Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Needs for Statistics on Immigration In this chapter we review why the United States needs accurate and timely statistics on immigration. As a panel of the National Research Council composed in part of academics with a strong empirical interest in migration studies, we could have easily provided lists of data needed for our own work, and those lists would be very long. We could also have attempted the difficult if not impossible task of identifying every person or organization with a real or perceived need for information, and their collective appetite for data would undoubtedly be enormous. Instead, we confine our attention in this chapter to the data that are necessary to the government's functions. We have grouped these data into three loose categories: (1) those that are required by law; (2) those necessary for the formulation, execution, and evaluation of public policy concerning the admission of aliens to the United States; and (3) those necessary for the management of government agencies, primarily the INS, whose mission is to apply the country's laws concerning the admission and residence of aliens. These categories are not mutually exclusive, but rather indicate the possible justifications for particular record-keeping processes and types of data. We conclude the chapter by discussing both the various obstacles to the collection of the necessary data and the ways in which they might be overcome. DATA REQUIRED BY LAW Aliens Other Than Refugees The population of the United States has its origins largely in immigration, and the present characteristics of the population still reflect these origins. One might expect, therefore, that the legislation controlling the flow of noncitizens across the country's borders and their residence in the country would require the regular collection and production of the information necessary to the policy makers and administrators responsible for the drafting, enactment, or implementation of that legislation. It is thus somewhat surprising to discover that the Immigration and Nationality Act (INA), the centerpiece of immigration law, contains virtually no direct references to the preparation of 26
27 statistics. The single provision on statistics, as such, is hidden away among the naturalization sections, specifically section 347 (8 USC 1485) The Attorney General is authorized and directed to prepare from the records in the custody of the Service a report upon those heretofore seeking citizenship to show by nationalities their relation to the numbers of aliens annually arriving and to the prevailing census populations of the foreign born, their economic, vocational and other classification, in statistical form, with analytical comment thereon and to prepare such report annually hereafter. A review of past performance (at least since 1952, when the present legislation was enacted), suggests that INS has met some of the requirements of this sentence through the publication, more or less annually, of a statistical yearbook and an annual report. Through 1978, these two reports were combined; since then, they have been issued separately, primarily because of severe difficulties in completing the assembly of the annual statistical data. For example, the statistical yearbook containing data for 1980 was issued in December 1983; that for 1981, in August 1984. The "analytical comment thereon" referred to in section 347, however, cannot be said to have been met adequately by the annual reports, which, in the words of INS's 1976 report, "depict Service efforts, programs, and achievements" but contain no analysis whatsoever of the status of immigrants or their relationship to the prevailing census populations of the foreign-born. The value of the annual report lies in providing readers (particularly policy makers in Congress) with a very abbreviated description of the diverse organizational units and functions within INS and in highlighting what INS considers to be its annual accomplishments. While this compilation includes some selected statistics concerning its operations, it seems to be a far cry from the intent of section 347. Given the paucity of statistics required by the INA, it is interesting to note that the immigration reform legislation that was almost, but not quite, passed by Congress late in 1984 (the Simpson-Mazzoli bill) would have mandated a manyfold increase in the number and scope of statistical and analytical reports. Those reports would have included, among other things: an analysis of the impact of implementing controls of the unlawful employment of aliens both on the employment, wages, and working conditions of legal U.S. workers and on the number of aliens entering the United States illegally; a report of the impact of general legal admissions over a 42-month period on the economy of the United States and on the employment of citizens and aliens; and a series of reports over a 5-year period on the alien population whose status would have been legalized, including their geographical origins, manner of entry, demographic characteristics, pattern of employment, and participation in social service programs. Although not specifically mandated for statistical purposes, the INA does contain language that results in the compilation of extensive record systems about people who, as immigrants, visitors, illegals, foreign students, etc., come into contact with the INS. Thus, section 290(a) of
28 the INA (8 USC; 1360) directs the establishment of a central index containing the names of all aliens admitted or excluded and "such other relevant information as the Attorney General shall require as an aid to enforcement of the immigration law." The law also requires that each person seeking a visa to enter the United States be registered through a State Department office abroad, as well as at the time of entry into the United States. The law further specifies that the following information be obtained: date and place of entry into the United States; activities in which the alien has been and intends to be engaged; length of time to be spent in the United States; and any police or criminal record. The management records compiled by INS are an essential source of information for the administration of its programs. They also have the potential to be important for planning new programs, evaluating old ones, and studying the impact of immigrants on the economy and social fabric, by providing a unique source of samples for longitudinal or other forms of surveys. To date, however, use of these records for any of these purposes has been very limited and their potential largely unrealized. Refugees The statistical and informational requirements contained in the Refugee Act of 1980 (P.L. 96-212) offer a startling contrast to the very limited requirements of the INA for data on immigrants. Whether reflecting a new awareness of the value of information or a legislative whim, the data demands of the Refugee Act of 1980 are much greater than those in the INA. Section 207(e), for example, requires that the President provide members of Congress with the following types of information: o A