NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Research Council (US) Committee on Policy Implications of International Graduate Students and Postdoctoral Scholars in the United States. Policy Implications of International Graduate Students and Postdoctoral Scholars in the United States. Washington (DC): National Academies Press (US); 2005.

Cover of Policy Implications of International Graduate Students and Postdoctoral Scholars in the United States

Policy Implications of International Graduate Students and Postdoctoral Scholars in the United States.

Show details

4Strategies for Improving Policy Decisions

Making sensible policy decisions about the flow of international students and scholars into the United States requires information about its benefits and costs—a sudden stop would be the most serious cost—and about how alternative policies can influence the flow. To the extent that the benefits and costs themselves depend on social and economic circumstances and policies, it is important to understand how the country can enhance the benefits relative to the costs of any given flow.

The benefits to the United States from the inflow of talented students and scholars are clear. Having access to a worldwide pool of talent leads to a higher-quality science and engineering (S&E) workforce than if the country had access only to domestic talent. The flow of international students and scholars also allows the United States to conduct research and education at lower cost than if the country had to rely exclusively on domestic talent. In addition, international students and scholars can help to form international research collaborations and to foster international understanding.

The costs of the flow are indirect. In the job market, a large supply of students and researchers depresses salaries and job opportunities and thus lowers the incentive for domestic students to enter S&E.1 If international students and scholars leave the United States, there is the risk that they will work in businesses that compete with those in the United States or use their knowledge in ways inimical to US security. A related but different risk is that international tensions or changes in world conditions will greatly reduce the inflow of overseas talent and diminish our educational and research leadership.

Our knowledge about the flow of international S&E students and scholars to the United States, although limited,2 allows us to predict with some certainty that in the extreme case a complete cutoff of the flow would create major problems for the US scientific and technologic enterprise. The impact would be felt rapidly by university graduate programs and by researchers who depend on graduate students and postdoctoral scholars. There would be a slower cumulative effect on hiring in industrial, government, and academic sectors.

It is harder to predict how modest changes would affect the scientific workforce, let alone the national well-being, or how larger changes would affect long-term outcomes. Universities might respond by placing more emphasis on the education and mentoring of domestic students. The country might, for example, respond to a drop in the number of international graduate students working as research assistants in laboratories by raising the incentives for US students to take such jobs, by recruiting more international postdoctoral scholars or immigrant scientists and engineers, or eventually by reducing the reliance of research laboratories on graduate students and postdoctoral scholars. Another possible response to increased wages would be to invest in labor-saving research technologies, such as high-throughput molecular biology equipment. With respect to industry, if the United States attracted fewer of the world's most talented scientists and engineers, US firms might shift more R&D activities overseas, maintaining their competitive edge through increased offshoring.


One reason for the uncertainties mentioned above is that the country lacks adequate data for measuring the international flow and career paths of foreign-trained S&E students and postdoctoral scholars (see Box 4-1). The need to improve data on immigration and emigration has been known for at least 20 years.3 As the flow of scientists and engineers into and out of the United States increased during the 1990s, studies reiterated and expanded on this urgent need.4

Box Icon

BOX 4-1

Improving Data Systems for Decision Making. Specific steps for improving data systems for the US S&E workforce are described in a recent report. The following high priority needs are relevant to international (more...)

An understanding of workforce trends is impossible without more frequent counts and timely publication of scientist and engineer populations and of the places in which they have been trained. For example, data on the S&E workforce from the Bureau of Labor Statistics establishment surveys include citizenship but not degree level, but the US Census Current Population Survey does have degree data. The counts of foreign trained scientists and engineers working in the United States produced by the National Science Foundational (NSF) throughout the 1990s were estimates based on the 1990 census; by the end of the 1990s, the foreign born share of scientists and engineers was substantially underestimated.5

