As consumers, workers, innovators, and entrepreneurs, immigrants help shape nearly every aspect of the economy and of society more broadly. Labor market consequences are perhaps the most visible and most debated economic concern, due to their direct impact on employment and wages. But product, housing, and capital markets are affected by immigration as well, as are some nonmarket activities and—as a result of human capital formation and innovation induced by high-skilled immigration in particular—the trajectory of long-run economic growth. This chapter discusses these economic impacts of immigration that take place beyond the labor market while recognizing that many outcomes associated with them are influenced by their interaction with changes in labor supply and demand over time.1
The chapter begins (Section 6.1) with a description of aggregate-level impacts: year-to-year changes in gross domestic product (GDP) or in GDP per capita, driven by expansion of the labor force and physical capital, as well as production-technology adjustments, responding to immigrant flows. Borjas (2013), in considering the impact of immigration on overall economic activity, estimated that the presence of immigrant workers—the stock of authorized and unauthorized foreign-born workers—in the labor market makes the U.S. economy an estimated 11 percent larger each year
1Nathan (2014), surveying what is only a fairly recent literature, organized these “wider economic impacts of immigration” (beyond labor markets) into a dynamic framework encompassing the production and consumption sides of the economy; the focus of his literature review is on the role of high-skilled immigrants.
(which amounts to around $2 trillion in GDP in 2016). As a percentage of the overall economy, annual GDP growth directly attributable to the labor of recent immigrant inflows is much smaller, and it mostly accrues to immigrants themselves.
However, when factors beyond those directly attributable to labor force expansion are considered—for example, the contribution of immigrants to capital formation, entrepreneurship, and innovation, which also shape the way and the pace at which growth unfolds—expansion of the aggregate economy attributable to new arrivals becomes much larger. Recent immigrants have higher patenting rates than natives due to their concentration in science and engineering and to their disproportionate representation among highly educated workers. One would expect this increased innovation to exert a positive externality on the productivity of natives, very likely raising per capita GDP growth. Peri (2012) performed a state-level analysis of the impact of immigration on total factor productivity. Using historic settlement patterns (similar to Hunt and Gauthier-Loiselle, 2010) and distance to the Mexican border as instruments to control for endogeneity of immigrants’ locational choices, he found immigration increased total factor productivity by promoting efficient task specialization. Similarly, Peri et al. (2015a) found that cities experiencing a greater immigration of science and engineering workers (broadly defined as the share of a city’s total employment comprised of foreign science, technology, engineering, and mathematics [STEM] workers) increased the productivity of college labor.
Evidence also exists (see Borjas, 2001; Cadena and Kovac, 2016; Somerville and Sumption, 2009) that immigrants locate in high labor demand/high wage areas for the skills they possess and are more willing than natives to relocate in response to changes in labor market conditions. This tendency may reduce friction and slack in labor markets by reallocating labor in a way that helps equalize compensation across geographic areas (see discussions in Chapter 5 of problems for spatial approaches to measuring wage effects of immigration).
Sections 6.2 and 6.3 discuss the consequences of immigration in specific sectors of the economy where the foreign-born population share is high and consider the influence of immigration on consumer prices and cost of living. Increases in the share of low-skilled immigrants in the labor force appear to have reduced, over time, the prices of immigrant-intensive services such as child care, eating out, house cleaning and repair, landscaping and gardening, taxi rides, and construction. Most of these services are “nontradable,” which means they must be produced and consumed in the same geographic area. The decrease in prices is found to be driven by lower wages paid by those hiring in labor markets populated by low-skilled workers of Hispanic origin, particularly those with relatively low English proficiency and/or who are not legally authorized to work (Baghdadi and Jansen, 2010; Cortés,
2008). Through lower prices, low-skilled immigration creates positive net benefits to users of these services. Furthermore, the availability of low-cost, flexible housekeeping and child care services provided by the foreign-born appears to have allowed women in high-salary jobs to increase their work hours (Cortés and Tessada, 2011).
Housing is a specific sector in which immigrants play an important role. On the supply side, immigrants are disproportionately represented in construction industries (see Chapter 3). Their addition to the labor force may reduce the cost of construction and maintenance services. However, new arrivals also provide a major source of housing demand and, by raising both prices and rents, generate a potential windfall for native owners of housing. Studies of U.S. metropolitan areas have detected this demand-driven impact on the price of housing services. Saiz (2007) estimated that an inflow of legal immigrants equal to 1 percent of the total population would be expected to lead to an increase of about 1 percent for both rents and housing values. Ottaviano and Peri (2012) arrived at similar results.
Section 6.4 shifts the focus from primarily short-term economic impacts created by new immigrant flows to impacts on long-term growth. The emphasis is on technological innovation and human capital formation—viewed here as interacting, or endogenous, components of the evolving economy rather than as factors determined outside the process, or exogenously—as engines of growth that takes place over decades, not years.2 The potential effects of immigration are assessed using growth models in which future productivity and income growth are determined by investments in human capital and technological innovation.
The difference between the measured economic outcomes generated by endogenous growth models, as opposed to models in which growth is exogenous to the economy, may be significant. The recent endogenous-growth literature suggests that estimates of productivity and wage impacts of immigration can be either larger or smaller than those derived when static conditions are assumed, depending largely on the extent to which new immigrants contribute to human capital formation and innovation. In particular, this literature finds that the positive effects associated with high-skilled immigration and the negative effects associated with low-skilled immigration are amplified when viewed in a long-run endogenous growth context. These results are compatible with evidence about the educational achievement of descendants of immigrants (Chapters 2, 3, and 8). The endogenous growth models also predict that complementarities between immigrants and natives in knowledge production lead to increases in the
2 In econometric models, exogenous variables are not systematically affected by changes in the other variables of the model, whereas endogenous variables are at least in part determined by other variables or latent factors that affect them both.
rate of per capita income growth, not just increases in the level of national income (economic activity).
Some endogenous growth models are also consistent with empirical evidence suggesting that the proportion of high-skilled workers immigrating to the United States (as well as to other major receiving countries), relative to total immigration flows, has been increasing in recent decades to the point where, in some sectors, their skill levels already match or surpass those of natives (Ehrlich and Kim, 2015). In terms of their contribution to innovation and average human capital formation, the impacts of immigration that play out in the long run also operate over transitionary phases and can appear within one generation. Consider, for example, the educational attainment of the children of relatively high-skilled immigrants, which on average outpaces that of their parents and of the native-born population. Estimated medium-run effects on average wages in the population (such as after 10 years) observed in the literature (see Chapter 5) are by and large consistent with many of the predictions from endogenous growth models.
Economic activities that take place beyond the market, such as in-home production, or in markets that operate on the fringes of taxing authorities, are discussed at the end of the chapter, in Section 6.5. If immigrants devote more time to nonmarket work such as caregiving and housework than do natives—and data from time-use surveys suggest that this may indeed be the case (Ribar, 2012)—or are more likely to be employed in sectors where informal work arrangements are common, reliance on conventional sources of wage and employment data and on GDP measures will result in incomplete assessments of the impact of immigration on the economy.
The size of a market economy is a function of the total number of workers, the stock of physical capital, and the average factor output, or productivity. Immigration directly adds to the size of the economy by increasing the population and workforce; it also affects the composition of the population in a number of ways, including age, gender, and education. The presence of immigrant workers (authorized and unauthorized) in the labor market has made the U.S. economy much larger—perhaps 11 percent larger, an increase equivalent to $1.6 trillion of GDP in 2012 (Borjas, 2013). Extrapolating, in 2016 this contribution to GDP is about $2 trillion. This makes sense intuitively, as the stock of foreign-born workers in the labor market, which has accumulated over many decades, is large. According to Bureau of Labor Statistics (BLS) data on labor force characteristics, there were 25.7 million foreign-born persons ages 16 and older
participating in the labor force in 2014, representing 16.5 percent of the total U.S. workforce.3
Quite distinctly from these contributions by the stock of immigrants, it is also of interest to know how much the annual flow of new immigrants contributes to economic growth. Under normal circumstances, the annual flow of foreign-born workers into most countries is small relative to the overall population. It is therefore unsurprising that studies focusing on short-run wage and employment impacts (such as those reviewed in Chapter 5) would imply increases in GDP attributable to recent immigration that are quite small when measured as a share of the total economy. In addition, the benefit accruing to U.S. natives (the immigration surplus discussed at length in Chapter 4) is typically estimated to be a small piece of this already small overall impact. Borjas (1995b) found that the foreign-born added about 0.1 percent to the portion of GDP accruing to the native-born. Borjas et al. (1997) and Johnson (1997) found somewhat higher and lower impacts respectively, but the differences do not change the conclusion that the contribution is practically undetectable in aggregate (GDP) data. Based on this and related literature, The New Americans (National Research Council, 1997, p. 153) concluded (in the context of the 1980s and 1990s): “Overall, barring sizable immigration-induced economies or diseconomies of scale, the most plausible magnitudes of the impact of immigration on the economy are modest for those [natives] who benefit from immigration, for those who lose from immigration, and for total GDP. The domestic gain . . . may be modest relative to the size of the U.S. economy, but it remains a significant positive gain in absolute terms.”
While aggregate annual impacts are small, immigration can nevertheless make a significant contribution to economic growth, especially since immigrants are disproportionately of working age and significantly boost employment growth. Consider how different the U.S. growth path would be had all immigration been cut off 10, 20, or 30 years ago: Clearly GDP would be much smaller, and perhaps per capita GDP would be as well—in no small part because the United States would have an older population with a considerably lower percentage of individuals active in the workforce (Myers et al., 2013).4 Over the long run, foreign-born inflows have a
3 See http://blogs.bls.gov/blog/tag/foreign-born [November 2016]. The concentration of immigrants varies greatly by geographic location and economic sector. In some cases, immigrants may even supply all of a business’s labor and create all of its demand. A restaurant in an enclave that hires only foreign-born workers and where all its customers are from the same community may have little to no effect on native wages and employment, while obviously contributing to a larger national economy.
