. Many western and northern European nations are pursuing popula- tion policies that could signifi cantly increase their populations in the coming years and hence increase their countries’ total ecological footprint, which is a function of the size of their population and their per capita consumption of resources (Lin et al. 2019). Ceteris paribus, more people mean more consumption and pollution, more greenhouse gases emitted, etc. (IPCC 2014; Díaz et al. 2019). There is abundant evidence that humanity’s current biophysical demands are already far in excess of what is ecologically sustainable (Ripple et al. 2017) and this problem must be tack- led both by decreasing overconsumption and ending population growth. Other EU countries, whose fertility levels could lead to shrinking populations and hence help lessen their environmental footprints, are instead bolstering their populations through historically high immigration levels. Germany, Italy and Spain are important examples. Spain, with a current TFR of 1.34, is on track to increase its population by 8.5 million by 2100. With the same low current TFR, Italy’s population will decrease by only 3.5 million people this century under the status quo. Accept- ing rapid population decline is probably not to be expected of any nation’s lead- ers. But accepting some population decrease in wealthy, crowded societies may be advisable, given the environmental benefi ts of smaller populations (Matanle 2017; Pilling 2018). 7 Conclusion In developing new policy-based population projections, we have shown that some policy choices have the potential to substantially increase future European pop- ulation numbers. Both family support and migration policies will infl uence these numbers, with migration policies likely having the greatest impact. The potential increases under more expansive policy scenarios would not be as large, either as a percentage change or in absolute numbers, as those that occurred in India and China during the past century, or those being predicted for Africa during this one. This fact should not mislead us. Increasing Europe’s population by even 100 million more people would be a very big “ask” on nature; decreasing it by 100 million peo- ple would be a substantial gift. As a matter of de facto environmental policy, demo- graphic decisions will be at least as important as any other policy courses charted by EU nations this century, since they impact all aspects of a country’s environmen- tal situation. In a recent working paper, we use these projections to explore popula- tion policies’ potential impacts on future efforts to reduce the EU’s greenhouse gas emissions and preserve its biodiversity (Cafaro/Dérer 2018). Policy-based Population Projections for the European Union • 195 Precisely how demographic policies impact such efforts demands further study (Rust/Kehoe 2017). So do population policies’ impacts on economic equality and social solidarity. By tying our population projections more explicitly to actual policy options than previous efforts, we hope to facilitate other researchers’ efforts to ask and answer such questions. By providing a full range of population projections for all the countries of the EU, we believe we have made it easier for concerned citizens from these countries to do so, as well. According to the UN Human Development Index, European nations are among the world leaders in terms of health and longevity; in wealth and sharing that wealth equitably; in securing honest governments that serve their citizens and uphold the rule of law; and in promoting tolerance and ensuring human rights (United Nations Development Programme 2018). But the EU’s demographic challenges and opportu- nities are complex. Which population policies will best sustain fl ourishing societies remains to be determined. Getting this right