Family-friendly and human-capital-based immigration policy

Shifting the focus from immigrants’ initial earnings to their propensity to invest in human capital

College of William and Mary, USA, and IZA, Germany

National Foundation for American Policy, USA, and IZA, Germany

one-pager full article

Elevator pitch

Immigrants who start with low earnings, such as family-based immigrants, experience higher earnings growth than immigrants who are recruited for specific jobs (employment-based immigrants). This occurs because family-based immigrants with lower initial earnings invest in human capital at higher rates than natives or employment-based immigrants. Therefore, immigrants who start at low initial earnings invest in new human capital that allows them to respond to the ever-changing needs of the host country’s economy.

Family visas associated with lower initial
                        earnings but faster earnings growth

Key findings


Immigrants who enter the US on a family visa often have low initial earnings, but also high earnings growth; compared to natives or other immigrants, they invest more in new human capital, because their opportunity costs are low, and new host-country human capital often makes their home-country skills more usable.

The willingness of family-based immigrants to invest in new skills makes the host-country labor market more flexible to employer needs and enables some immigrant entrepreneurs to provide new types of goods and services.

Family-friendly immigration policies may aid in recruiting the most highly skilled migrants as they are concerned about immigration of their families and may have employment options in several different countries.


Uncertainty about their ability to remain in their host country reduces immigrants’ incentive to invest in new human capital.

Because immigrants often bring unusual combinations of home- and host-country skills to their jobs, countries with relatively inflexible occupation requirements may discourage immigrant human capital investment.

Author's main message

Family-based immigrants usually do not enjoy the immediate high demand for their skills that employment-based immigrants do, but they experience much higher rates of earnings growth. Their entire earnings path is a better indicator of the value of their migration, both to the immigrant and to the host country. Their high rates of investment in new skills helps to make their host country’s labor market more flexible to changing needs, and leads to entrepreneurial creation of new goods and services. Thus family-based immigration is a valuable component of any national immigration strategy, offering different types of benefits to the host economy.


A diversity of ever-changing immigrant admission policies exists among developed countries. This time-variant collage of policies may be broadly envisioned as a grid with varying weights assigned to three major immigrant-selection strategies: (i) the admission of people with specific skills to fill specific jobs (employment-based immigration); (ii) admission on the basis of general measures of human capital, such as an applicant’s schooling level (human-capital-based immigration); and (iii) the admission of people, irrespective of either specific or general human-capital qualifications, who have family members in the host country (family-based immigration).

Refugees are a separate category whose numbers, at least in principle, are determined not by national immigrant strategies but by humanitarian need. Although refugees have much less control over the decision to migrate than family immigrants, they are similar in that their skills often do not initially transfer to their host country’s labor market.

One part of the current US admission program fits squarely into the first category, employment-based immigration. Immigrants (at both high and low levels of education) with specific skills are admitted to fill specific jobs. The very nature of their admission—based on an employer’s willingness to participate in a labor certification process—guarantees that employment-based immigrants have skills that are immediately valued in the host-country’s labor market. Nevertheless, the predominant US admission strategy since 1965 has been family-based immigration. Indeed, the US has the most family-based immigration system of all developed nations.

Most economists think that the US economy would benefit by adopting an immigrant admission system that is less family-oriented and more employment and human capital based. They argue that employment-based and human-capital-based immigrants productively contribute to their host countries’ economies, while family-based immigrants and refugees do not, their admission being justified solely on humanitarian grounds.

In considering which type of admission system is best for an economy, much attention is paid to whether immigrant earnings equal those of natives with similar levels of human capital. Desirable immigrants are those who, given their levels of human capital, rapidly adjust to the host-country labor market. From this rapid-assimilation perspective, an employment-based admissions strategy is best, since research shows that only employment-based immigrants initially earn on par with similarly educated natives. In contrast, immigrants who are not selected to fill specific jobs initially earn substantially less than employment-based immigrants [1], [2], [3], [4].

There is, however, great difficulty in identifying which specific skills a nation should recruit through immigration. This challenge has led some scholars to argue that an admission strategy based on general levels of human capital (e.g. level of schooling) better serves the ever-changing, hard-to-predict needs of a dynamic economy than a “specific skills for specific jobs” approach. However, even immigrants with advanced degrees sometimes have initial difficulty using their credentials or using their skills in their host country’s labor market. For this reason, at least a portion of immigrants selected for their human capital may have incentives for human capital investment similar to those of family immigrants.