A particular weakness in data concerns the postdoctoral population, in which international researchers make up over half the academic workforce and from which US S&E researchers recruit globally for the laboratory workers that they need. The NSF Graduate Student and Postdoctorates Survey does not collect demographic information, and the Survey of Doctoral Recipients does not include scholars who earned their PhDs outside the United States. NSF's 1997 Survey of Doctoral Recipients had a special section on postdoctoral scholars, but the next similar section is scheduled to appear 10 years later, in 2007. The 1995 retrospective career-history questions that allowed analysis of the duration of the postdoctoral experience are not yet scheduled to be updated. Policies crafted for postdoctoral scholars in the interim are therefore based on outdated assumptions. In the 2003 2004 academic year, Sigma Xi launched a National Postdoctoral Survey.6 Findings from that survey provide information on postdoctoral scholars previously not available, but it, too, has limitations. The response rate was less than 40 percent, and it is not clear whether the survey will be repeated in successive years to provide longitudinal data.

The inadequacy of data on international graduate students and postdoctoral scholars limits our understanding of the composition of the S&E workforce and of how that workforce might respond to economic or political changes. Moreover, the lack of timeliness and coverage of data on US-trained and internationally-trained scientists and engineers hinders the examination of trends and relationships between student flows, enrollments, economic cycles, and other factors. Congress and administrative agencies need better data and more analysis to craft better policies.


The United States needs a new system of data collection to track international student flows and to understand the dynamics and effects of shifting sources of talent. Pulling together existing country-specific data is challenging. A thorough analysis of 25 surveys on the labor market for students and early-career researchers carried out in 21 countries revealed that diverse collection methods often preclude comparison of data.7 By and large, surveys have been developed to serve national needs and fail to capture such characteristics as international experience and mobility.8 Clearly, more rigorous and normalized data systems for tracking international graduate students and postdoctoral scholars are required. The United States should partner with other nations to create a truly global system.

One possible model is the US-based international balance-of-trade account for commodities. The US Department of Commerce maintains TradeStat Express, a Web site with extensive data on US merchandise trade that allows the tracing of comprehensive global patterns.9 Information is compiled from forms and automated reports required by US law to be filed with the US Customs Service and then transmitted to the Census Bureau for virtually all shipments leaving (exports) or entering (imports) the United States (see Box 4-2). To ensure comparability of collected data, the Harmonized Commodity Description and Coding System has been adopted by most countries.

Box Icon

BOX 4-2

Collection of US Foreign Commerce and Trade Statistics. Information on exports of merchandise from the United States to all countries, except Canada, is compiled from copies of Shipper's Export Declarations (SEDs) (more...)

Can such an international tracking system be conceptualized for students and scholars? The Institute for International Education, in collaboration with the British Council and IDP Education in Australia, produced The Atlas of Student Mobility (see Box 4-3), which gives a first glimpse of international student flows. But the picture will need to be refined, particularly with respect to mobility of students in different disciplines and degree levels, mobility of postdoctoral scholars, career opportunities in different countries, and impact of mobility on workforce productivity. To obtain compatible international data on those issues, systems to capture data on student and postdoctoral-scholar mobility will need to be implemented and harmonized between countries.

Box Icon

BOX 4-3

Project Atlas. One recent project, The Atlas of Student Mobility, illustrates how data on international students can be effectively collated and presented. The atlas synthesizes information from 21 main destination (more...)

The atlas has initiated a transnational collaboration, Project Atlas, and has led to agreements on several definitions and constructs. For example, it was agreed that international student should be defined as “a person who physically moves from his or her place of residence for the purposes of study” with nonimmigrant, nonpermanent residence status.10 Similar harmonization efforts for R&D data are under way among Organization for Economic Co-operation and Development (OECD) countries (see Box 4-4) to build on the 1993 Frascati Manual to function as a “basic international source of methodology for collecting and using R&D statistics.” The manual includes a section titled “Measurement of Personnel Devoted to R&D”.11 OECD plans to develop a definition of postdoctoral scholars and to increase the development of data by fields of science.12

Box Icon

BOX 4-4

International Tracking of Doctorates. Recently, the Organisation for Economic Co-operation and Development (OECD), Eurostat, and UNESCO's Institute for Statistics (UIS), sponsored by the US National Science Foundation, (more...)


