Many Western countries began immigration policies without feeling any need to monitor the long-term results. Indeed, many considered that immigration was an expedient response to labour shortages, and that the labourers, such as Turkish guest-workers in Germany, would probably eventually want to return home with their earnings at retirement. The United Kingdom seems to have begun its largest and most transformative policy in a typical fit of absent-mindedness. Workers were needed by London Underground, and that influx from the Caribbean began a de facto immigration policy, to which permissive legislation was added later, giving rights to all Commonwealth countries. In some sense it seemed a temporary expedient, and nothing more. Travel was by steamer, things happened slowly, and immigration proceeded steadily.
Western governments have rarely carried out large scale and detailed analyses of the benefits and costs of immigration with country-of-origin comparisons. Even now, it is hard to obtain good quality data on immigrants from different countries. For some host countries, even carrying out such analyses was deemed unseemly: immigration was seen as a fundamentally good policy, which should not be questioned. France was an example of a country which deliberately did not study such matters, because every citizen was a citizen, and comparisons are odious in the eyes of the Republic. As a result, French citizens are free to imagine the worst.
However, there are many general studies of immigrants’ contributions to the labour force, usually concluding that young immigrants are net contributors. Only studies which do the accounts over the entire lifetime (i.e. from the moment of birth or immigration to the moment of death or repatriation) give a full picture, in that young people become old, and require more services later in life. Many studies are often restricted to those coming specifically to work in the host country, and don’t include students and asylum seekers.
A typical finding is given in a 2014 OECD study:
Immigrants are thus neither a burden to the public purse nor are they a panacea for addressing fiscal challenges. In most countries, except in those with a large share of older migrants, migrants contribute more in taxes and social contributions than they receive in individual benefits. This means that they contribute to the financing of public infrastructure, although admittedly to a lesser extent than the native-born.
In contrast to these generally positive claims, albeit they admit some shortcomings, it is a welcome surprise to find that researchers in one country have been able to gather extensive official data from Statistics Netherlands which allows costs and benefits to be studied carefully in the long term, with results tabulated according to reasons for immigration, and by country of origin. At the moment this study of immigration to the Netherlands is in Dutch, but there is an English summary, and a full English translation is being prepared. If you want to contribute to that, so that the study might be read by English speaking governments, then the donation link is given here:
“The Borderless Welfare State” Jan van de Beek, Hans Roodenburg, Joop Hartog and Gerrit Kreffer, 2021, Amsterdam School of Economics.
(English abstract begins on page 19)
The authors stick closely to the economics of immigration, and report their findings in financial terms, without going into any deeper causes.
The report aimed to answer two questions:
1) What are the fiscal costs and benefits of immigration by migration motive (labour, study, asylum & family migration) and by region of origin?
2) To what extent can immigration provide a solution to the ageing population in the Netherlands?
The current report is an update of the Public Sector chapter of the report Immigration and the Dutch Economy (2003) by the Netherlands Bureau for Economic Policy Analysis (CPB). Both reports deploy the method of generational accounting to calculate the net contribution – revenues minus expenses – of immigrants to public finances, measured from the moment of their immigration to the time of repatriation or death. This net contribution is the key concept of the current study.
The study uses microdata from 2016 provided by Statistics Netherlands. These are very detailed, anonymized data of all 17 million Dutch residents, including about two million people with a first-generation migration background and almost two million people with a second-generation migration background.
It is rare to have a dataset of this quality, completeness, and size. The authors took 2016 data and studied 23 cost/benefit items. They then used detailed estimates of the future (post 2016) development of those 23 items used by the Netherlands Bureau for Economic Policy Analysis (CPB) in their population ageing studies.
The authors find that:
The rapid pace of immigration into the Netherlands has greatly increased the Dutch population, but not the sustainability of the Dutch welfare state. Of the 17 million Dutch inhabitants at the end of 2019, 13% were born abroad (first generation) and 11% were children of immigrants (second generation).
Currently, per capita expenditures on immigrants are significantly higher than on indigenous people in areas such as education, social security and benefits. Moreover, immigrants pay fewer taxes and social security premiums, which further lowers their net fiscal contribution. The current study looks back at past data as well as forecasts from Statistics Netherlands to calculate the total cost of immigration for the next two decades if policy remains unchanged.
Figure 2.3 gives the immigration history of the Netherlands in a snapshot.
Since 1900 there have always been some Westerners coming in to the country, but non-Westerners started coming in the 1970s and it is predicted (“prognose”) that by 2060 their higher birth-rates and continued immigration will make them about 23% of the population.