is important in Europe – and beyond. A world caught up in an economic system built around the uncritical pursuit of growth and bumping up against ecological limits (O’Neill et al. 2018) needs to see that the socio-economic challenges of aging and shrinking societies are inevitable but man- ageable (Götmark et al. 2018) and that international solidarity can be combined with demographic responsibility (Cafaro 2015). As we have shown, most EU countries are well placed to stabilize or slowly re- duce their populations. Europeans’ great contribution in the twenty-fi rst century could be to model successful societies that do not depend on continued growth, but that instead prioritize societal well-being and an acceptance of limits (Kallis 2018). We hope these population projections are a useful tool for exploring such questions and ultimately for imagining and creating such societies. Acknowledgements The authors thank the Department of Philosophy, Colorado State University, the De- partment of Biological and Environmental Sciences, University of Gothenburg, and the Global Challenges Foundation, Stockholm, for their generous support. References Adserà, Alicia 2004: Changing Fertility Rates in Developed Countries. The Impact of La- bor Market Institutions. In: Journal of Population Economics 17,1: 17-43 [doi: 10.1007/ s00148-003-0166-x]. Azose, Jonathan J.; Ševčíková, Hana; Raftery, Adrian E. 2016: Probabilistic Population Projections with Migration Uncertainty. In: Proceedings of the National Academy of Sciences 113,23: 6460-6465 [doi: 10.1073/pnas.1606119113]. Balbo, Nicoletta; Billari, Francesco C.; Mills, Melinda 2013: Fertility in Advanced Societ- ies: A Review of Research. In: European Journal of Population 29,1: 1-38 [doi: 10.1007/ s10680-012-9277-y]. Booth, Heather 2006: Demographic Forecasting: 1980 to 2005 in Review. In: Interna- tional Journal of Forecasting 22,3: 547-581 [doi: 10.1016/j.ijforecast.2006.04.001]. • Philip Cafaro, Patrícia Dérer196 Cafaro, Philip 2015: How Many Is Too Many? The Progressive Argument for Reducing Immigration into the United States. Chicago: University of Chicago Press. Cafaro, Philip; Dérer, Patricia 2018: New Policy-Based Population Projections for the European Union. The Overpopulation Project. Göteborg, Sweden [https://overpopula- tion-project.com/wp-content/uploads/2019/01/new-policy-based-population-projec- tions-for-the-european-union.pdf, 27.08.2019]. Connor, Phillip; Krogstad, Jens Manuel 2018: Many Worldwide Oppose More Migration – Both into and out of Their Countries. Pew Research [http://www.pewresearch.org/ fact-tank/2018/12/10/many-worldwide-oppose-more-migration-both-into-and-out-of- their-countries/, 18.12.2018]. Craveiro, Daniela et al. 2019: Back to Replacement Migration: A New European Perspec- tive Applying the Prospective-Age Concept. In: Demographic Research 40: 1323-1344 [doi: 10.4054/DemRes.2019.40.45]. d’Albis, Hippolyte; Boubtane, Ekrame; Coulibaly, Dramane 2018: Macroeconomic Evi- dence Suggests That Asylum Seekers Are Not a “Burden” for Western European Countries. In: Science Advances 4,6 [doi: 10.1126/sciadv.aaq0883]. De Beer, Joop et al. 2010: DEMIFER – Demographic and Migratory Flows Affecting Eu- ropean Regions and Cities, Final Report [https://www.espon.eu/demifer, 29.01.2019]. Díaz, Sandra; Settele, Josef; Brondizio, Eduardo 2019: Summary for Policymakers of the Global Assessment Report on Biodiversity and Ecosystem Services – Unedited Ad- vance Version. Intergovernmental Science-Policy Platform on Biodiversity and Eco- system Services (IPBES) [https://www.ipbes.net/system/tdf/ipbes_7_10_add.1_en_1. pdf?fi le=1&type=node&id=35329, 21.10.2019]. European Commission 2005: Confronting Demographic Change: A New Solidarity be- tween the Generations. Green Paper [https://eur-lex.europa.eu/legal-content/EN/TXT/ PDF/?uri=CELEX:52005DC0094&from=EN] [https://eur-lex.europa.eu/legal-content/ EN/TXT/PDF/?uri=CELEX:52005DC0094&from=EN, 26.11.2018]. European Commission 2014: Population Ageing in Europe: Facts, Implications, and Policies [https://ec.europa.eu/research/social-sciences/pdf/policy_reviews/ kina26426enc.pdf, 27.08.2019] [doi: 10.2777/60452]. European Greens 2019a: Manifesto – Time to Renew the Promise of Europe [https:// europeangreens.eu/manifesto, 14.08.2019]. European Greens 2019b: Priorities for 2019: What European Greens Fight For [https:// europeangreens.eu/priorities-2019-what-european-greens-fi ght, 14.8.2019]. Eurostat 2015: Main Tables of Population Projections [https://ec.europa.eu/eurostat/ web/population-demography-migration-projections/population-projections-/main- tables, 12.02.2019]. Eurostat 2019: Main Tables – EUROPOP2018 [https://ec.europa.eu/eurostat/web/pop- ulation-demography-migration-projections/population-projections-/main-tables, 21.10.2019]. Eurostat Database 2015: Population on 1st January by Age, Sex and Type of Projection [Proj_15npms] Last Update: 05-02-2019 [https://ec.europa.eu/eurostat/data/database, 17.06.2019]. Federal Statistical Offi ce of Germany 2015: Germany’s Population by 2060: Results of the 13th Coordinated Population Projection [https://www.destatis.de/EN/Publi- cations/Specialized/Population/GermanyPopulation2060_5124206159004.pdf?_ _ blob=publicationFile, 15.06.2018]. Policy-based Population Projections for the European Union • 197 Frejka, Tomas 2008: Summary and General Conclusions Childbearing Trends and Poli- cies. In: Demographic Research 19,2: 5-14 [doi: 10.4054/DemRes.2008.19.2]. Frejka, Tomas; Gietel-Basten, Stuart 2016: Fertility and Family Policies in Central and Eastern Europe after 1990. In: Comparative Population Studies 41,1: 3-56 [doi: 10.12765/CPoS-2016-03en]. Gauthier, Anne H. 2013: Family Policy and Fertility: Do Policies Make a Difference? In: Buchanan, Ann; Rotkirch, Anna (Eds): Fertility Rates and Population Decline. Palgrave Macmillan Studies in Family and Intimate Life. London: Palgrave Macmillan: 269-287 [doi: 10.1057/9781137030399_16]. Götmark, Frank; Cafaro, Philip; O’Sullivan, Jane 2018: Aging Human Populations: Good for Us, Good for the Earth. In: Trends in Ecology & Evolution 33,11: 851-862 [doi: 10.1016/j.tree.2018.08.015]. Harmon, Nikolaj A. 2018: Immigration, Ethnic Diversity, and Political Outcomes: Evi- dence from Denmark. In: The Scandinavian Journal of Economics 120,4: 1043-1074 [doi: 10.1111/sjoe.12239]. Hazan, Moshe; Zoabi, Hosny 2015: Do Highly Educated Women Choose Smaller Fami- lies? In: The Economic Journal 125,587: 1191-1226 [doi: 10.1111/ecoj.12148]. Hilgeman, Christin; Butts, Carter T. 2009: Women’s Employment and Fertility: A Wel- fare Regime Paradox. In: Social Science Research 38,1: 103-117 [doi. 10.1016/J.SSRE- SEARCH.2008.08.005]. Hong, Sung Hyo; Sullivan, Ryan 2016: The Effects of Subsidies for Childbearing on Mi- gration and Fertility: Evidence from Korea. In: The Singapore Economic Review 61,04: 1550040 [doi: 10.1142/s021759081550040x]. IPCC 2014: Climate Change 2014: Mitigation of Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Working Group III Contribution to the Fifth Assessment Report of the Inter- governmental Panel on Climate Change. Cambridge [doi: 10.1017/CBO9781107415416]. Istat 2018: Il Futuro Demografi co Del Paese. In: Istat.It 2065: 1-30 [https://www.istat.it/ it/fi les/2018/05/previsioni_demografi che.pdf, 21.10.2019]. Jakovljevic, Mihajlo; Laaser, Ulrich 2015: Population Aging from 1950 to 2010 in Sev- enteen Transitional Countries in the Wider Region of South Eastern Europe (Original Research). In: SEEJPH 3 [doi: 10.12908/SEEJPH-2014-42]. Kallis, Giorgos 2018: Degrowth. Columbia University Press. Kalwij, Adriaan 2010: The Impact of Familiy Policy Expenditure on Fertility in