Data problems aside, the most important reason for our uncertainty about benefits and costs of international student flow is the paucity of analysis of the major behavioral and market factors that shape student and scholar decisions to come to the United States for training and decisions to remain here thereafter. Similarly, little is known about the interaction between the flow of international talent into the United States and the decisions of US citizens or permanent residents to choose S&E careers. Even if a frequent, comprehensive census of the S&E population were done, more is needed than the numbers of students and graduates if we are to understand the factors that influence decisions and how policies can affect outcomes.

The difficulty lies not in choosing the factors at work in the supply of talent but in measuring the effects of the factors. For example, the desire of international students to study in the United States depends on such factors as how the United States education system compares with others; the amount of support available in grants, teaching assistantships, and research assistantships; the probability that studying in the US will lead to permanent residence and possibly citizenship or employment with US firms; and a host of amorphous factors, such as the perceived safety of the country and one's attitude toward the United States. We do not know the degree to which the flow of international students is correlated with those factors, because there has been relatively little analysis of student decisions. Despite the general concern about the decline in the attractiveness of the United States as a destination for students and scholars, the present committee has found no definitive studies on the quantitative effects of policy changes, such as increased security measures or increased fees, on the entry and exit of graduate students and postdoctoral scholars.

For all their importance, moreover, the decisions of foreign-born scientists and engineers with degrees from either US or overseas institutions to immigrate to the United States have not been analyzed in depth. We witnessed a huge influx of immigrant scientists and engineers in the 1990s and saw that the stay rate of foreign-born PhDs in S&E rose during roughly the same period. Were the influx and stay rate primarily responses to the booming US job market of the 1990s, or did they reflect longer-term developments, such as the increased production of S&E graduates overseas, which created a larger global supply of scientists and engineers?

Another analytic weakness is in the understanding of why US students decide to choose S&E careers. Relatively low compensation and long duration of postdoctoral appointments may be deterrents to an S&E career, but these conditions are prevalent only in some fields. The abundance of foreign-trained graduate students undoubtedly contributes to low postdoctoral compensation at least in universities, but we do not know the relative importance of postdoctoral pay in influencing decisions compared with research-career prospects, the length of graduate study, and the attractiveness of options outside S&E research. Nor are there good estimates of how much postdoctoral stipends might rise if the supply of postdoctoral scholars diminished. Answering such questions requires regularly gathered information about annual changes in pay, benefits, and career aspirations of postdoctoral scholars and about opportunities abroad, and it requires creative analysis of the information.

The primary factor affecting students' decisions to major in S&E fields and whether they pursue graduate study may not be postdoctoral compensation but rather their perception of whether S&E degrees will lead to satisfying research careers. Available studies indicate that students are sensitive to labor-market conditions,13 but where students and scholars learn about the job market, the alternatives they consider, and the effects of their schooling and mentoring experiences on career choices are not well understood. Statistical investigations of the supply behavior of domestic and international students, of the demand behavior and wage responsiveness of US employers to changes in the supply of foreign talent, and of the responsiveness of talented US students to changes in incentives are needed to create an evidence-based policy. Although the charge to this committee does not include an examination of why domestic students enter or avoid S&E, any assessment of policies regarding international students requires knowledge of the decision making of US citizens and permanent residents. The effect of any loss of international talent will be very different if the supply of domestic talent is highly responsive to incentives—monetary or instructional—from what it will be if the supply is barely responsive to incentives.