4 A recent working paper by Maestas et al. (2016) examines the effect of an aging population on per capita output at the state level in the United States. They found that per capita GDP growth during the period 1980-2010 was 9.2 percent lower than it would have been had the
compounding effect that potentially influences economic and fiscal trends in profound ways. As a result, the Congressional Budget Office and other organizations are interested both in estimating how immigration flows impact GDP and in the fiscal picture for various scenarios of the volume and composition of immigration (e.g., legal status, skill mix).5
Conclusions such as the one cited above from The New Americans—reflecting estimates derived from a static framework that typically only accounts for the direct labor share of income and the immigrant and native-born shares of the labor force—are being reconsidered in light of evidence that immigrants may increase the productivity of some natives. When factors beyond those directly attributable to labor force expansion are considered—specifically, those effects created indirectly through higher savings, investment, and capital flows—expansion of the aggregate economy attributable to new arrivals becomes larger. Ben-Gad (2008) analyzed the impact on the United States of absorbing an additional 60,000 immigrants per year over the course of a decade. If all these additional immigrants have college degrees, per capita GDP would rise by 0.15 percent at the end of the first decade. Ultimately, as the capital stock continues to adjust, per capita GDP would increase by a further 0.105 percent in the decades that follow. If none of the additional immigrants have college degrees, the additional inflow ultimately lowers per capita GDP by 0.09 percent, though natives still benefit from an immigration surplus.6
Yet all these studies, whether static (Borjas, 1995b; Borjas et al., 1997; Johnson, 1997) or dynamic (Ben-Gad, 2008), fall within the neoclassical economics tradition. Different types of labor combine with physical capital to produce output using a predetermined technology. This framework does not exclude analysis of long-run growth as the technology evolves over time; however, there is no sustained immigration-induced technological change. For example, what happens if immigrants themselves change the technology? As detailed in Section 5.6, patenting activity by foreign-born
population not aged, with two-thirds of this reduction attributable to slower growth in the labor productivity of workers and about one-third attributable to slower labor force growth. Given current population projections, their results imply that “annual GDP growth will slow by 1.2 percentage points this decade and 0.6 percentage points next decade due to population aging” (Maestas et al., 2016). This aging effect would be even more pronounced without the influence of the immigrant population, which is relatively younger than the native-born population.
6 Studying the Canadian case, Dungan et al. (2013) used a macroeconometric forecasting model to simulate “the impact on the Canadian economy of a hypothetical increase in immigration.” They found generally positive impacts on real GDP and GDP per capita, aggregate demand, investment, productivity, government expenditures, taxes, and especially net government balances, with essentially no impact on unemployment.
college graduates is estimated to have increased U.S. GDP by 1.4-2.4 percent over the decade of the 1990s (Hunt and Gauthier-Loiselle, 2010). Although the overall macroeconomic impact of immigration that takes place in a given year is modest compared to other factors, the compounding role of foreign-born innovators and other kinds of workers becomes significant to long-run economic growth.
Finally, beyond the impact of immigration on total or per capita GDP, there may be effects on the distribution of income. The flow of immigration typically alters the skill and occupational composition of a country’s workforce. If immigrants disproportionately increase the size of the lowest earnings quintiles, their addition to the population will raise overall inequality by any measure (such as a Gini index). The same logic holds for measures of poverty rates. Moreover, if immigrants are concentrated in the lowest and in the highest education groups, as is the case in the United States, this change in the composition of the population increases measured wage inequality, although such an accounting does not take into account any (positive or negative) effects of immigration on native-born workers. Analyses of the U.S. economy (e.g., Blau and Kahn, 2015 and Card, 2009) have found this direct compositional effect to be very small. That said, Card (2009) did find the effects of recent immigrant inflows on overall wage inequality in the population (including natives and immigrants) to be somewhat larger than the impact on the relative wages of U.S. natives, “reflecting the concentration of immigrants in the tails of the skill distribution and higher residual inequality among immigrants than natives” (Card, 2009, p. 1). Overall, however, Card found that immigration still accounted for only a small share (5%) of the increase in wage inequality in the United States from 1980 to 2000.
Wage inequality could also be affected when immigration impacts the wages of natives (as described in detail in Chapters 4 and 5). If, for example, immigration increases the relative supply of low-pay, low-skilled workers and there is only a partial offsetting increase in demand for goods and services they produce, the pay of low-wage workers will fall relative to that of high-wage workers—leading to an increase in measured inequality. If low-skilled immigrants competing with natives are, at the same time, complements to business owners and high-skilled workers at the high end of the income distribution, the wages of the latter two groups may rise. Such wage changes would exacerbate inequality, which is already growing due to the increasing demand for high-skilled labor that has taken place since the 1970s. In addition, international trade during this period may have put downward pressure on demand for and wages of workers in medium- and low-skilled sectors.
Although immigration flows over any given year or quarter have a minimal impact on overall levels of economic activity as measured by GDP or GDP growth, certain sectors or regions may be disproportionately affected. As documented in detail in Chapter 3, foreign-born workers are more likely than native-born workers to be employed in low-wage service sector occupations and less likely to be employed in management, professional, and sales occupations. They are also more likely to be employed in goods-producing sectors such as construction, agriculture, and manufacturing. This occupational and industrial sorting primarily reflects the disproportionate presence of foreign-born workers at the low-education end of the skill spectrum; it also represents different skill sets (e.g., English-language proficiency) that are at least partly independent of years of schooling.
Although the foreign-born have historically been concentrated in construction, farm, and service-sector jobs, they are playing an increasing role in high-skilled occupations, many of them in STEM fields.7 Research into this trend has found evidence of clear links between high-skilled immigration, entrepreneurship, and innovation in high tech sectors (Kerr, 2013b; Kerr and Kerr, 2011). This line of research, summarized in Chapter 5, supplements work on more traditional ethnic entrepreneurship focusing on small business formation in nontradable sectors such as lower-end retailing and restaurants (Kloosterman and Rath, 2001). As covered elsewhere in this report (Chapter 3 and Section 6.5 below), the higher-education sector in the United States is generating graduates who help meet the demand for high-skilled workers. In 2013, the number of foreign students attending U.S. universities was around 820,000 (F-1 visas); the number of students from China alone was about 200,000 (up from 16,000 in 2003). A large percentage of these foreign-born students, particularly at the graduate level, are enrolled in STEM fields.
Just as the occupational sorting of immigrants has resulted in the concentration of foreign-born workers in certain sectors, both low- and high-skilled, migration flows have also been spatially clumped. Many entry-level service sector jobs are located in urban areas with prior immigration, which draws low-skilled immigrants. Highly educated foreign scientists and
7Lofstrom and Hayes (2011) analyzed earnings differences between H-1B visa holders and U.S.-born workers in STEM occupations and found little evidence that the visa holders were paid less than natives of similar age and education. Hunt (2015) examined immigrant and native skills and wages in U.S. computer and engineering labor markets and found that immigrants earned higher wages on average due to higher average levels of education. The wage advantage was larger for computer workers than for engineering workers, possibly due to greater returns on English proficiency for the latter. Occupation-based samples of the American Community Survey reveal larger wage differentials between the two groups than do education-based samples.
engineers also tend to locate in cities where clustering of human capital and more efficient migrant-native task specialization are facilitated (Peri et al., 2015a). Hall et al. (2011, p. 1) reported that in 44 of the nation’s 100 largest metropolitan areas—including large coastal cities such as San Francisco and Washington, DC—college-educated immigrants outnumber immigrants without high school diplomas by at least 25 percent. The low-skilled destinations, which have the reverse distribution, tend to be in the Great Plains and in the border states of the West and Southwest.
Beyond cities, there are examples of immigrants reversing the fortunes of declining regions or helping to fuel growth in small towns. Immigrants typically flow to these towns in response to employment opportunities, such as meat packing or poultry processing. Carr et al. (2012) documented the rise of Hispanic boomtowns and examined rural populations where decline had been slowed and even reversed from an infusion of new immigrants.8Hong and McLaren (2015) explored this potential “shot in the arm for local economies,” focusing mainly on the labor market impact of consumer demand for local services. They found that the bump in consumption can “attenuate downward pressure from immigrants on non-immigrants’ wages, and also benefit non-immigrants by increasing the variety of local services available.”9 Using Decennial Census data from 1980 to 2000, the authors found evidence that, due to these effects, immigrants did in some cases raise native workers’ real wages. They also found an employment effect: specifically, that each immigrant created 1.2 local jobs for local workers, most of them going to native-born workers. Sixty-two percent of these jobs are in nontraded services; that is, where the good or service must be produced and consumed in the same local area.
Explaining why net migration patterns are most likely to affect markets for nontradable goods, Mazzolari and Neumark (2012) noted that, for some kinds of goods and services, trade is impractical due to high cost of transportation, short shelf life (e.g., restaurant meals), fixed location of output (e.g., landscaping), or even for legal or security reasons (e.g.,
8Carr et al. (2012) cite a number of factors that incentivized new arrivals to reside in specific locations where allowed by enforcement and where jobs could be found which, in the process, contributed to the creation of immigrant “boomtowns.” Among these factors were the Immigration Reform and Control Act of 1986; anti-immigrant legislation (e.g., California’s Proposition 187); militarization of the border; transformation of the meat packing industry; and concentration of oil, timber, furniture, carpeting, textiles, and other nondurable manufacturing. See an issue brief from the Immigration and the States project of the Pew Charitable Trusts for data showing the impact of immigrants slowing population decline in some counties (http://www.pewtrusts.org/en/research-and-analysis/issue-briefs/2014/12/changing-patterns-in-us-immigration-and-population [November 2016]).
national defense). These constraints create niches for low-skilled workers in occupations serving these markets.