There is a very big difference in financial contribution according to the reason for gaining entry to the country:
Those coming to take up a job generate a positive net contribution of, on average, €125,000 ($152,500) per immigrant.
Those coming to study cost €75,000 ($91,500) per immigrant.
Those entering for “family formation” or “family reunification” cost €275,000 ($335,500) per immigrant. (Get one immigrant, then get their marriage partner, then get other family members including elderly parents).
Asylum seekers cost €475,000 ($579,500) per immigrant.
There are also considerable differences by region of origin. On average, Western immigrants make a positive contribution of €25,000 ($30,500), while non-western immigrants cost nearly €275,000 ($335,500). Within the categories of Western and non-Western there is, however, much variation.
Immigration from most Western regions usually has a positive fiscal impact. Immigrants from Japan, North America, Oceania, the British Isles, Scandinavia, and Switzerland, in particular, make a significant positive contribution of roughly €200,000 ($244,000) per immigrant. On the other hand, immigration from Central and Eastern EU-member states costs about €50,000 ($61,000). Immigration from former Yugoslavia and the former Soviet Union mainly concerns asylum seekers, who cost much more at €150,000 ($183,000).
Table 0.2 gives detailed results.
The authors comment on the differences between immigrants from different countries as follows:
Immigration from non-Western regions is usually unfavourable for public finances. This applies especially to the areas of origin Caribbean, West-Asia, Turkey and North, Central and West Africa with net costs ranging from €200,000 to €400,000 per immigrant, and Morocco, the Horn of Africa and Sudan with net cost of €550,000 to €600.000 per immigrant. By way of comparison: an average Dutch native is roughly ‘budget-neutral over’ his or her life.
For all migration motives, Western immigrants seem to ‘perform better’ than non-Western immigrants. The difference is approximately €125,000 for labour and study migrants, and €250,000 for asylum and family migrants.
In isolation, only two categories seem favourable for Dutch public finances; labour migration from western countries (except Central and Eastern European countries), Asia (except the Middle-East) and Latin America, as well as study migration from the EU. All other forms of immigration are at best budget-neutral or have a considerable negative fiscal impact. The highest net costs apply to asylum migration from Africa. It should be noted that study and labour immigration usually come with family migration, which may have a considerable negative impact on the combined net contribution.
Almost by definition, the Dutch citizenry are budget neutral over a lifetime, but if the Dunedin data are anything to go by, it will be because 20% of citizens need 80% of the financial support. Call them the Very Needy. The problem with predominantly low-skilled immigration in a welfare state is that you add too many people to the “Very Needy” benefit-receiving category.
The remaining 20% of benefits will go to the middle 40% of the population: people who need support only every now and then. Call them the Trampoline People: they bounce back.
So, to summarise what welfare states are like (using the Dunedin studies as a prime example): 20% of citizens get 80% of the money; 40% of citizens get the remaining 20% of the money; the top 40% contribute a lot but get nothing back, except perhaps the satisfaction of having helped the other 60% of their fellow countrymen. In fact, they may get nothing back but still be deeply resented, but that is another story. Call them Altruists, or Suckers, or the Blood Sucking Elite, depending on your perspective.
The authors then turn to the crucial matter of second-generation outcomes. Clearly, if everything goes well in the second generation, first-generation costs may seem worthwhile. Here are the main findings:
For only a limited number of countries and regions is the second generation as well or better integrated than natives. On a scale on which the natives are 100% integrated, some immigrants are even more successful. These are mainly the Scandinavian countries and Switzerland (110%) and some East Asian countries, notably Japan (128%), China (115%) and the Asian Tigers (South Korea, Hong Kong, Taiwan and Singapore, 104%).
Those first-generation migrants who cost the country a lot, produce second generations who either also cost a lot, or at best are only budget neutral, such that the net present value of future generations will not pay back the cost incurred by that first generation. Things have not “changed for the better” as those in favour of immigration often assert. Certainly, on average the second generation in many respects performs better than the first, but this doesn’t offset the costs of the first generation for those groups with net costs for the first generation. A simple reason is that for the first generation the costs of education and youth-related expenditures etc. mainly have already been paid by the source country, and for the second generation those costs are paid by the host country. The second generation includes those immigrants who marry among their own group (two parents with immigration background) and those who have mixed background (one local with one immigrant).
The Dutch CITO tests are administered at age 12. The Cito exam (Cito-toets) is an independent assessment of final year Dutch primary school pupils. It is given in February and an estimated 167,000 children sat the exams in 2021.