Western Europe. In: Source: Demography 47,2: 503-519 [doi: 10.1353/dem.0.0104]. Kim, Young-Il Albert 2014: Lifetime Impact of Cash Transfer on Fertility. In: Canadian Studies in Population 41,1-2: 97 [doi: 10.25336/P64S52]. Kupiszewski, Marek; Kupiszewska, Dorota 2011: MULTIPOLES: A Revised Multiregion- al Model for Improved Capture of International Migration. In: Stillwell, John; Clarke, Martin: Population Dynamics and Projection Methods. Dordrecht: Springer Nether- lands: 41-60 [doi: 10.1007/978-90-481-8930-4_3]. Lanzieri, Giampaolo 2017: Technical Note: Summary Methodology of the 2015-Based Population Projections [http://ec.europa.eu/eurostat/cache/metadata/Annexes/proj_ esms_an1.pdf, 12.02.2019]. Legrain, Philippe 2014: Immigrants: Your Country Needs Them. New Jersey: Princeton University Press. • Philip Cafaro, Patrícia Dérer198 Lin, David et al. 2019: Working Guidebook to the National Footprint and Biocapacity Accounts. Oakland: Global Footprint Network Report [https://www.footprintnetwork. org/content/uploads/2019/05/National_Footprint_Accounts_Guidebook_2019.pdf, 14.08.2019]. Luci-Greulich, Angela; Thévenon, Olivier 2013: The Impact of Family Policies on Fer- tility Trends in Developed Countries – L’infl uence Des Politiques Familiales Sur Les Tendances de La Fécondité Des Pays Développés. In: European Journal of Population 29,4: 387-416 [doi: 10.1007/s10680-013-9295-4]. Lutz, Wolfgang et al. 2019: Demographic Scenarios for the EU – Migration, Population and Education. Luxembourg [doi: 10.2760/590301]. Lutz, Wolfganget al. 2018: Demographic and Human Capital Scenarios for the 21st Cen- tury. Publications Offi ce of the European Union, Luxembourg [doi: 10.2760/41776]. Marques, Alexandra et al. 2019: Increasing Impacts of Land Use on Biodiversity and Carbon Sequestration Driven by Population and Economic Growth. In: Nature Ecology and Evolution 3,4: 628-637 [doi: 10.1038/s41559-019-0824-3]. Matanle, Peter 2017: Towards an Asia-Pacifi c “depopulation Dividend” in the 21st Cen- tury Regional Growth and Shrinkage in Japan and New Zealand. In: The Asia-Pacifi c Journal 15,6 [http://eprints.whiterose.ac.uk/114351/%0Ahttp://eprints.whiterose. ac.uk/]. Mishtal, Joanna Z. 2009: Understanding Low Fertility in Poland: Demographic Con- sequences of Gendered Discrimination in Employment and Postsocialist Neo- liberal Restructuring. In: Demographic Research 21,20: 599-626 [doi: 10.4054/ DemRes.2009.21.20]. O’Neill, Daniel W. et al. 2018: A Good Life for All within Planetary Boundaries. In: Nature Sustainability 1,2: 88-95 [doi: 10.1038/s41893-018-0021-4]. OECD 2011: Doing Better for Families. Paris: OECD Publishing [doi: 10.1787/9789264098732- en]. Pilling, David 2018: The Growth Delusion. Bloomsbury Publishing. Pollmann-Schult, Matthias 2018: Parenthood and Life Satisfaction in Europe: The Role of Family Policies and Working Time Flexibility. In: European Journal of Population 34,3: 387-411 [doi: 10.1007/s10680-017-9433-5]. Raftery, Adrian E.; Alkema, Leontine; Gerland, Patrick 2014: Bayesian Population Pro- jections for the United Nations. In: Statistical Science 29,1: 58-68 [doi: 10.1214/13- STS419]. Rees, Philip et al. 2012: European Regional Populations: Current Trends, Future Path- ways, and Policy Options. In: European Journal of Population / Revue Européenne de Démographie 28,4: 385-416 [doi: 10.1007/s10680-012-9268-z]. Ripple, William J. et al. 2017: World Scientists’ Warning to Humanity: A Second Notice. In: BioScience 67,12: 1026-1028 [doi: 10.1093/biosci/bix125]. Rogers, Andrei 1990: Requiem for the Net Migrant. In: Geographical Analysis 22,4: 283- 300 [doi: 10.1111/j.1538-4632.1990.tb00212.x]. Rust, Niki; Kehoe, Laura 2017: A Call for Conservation Scientists to Empirically Study the Effects of Human Population Policies on Biodiversity Loss. In: The Journal of Popula- tion and