Policy decisions that affect graduate students also affect supplies of doctorates and postdoctoral scholars several years in the future,14 so such policies should be flexible enough to allow for changes in the economy and in demand for S&E workers. But predicting workforce trends is highly problematic.15 A few years ago, the country was warned of a shortage of information-technology (IT) scientists, so Congress raised the ceiling on H-1b visas to allow a larger flow of qualified IT workers. Not only was the decision based on faulty information, it also was quickly outdated by changes in the market. Today, in part because of the abundance of programmers and computer experts trained in other countries, US enrollments in computer science at the bachelor's level are down sharply. Unemployment rates for computer programmers are high, electrical and electronic engineers face a more difficult labor market than most other professionals, and talk of a shortage has diminished considerably.

Over the years, various institutions and agencies have issued forecasts of shortages or surpluses of scientists and engineers. The accuracy of those forecasts has been weak for three reasons: labor-market researchers lack timely and comprehensive data; many forecasts are issued by groups with a vested interest in the outcome, that is, those who are predisposed to the belief that there should be more or fewer participants;16 and labor markets can be rapidly affected by various exogenous variables, such as economic expansion or recession, federal budget priorities, war, and immigration policies that are hard to forecast.17 In the present context, for instance, a slowdown in the growth of federally sponsored R&D could shrink the nation's demand for scientists and engineers, depressing the job market and reducing both international- and domestic-student enrollments. A commitment to double NSF's budget over some period, in contrast, would probably have the opposite effect.

Given those problems, this study has not attempted to forecast labor-market conditions for scientists or engineers. Instead, it has discussed some of the short term variables that influence the flow of international scientists (Chapter 2) and the global context in which more nations strengthen their S&E capacity and compete for the best students (Chapter 3).


Input of two kinds is required to improve policy responses to the flow of international students and researchers to the United States. The first kind of input is more and better data specifically designed to answer key analytic questions about the function of the labor market for the S&E workforce. The second kind is the results of rigorous labor-market analyses that can be used to help understand the nation's needs for S&E workers, address the repeated claims of shortages of scientists and engineers, develop strategies that attract high ability US students to S&E, and assess the costs and benefits of such strategies.18

Decision-making does not come to a halt in the absence of adequate data or in the absence of adequate modeling of processes. It proceeds with inadequate data and with uncertainty about the effects of policies. In such a situation, it is useful to imagine a series of scenarios and their likely consequences and to at least think through if not simulate counterfactual scenarios—scenarios that involve conditions different from those we see today—and lay out alternatives to decisions under consideration. One can imagine extremely favorable and unfavorable scenarios for the impact of foreign graduate students and postdoctoral scholars on the US S&E enterprise. Figure 4-1 illustrates the inputs and outputs of the US higher-education system for international students. Input filters that may reduce their participation include admission decisions that favor domestic students,19 visa fees, and security screens. Factors that may favor their participation are networks of former students who have made the United States their home—the diaspora effect—in addition to a number of pull factors, discussed in Chapter 3. Once students complete their degrees, they may stay, go home and collaborate, or leave and compete.

FIGURE 4-1. Outcomes model illustrating the inputs and flows of inernational students (IS), domestic students (US), and postdoctoral students (PD).


Outcomes model illustrating the inputs and flows of inernational students (IS), domestic students (US), and postdoctoral students (PD).

In an unfavorable scenario, large numbers of international students and postdoctoral scholars would discourage US students from participation, and a sudden cutoff of the international flow would in the short term leave the United States with a substantial deficit of researchers and technical personnel. Once home, the returnees would contribute to industries that compete with US based industries or take jobs that are offshored from US multinationals, shifting the locus of scientific and technologic leadership overseas while the United States struggled to replenish its S&E workforce by developing domestic talent. We are nowhere near that extreme, but it may be useful to keep it in mind as a worst-case scenario.