The examinations have been given for 35 years and 90 percent of pupils are given 290 multiple choice questions testing their Dutch and comprehension skills, mathematics, world orientation (which involves geography, biology and history) and study skills. The authors point out that the “Cito test” at the end of primary school is not an intelligence test, but an assessment tool for the so-called 8th grade school recommendation. On this basis children will be advised to go to either more academic or more vocational schools, according to their scores.
Despite Cito not being a formal intelligence test, on the basis of extensive UK data on secondary school intelligence tests at age 11 and national scholastic examinations at age 16, correlations as high as 0.8 might be expected between the two. Scholastic tests early in schooling are usually good predictors of later exam success. So, though Cito is not an intelligence test, it is a good predictor of later educational progress, and a good indicator of general mental ability.
Furthermore, the authors find a strong relationship between education and net lifetime contribution to the Dutch treasury. A higher level of education on average implies a higher net contribution. In the same vein, a higher score on the so-called ‘Cito test’ – a 50-point student assessment scale for primary education – is correlated with a higher net contribution to Dutch state coffers. The lead author points out that if you make a map of the world showing the net fiscal contribution immigrants make (as shown above), and another map of the world showing the Cito scores of immigrants from those countries, the two maps correlate so strongly it is hard to tell the difference.
The Cito scores of the 1st generation are good predictors of success as an immigrant.
The Cito scores of the 1st generation correlate:
- .78 with lifetime net fiscal contribution of the 1st generation
- .86 with cito scores of the 2nd generation
- .72 with cito scores of the 3th generation
- .72 with lifetime net fiscal contribution of the 2nd generation
- .80 with the integration-indicator for the 2nd generation
The authors also find that brighter immigrants are more likely to be able to re-emigrate, presumably because they can even further better their circumstances. The emigration rate after 10 years correlates r = .86 with the Cito scores (page 196).
Cito scores are strongly correlated with the educational level of the first generation, with the Chinese as a notable exception, in that first generation poorly educated Chinese parents led to high Cito scores for the first, second and third generation.
Selecting on the educational level of the first generation is thus the best way to select for positive results for the host country’s treasury (provided one has a sound method of assessing the real value of ‘foreign diplomas and educational certificates’). As a rule of thumb: immigrants on average should have at least bachelor level (“hbo-bachelor” in the Dutch educational hierarchy) in order to show a neutral or positive net fiscal contribution. This is lower and more usual qualification for those not going to university.
The educational attainments of immigrants are very important. Immigrants with at most primary education cost €400,000, immigrants with a master’s degree contribute € 300,000.
As regards their test data, they say:
Furthermore, a robust correlation exists between net contribution and scores on the so-called ‘Cito test’, a 50-point student assessment scale for primary education. For natives, lifetime net contributions range from roughly –€400,000 for the lowest Cito score to +€300,000 for the highest Cito score. For people with a second-generation migration background, a similar correlation exists, though at a considerably lower level.
There are considerable differences in Cito scores between regions of origin and also between migration motives. At group level, the net contribution of first-generation and second-generation immigrants is closely related to the Cito scores. The Cito scores and the educational level of the second generation are strongly related to the educational level of the first generation. Partly for this reason, any educational attainment of the first generation affects the net contribution of the second generation. ‘Mixed relationships and marriages’ have a considerable positive effect on the Cito scores and thus, the net contribution of the second generation.
So, the Netherlands will not get richer by immigration. Might new immigrants at least boost the birth rate sufficiently to help compensate for an aged population?
For almost half a century, the average number of children per woman has been around 1.7, well below the replacement level of approximately 2.1 needed for a constant population size. This causes every new generation to be smaller in size than the previous one, resulting in a dwindling share of young people in the population, a process sometimes referred to as “dejuvenation”. It also leads to an increase in the so-called grey pressure, that is, an increase in the ratio between the over-65s and people in the age group 20 to 65 years.
In line with the literature, this study found that solving dejuvenation by immigration resembles a pyramid or Ponzi scheme. A simulation shows that ever-increasing numbers of immigrants are needed to keep the Dutch grey pressure at the 2020 level. This results in significant population growth: 35 million inhabitants by the year 2060, 75 million at the end of this century, and half a billion by the year 2200.
Immigration does not provide a stable solution to population ageing because the underlying problems of low fertility and dejuvenation are not resolved. On average, fertility of immigrants is below the replacement level as well, partly because women from high fertility groups adjust their fertility downwards over time, and partly because immigrants from most countries in the Americas, Europe and Eastern Asia already have low fertility rates.