There is evidence that immigrants are more responsive than natives to regional differences in labor demand, a factor that makes labor markets more efficient because workers flow to where wages are rising (Borjas, 2001; Somerville and Sumption, 2009). Controlling for endogeneity of destination decisions, Cadena and Kovac (2016) culled evidence from the Great Recession to conclude that “Mexican-born immigrants’ location choices in the U.S. respond strongly to changes in local labor demand, and that this geographic elasticity helps equalize spatial differences in labor market outcomes for low-skilled native workers, who are much less responsive” (Cadena and Kovac, 2016, p. 257).
Borjas (2001) examined the role of immigrants in improving labor market efficiency and found that immigration “greases the wheels of the labor market by injecting into the economy a group of persons who are very responsive to regional differences in economic opportunities” (Borjas, 2001, p. 4). The paper explored empirically and theoretically how labor market efficiencies gained from immigrants clustering in higher-wage regions raises GDP, relative to what would have been observed if immigrants had simply replicated the geographic sorting of the native population. Analyzing Decennial Census data for the period 1950-1990, Borjas found evidence that geographic sorting of immigrants reflected interstate wage differences. New immigrants were found to be more likely to locate in states that offer the highest wages for the category of skills that they possess. In other words, new immigrants “make up a disproportionately large fraction of the ‘marginal’ workers who chase better economic opportunities and help equalize opportunities across areas” (Borjas, 2001, p. 2). If the foreign-born respond to increasing wage differentials by moving toward relatively higher paying regions, they may help fill labor demand in expanding industries (such as health care10) driven by an aging population or other factors. Borjas also found evidence of greater wage convergence across geographic regions during high-immigration periods. However, at the low-skilled end of the labor spectrum, Orrenius and Zavodny (2008) found evidence that, during the 1994-2005 period, some immigrants “may have been discouraged from settling in states that set wage floors substantially above the federal minimum.” This indicates that, in some cases, immigrants locational choices are more closely linked with job opportunities (employment growth) than with wages. Cadena (2014) corroborated this hypothesis. He found that, over the 1994-2007 period, recently arrived low-skilled immigrants selectively located in states that had not increased their minimum wage levels, suggest-
ing a sensitivity among workers to the potential for subsequent “disemployment effects” that could be induced. One conclusion reached by the author is that these locational choice patterns may diffuse any negative wage impacts affecting established workers in immediately affected local labor markets throughout the country.
One barrier to this kind of efficient allocation of new workers, foreign-born or otherwise, relates to cost of living (i.e., real wages). Hsieh and Moretti (2015) found that homeowners in high-wage cities have an incentive to restrict housing supply through regulatory means. Studying the contributions of individual U.S. cities to national GDP growth, they showed that worker productivity was increasingly variable across cities, reflecting “an increasingly inefficient spatial allocation of labor across U.S. cities” (Hsieh and Moretti, 2015, p. 1). Part of this variability was tied to housing prices (and policies). They found that the main effect of fast productivity growth in cities like New York, San Francisco, and San Jose—cities with booming high-tech and finance industries—was an increase in local housing prices and local wages, not in employment. In the presence of strong labor demand, tight housing supply effectively limited employment growth in these cities. In contrast, “the housing supply was relatively elastic in Southern cities. Therefore, total factor productivity growth in these cities had a modest effect on housing prices and wages and a large effect on local employment” (p. 34). This constraint means that not all workers, including immigrants, have the option of locating in the most productive cities.11 It may also partly explain the shift since the 1990s in immigrant location patterns from traditional gateways such as California, Florida, Illinois, New Jersey, and New York to states in the Southeast, Rocky Mountain West, and Pacific Northwest (Pew Charitable Trusts, 2014a).
In previous chapters, the size and direction of wage impacts driven by immigration was shown to be highly context-dependent, varying by size and duration of inflow, the skill mix of workers, and sector. For similar reasons, immigration has an ambiguous theoretical effect on the relative prices of different goods and services. The same economic change—an increase in the supply of workers—that can lower wages and production costs can also lower prices, particularly in labor-intensive sectors (Baghdadi and Jansen,
11 It may also reflect the possibility that high-skilled labor markets are more “national” in scope while low-skilled labor markets are more local.
2010; Cortés, 2008). However, immigrants are consumers as well as producers. And, although their average purchasing levels and patterns will not exactly mirror those of the rest of the population due to a range of factors, including the sending of remittances (see Box 6-1), immigrants contribute to the demand for goods and services, creating a potentially offsetting channel
through which market dynamics may be affected.12 For example, foreign-born workers in the construction industry may lower the cost of producing new owner-occupied or rentable housing if they reduce wages—Current
Population Survey data indicate they constitute about 25 percent of all workers in construction industries. However, because they also demand units in which to live,13 the impact on final prices is ambiguous.
Though the evidence is somewhat limited, the intensive infusion of lower-cost foreign-born labor into certain occupational sectors would be expected to reduce prices, benefiting consumers who purchase these goods and services (see Section 4.4). Among the foreign-born, unauthorized workers may do disproportionately more to reduce prices because they earn less than otherwise comparable authorized workers, foreign- or native-born. Benefits in the form of reduced costs of living created by lower prices to consumers should, as noted above, be largely restricted to nontraded services. Child care, eating out, house cleaning and repair, landscaping and gardening, taxi rides, and construction are a few examples of goods or services that must be produced and consumed in the same geographic location and for which prices are most likely to be affected by local availability of different pools of labor. While international trade allows production processes to be transported to where low-cost labor is located, immigration allows the low-cost labor to be brought to where production takes place. If one takes metropolitan areas as the unit of observation, the local concentration of low-skilled immigrants working in traded industries would be expected to have little to no impact on prices, at least at the local level (Cortés, 2008, p. 383).14
Using microdata from the BLS Consumer Price Indexes, Cortés (2008) estimated variation in prices across cities and over time in relation to the proportion of low-skilled immigrants in the working population. She found that, overall, a “10 percent increase in the share of low-skilled immigrants in the labor force of a city reduces prices of immigrant-intensive services, such as gardening, housekeeping, babysitting, and dry cleaning, by approximately 2 percent” (Cortés, 2008, p. 382).15 Over the period 1980-2000, this translated into a decrease in the prices of immigrant-intensive services by a city average of 9 to 11 percent. She found the decrease in prices to be
14 There have been a few studies examining the impact on aggregate prices. For example, in their conclusions, Blanchflower et al. (2007) concluded “. . . at present it appears that A8 [visa] immigration has tended to increase supply by more than it has increased demand in the UK (in the short run), and thereby acted to reduce inflationary pressure.” Bentolila et al. (2007), using Spanish data, found high levels of immigration into Spain had a negative impact on inflation (0.9 percentage points per year), which helped to bring down the overall unemployment rate by almost 7 percentage points over the period 1999-2006.
15 This finding provides some supports for the idea that low-skilled, foreign-born workers largely compete with one another. In the year 2000, 60 percent of high school dropouts were native born, yet they made up only a quarter of dropouts working in the gardening and housekeeping sectors (Cortés, 2008, p. 389).
driven by decreased wage bills for employers, mainly those who had hired immigrants who competed directly in the labor market populated by individuals of Hispanic origin with relatively low English proficiency.
Next, Cortés used BLS Consumer Expenditure Survey data to identify which groups have the highest propensity to consume goods and services produced by sectors making intensive use of immigrant labor. The overall effect on consumption baskets was found to be largest for high-income households, who are more likely than low-income households to consume products such as child care, landscaping, and restaurant meals that are immigrant-intensive in production. Immigrants working in child care and other household services influence labor market dynamics (and patterns of consumption) in a particularly important way. The lower cost of these services made possible by the increased supply of labor for their provision has allowed native and of course some immigrant families with comparatively high levels of education and income to outsource them. As a result, individuals in these households are able to redirect labor toward higher-earning market occupations. This link is investigated in Cortés and Tessada (2011) who, using Decennial Census data to track immigrant cohorts of the 1980s and 1990s, examined how low-skilled immigration affects the labor supply of highly educated women in the United States. They found a striking correspondence between the availability of low-cost, flexible housekeeping and child care services provided by the foreign-born and increases in the number of hours worked by women in high-salary jobs.16
Of course, lower-income households also benefit from reduced prices of clothing, housing, food, etc.; however, they (especially recent immigrant cohorts) have also been empirically shown to bear the brunt of any negative wage impacts associated with new immigration that occurs. As a result, the overall (net) effect on economic well-being is negative for some and positive for others, unless immigration-induced technological progress is sufficient to raise all wages. As put by Cortés (2008, p. 414):
The low-skilled immigration wave of the period 1980-2000 increased the purchasing power of high-skilled workers living in the 30 largest cities by an average of 0.32 percent and decreased the purchasing power of the typical native high school dropouts by a maximum of 1 percent and of Hispanic low skilled natives by 4.2 percent.
These findings support the conclusion that, through lower prices, low-skilled immigration created positive net benefits to the U.S. economy during
16 For natives switching time from nonmarket work to market jobs, the reduction in their own home production does not count against GDP, whereas their new work, and that of workers they hire to do the same home tasks, is included in GDP. So, if measured by GDP only, the overall increase in the value of economic activity may be overstated.
the last two decades of the 20th century, while also generating a redistribution of wealth from low- to high-skilled native-born workers.
Immigration and the Housing Sector
Immigration significantly impacts local housing markets by contributing to the demand for apartments and single-family homes. If there is a resultant increase in home prices, then this raises the wealth of current homeowners. According to the Census Bureau, housing wealth (home equity) accounted for about 25 percent of total wealth in U.S. households in 2011.17 On the other hand, higher prices reduce housing affordability for potential home buyers. Housing expenses, including utilities and furnishings, account for 33.6 percent of average household consumer spending, with direct shelter expenses of 19.7 percent, compared to 17.6 percent of spending allocated to transportation and 12.9 percent for food. Spending on shelter is moderately higher in absolute terms for homeowners than renters, and spending by homeowners on utilities, supplies, and furnishings is considerably higher than it is for renters (U.S. Bureau of Labor Statistics, 2015b).