description of the nature of the world refugee situation; o A description of the number and allocation of the refugees to be admitted and an analysis of conditions within the countries from which they come; o An analysis of the anticipated social, economic, and demographic impact of their admission to the United States; and o A description of the proposed plans for their movement and resettlement. Apparently recognizing that properly presented data can contribute to the decision-making process, Section 207 also requires: "to the extent possible, information described in this subsection shall be provided [to members of Congress] at least two weeks in advance of discussions in person by designated representatives of the President." It may be noted in passing that the specified information requirements are much more policy related than management related. Such detailed specification of information needs within legislation ensures that the task is assigned a high priority and that, invariably, funds and competent staff become available for its accomplishment. Such certainly seems to be true in the case of refugee statistics produced by the Office of Refugee Resettlement (ORR) in the Department of Health and Human Services. Even during the current period of extreme budget
29 stringency, the statistical activities of this off ice have survived (albeit at less than an optimum level) . As a final note, the 1982 amendments to the Refugee Act (P. L. 97-363, Oh. 2, See . 412~2) ~ mandate that the director of the ORR "compile and maintain data on secondary migration of refugees within the United States and, by state of residence and net tonality, on the proportion of refugees receiving cash or medical assistance--e/c." The value of such information for policy evaluation, by indicating the direct costs of admitting refugees and which states bear them, is obvious. Less obvious but also important is the value of such information for understanding the process of social and economic adjustment of refugees. DATA NEEDED FOR PUBLIC POLICY As outlined in the preceding section, the relevant law does not require, except for refugees, extensive data on noncitizens coming to or residing in the United States. Therefore, in this section we start from first principles to ask what data should be required in order to formulate, execute, and evaluate public policy on immigration. At the most simplistic level, the issues of immigration policy can be boiled down to two questions: How many aliens should be admitted for long- or short-tenm residence, and who, in terms of origins, skills, and ties to the United States, should they be? Even a casual review of recent debates about immigration reform reveals that arguments revolve largely around the pocketbook issue of whether and why immigration is beneficial to the nation. Hence, the first thing policy makers need to know is who benefits from immigration and who is hurt by it. A common assumption is that immigrants gain and current residents, particularly those who hold low-paying jobs, lose; however, this assumption has so far been addressed only by rhetoric, not by data. In addition to identifying those groups who gain and those who lose, it is necessary to quantify the size of the benefits and costs. Economic benefits and costs (wages, unemployment, taxes) are perhaps the easiest to quantify, but the cost-benefit calculation should not be so narrowly focused. Immigration also has social, psychological, and political effects--for example, the work of foreign-born but U.S.-resident Nobel Prize laureates can be evaluated in terms of the economic value of their contribution, the spin-off benefits on their coresearchers, and the morale-boosting effects of their presence in the country. Analysis should not be limited solely to nationally aggregated data. Although immigration policy is set at the national level, immigrants live in local communities and contribute to public coffers through taxes at the federal, state, and local level; they may not, however, consume public services provided by these governments in amounts equal to their contributions. Hence, as in the case of military installations, which have benefits (national defense, contributions to local revenues) and costs (support from central government taxation, consumption of local services such as public education) that are both national and local, a federalism issue may arise. A time dimension also is involved: immigrants do not adapt instantly to a new culture, and their costs and
30 contributions to society can be expected to change as the duration of their stay in the country increases. The study of such changes over time, and the process of assimilation of immigrants in general, requires longitudinal rather than cross-sectional data. U.S. policy for the admission of aliens for long-term residence appears, on the basis of current legislation, to have three main objectives: admitting aliens for the purposes of reunifying families, admitting workers with particular skills that are in short supply domestically, and accepting a share of worldwide refugees consistent with foreign policy objectives, altruistic sentiments, and absorptive capacity. To evaluate whether these objectives are being met, empirical evidence is needed. With adequate data, one could study the extent to which families are united and stay united, and also judge whether immigrants do supply needed labor in areas (occupational and geographical) of shortage. Whether the country accepts its fair share of refugees is an issue that cannot be answered by national statistics alone. To answer this question, one would need as a minimum international data on the number of refugees in need of resettlement and on the refugee resettlement capacity of, and admissions by, other nations. Data on family structure and on labor market dynamics could therefore indicate whether the apparent objectives of current immigration policy are being met. Even with such data, however, policy makers would stil 1 need some way to set immigration levels. Few would dispute the proposition that some immigration is beneficial but that totally unrestricted entry would have unacceptable consequences. Clearly, however, such extremes are of little help in deciding how many aliens should be admitted for long-tenm residence each year. A more rational approach would use data to determine and compare the benefits and costs of different levels and mixes of immigration. The debate in Congress in 1984 on revisions to the present immigration law provides a c tear example of unmet data needs for policy. The requirements of the legislation, from amnesty provisions and employer sanctions for illegal aliens to visa waivers for temporary visitors, called for empirical evidence to set a common framework for the discussion. Regrettably, the record of the debate is replete with concerns over the lack of factual information. In a most unusual action, Congress noted its concerns by inc. luding the following language in its authorization of the fiscal 1985 appropriation for the Justice Department (U.S. Congress, 1984~: The Committee is deeply concerned about the unavailability of accurate and current statistical information on immigration matters. The Committee notes in this regard that INS . . . has ignored the statistical needs of Congress. . . . The Committee frequently has been frustrated in its attempts to obtain the statistics needed to assess the impact of legislation it is considering. Addressing any of the policy questions posed above is rendered far more difficult by the fact that immigrants in the United States do not consist only of those who are living here legally. Information on both