This is a sobering finding. Immigration is not making the country richer, nor is it sustaining the original citizens. It will merely make them minorities in their over-populated country. Might a way out of this dilemma be found by getting the larger immigrant population to fund healthcare and pensions for the elderly?
Due to the increasing ageing of the population (the grey pressure), the costs for healthcare and state pensions are rising rapidly. The current study challenges the idea that it is possible to absorb the additional costs of an ageing population through immigration.
Immigration as a means of absorbing the costs of an ageing population encounters the same practical objection as the strategy of immigration as a demographic solution to the ageing population, being strong population growth. A simulation shows that closing a permanent financial gap in public finances of 2.5% of gross domestic product, by admitting labour migrants with high economic potential, would lead to additional population growth of 7.2 million inhabitants in the period 2020-2080. In addition, mass recruitment of high-potential migrants may prove difficult in practice, as most high net contributors currently come from countries that are themselves grappling with a rapidly aging population and/or trying to attract highly skilled immigrants.
So, it would seem that this is another non-starter. Talent is what counts, and not every country can win the war for talent.
The authors point out that, given current costs and the projections for future immigration 2020-2040, the total cost of immigration will continue to rise, by an extra €600 billion ($732 billion). This is a projection, but it seems likely to be an accurate one, since cultural explanations of assimilation as a learning process assume that first generation immigrants will adapt relatively quickly and the second generation will complete any remaining steps required to blend with the host population. That is to say, if it has not happened in the second generation, there seems to be little justification for assuming it will happen in the third generation.
A slight increase in asylum seekers would lead to a significant increase in future costs. If the volume of asylum migration from West Asia and Africa kept pace with the population growth there, this alone would result in additional costs of €64 billion ($78 billion).
Less costly outcomes would require a fundamental policy change. If labour migration came only from Western countries (not including Central and Eastern Europe), from Latin America and Asia (except the Middle East), and if there was also a 50% reduction in family migration and a 90% reduction in asylum migration, then total immigration would not even be absolutely budget neutral. This scenario is highly selective. And would require changes in international treaties, such as the UN Refugee Convention.
The authors venture some policy implications:
The net costs of immigration to the government are considerable, and projections show they will consume a steadily increasing portion of the annual government budget. These costs are mainly due to redistribution through the welfare state. Continuation of the current level of immigration and current arrangements of the welfare state increases pressure on public finances. Downsizing the welfare state and/or curtailment of immigration will then be inevitable.
Immigration does not appear to be a solution to the ageing population either. In essence, ageing is mainly dejuvenation due to a low fertility rate. The only structural solution for that is an increase in the average number of children per woman in the Netherlands to approximately 2.1.
As this report demonstrates, educational attainment of the first generation, and school success of children of the first and second generation are quintessential for positive outcomes.
A more structural approach is to monitor the costs of current immigration flows and to keep an account of the outstanding claims that immigrants have on the treasury. This brings us to the value of periodic surveys of net contributions. The Dutch government has not published data on net contributions to public finances of migrants since 2003. We can only guess the reasons for this. Hopefully, this research will make it clear that this information is necessary for the foundation of policies and insight into future government spending.
Immigrants that make on average a significantly negative contribution to Dutch public finances are mainly those who exercise the right to asylum, especially if they come from Africa and the Middle East. The latest UN population forecast shows that the total population in these areas will increase from 1.6 billion to 4.7 billion by the end of this century. It is not implausible that the migration potential will at least keep pace. Migration pressure, in particular on the welfare states in North-Western Europe, will therefore increase to an unprecedented degree. This raises the question of whether maintaining the open-ended arrangement enshrined in the existing legal framework is a realistic option under these circumstances.
As far as I know, this is the only published study on the costs and benefits of immigration conducted on detailed official financial statistics for an entire national population and tracing everything back to the country of origin of immigrants, and also for their reason for emigrating.
It is a considerable contribution, in that it puts the many claims which have been made about immigration to an evidence-based test. It is a sober economic study, restricting itself to the data, but venturing a cultural and educational explanation for the costs of immigration. They mention World Values Survey data such as Traditional vs Secular/Rational values, and Survival vs Self-Expression values. They also mention World Values Survey clusters of immigrants largely into religious/belief groups. The authors have backgrounds in mathematics, anthropology, economics, sociology and demography, but make it clear they are not psychometricians.
It is astounding that there are few comparable publications on detailed national data, given that immigration is, de facto, a major policy in the West. If anyone seriously argues that these results are specific to the Netherlands, then they should back up their claim by referencing studies of commensurate quality and detail drawn from several other nations.