Demand for housing is a direct function of the rate of household formation, which depends on a host of demographic factors including immigration but also depends on the health of the economy. Household formation is defined as the net change in the number of households in a given period, also equivalent to the net change in occupied housing units (owned and rented combined). One inhibitor of the economic recovery following the Great Recession is that household formation has proceeded at less than one-half of its normal rate since 2007, eliminating the growth in spending that accompanies it (Paciorek, 2013).18 A decline in household formation is consistent with a delay in family formation, but it can also signal increased doubling up of individuals in shared housing who otherwise would have lived in separate units. The failure to increase occupancy of more housing units has its greatest impact on the construction industry, which tends to be sensitive to the business cycle.
Immigrants have accounted for roughly one-third of household formations during the last two decades. In the decade of 2000-2010, even though the pace of new immigrant arrivals was somewhat reduced, immigrants
18 “This persistent weakness in the housing market has also contributed to the slow pace of the overall economic recovery. For example, the direct contribution of residential investment to annual GDP growth frequently reached 1 to 1.5 percentage points in recoveries prior to the mid-1980s. During the 3 years subsequent to the end of the recession in the second quarter of 2009, the contribution of residential investment to GDP averaged close to zero” (Paciorek, 2013, p. 2).
still accounted for 32.6 percent of the nation’s household formations, partly because native-born household formation was contracting (Myers and Pitkin, 2013). The children of immigrants—the second generation—also add to household formation and the demand for housing. A study by the Harvard Joint Center for Housing Studies used tabulations of 1994 and 2014 Current Population Survey data to estimate that second-generation immigrants accounted for the largest share of growth in households among the under-30 cohort during the last 20 years (Masnick, 2015).
The long-term effects of immigrants in the housing market have been documented in a series of studies. The ever-rising share of household growth (owners and renters combined) accounted for by immigrants in recent decades in turn contributed to the demand for homes. Masnick (2015) calculated that the immigrant share of all owner-occupied units increased from 6.8 percent in 1994 to 11.2 percent in 2014. The immigrant share of homeowner growth rose from 10.5 percent in the 1980s to 20.9 percent in the 1990s, then to 39.2 percent in the 2000s, and is projected to be 35.7 percent in the 2010s (Myers and Liu, 2005; Myers and Pitkin, 2013). The immigrant share of rental unit growth was 26.4 percent in the 1980s, 60.4 percent in the 1990s, and 31.7 percent in the 2000s; it is projected to be 26.4 percent in the 2010s. The unusually high immigrant share of rental unit growth in the 1990s is attributed to an upswing in immigration in that decade, combined with a downswing in the population growth of native-born young adults, due to the arrival in adult years of the undersized cohort known as Generation X (those born from the mid-1960s to the early 1980s). Similarly, the high share of homeowner growth in the 2000s attributed to immigrants stemmed from advancement of that relatively small native-born cohort into prime home-buying ages, combined with the advancement of immigrants from rental to home-owning status. In the current decade, native-born homeowners are continuing to lag, with immigrants again supplying a large share of the growth that upholds house values.
Immigrants from Asian countries are observed to have higher homeownership rates than immigrants of Hispanic origin, and both rates have been lower than rates for other demographic groups, even after controlling for income (Alba and Logan, 1992; Coulson, 1999).19 However, one of the strongest findings in the immigrant housing literature is that immigrants advance rapidly into homeownership the longer they reside in the United States, with especially steep gains among Hispanics, who start from lower levels (Myers and Lee, 1998). The research indicates that the gains for the
19 Estimates based on the Current Population Survey Annual Social and Economic Supplement (IPUMS) indicate that homeownership rates (the number of owner-occupied housing units divided by the total number of occupied housing units) for Asian and Hispanic groups in 2010 were about 59 and 48 respectively, compared to around 68 for the nation as a whole.
housing market from new immigrant arrivals continue to increase for three decades after their arrival.
The discussion above suggests that immigration, like any increase in the population, has the potential to drive up an area’s house prices because, at least in the short run, the supply of housing is relatively inelastic. This is beneficial for homeowners and those who derive income from renting out accommodations. For natives who do not already own homes, whether they plan to continue renting or aspire to eventually purchase a home, this represents an increase in the cost of living. Ottaviano and Peri (2005) and Saiz (2003, 2007) found that the price of housing in metropolitan areas was systematically positively correlated with immigration. Saiz (2003) found strong evidence that the Mariel Boatlift influx of immigrants had a pronounced impact on the Miami housing market for several years following the event. Using a difference-in-difference approach common to spatial wage studies (covered in Chapter 5), he found that the unexpected shock to housing demand caused short-run rental prices in Miami to increase by 8-11 percent more than those for comparable housing markets.
Studies of a more general set of U.S. metropolitan areas have also found this demand-driven impact on the price of housing services. Saiz (2007) estimated that an inflow of legal immigrants equal to 1 percent of the total population would be expected to lead to an increase of about 1 percent for both rents and housing values.20Ottaviano and Peri (2012) found the increase in housing prices from a similar event to be between 1.1 percent and 1.6 percent. Because immigrants tend to locate in cities with faster wage (and possible housing price) growth, analyzing local labor market impacts of immigration on native outcomes without controlling for city characteristics will bias estimates. Vigdor (2013) examined the contribution of immigrants in the creation of housing wealth in places like New York City, particularly in downtown neighborhoods, while also showing how prices have stabilized in Rust Belt cities. The study, conducted using county-level data spanning 1970 to 2010, found “the most pronounced impact of immigration on housing values was in thriving Sun Belt cities that remain affordable and in declining Rust Belt cities where immigration acts as a barrier against even greater declines in home values” (Vigdor, 2013).
It should be noted that data limitations make housing price studies difficult in part because most of the data used must be aggregated to at least the metropolitan-area level. If immigrants cluster in specific neighborhoods within metropolitan areas, then analyses using Decennial Census
20 For the United States, Saiz (2007) and Saiz and Wachter (2011) found that immigration raises rents and housing values in destination cities, with population and rents rising in proportion. Within cities, the most immigrant-dense neighborhoods saw relatively slower price increases, an effect the authors attributed to native exits and increased urban-level segregation.
data will dilute effects of immigration on housing prices because the data are aggregated for the entire metropolitan area. Thus, data that can be further disaggregated to the area of individual neighborhoods or smaller levels are needed to accurately assess the impact of immigration on housing. Saiz and Wachter (2011) used track-level data from the Decennial Census to show that, even in the presence of an overall positive relationship between housing values and in-migration at the metropolitan area level, a negative relationship often emerges at the neighborhood level. This observation is indicative that immigration may be inducing sorting across neighborhoods as opposed to across metropolitan areas (this sorting still has distributional consequences). Findings by Cascio and Lewis (2011) based on school district level data sources suggest that the negative relationship between in-migration to neighborhoods and housing values may be partly accounted for by parents’ housing choices based on preferences regarding the ethnic composition of public schools.
As discussed in detail in Chapter 5, much of the research on the economic impact of immigration, such as that focusing on labor market effects, takes a somewhat short-run perspective. While these analyses typically distinguish time durations too short for firms to adjust capital from durations sufficient to allow such adjustments (“the long run”), the focus of the latter is still typically not on periods of history long enough to follow determinants of economic growth. So a further distinction must be made between analyses examining “long-run” changes in wages and employment (as “the long run” was defined for the purposes of Chapter 5) and analyses of the sources of growth in an economy. The latter are concerned with the impact of immigration on trends in GDP growth that unfold over decades, not years.
Solow (1956) famously devised a model in which growth in an economy’s total output derives from accumulation of the factors of production.21 As a nation’s capital and labor inputs expand—and, crucially, technological progress occurs—economic growth is generated. Factor accumulation alone cannot sustain growth in per capita income; technological progress is needed to overcome diminishing marginal returns to variable factors of production. The contributions of expanding labor and capital are directly accounted for in the production function, while the effects of technological change enter as a residual. The growth in total output is thus
21 Presentation of Solow’s “neoclassical” growth model can be found in any good macroeconomic text such as Mankiw (2008). Bodvarsson and Van den Berg (2009) presented a graphical representation of the Solow model in the context of the economics of immigration.
accounted for by the growth in the supply of inputs, subject to the depreciation rate of capital, and by growth in total factor productivity due to growth in technology—all determined exogenously.
As explored in detail in Chapters 4 and 5, this aggregate production function framework linking inputs to outputs is a foundational feature of much of the empirical literature measuring the effect of immigration on wages and employment;22 it provides a method of combining workers of different skills in order to evaluate competitive effects as well as cross-skill complementary effects of immigrants on wages (Ottaviano and Peri, 2012). The structural model studies using the factor proportions approach reviewed in Chapter 5 largely follow Solow by assuming a constant elasticity of substitution baseline production function.
A shortcoming of early growth models was that, once an economy had accumulated a level of physical capital sufficient to meet needs dictated by the current production technology, any further economic growth was predicated on improving that technology, which was exogenously determined. More recent models have introduced investments in knowledge—for example, human capital and innovative activity—to provide a mechanism with which to account for economic growth within the processes modeled, in effect connecting growth to internal forces within the economy and thereby making it endogenous (with respect to that model). In essence, endogenous growth models start where the Solow-type growth model ends. With human capital or other knowledge recognized as critical and controllable factors, people’s ideas and innovations become a component of technology that is subject to deliberate investment decisions; this treatment, in turn, allows the model to project self-sustaining and persistent long-term growth in both per capita and aggregate output.