In a favorable scenario, highly skilled international graduate students and postdoctoral scholars enter S&E positions. Those who stay in the United States become permanent residents and citizens; those who eventually return home enter valuable collaborations with US colleagues and become informal ambassadors who communicate the democratic values of scientific research and of the United States. They order US products for their businesses and provide expertise for local divisions of US industries. At the same time, the United States might use graduate-student fellowships, higher postdoctoral pay, and other incentives to increase the flow of the best US talent into S&E. We are not at this extreme either, but it may be useful to keep it in mind as a best case to work toward.


Input of two kinds is required to improve policy responses to the flow of foreign-born students and researchers to the United States. The first kind of input is more and better data specifically designed to answer key analytic questions about the function of the labor market for the S&E workforce. The second kind is the results of rigorous labor-market analyses that can be used to help understand the nation's needs for S&E workers, address the repeated claims of shortages of scientists and engineers, develop strategies that attract high ability US students to S&E, and assess the costs and benefits of such strategies.20

What can be said is this: there are both benefits and costs to having international graduate students and postdoctoral scholars in the United States. The benefits include increasing the S&E talent pool in the United States, enhancing and diversifying the academic community, lowering the cost of doing research, and enhanced international research collaborations. In short, talented graduate students and postdoctoral scholars constitute a critical input for our knowledge-driven economy. At the same time, having such an open supply of talent has significant costs. It affects job opportunities for all students. Restrictions on international travel or exchange can rapidly affect US research capabilities. International collaborations may lead to enhanced international competition.

At present, the strategy of the United States is to draw heavily on and profit from the international talent pool. However, increased security regulations are restricting entry of prospective international students and scholars and restricting the fields in which they may study. Other nations are fortifying their S&E infrastructure and competing for the best students and scholars. It is in this context that the United States needs to craft policies to maintain its current quality and effectiveness in S&E, including encouraging the interest of domestic S&E students, at the same time that it minimizes the barriers to mobility for international students. Given the lack of control over exogenous events, policies should be crafted to ensure that S&E institutions and the labor force develop enhanced flexibility to respond quickly to changing conditions.



See for example, George Borjas. 2004. Do Foreign Students Crowd out Native Students from Graduate Programs? (Working Paper Number 10349.) Cambridge, MA: National Bureau of Economic Research; Mark Regets. 2001. Research and Policy Issues in High-Skilled International Migration (DP No. 366). Bonn: IZA.


See Grant Black and Paula Stephan. 2005. “The importance of foreign PhD students to US science.” In: Science and the University, eds. R. Ehrenberg and P. Stephan. Madison, WI: Universty of Wisconsin Press (forthcoming); Jeffrey Mervis. 2004. “Many origins, one destination.” Science 304:1277. This special section reviews experiences of foreign born scientists and engineers working in the United States. Mervis writes, “For all its importance, the relationship between the domestic and foreign born scientific workforce remains an understudied topic.”


National Research Council. 1985. Immigration Statistics: A Story of Neglect. Washington, DC: National Academy Press.


National Research Council. 1996. Statistics on U.S. Immigration: An Assessment of Data Needs for Future Research. Washington, DC: National Academy Press; National Research Council. 1999. Measuring the Science and Engineering Enterprise: Priorities for the Division of Science Resources Studies. Washington, DC: National Academy Press. This study focused on the Science Resource Statistics division of the National Science Foundation and urged sufficient funding to “continue and expand significantly its data collection and analysis.”


The issue of timeliness is addressed by the NSF as follows: “Because the NSF's demographic data collection system cannot refresh its sample of individuals with S&E degrees from foreign institutions (as opposed to foreign born individuals with a new US degree, who are sampled) more than once per decade, counts of foreign born scientists and engineers are likely to be underestimates.” National Science Board. 2004. Science and Engineering Indicators 2004 (NSB 04-1). Arlington, VA: National Science Foundation, p. 3-33.