Hence, a different perspective emerges when the likely causes of low initial earnings (adjusted for schooling levels) are considered and when the relationship between immigrants’ entry earnings and earnings growth is analyzed. This perspective challenges the notion that low initial immigrant earnings represent an inefficient use of immigrant human capital. Instead, it suggests that family-based immigrants do contribute to the long-term economic productivity of their host countries. More fundamentally, in countries with flexible labor markets, immigrant economic adjustment confers significant economic benefits.

Discussion of pros and cons

Immigrant entry earnings and earnings growth

Following a shift in the 1960s from an immigration policy favoring west-European immigration to a predominantly family-based policy, the education-adjusted entry earnings of US immigrants fell precipitously. Accompanying this decline was an increase in immigration from developing countries.

Post-1965 immigrants to the US from regions with high levels of economic development had initial earnings that approached or exceeded the earnings of similarly educated and experienced US natives. Although the trend has moderated over time, immigrants from developing countries have lower initial earnings than similar US natives or immigrants from more developed economies [5]. They are less likely to emigrate from the US, to their home or to some other country, than immigrants from more developed countries.

A plausible explanation for the low entry earnings and higher permanence of immigrants from developing countries is that limited opportunities in these countries make it worthwhile for individuals to migrate to developed countries, even when their source-country skills do not fully transfer to the new labor market and immigration entails substantial post-migration investment in new skills and credentials. More generally, people who are constrained from pursuing long-term dreams in their origin countries—whether the constraints be war, gang violence, discrimination, limited professional opportunities, an inflexible labor market, restrictions on adult education, rigid social structures, or limited opportunities for their children—would be more likely to migrate, even if the migration entailed substantial investment in new human capital that may include pursuing a new line of work. Such individuals would embark on these pursuits only if the benefits from investing in new human capital could be reaped. Embedded in their decision to migrate is the decision to stay in the adopted country.

The US decline in immigrant entry earnings adjusting for education and age has been interpreted as a decline in immigrant labor market quality and attributed to the change in the mid-1960s to a family-based immigration policy.

Yet, theoretically, one would expect a higher propensity to invest in human capital for immigrants whose (adjusted) initial earnings are low. Immigrants whose source-country skills do not transfer fully to the new labor market will—due to their lower wages—face a lower opportunity cost of human capital investment. In other words, the time they spend learning new skills (instead of fully devoting their time to earning) is less costly for them than for natives or for immigrants with highly transferable skills, such as employment-based immigrants. Their undervalued source-country skills are also useful for learning new skills: people who have learned one set of skills have learned how to learn, and common elements between old and new skills aid learning.

Combining a low opportunity cost of human capital investment with the learning potential of undervalued human capital increases the incentive to invest in human capital. This higher incentive pertains not only to human capital that restores the value of an immigrant’s source-country human capital (the foreign-born chemical engineer who learns English so that he can be a chemical engineer again) but to new human capital in general. Natives well launched in their careers or immigrants with skills that are immediately valued would be reluctant to undertake training in another field, even if the training complemented their current work and ultimately facilitated better-paid work. The low opportunity cost for immigrants who could not initially transfer their source-country human capital—paired with the value of that human capital for learning new skills—makes further training an attractive option. Immigrants, especially those who lack host-country-specific skills, such as family-based immigrants and immigrants admitted for their general levels of human capital, should be more likely to invest in new skills and do so over a longer period than otherwise similar natives or employment-based immigrants.

Most labor economists believe that earnings growth reflects human capital investment. Immigrants who start at low (high) earnings relative to natives with similar levels of schooling have high (low) earnings growth. Furthermore, decreases (increases) in the initial earnings of immigrants are associated with increases (decreases) in earnings growth [5], [6]. Figure 1 shows this phenomenon for US immigrants. Despite the fact that immigrants’ entry earnings decreased with the advent of family-based admissions, their earnings growth increased to such an extent that ten years later the earnings of the more recent cohort equaled the relative earnings that earlier cohorts had achieved after ten years. The convergence occurs because the more recent cohort, with lower relative entry earnings, had higher earnings growth [5].

As adjusted entry earnings fell, earnings
                        growth increased

Similar results emerge following individual earnings with longitudinal survey or administrative record data [7], [8]. Regardless of where immigrants begin, their earnings converge with time, resonating with the prediction that low-skill-transferability immigrants invest more in human capital, and hence experience higher earnings growth, than high-skill-transferability immigrants (employment-based immigrants). Studies of Canadian immigrants also reveal an inverse relationship between immigrant entry earnings and earnings growth [9] and, in both Canada and the US, family admissions are associated with lower entry earnings, relative to employment-based immigrants, but higher earnings growth [1], [2], [3], [4].