Models such as those developed by Romer (1990) and Lucas (1988) shift the analytic emphasis from factor accumulation to increases in productivity by allowing the growth rate of technological progress to be determined within the system rather than exogenously. Barro (1991) and Mankiw et al. (1992) also refined empirical growth modeling by adding the concept of human capital—which includes the knowledge, skills, and experience possessed by individuals—as a factor of production. Since human capital is largely unobserved, Barro used level of academic achievement (education) as a proxy and found the variable to be statistically significant and positively related to economic growth over time. Similarly, Baumol (1993, pp. 259-260) concluded that “. . . so far as capital investment, education, and the like are concerned, one can best proceed by treating them
22 For example, a nested, constant elasticity of substitution production function was used in Borjas and Katz (2007); Borjas et al. (1997); Card (2009); D’Amuri et al. (2010); Ottaviano and Peri (2012); and other studies.
as endogenous variables in a sequential process—in other words, these variables affect productivity growth, but productivity growth, in turn, itself influences the value of these variables, after some lag. These endogenous influences are, then, critical components of a feedback process.”
One motivation behind endogenous growth modeling is to reveal how human capital—specifically the generation of new ideas through research and development (R&D) that create new products and production processes—advances the technological frontier and translates into productivity gains (Lucas, 1988; Romer, 1990). Competition is also created among firms when entrepreneurs create businesses around new ideas (Aghion et al., 2009; Schumpeter, 1950). This reassessment of economic growth processes is potentially very important for characterizing the economic contributions of the foreign-born because immigrants bring with them, and acquire, levels of human capital that are different from those of the general population.
“Endogenizing” human capital into the growth model can allow for consideration, as discussed in Section 5.5, of how innovation and entrepreneurship injected into an economy by immigrants may alter total factor productivity and, in turn, long-term growth in economic output. Peri et al. (2014) found, for example, that STEM workers (foreign- and native-born) may have accounted for 30-50 percent of all U.S. productivity growth between 1990 and 2010. Within endogenous growth frameworks, immigration provides labor and human capital factor growth—the working-age population in countries like Germany and Japan would actually be shrinking (or, in the case of the latter, shrinking more) without it—as well as other forms of capital such as financial, social, and cultural capital. Skilled migrants especially may influence drivers of productivity such as entrepreneurship (discussed in Section 6.4), investment, and innovation (Section 5.5 provides support for this).
The Main Ideas Underlying the Endogenous Growth Concept
The essence of endogenous growth theory is that the persistent and largely uninterrupted growth in per capita income in the United States and other developed economies over the past 170 years or so can be explained as the outcome of continuous investments in human capital and knowledge formation, or in direct innovative activity at the firm and industry levels, which serve as engines of advancement in total factor productivity and per capita income.23 While the literature varies in terms of the way that the
23 Though such growth is ordinarily accompanied by investment in and accumulation of physical capital as well, models that rely solely on the physical capital channel either cannot bring about sustained growth over long time periods or generate empirically implausible predictions.
mechanics and motivating forces of this process are identified, all endogenous growth models share two basic characteristics. One is that the process can be self-sustaining because of continuing investments that individuals, families, and firms make in the formation of human capital and associated physical capital. The other is that the process is invariably aided by knowledge spillover effects24 and related economies in the process of knowledge formation or technological innovations that bring about not just a self-sustaining level of productivity and (per capita) income, but a continuous rate of growth. Transitions from lower stages of economic development into regimes of continuous productivity growth occur endogenously within the economy through optimal allocation of productive resources into learning, education, basic science, and R&D, rather than exclusively through discrete technological breakthroughs that occur randomly and unpredictably in a way that is largely exogenous to the economy.
Innovation and knowledge formation occur not just through investments by the native population; they can be affected by immigration as well. While most of the theoretical literature on endogenous growth has so far been formulated in a closed-economy set-up, there is a fledgling strand, described below, that is exploring the relevance of immigration to knowledge formation in an open-economy setting. Skilled immigrants contribute to knowledge formation through their own acquired knowledge as well as via “diversity effects” in knowledge formation, as modeled in Ehrlich and Kim (2015).25 As noted by Hanson (2012), the flow of innovation is constrained by the supply of talented scientists, engineers, and other technical personnel; immigration helps relax this constraint, both in theory and in practice:
Each year, U.S. universities conduct a global talent search for the brightest minds to admit to their graduate programs. Increasingly, foreign students occupy the top spots in the search. Data from the National Science Foundation’s Survey of Earned Doctorates show that between 1960 and the late 2000s, the share of PhDs awarded to foreign students rose from one fifth to three fourths in mathematics, computer science, and engineering; from
24 Knowledge spillover effects are those that create impacts beyond the entity in which they occurred—for example, when knowledge or ideas accumulated by a specialized or geographically concentrated group of agents stimulate knowledge formation in others through interaction among agents within an organization or through transmission of knowledge across various communication and networking channels outside an organization.
25 At a highly aggregated (national) level, Alesina et al. (2016, p. 101) found that greater diversity of the skilled workforce (defined by people’s birthplaces) “relates positively to economic development (as measured by income and TFP [total factor productivity] per capita and patent intensity) even after controlling for ethno-linguistic and genetic fractionalization, geography, trade, education, institutions and origin-effects capturing income/productivity levels in the immigrants’ home countries.”
one fifth to three fifths in physical sciences; and from one fifth to one half in life sciences (Hanson, 2012, p. 26).
This process contributes to U.S. economic growth due to the fact that many foreign students stay after completing their schooling; for example, Finn et al. (2005) found that almost two-thirds of foreign-born students in science and engineering fields remained in the United States a decade after they earned their doctoral degree.
Approaches to Modeling the Mechanics of Endogenous Growth
The two main approaches used to identify the engines of economic growth in this literature are the human-capital-based models and the R&D-based technology-production models. Models using either approach replace the assumption that the technology is exogenous with one in which the economy can grow endogenously through deliberate investments in infrastructure and basic science by individuals, private firms, and the government.
The human-capital-based approach focuses on investments in human knowledge, cognitive skills, and higher education, along with other determinants of human capital (fertility, health, population size). Individuals and families invest in their own or their offspring’s learning capacity and knowledge formation.26 Such knowledge production can lead to self-sustaining, long-term growth in total factor productivity and per capita income on the assumption that “knowledge is the only factor of production that is not subject to diminishing returns” (see Clark, 1923, p. 120).
The technology production approach focuses on technological innovations that are driven by profit-maximizing firms investing in R&D and competing over innovations that yield higher-quality products and production processes or greater variety and superior quality of new goods, innovations that lead to self-generating expansion in real output per capita and individual welfare.27 This technology production is generally assumed to
26 This may be motivated by economic, altruistic, and related intergenerational objectives. The literature following this approach includes the path-setting contributions by Lucas (1988), Becker et al. (1990), and other studies included in Ehrlich (1990), which were based on dynastic-type models of investment in general human capital and fertility. Further expansions by Ehrlich and Kim (2007), Ehrlich and Lui (1991), and Galor and Moav (2004) used overlapping-generations frameworks to identify the role of additional factors that motivate individuals and parents to invest in the education, skill, and health of their children as well as complementary factors of production that enhance human capital formation and economic growth.
27 The literature following this approach, which includes Romer (1986, 1990) and Stokey (1988), emphasizes profits and rewards to innovators, as well as the market structure within which innovations are produced, as motivating forces influencing investment in innovation and growth.
be subject to economies of scale in R&D production. Even this literature, however, recognizes human capital formation as a critical factor that contributes to innovation.
The bulk of the endogenous growth model literature consists of closed-economy models, which means they do not account for trade and immigration. (The Technical Annex to this chapter, Section 6.8, illustrates the mechanics through which endogenous growth can occur in closed-economy models.) They do, however, emphasize the specific role higher education plays in the development process, essentially because tertiary education is more likely to contribute innovative ideas that enhance scientific and entrepreneurial innovations. Moreover, higher-skilled workers and inventors can generate knowledge spillover effects that enhance knowledge formation and the productivity of lower-skilled workers with whom they interact in production and job training. There is indeed a general recognition in both the literature on innovation and the endogenous growth literature that the processes of knowledge formation, innovation, and economic growth are enhanced not just by individuals’ own educational investments but also by the spillover effects conferred by the interaction within and across different skill groups, and thus also by the average skill level and educational attainments in the population.
Beyond the closed-economy models, there is a nascent literature on endogenous growth that adopts an explicit or implicit open-economy setting that allows for the role of immigration in enhancing either R&D/new goods production or human capital formation. The product-innovation-based models of an open-economy focus on the potential contribution of immigrants to the scale of the labor force employed in the R&D sector of the economy through various channels. For example, Lundborg and Segerstrom (2000, 2002) developed two versions of an open-economy model with two trading countries (either “North-North” with two rich countries, or “North-South” with a rich and a poor country) in which self-sustaining growth occurs through continuous product innovation. Firms in both countries compete to become leaders in introducing improved quality products, which are then adopted by consumers in both countries through trade. Growth is measured in terms of real consumption or utility from quality product innovations. Since all products are available to consumers in both countries, both countries share the same growth rate.
The R&D production function in this “quality ladder” model is subject to scale economies, so the equilibrium rate of growth in consumer utility is determined by the size of the labor force engaged in R&D production. Immigration matters in these models simply because it increases population and labor force size. When immigration occurs, the productivity gains enjoyed by the receiving country are offset by productivity losses in the sending country. Where the countries have similar production technolo-
gies but different population endowments and wages, (as in Lundborg and Segerstrom, 2000), there are efficiency gains from workers migrating from the more populated to the less populated country. In Lundborg and Segerstrom (2002), where the North has superior R&D production technology and wages are initially higher, immigration is again treated as an exogenous variable that is determined through the imposition of quotas. In both cases, immigration leads to more efficient production and world output rises. However, immigration reduces the welfare of the receiving country’s workers in the case where natives’ wages fall.
Another example of an innovation-based model is from Drinkwater et al. (2007), who adopted a model with R&D production serving as the engine of growth. The economy in this model consists of three sectors producing ordinary manufacturing goods and R&D output consisting of blueprints for new varieties of goods. Unlike the Lundborg and Segerstrom (2000, 2002) models, this model recognizes two types of workers—skilled and unskilled—as well as physical capital, and employment in R&D is assumed to be relatively skill-intensive. Self-sustaining growth in income occurs as a result of external economies generated by the “density” of new product varieties: the ratio of new products relative to the economy’s population, rather than population size itself. The authors call this density of new product varieties “knowledge capital.”