Sustainability 1,2: 53-65. Sobotka, Tomáš 2003: Re-Emerging Diversity: Rapid Fertility Changes in Central and Eastern Europe after the Collapse of the Communist Regimes. In: Population: An Eng- lish Selection 58,4/5: 451-485 [doi: 10.3917/pope.304.0451]. Policy-based Population Projections for the European Union • 199 Stover, John; Kirmeyer, Sharon 2007: DemProj: A Computer Program for Making Popu- lation Projections. Washington DC. [http://data.unaids.org/pub/manual/2007/dem- proj_2007_en.pdf, 19.11.2018]. Stover, John; McKinnon, Robert; Winfrey, Bill 2010: Spectrum: A Model Platform for Linking Maternal and Child Survival Interventions with AIDS, Family Planning and Demographic Projections. In: International Journal of Epidemiology 39: i7-i10 [doi: 10.1093/ije/dyq016]. Thévenon, Olivier; Gauthier, Anne H. 2011: Family Policies in Developed Countries: A ‘Fertility-Booster’ with Side-Effects. In: Community, Work & Family 14,2: 197-216. [doi: 10.1080/13668803.2011.571400]. United Nations 2017a: World Population Prospects: The 2017 Revision, Methodology of the United Nations Population Estimates and Projections. ESA/P/WP.250. In: World Population Prospects The 2017 Revision. New York: United Nations [https://popula- tion.un.org/wpp/Publications/Files/WPP2017_Methodology.pdf, 12.02.2019]. United Nations 2017b: World Population Prospects The 2017 Revision Key Findings and Advance Tables ESA/P/WP/248. In: World Population Prospects The 2017 Revision: 1-46. United Nations Development Programme 2018: Human Development Indices and Indi- cators 2018. Statistical Update. New York: UN [doi: 10.18356/656a3808-en]. Wilson, Tom; Rees, Phil 2005: Recent Developments in Population Projection Methodol- ogy: A Review. In: Population, Space and Place 11,5: 337-360 [doi: 10.1002/psp.389]. Wittgenstein Centre for Demography and Global Human Capital 2018: Wittgenstein Cen- tre Data Explorer Version 2.0 (Beta) [http://dataexplorer.wittgensteincentre.org/wcde- v2/, 17.06.2019]. Ž iž ek, Slavoj 2016: Against the Double Blackmail: Refugees, Terror and Other Troubles with the Neighbours. London: Penguin. Date of submission: 20.02.2019 Date of acceptance: 29.08.2019 Philip Cafaro (). Colorado State University, Department of Philosophy. Colorado, USA. E-mail: philip.cafaro@colostate.edu URL: www.philipcafaro.com Patrícia Dérer. University of Gothenburg. Department of Biological & Environmental Sciences. Gothenburg, Sweden. E-mail: patricia.derer@bioenv.gu.se URL: https://bioenv.gu.se/english/staff?languageId=100001&userId=xderpa Published by Prof. Dr. Norbert F. Schneider Federal Institute for Population Research D-65180 Wiesbaden / Germany 2019 Managing Editor Prof. Philip Rees Dr. Katrin Schiefer Copy Editor Julia Luther Editorial Assistant Beatriz Feiler-Fuchs Wiebke Hamann Layout Beatriz Feiler-Fuchs E-mail: cpos@bib.bund.de Scientifi c Advisory Board Karsten Hank (Cologne) Michaela Kreyenfeld (Berlin) Marc Luy (Vienna) Natalie Nitsche (Vienna) Peter Preisendörfer (Mainz) Zsolt Spéder (Budapest) Rainer Wehrhahn (Kiel) Comparative Population Studies www.comparativepopulationstudies.de ISSN: 1869-8980 (Print) – 1869-8999 (Internet) Board of Reviewers Martin Abraham (Erlangen) Laura Bernardi (Lausanne) Hansjörg Bucher (Bonn) Claudia Diehl (Konstanz) Andreas Diekmann (Zurich) Gabriele Doblhammer-Reiter (Rostock) Jürgen Dorbritz (Wiesbaden) Anette Eva Fasang (Berlin) E.-Jürgen Flöthmann (Bielefeld) Alexia Fürnkranz-Prskawetz (Vienna) Beat Fux (Salzburg) Joshua Goldstein (Berkeley) Sonja Haug (Regensburg) Hill Kulu (Liverpool) Aart C. Liefbroer (The Hague) Kurt Lüscher (Konstanz) Emma Lundholm (Umeå) Nadja Milewski (Rostock) Dimiter Philipov (Vienna) Roland Rau (Rostock) Tomáš Sobotka (Vienna) Jeroen Spijker (Barcelona) Olivier Thévenon (Paris) Helga de Valk (Brussels) Heike Trappe (Rostock) Michael Wagner (Cologne)