31 illegal and legal migrants is therefore required. Data on the latter are far easier to obtain than data on the former, since those here illegally have little incentive to cooperate with agencies that gather data. Two concluding comments are in order here. First, we recognize that immigration is an intensely emotional issue. Immigration policy is set through a political process that requires compromise--as does all policy--and the result may well be a policy that appears to lack rational underpinnings. The availability of better data is unlikely to remove emotional issues from national debates about immigration. Yet we would argue that data are essential to these debates by documenting to some extent what the effects of particular policy alternatives are likely to be. Second, we recognize that the history of the use of statistics in setting immigration policy is not a happy one. On too many occasions, those involved in making policy could not have drawn on statistical data because the necessary data did not exist. But even on those occasions when data were used, the results sometimes have been disappointing, if not disgraceful. One of the most infamous cases occurred in the 1920s. The first numerical ceilings on immigration from outside the Western Hemisphere were set in 1921 on the basis of the origins of the people enumerated in the 1910 census. In 1924, the basis of the national origins law was changed, not to the more recent 1920 census but back to the 1890 census. This change was made to ensure more immigrants from northwestern Europe and fewer from southeastern Europe and was based on misinterpretation of data collected from World War I draftees (Brigham, 1923~. Despite these reservations, however, there can be little disagreement with the notion that good-quality, reliable immigration data, objectively and competently analyzed, are likely to enlighten debates about policy and consequently to improve policy outcomes. DATA NEEDED FOR PROGRAM MANAGEMENT Effective program management requires timely information about various operational aspects of the program. Continuing series of such information enable managers, policy makers, and others to measure their accomplishments (whether they be positive or negative), as defined either by the policy makers or by the managers. Thus, if a program involves admitting numbers of immigrants specified by Congress, program information must be developed, by category of admission, to enforce the limit. Program managers may also want information on method and place of arrival, in order to ensure the most efficient placement of personnel, and on the time required to process immigrants, in order to measure productivity, develop staffing profiles, or justify funding requests. Another element of program management for which data are collected is evaluation, to find out whether a program is achieving its objectives. This requirement may be specified by Congress or initiated by a program agency itself, and its scope and form can vary widely. Although in many cases evaluation may draw on regular program information, in other cases an evaluation may require additional data collection. Programs aimed at helping refugees to settle in this country provide a simple example of such evaluation: Does the assistance reach those most in need and does
32 it accomplish its objectives of easing the transition period and speeding up the process of assimilation whereby refugees become self-supporting? A further area of need for accurate data is for population accounting purposes. The census is taken every 10 years, and the Census Bureau is charged with counting all persons normally resident in the United States, regardless of their legal status. Between censuses, the Bureau prepares estimates of current population by adding births and in-migrants and subtracting deaths and out-migrants from the most recent census count. Births and deaths are counted with considerable accuracy through the national system of registration of vital events, but the population between census dates can change more through net migration than through natural increase (births minus deaths). One would not want to establish an elaborate statistical data collection network merely to know the total number of people living in the country in years between the censuses, but where they live is a matter of great importance, since billions of federal dollars are distributed to states and local areas partly on the basis of population counts or estimates. Hence, the interest of mayors and governors in having their populations estimated accurately is strong. Net international migration can be an even more important component of local area population change than it is of population change at the national level, but it is also even more difficult to measure. Given the number of different agencies involved in administering programs concerned with international migration, it is not surprising that the information needs vary widely. For example, immigrants come into contact initially with State Department representatives throughout the world; with the INS, whom they encounter at U.S. borders; with the federal, state, and local agencies to whom they may turn when in need; with schools, which attempt to provide language capability and citizenship training; with the Department of Labor and related state agencies, which may provide job training or jobs. Refugees, asylees, visitors, and short-term workers all find themselves involved with government agencies and private voluntary agencies before or during their stay in the country, and each such contact invariably becomes noted in the information systems developed by these respective agencies for managing their operations, assessing their successes and failures, preparing requests for funding, and responding to questions from those at all levels of responsibility or from the public. In view of such diversity, it is clearly beyond the scope of this report to attempt to delineate all the information needed for program management. Above all, program management should be seen as a continuing, ever changing activity that generates information needs at all levels, from the national level in Washington to the first-level supervisor or individual Border Patrol officer at a small port of entry on the Canadian border in Maine. All such needs must be considered in designing a program information system. What we wish to emphasize here is that program management requires the maintenance of information systems, the primary function of which is administrative, but which may collect and classify data of use for other, more general purposes.