The focus of the Drinkwater et al. study (2007) is on how immigration, treated as exogenous, affects the receiving country’s long-term growth and the net benefit to natives in that country: the “immigration surplus.” Calibrated simulation runs of the model indicate that if immigration involves exclusively high-skilled migrants, the growth rate of real income rises due to an increase in skill-intensive R&D activity. In contrast, the net real income benefits to natives were negative if immigration was exclusively low skilled. These results derived from simulations in which skilled labor and physical capital were assumed to be substitutes in production, but the same qualitative results were obtained in simulations where the two factors were modeled as complementary. The welfare implications remain the same when measured in utility terms, rather than real income terms, in the two illustrated cases in which immigrants were exclusively high skilled or low skilled.
The human-capital-based models focus on the channels through which human capital formation and migration contribute to growth. Zak et al. (2002) developed an overlapping-generations model in which growth is enabled through human capital formation. Children’s human capital grows if parents choose to lower fertility, which varies as a function of household income. The economy may be in one of three possible development states: a “poverty trap,” a “middle-income trap,” or a balanced-growth equilibrium
path.28 The prospect of growth depends on the economy’s initial distributions of human capital among natives and immigrants, its initial levels of physical capital, and its “political capacity.” All inputs must be sufficiently above a threshold level to enable reaching the balanced growth path. Simulation runs using this model indicate that migration can enhance the level of the growth equilibrium path in the receiving economy only within specific bounds. If the migration inflow is sufficiently high or the human capital of immigrants relative to natives is sufficiently low, the development trajectory of an initially growing economy can reverse, starting a slide toward the poverty trap. But high-income receiving countries are more likely to benefit from a skill distribution of migrants that is skewed toward high levels of human capital. More generally, the model implies that, while skilled immigration can favorably affect the rate of convergence to a balanced growth path or the likelihood the latter occurs, it does not affect the economy’s growth rate if the economy is already in a growth equilibrium.
Ehrlich and Kim (2015) added a new dimension to the human-capital-based endogenous growth model that allows for international labor mobility by treating the flow of immigrants and their skill composition, as well as human capital formation, income distribution, and economic growth, as endogenous variables. To this end, they pursue an open-economy model recognizing two interacting countries—destination and source—as well as two types of workers: skilled and unskilled. For analytical convenience, these workers are assumed to be employed exclusively in two sectors producing high tech and low tech consumer goods, respectively. The goods production functions exhibit constant returns to scale in effective labor hours, but they are also subject to external effects that are decreasing in the quantity of workers but increasing in the average worker’s human capital due to workplace interactions among workers. The model recognizes both fertility and investment in human capital to be endogenous variables that are determined by parents within each skill type. To derive globally balanced growth equilibrium paths in both countries, the skilled and unskilled natives and immigrants are linked through spillover effects in knowledge production across skill group within each country, as well as across the same skill groups across the receiving and sending countries. The model offers theoretical propositions and supporting empirical evidence showing that a skill-biased technological shock (SBTS), can, for example, lead to a higher skill composition in the migration to receiving countries and that such induced migration can contribute to a higher balanced growth path of per capita income while also moderating the increase in the level of income inequality within receiving countries, both of which occur as a result of the SBTS. In an extended model, the authors allow human capital formation
to also benefit from “diversity effects” due to complementarities between immigrants and natives in knowledge production.
Empirical Evidence of the Role of Human Capital in Migration and Growth
The treatment of immigration flows and the skill distribution of natives and immigrants as exogenous variables is common to the above-described endogenous growth literature. The work by Ehrlich and Kim (2015) differs in that it treats both the growth prospect and the distributions of skill types and human capital attainments in receiving and sending countries and among immigrants as endogenous outcomes of underlying exogenous parameters, including those affecting the production and transmission of knowledge and the costs of parental investments in the quantity and human capital attainments of children. The model can therefore offer testable implications about the impact of changes in the volume and skill distribution of migration flows and population shares, as well as the impact of these changes on the global economy’s balanced growth path.
A plausible scenario in Ehrlich and Kim (2015) that leads to testable implications is one in which a skill-biased technological advance occurs either in just the receiving country or in both the receiving and sending countries simultaneously. A real-world example is the information technology revolution that started in the 1970s, became widely spread around the world in the following decades, and is still continuing. Analytical considerations and calibrated numerical simulation in Ehrlich and Kim (2015) imply that such technological advances generate a higher rate of human capital formation and full-income growth, as well as a generally rising level and share of skilled migrants relative to both the migrant and native populations in the receiving countries.
The latter implications have been tested against data from two international panels reporting the skill composition of migrant populations, indicated by college educational attainments: (1) a World Bank panel assembled by Schiff and Sjoblom (2008), including data on the 6 major receiving countries of immigration from 190 sending countries over the period 1975-2000; and (2) a 2013 panel assembled by the Institute for Employment Research, Nuremberg, Germany, (Institut für Arbeitsmarkt- und Berufsforschung, IAB), which contains data on the same receiving countries over the period 1980-2010. Both panels use aggregate census data on immigration assembled by each of the receiving countries. The metric for high skill employed in these panels is having “at least some tertiary education” (13-plus years of schooling).
Regression analysis conducted by Ehrlich and Kim (2015) based on the World Bank data for five of the six major receiving countries29—Australia, Canada, France, the United Kingdom, and the United States—indicates that the high-skilled component of net migrant flows from the 190 sending countries has indeed been continually rising over the entire sample period.30 Detailed raw data from both the World Bank and IAB panels confirm this pattern for each of the receiving countries. Moreover, while the native-born populations in the five receiving countries have experienced a rising trend in the same skill composition measures over the same period, the rise in the skill level of migrant populations has exceeded that in the native-born populations for most of them.
These findings from the World Bank and IAB panels are corroborated by more recent and detailed data from the U.S. Census Bureau. As Table 6-1 indicates, the percentage of the foreign-born population in the United States with bachelor and higher degrees has been generally rising by year of entry of immigrants even before 1970. For those entering the United States over that table’s most recent period (2000-2012), the percentage of immigrants with a college degree or higher (32.9%) exceeds that of the native population in 2012 (31.3%, as shown in Table 6-2). As Table 6-2 also shows, in 2012 the percentages of Asian and European immigrants with bachelor and higher degrees were substantially higher than the percentage of the native-born population, while the percentages of Latin American (all) and Mexican immigrants with bachelor and higher degrees were substantially lower. A similar trend is found using Decennial Census data assembled by Smith (2014b) for the average years of schooling over an even longer period: 1940-2010 (see Table 6-3). By this measure, while the average years of schooling of all foreign-born entrants is still below that of natives in 2010, the gap has been narrowing over time, with Asian and European migrants’ average years of schooling again exceeding that of the native-born population.
The Immigration Surplus in Endogenous Growth Models
The endogenous growth paradigm, which focuses on the long-term dynamic implications of immigration, also offers new insights concerning the measurement of the net economic costs and benefits to natives associated with immigration—what the literature has often termed the “immi-
29 The sixth country, Germany, was excluded due to absence of relevant time series data for a reunified Germany prior to 1990.
30 This pattern was derived from fixed effects models regressing changes in migrant population stocks on GDP (in cubic transformation) in destination countries, GDP in source countries, and other standard correlates.
gration surplus.” The standard approach for measuring the immigration surplus is based on a static framework in which the capital stock is a given constant, production of output is subject to constant returns to scale, and the economy is competitive. The surplus is then assessed as the difference between the increased output, which by definition is equal to the income of all natives in the economy (workers and owners of capital) resulting from migration, and the reduced labor wages of native workers brought about by the increased labor supply due to migration (see the simple models described in Chapter 4). Variations in this standard approach include allowances for different labor skills and possible discrete shifts in the economy’s capital stock that may accompany the migration increase. The immigration surplus thus measured is positive, but small—typically less than 1 percent of GDP (see Borjas, 1995b, and Chapter 4).
The difference between the measures of the conventional immigration surplus generated in static models or in dynamic models with exogenously determined growth and those based on the endogenous growth paradigms is that the latter account for the way immigration interacts with the economy’s human capital formation and self-sustaining growth. While it may seem that, by comparison, the measures derived in an endogenous growth context would always result in larger positive magnitudes than the static measures, the literature surveyed below indicates that this is not necessarily the case. Indeed, two of the studies—Drinkwater et al. (2003) and Ehrlich and Kim (2015)—computed the immigration surplus using numerical simulations and found that the estimates can be either larger or smaller than those derived under static conditions, depending on assumptions regarding the mix of high- versus low-skilled immigrants.
The Drinkwater et al. (2003) study provides estimates of the immigration surplus using both a baseline model, where no complementarities between skilled labor and physical capital are assumed, and an alternative model where such complementarities are allowed (the results for the alternative model are shown in parentheses below). If migration is restricted to include exclusively high-skilled migrants, it can result in a dynamic immigration surplus as high as a 3.6 percent (4.3%) increase in the steady-state consumption equivalent for a representative household in the destination country, compared to as low as a 0.33 percent (0.55%) increase in the static case. In contrast, if immigration is restricted exclusively to unskilled migrants, the dynamic immigration surplus becomes negative—as low as −3.5 percent (−4.0%) of the consumption equivalent of the representative household, as opposed to a positive level of 0.18 percent (0.04%) in the static case. Each skill group in the destination country gains less than the representative household when immigration is exclusively by the same skill group, but the change affects more heavily the unskilled group in both the dynamic and static cases. The immigrant surplus magnitudes of
|Population (total or by degree attainment)||Total||Year of Entry|
|2000 or later|
|High School and Above||24,477||71.7||7,607||71.1|
|Bachelor’s and Above||9,943||29.1||3,491||32.9|
|Master’s and Above||3,826||11.2||1,450||13.7|
NOTE: In 2012, the percentages of the native population that had attained high school and above, bachelor’s and above, master’s and above, and doctorate were 90.9%, 31.3%, 11.1%, and 1.5%, respectively.