Sigma Xi contacted 22,178 postdoctoral scholars at 46 institutions, including 18 of the 20 largest academic employers of postdoctoral scholars and NIH. 8,392 (38 percent) responded; 6,775 (31 percent) made it all the way to the end of the 100-question survey. Response rate was increased substantially when a local postdoctoral association was involved (87 percent) and decreased when institutional review boards did not allow Sigma Xi to send multiple reminders (17 percent).


Isabelle Recotillet. 2003. Availability and Characteristics of Surveys on the Destination of Doctorate Recipients in OECD Countries. Statistical Analysis of Science, Technology and Industry Working Paper 2003/9. Paris: OECD. Available at .


Martin Schaaper. 2004. “OECD methodology for tracking doctorate holders and measuring international mobility of HRST.” Paper prepared for the sixth Ibero-American and Inter-American Workshop on Science and Technology Indicators, Buenos Aires, September 15-17, 2004. Available at ; see also: Laudeline Auriol. 2004. “Why do we need indicators on careers of doctorate holders?” OECD Workshop on User Needs for Indicators on Careers of Doctorate Holders, Paris, September 27, 2004. Available at .


TradeStat Express home page .


Adria Gallup-Black. 2004. “International student mobility: Project Atlas.” International Higher Education 37:10-11.


OECD. 1999. Main Definitions and Conventions for the Measurement of Research and Development, 5th Edition Paris: OECD. Available at .


Recotillet. Ibid. 2003.


Richard Freeman. 1971. The Market for College-Trained Manpower. Cambridge, MA: Harvard University Press; Richard Freeman. 1980. “Employment opportunities: The doctorate manpower market.” Industrial and Labor Relations Review 33(2):185-196; and Richard Esterlin. 1993. “Prices and preferences in choice of career: The switch to business, 1972-1987.” Discussion Paper 2-1, Williams Project on the Economics of Higher Education, available at .


As in previous surveys, students receiving doctorates in 2003 took on the average 7.5 years to complete their degree requirements. National Opinion Research Center. 2005. Doctorate Recipients from United States Universities: Summary Report 2003. Chicago, IL: NORC. .


National Research Council. 2000. Forecasting Demand and Supply of Doctoral Scientists and Engineers: Report of a Workshop on Methodology. Washington, DC: National Academy Press.


See, for example, Michael Teitelbaum. 2004. “Do we need more scientists?” In: The U.S. Scientific and Technical Workforce: Improving Data for Decisionmaking, eds. T. K. Kelly et al. Arlington, VA: RAND Corporation, pp. 11-20; Donald Kennedy, Jim Austin, Kirstie Urquhart, and Crispin Taylor. 2004. Supply without demand. Science 303:1105; Daniel S. Greenberg, “What scientist shortage?” Washington Post, May 19, 2004, p. A23.


Paula E. Stephan. 2004. “What data do labor market researchers needs? A researcher's perspective.” In: The U.S. Scientific and Technical Workforce: Improving Data for Decisionmaking, eds. T. K. Kelly et al. Arlington, VA: RAND Corporation, p. 45.


Richard B. Freeman. 2004. “Data! Data! My kingdom for data! Data needs for analyzing the S&E job market.” In: The U.S. Scientific and Technical Workforce: Improving Data for Decisionmaking, eds. T. K. Kelly et al. Arlington, VA: RAND Corporation, p. 33.


Gregory Attiyeh and Richard Attiyeh. 1997. Testing for bias in graduate school admissions. Journal of Human Resources 32(3):524-48. The authors found a 5:1 bias in admissions favoring domestic students over international students. See discussion in Chapter 1.


Richard B. Freeman. 2004. “Data! Data! My kingdom for data! Data needs for analyzing the S&E job market.” In: The U.S. Scientific and Technical Workforce: Improving Data for Decisionmaking, eds. T. K. Kelly et al. Arlington, VA: RAND Corporation, p. 33.

Copyright © 2005, National Academy of Sciences.
Bookshelf ID: NBK37572


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (1.4M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...