Does the higher earnings growth simply reflect language learning?

Language is an obvious example of a skill that often increases with years in the host country and that makes it easier for immigrants to use their home country’s human capital. However, it would be a mistake to view this as the only important complementarity between the human capital acquired in each country. Several observations suggest that much more than host-country language acquisition fuels the high earnings growth of immigrants with low initial earnings.

First, if the inverse relationship between initial earnings and earnings growth was mostly explained by learning the host-country language, then immigrants’ initial earnings should be mostly explained by variations in host-country language proficiency. Yet, several variations in US immigrants’ initial earnings (adjusting for schooling and age) defy an English-language-proficiency explanation. For instance, the education-adjusted entry earnings of many post-1960 cohorts from Korea, India, China, and the Philippines are similar despite enormous variation in their English proficiency. The common link among these countries is not a similar level of English proficiency, but rather a lower level of economic opportunity vis-à-vis the US.

Human capital investment takes myriad forms and is difficult to measure. Nevertheless, what can be measured further suggests that much more than language acquisition fuels the high earnings growth of US and Canadian immigrants. Canadian immigrants have higher rates of occupational change, and at older ages, than natives; the timing of the occupational changes across year-of-entry cohorts suggests that these changes are responses to an evolving demand for different types of labor market skills [10]. Examining schooling that leads to a high school diploma or a college degree (and thus is unlikely to be schooling meant to improve English skills per se), US adult immigrant men and women (who predominantly enter the US via family admissions) are much more likely than natives, at all ages, to be enrolled in school [6]. The greater propensity of immigrants versus natives to pursue schooling is particularly large for immigrants who have decided to stay permanently in the US—foreign-born naturalized citizens—and it is particularly high for family-based immigrants. At age 25, the propensity of a family-based immigrant to attend school is two and a half times greater than for an employment-based immigrant; at age 40, it is twice as high; and at age 50 it is one and a half times greater. Figure 2 presents these relationships for both younger and older immigrants.

Estimated in-school rates for immigrants
                        by age and probability of entering the US on a family-based visa

Family admissions and the admission of highly educated immigrants

The debate about immigrant admission policies is presented as a dichotomy: admit highly educated immigrants (and immigrants with specific skills) or admit immigrants with kinship ties. Yet highly educated immigrants have families too. Will they be more likely to choose a country where their siblings, parents, and adult children are also welcome, or one where only certain family members can follow? When immigrant scientists and engineers are asked why they moved to the US, for instance, family reasons dominate [11]. A family-friendly policy may thus help attract highly educated immigrants.

US Census data matched to admissions data reveal that immigrant education levels positively correlate with the percentage of employment-based immigrants. This is not surprising, given that professionals, scientists, and artists of exceptional ability are admitted via one employment category. But immigrant education levels also increase with sibling admissions [8]. This suggests that employment-based immigrants are followed by their siblings, who gain admission via the siblings’ preference. If the employment-based immigrants are highly educated, their siblings are likely to be highly educated too.

Siblings following in the footsteps of well-educated, employment-based immigrants would be most prevalent among immigrants from less developed countries with more limited opportunities for the highly educated. All other things being equal, a college-educated sibling of a US immigrant from Germany, for example, would be less likely to migrate to the US than the college-educated sibling of a Mexican immigrant. Indeed, education and percentage of siblings also immigrating are negatively associated for US immigrants from Europe, but positively associated for immigrants from Asia and Central and South America [4].

The effect of family- versus employment-based immigrants on the host region’s labor market

A concern in immigration policy is how the selected immigrants will affect the earnings of natives. Economic theory suggests that the greater the extent to which immigrant workers substitute for native-born workers, the greater the potential for harmful effects. If low skill transferability promotes immigrant employment that is distinct from native-born employment then low-skill-transferability immigrants (kinship-based immigrants, refugees, and people admitted based on their level of schooling) may pose less of an economic threat to natives than high-skill-transferability immigrants (employment-based immigrants).

Entrepreneurship may be one way that this occurs. For example, for the two largest US immigrant groups, Asians and Hispanics, employment-based admissions negatively affect or have no effect on the propensity for immigrants to be self-employed. In contrast, sibling admissions for these two groups have a large positive impact on immigrant self-employment that overshadows the effects of all other explanatory variables [4].