SOURCE: U.S. Census Bureau, Current Population Survey, Annual Social and Economic. Supplement, 2012, Table 2.5. Available: https://www.census.gov/data/tables/2012/demo/foreign-born/cps-2012.html [May 2017].
|Population (total or by degree attainment)||Total Foreign-Born||Natives|
|High School and Above||24,477||71.7||154,826||90.9|
|Bachelor’s and Above||9,943||29.1||53,348||31.3|
|Master’s and Above||3,826||11.2||18,904||11.1|
SOURCE: U.S. Census Bureau, Current Population Survey, Annual Social and Economic Supplement, 2012, Table 2.5. Available: https://www.census.gov/data/tables/2012/demo/foreign-born/cps-2012.html [May 2017].
|World Region of Birth|
|Population by Birth Origin and Years in U.S. (for foreign-born)||2010||2002||1996||1990||1980||1970||1960||1950||1940|
|1-5 years in U.S.||12.53||12.32||11.73||11.65||11.25||10.36||9.95||NA||8.90|
|1-5 years in U.S., Asian||14.07||14.73||13.13||12.90||12.50||13.46||12.08||NA||10.44|
|1-5 years in U.S., European||14.05||14.61||14.65||13.63||12.11||10.35||10.32||NA||8.95|
|1-5 years in U.S., Hispanic||10.33||9.84||8.41||9.14||8.26||8.40||7.23||NA||7.25|
|1-5 years in U.S., Mexican||9.56||8.53||7.52||7.83||6.33||5.93||4.58||NA||6.06|
SOURCE: Smith (2014b, Table 4.3); based on Decennial Census data, 1940-1990, and March Current Population Survey for 1996-2000.
the opposite effects become even larger when the Drinkwater et al. (2003) simulations allow for a complementary relation between skilled labor and physical capital (the corresponding percentage changes are shown by the parenthetical figures above).
In the Ehrlich and Kim (2015) benchmark model, the immigration surplus generated by the endogenous increase in immigration is also found to be higher for the average household of natives in the destination country, but it reflects opposite net gains to skilled and unskilled native households (thus generating distributional effects similar to those derived in Drinkwater et al., 2007). Skilled households gain less than the average household and unskilled households gain more. Specifically, in the Ehrlich and Kim model, the percentage change in the full income per capita (FIPC) experienced by natives in the destination country following a SBTS is measured using two scenarios: (a) when the skill composition of immigrants at the destination country is free to adjust following the SBTS, and (b) when the skill composition is confined by an immigration policy restricting it to remain fixed at its initial equilibrium steady state. The percentage difference in the natives’ FIPC in scenario (a) versus (b) accounts for the net benefits from the unrestricted migration scenario relative to the restricted migration scenario, which in this model is the immigration surplus. Ehrlich and Kim estimated this immigration surplus to be 1.48 percent of the natives’ FIPC at the end of a 15-generations period, which is equivalent to a modest 0.003 percentage point gain in the average annual growth rate of FIPC over that period. This long period is selected for illustration as it approximates the period over which the economy approaches a new steady state. Note, however, that the rise in the FIPC under these conditions, as well as under the conditions of the simulations reported below, already appears after the first generation following the SBTS and continues over the entire transition phase leading to a new steady state.31
Ehrlich and Kim (2015) also simulated the immigration surplus under two alternative scenarios: (a) when the skill composition of immigrants is freely determined in an initial equilibrium steady state at the destination country, and (b) when the destination country disallows altogether the migration of either skilled or unskilled migrants. Here, if skilled migration is disallowed, the difference in FIPC between the unrestricted and restricted
31 Tables 3 and 4 in Ehrlich and Kim (2015) illustrate the magnitudes of the immigration surplus over 5 and 10 generations, as well as the 15-generation period. It is interesting that the percentage changes in the natives’ initial FIPC in the generations immediately following the SBTS are estimated to be slightly higher than those after 15 years. But the immigration surplus thus measured is “partial”: it captures the net benefits from additional unrestricted immigration following an SBTS, starting from positive values, rather than zero values of skilled and unskilled migration following the SBTS. The latter are captured by the estimates based on the alternative scenarios in the following paragraphs.
immigration scenarios is significantly more pronounced in the benchmark case, where it amounts to a cumulative gain in the natives’ initial FIPC of 79.8 percent after 15 generations, equivalent to a 0.376 percentage point gain in the average annual growth rate of FIPC in the unrestricted immigration scenario relative to the restricted immigration scenario over this period. However, the opposite outcome occurs when the destination country disallows any unskilled migration. In this case, natives experience a gain of 33.0 percent in FIPC in the restricted immigration scenario relative to the unrestricted scenario (i.e., an immigration surplus of −33.0%) after a period of 15 generations, or a change in the average annual growth rate of FIPC of −0.058 percentage points per annum over this period.
Larger estimates of the immigration surplus are computed in Ehrlich and Kim’s (2015) extended model, which allows for positive complementarities or “diversity effects” in knowledge production across natives and immigrants of the same skill group. For example, the immigration surplus in the case where all migration is disallowed in the destination country amounts to a persistent gain of 0.593 percent in the annual growth rate of FIPC after a 15-generations period.
Bear in mind that all the immigration surplus estimates reviewed in this section are theoretical and subject to limiting assumptions. They do indicate, however, that the long-term dynamic immigration surplus could far exceed its estimates based on static models, both on the up side and the down side. This realization opens up opportunities for immigration policies that could enhance the benefits of migration to both destination and source countries.
Immigrants, like their native-born counterparts, also contribute to the economy in ways that are not, or at least not fully, captured by market-based economic statistics such as GDP and employment rates.32 Much labor used in the provision of health, child, and elder care for family members or friends, for example, goes undetected in official statistics, as a substantial
32Becker (1991) observed that extensive, economically valuable work—from care activities to home maintenance—goes on inside the family but is largely unrecognized in conventional measures of economic output. The National Research Council (2005) report Beyond the Market explores in great detail methods for accounting for nonmarket economic activities in the areas of household production, investment in education, investment in health, selected government and nonprofit sector activities, and environmental assets and services.
amount of that valued activity is nonmarket in nature.33 Immigrant women play a particularly important role in housework and child care, whether done for their own families, working in informal arrangements (which may be market or nonmarket based) for others, or in formal employment. Female participation rates of immigrants in market work are on average lower than for native-born females, indicating that they may be engaged in more nonmarket production. Also, immigrants more often live “doubled up” or in extended family situations, raising the possibility of greater nonmarket production or a shifting of who is doing it (e.g., grandma watches the kids while mom works) relative to nonimmigrant households where child care and other services are more likely to be purchased in the market.
Because home-produced services do not involve market transactions, some of the economic benefits of family-based immigration policies may be underestimated or overlooked by conventional economic statistics. However, the American Time Use Survey has allowed researchers to begin examining immigrant-native differences in nonmarket work. Ribar (2012) provided a broad overview of immigrant time use, using data from this survey. His study confirmed that immigrant women in his dataset devoted more time to household production (caregiving and housework) than native-born women. He also found that they spent more time sleeping. Immigrant men spent more time in market work and less time performing housework, community activities, and leisure than did native-born men.34Vargas (2016) found that results vary considerably by country of origin, but over time, immigrant time use becomes more like that of natives.
Alesina and Giuliano (2007) examined time use patterns as well and found that, relative to population averages, strong family ties35 are associated with a higher number of hours spent in home production and lower labor force participation of women, as well as less reliance on the government for social insurance. Abrams (2013) discussed easy to overlook (and difficult to measure) benefits of family-based immigration policy and assessed the role of immigrants who may not participate in wage-paying labor but who nonetheless contribute in economically valuable ways by providing “unpaid care work in the homes of relatives who are participat-
33 A common illustrative example is an individual who marries his/her housekeeper. If the housekeeper’s wife/husband continues to clean the house, GDP decreases, even though the amount of economic activity remains the same.
34 Some of the redistribution of time for immigrants relative to natives may be attributable to the need of the former to engage in assimilation-related activities that are costly and take time; Hamermesh and Trejo (2013) explored this issue.
35 Strength of family ties is scored based on responses by individuals across 81 countries from the World Value Survey regarding “the role of the family and the love and respect that children are expected to have for their parents.”
ing in market labor, sometimes even making such market participation possible” (Abrams, 2013, p. 21).
Nonmarket activities in the sphere of home production are different from labor that takes place outside the household where immigrants are paid but their compensation is not reported through official channels. Low-skilled immigrants work in a range of sectors where their labor is more likely to be “off the books” and hence untaxed. Occupations for which this may be true (but not always) include house cleaning and babysitting services, home repair, landscaping, and many others.36
As noted by Bohn and Owens (2012), informal sector employment—defined in their analysis as paid work that would have been taxable if it had been reported to the tax authorities—is thought to be large and growing. Bohn and Lofstrom (2013) found self-employed, “likely unauthorized” men to be especially concentrated in a handful of industries and occupations—about 46 percent of this group worked in construction while another 17 percent worked in landscaping.37 Much unreported work, but not all, takes place in “markets” and shows up in GDP. Studies of employment arrangements estimate that over half of the unauthorized immigrants in the United States pay income and payroll taxes through employers withholding from their paychecks or by the immigrants filing tax returns (Congressional Budget Office, 2007).
Some work that takes place informally does so without employment protections, health insurance, Social Security, and other worker benefits. Unregulated work is often connected to immigration through the growth of ethnic economies and because some immigrants lack documentation to work legally in the United States. Across states and over time, there is a relationship between the sizes of informal economies and changing rules and processes for immigrants to attain legal status; enforcement is also a factor. Bernhardt et al. (2009)38 presented evidence from qualitative fieldwork and the 2008 Unregulated Work Survey about how unauthorized status can play out in the workplace and its correlation with higher rates of
36Haskins (2010) reviewed the literature examining reasons why tax evasion is prevalent, using analysis of Internal Revenue Service data plus qualitative interviews with Filipina nannies in the Washington, D.C., area.