33 COORDINATING POLICY AND PROGRAM DATA There is a range of users of statistical data and the products of information systems in the immigration field. At one end we find officials in the legislative and executive branches of national government who make and execute immigration policy. Their need is for analysis of aggregated, summary information: information that gives insight into and understanding of demographic, economic, or geographic relationships, such as the age-sex-family status of entering immigrants, the occupations and incomes of those entering for temporary work, the current labor force status of recent refugees, and how efficiently program activities are functioning. In the middle we find those who run the agency offices that deal directly with immigrants, in either service or law enforcement roles; they are largely involved with the effective day-to-day management of services offered and activities performed--using scarce resources to accomplish tasks. At this level, the need also is for aggregated information, but of a somewhat different nature. Generally limited in detail or characteristics, the data consist of summaries--of the number of clients processed in a particular time or similar measures used to measure productivity, accomplishment, adherence to statutory or administrative limitations, and a host of similar managerial controls. At the other end are the program staff who have daily contact with aliens, whose primary need for information is access to records about individuals. The need at this level is for speedy access to information specific to an individual: officials need information to reach a decision and carry out rapidly the service or enforcement activity in the light of all the information available concerning an individual, and subsequently need to update the individual's record to reflect the action taken. Any other information collected at this level is often viewed as an unnecessary burden and an interference with the primary task. A conflict of purpose exists because many of the statistics used for policy or management purposes are initially generated by the program staff, who have little if any stake in or understanding of how records of their individual-level transactions, in a more aggregated form, are used by agency heads and policy makers. For aggregated data to be accurate and timely, an optimal information system would be designed so that the statistical operations required of the program staff impose little additional burden and either benefit them directly (e.g., by also providing useful operational information) or are understood as important to the mission of their agency. An overly simple but nevertheless instructive way to visualize such a system would be to imagine that each employee having contact with aliens has the use of a computer terminal. By accessing a computerized data file that contains relevant information on each alien, the employee retrieves information necessary to making a decision; appropriate additions or changes can be made to the file immediately, thus avoiding endless proliferation of forms containing much duplicated information. Since all files are stored and all transactions are recorded electronically, those in middle- or top-level management positions can simply request and obtain directly from their own terminals the
34 aggregated information they need for effective management or to answer the questions of policy makers. Such a system would clearly require built-in safeguards to ensure confidentiality, to protect identifiers of individuals from unauthorized access, and to avoid illegitimate tampering with the records. At the aggregate level, the records would serve a number of purposes. An office manager could determine how many aliens agent X served or agent Y arrested in a particular period. Other questions that could be answered at various managerial levels would include: How many Iranian students are studying in the United States and where are they located? How many refugees admitted in April 1983 had active cases of tuberculosis? Where are these persons now? What is the age distribution of persons apprehended this year compared with last? Clearly, not all important questions can be answered with management records alone, but they would be much easier to answer if such a system were a reality. For example, suppose policy makers wanted to know how the cohort of Cuban refugees from the 1980 Mariel boatlift has fared in the United States: a special survey would have to be made since the desired information on current status would not be found in the management files. But if a good data system was in place, a sample for the survey could be drawn and located relatively easily from a master listing, including current or most recent address, of the population. Having identified the policy and program needs for accurate and timely data on international migration and briefly outlined the characteristics of a suitable information collection system, we turn next in Chapters 4 through 7 to examine what data are actually collected, by whom, and for what purposes. REFERENCES Brigham, Carl 1923 A Study of American Intelligence. Princeton, N.J Princeton University Press. U.S. Congress, House 1984 Department of Justice Appropriation Authorization Act, Fiscal Year 1985. Report to accompany H.R. 5468. Report 98-759. 98th Congress, 2d session. Available from the Superintendent of Documents, U.S. Government Printing Office.