As discussed previously, human capital acquired in immigrants’ original countries that does not transfer to the host country is useful for learning new skills. Low-skill-transferability immigrants, particularly those who are highly educated, are able to learn new skills and methods at a lower opportunity cost than observationally equivalent natives and employment-based immigrants. Thus, the immigration of people with low skill transferability, particularly the highly educated, should foster new business development by natives. Empirically, variation in concentrations of highly educated US immigrants across time and space positively correlates with business development by US natives. States with higher shares of college-educated immigrants in their workforce have higher rates of new entrepreneurship by US natives [12].

US Census data matched to admissions data also suggest that employment-based immigrants have a small but statistically significant negative effect on the employment opportunities of native-born white males, while family-based immigrants have a statistically significant positive effect on the earnings of US-born white and black Americans [13].

Limitations and gaps

The usefulness of immigrants lacking transferable skills to a given economy will depend upon the flexibility of its wage structure and the extent to which it embraces mid-career education and occupation changes. Without these ingredients, low-skill-transferability immigrants may not have a greater propensity to invest in human capital. Rigid structures governing wages and employment would also make finding employment a major challenge for immigrants who initially lack transferable skills.

The predominant methodology used by economists to measure immigrant economic assimilation biases downwards the earnings growth of immigrants who start with low adjusted earnings. The accurate depiction of how immigrants contribute to their host-country economies requires an empirical approach untethered by a priori assumptions [5].

In addition, more information is needed on immigrants’ human capital investment. Data that may be particularly useful include the New Immigrant Survey and the Survey of Income and Program Participation. More information is also needed on the effects of the skill acquisition that accompanies the high earnings growth of immigrants with low initial earnings on a nation’s economic productivity.

Summary and policy advice

Scholars and policymakers often focus on immigrants’ initial earnings and ignore earnings growth. This creates a distorted view of immigrants’ successes and economic contributions to their host countries. This focus promotes a policy preference for immigrants who fill an immediate labor market need. While not denying the value of this type of immigrant, other immigrants also succeed economically, and the ways they adapt to their host country’s labor market adds economic flexibility and innovation in ways fundamentally different from those that fill existing slots. Policies hostile to non-employment-based immigrants risk not only losing out on these sources of innovation, but also risk damaging a country’s ability to recruit employment-based migrants. People’s decisions to migrate are more complex than the single reason stated on their visas, and excessive “cherry-picking” will alienate many potential high-skilled immigrants with families.

The strong inverse relationship between immigrant entry earnings and earnings growth in the US suggests that policymakers should not be overly concerned about low initial earnings among immigrants with otherwise similar schooling levels. It further suggests that in countries with flexible labor markets and a societal openness to learning throughout life, immigrant economic adjustment confers broad economic benefits.

If immigrants’ human capital transfers easily to the host country, then their earnings profiles resemble those of similarly educated and experienced natives. The less transferable source-country skills are to the host-country labor market, the lower are the initial earnings of immigrants relative to otherwise similar natives. However, these same immigrants experience higher earnings growth, reflecting a higher propensity to invest in host-country human capital.

Those who immigrate to fill specific jobs, and are paid accordingly, have less of an incentive to invest in new human capital than immigrants lacking immediately transferable skills. As such, an immigrant admission policy designed to fill specific labor market needs may be less likely to promote a flexible labor force than a family-based or human-capital-based policy.

Because they initially lack specific skills valued by the host country’s labor market, family-based immigrants and immigrants selected for their education levels will be more likely to invest in new human capital than natives or employment-based immigrants with comparable education levels. High rates of human capital investment mean that these immigrants are more able to adapt to changing skill needs in the host country’s economy. When demand shifts require workers to learn new skills, immigrants who initially lacked host-country-specific skills will be more likely to pursue new opportunities via investment in new human capital than will natives or immigrants who initially had highly transferable skills.


The authors thank an anonymous referee and the IZA World of Labor editors for many helpful suggestions on earlier drafts. Previous work of the authors contains a larger number of background references for the material presented here and has been used intensively in all major parts of this article [4], [5], [6], [7], and Duleep, H., and M. Regets. “Immigrants and human capital investment.” American Economic Review 89:2 (1999): 186–191.

Competing interests

The IZA World of Labor project is committed to the IZA Guiding Principles of Research Integrity. The author declares to have observed these principles.

© Harriet Duleep and Mark Regets

evidence map

Family-friendly and human-capital-based immigration policy

Full citation

Full citation

Data source(s)

Data type(s)