37Bohn and Owens (2012) found that states with high concentrations of low-skilled male immigrants have higher levels of informal employment in the landscaping industry. Measuring informal work is difficult and requires case studies and specialized surveys (e.g., the National Day Labor Survey). Bohn and Owens used a residual method to estimate informal work in landscaping and other occupations. For construction, the residual was based on a total employment estimate based on “unofficial data”—for example, based on building permits and other information, minus a count of documented workers captured in an “official” source such as the BLS’s Quarterly Census of Employment and Wages for residential construction.
38 See http://www.nelp.org/content/uploads/2015/03/BrokenLawsReport2009.pdf [November 2016].
unemployment and labor law violations, including paying below-minimum wages.39 In addition to the potential for worker abuse, injury, and exploitation, another secondary economic effect of informal, unreported work is that employers may prefer immigrants to competing native workers when only the immigrants can be employed under arrangements in which payroll taxes are ignored and labor regulations are not observed.
Even in the context of formal labor markets, there is some evidence that immigrants are more likely to hold jobs characterized by poor working conditions or high risk than are natives. Based on individual-level data from the 2003-2005 American Community Survey and from the BLS, Orrenius and Zavodny (2009) found that foreign-born workers were employed in more dangerous jobs than were U.S.-born workers, “partly due to differences in average characteristics, such as immigrants’ lower English-language ability and educational attainment” (Orrenius and Zavodny, 2009, p. 535).
Informal work arrangements also carry fiscal implications when wages are not taxed or if the amount of wages taxed is smaller than it should be. A study of Los Angeles County by Flaming et al. (2005) indicated the substantial role of informal workers in the local economy: 679,000 in 2004, or roughly 15 percent of the county’s labor force. The report estimated that the informal economy in Los Angeles County generated an $8.1 billion payroll in 2004, which translated into a $1 billion reduction in Social Security taxes that would have been paid by employers and workers if it were formal work.40Flaming et al. (2005) estimated that Medicare taxes paid by employers and workers were reduced by $236 million for that year; California State Disability Insurance payments paid by workers were reduced by $96 million; unemployment insurance payments paid by employers were reduced by $220 million; and Workers Compensation Insurance payments paid by employers were reduced by $513 million. These estimates illustrate that, since wage transactions in the informal sector are not always taxed, the fiscal impact is negative (relative to equivalent taxed work). Although not all informal work is performed by unauthorized immigrants, and a minority of unauthorized immigrants are engaged in off-the-books employment, legalization of unauthorized immigrants would likely result in a reduction of untaxed labor in the informal market.
The overall impact of the informal economy on jobs, production, taxpaying status, and fiscal consequences is not a thoroughly studied topic.
39 Surveying unauthorized workers and hard-to-sample groups (where there is no sampling frame) often requires innovative methods such as respondent-driven sampling, which also means the data are not necessarily representative.
40 A considerable amount of money—estimated to be in billions of dollars—is also paid into the Social Security system that is associated with faulty Social Security numbers or Individual Taxpayer Identification Numbers. A rapid growth in the Social Security Earning Suspense File affects Social Security Trust Fund balances and, in turn, program costs and fiscal projections.
However it does appear that state-level immigration laws can play a role in pushing people off the books. Bohn and Lofstrom (2013) addressed the employment effects of state legislation on employment outcomes of low-skilled, unauthorized workers. Analyzing the impact of the 2007 Legal Arizona Workers Act—which allows the state to suspend or revoke the business licenses of employers found to have knowingly hired unauthorized workers—they found a lower probability of wage and salary employment and a higher rate of self-employment among this group. The size of the gray/underground economy may have been put on a different course after the September 11, 2001, terrorist attacks with changed laws and enforcement protocols; it is now also more difficult to get Social Security numbers, which, for example, are needed to work in many jobs.
There are many are other nonmarket impacts created by immigration, sometimes negative but often positive. These issues are not dealt with in any detail in this report, but they are covered elsewhere: The impact of immigration on population health, crime (Castañeda et al., 2015; National Academies of Sciences, Engineering, and Medicine, 2015), and subjective well-being of individuals (Polgreen and Simpson, 2011) are just a few examples. Also, The New Americans (National Research Council, 1997, pp. 98-99) discussed how immigration contributes to population growth and congestion in destination countries, which places demands on the environment and infrastructure. However, that report also notes that immigration is primarily distributive, since an immigrant is leaving one place (relieving congestion) and moving to another (adding to congestion).
The economic impact of immigration extends well beyond the wage and employment interactions reviewed in Chapter 5. With so much focus in the literature on the labor market (and much of this, on the short run), other critical issues—such as the role of immigrants in contributing to aggregate demand, in affecting prices faced by consumers, or as catalysts of long-run economic growth—are sometimes overlooked by researchers and in the policy debates. In fact, by construction, many of the labor market analyses reviewed in Chapter 5 net out the kinds of economic effects that have been discussed in this chapter, many of which are positive, in order to identify direct, short-run wage and employment impacts.
The contributions of immigrants to the labor force reduce the prices of some goods and services, which benefits consumers in a range of sectors including child care, food preparation, house cleaning and repair, and construction. Moreover, new arrivals and their descendants also provide a major source of demand in sectors such as housing, benefiting residential real estate markets. To the extent that immigrants flow disproportionately
to where wages are rising and local labor demand is strongest, they help equalize wage growth geographically, making labor markets more efficient and lowering slack.
Immigration also contributes to the nation’s economic growth. Most obviously, immigration supplies workers, which increases GDP and has helped the United States avoid the fate of stagnant economies created by purely demographic forces—in particular, an aging (and, in the case of Japan, a shrinking) workforce. Perhaps even more important than the contribution to labor supply is the infusion by high-skilled immigration of human capital that has boosted the nation’s capacity for innovation and technological change. The contribution of immigrants to human and physical capital formation, entrepreneurship, and innovation are essential to long-run sustained economic growth. Innovation carried out by immigrants also has the potential to increase the productivity of natives, very likely raising economic growth per capita. In short, the prospects for long-run economic growth in the United States would be considerably dimmed without the contributions of high-skilled immigrants.
In Part III of this report (Chapters 7 through 10), the panel turns to another key component of immigration that must be considered alongside labor market and other economic impacts in order for policy assessment to be comprehensive: the fiscal impact created by the new arrivals.
The basic mechanism through which endogenous growth occurs can be illustrated using human-capital-based models. The perception of human capital or human knowledge as the economy’s engine of growth stems from a wide agreement in economics that knowledge is the major force affecting productivity growth and the only reproducible economic asset that is not subject to diminishing returns (paraphrasing Clark, 1923). This thesis can be supported by examining the critical role played by the accumulated stock of past knowledge in transmitting and facilitating the acquisition of new knowledge, which offsets any diminishing returns from investment in the latter. In Lucas (1988), this process operates implicitly because the investor (or productive enterprise) is infinitely lived. In the Becker et al. (1990) dynastic model and in the Ehrlich and Lui (1991) overlapping generations model, knowledge formation occurs through the transfer of knowledge from finitely lived young parents to offspring via the following human capital production function:
(1) Ht +1= A (He + Ht) (ht)α
where Ht and Ht + 1 measure the human capital acquired by the parent generation (t) and the offspring (t + 1), He denotes an endowed productive capacity measured in units of human capital; A denotes the technology of knowledge transmission from the parent generation to that of the offspring; and ht is the share of total productive capacity, (He + Ht), or “full income” that young parents devote to promote the acquisition of human capital by their kids. (For simplicity, total productive capacity can be assumed to equal full income if human capital is taken to be the only asset underlying the production of goods and has a neutral effect on the productivity of labor and capital, and thus on the capital/labor ratio.) While the investment share of full income ht can be subject to diminishing returns (if α < 1) with no loss of generality, the stock of new human capital, Ht + 1, is assumed to be linearly related to old human capital, Ht, consistent with Clark’s assumption that human capital as a productive asset is not subject to diminishing returns. The contribution, ht, of the parent generation to the acquisition of future knowledge, Ht + 1, is thus seen as the sine qua non for innovation and technological advance.
The assumed production function illustrates the role of intergenerational spillover effects in achieving the growth of innovative human capital. Absent any link between the generations, human capital would be essentially stagnant. But, by this formulation, whether innovative production capacity can actually grow over time crucially depends on the size of investment in new knowledge capital chosen by generation t. This can be illustrated as follows: if α = 1 and the value of ht is assumed to be a constant fraction of total production capacity h* that can generate a continuous growth in future production capacity, then by equation (1), the growth evolution equation would be:
(2) (He + Ht + 1)/(He + Ht) ≡ (1 + gt) = Ah* + [He/(He + Ht)]
which implies that if t approaches an infinite value, the last term in equation (2) will disappear and the growth rate of full income will be given by the term Ah*. This term indicates that a steady state of continuous growth in total productive capacity, i.e., g > 0 in equation (2), can be attained only if investment in human capital, h* reaches a threshold level h* > 1/A. By contrast, a value of h* ≤ 1/A can be shown to yield a stagnant equilibrium.41
The equilibrium steady state of long-term growth, g* > 0, is thus essentially a function of the optimal investment parents choose to make in the human capital of their children, h*. The conditions that determine this
41 The solution of the difference equation (1) is given by: He + Ht = (Ah*)t[H(0) + He] + [(Ah*)t – 1]He/(Ah* – 1). Thus, growth can occur if and only if Ah* > 1. If Ah* < 0, e.g., He + Ht = He/(1 – Ah*).
level are a function not just of the technology of knowledge production and transfer but also of the altruistic preferences of parents and the relative costs motivating them to choose between quantity and quality of children and their own consumption, as well as the financing constraints limiting their ability to invest.
This page intentionally left blank.