Click for next page ( 2


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 1
- ~ - Overview and Major Recommendations The minute you walk inside the room, you sense that you've walked into a metaphor, a huge metaphor, for what ails the world of immigration statistics. The room is an immense space inside the district headquarters of the Immigration and Naturalization Service near Wall Street in New York City. It is called the File Room; the simplicity of the name belies the complexity of its contents. The room is as wide as a tennis court and as long as a football field, with sunlight streaming in along the long 300-foot side and the interior aisles between the files lighted by strings of neon. To walk around this room is amazing as a spatial experience. Along one wall are endless banks of metal drawers containing cards for immigration files there's no longer room for. They've been sent to purgatory at a Federal Records Center in New Jersey; some 25,000 are called back for use each year. Near the door are stacks of cardboard boxes and rows of shiny shopping carts full of files in transit. A wall of metal bins distinguishes between "JFK," "Seaport," and other origins. But all these trappings of mundane bureaucracy are mere shrubbery compared with the centerpiece of the File Room. What occupies most of the vast space is the "A-File." This single room contains the immigration records of 2.1 million people and to see them all in one place is to see a startling contradiction. Superficially, the A-File radiates modernity. It is housed in huge wall segments called Electric Compack-to-Files, 24 on one side of the center aisle, 34 on the other. The system looks as up-to-date as the nomenclature. The file segments are metal, handsomely painted in designer pastels of blue, salmon, tangerine, and they move from side to side, with a hum, at the push of a button. The A-File, alas, only looks modern. Its packaging covers a Dickensian interior, for inside, the system for record-keeping is no better, indeed is clumsier, than that used by clerks of the Cratchit era. Precisely because the file segments do move from side to side, to save space, they block each other. When one clerk uses a seven-shelf segment, she blocks another clerk who needs to consult the adjacent one: mechanical gridlock. Typically, it takes 20 minutes to get traffic - moving again. Even then, for all the electric modernism, the contents, when you can get to them, are still paper, the paper is still sometimes wrong, often 1

OCR for page 1
2 incomplete and always clumsy, and it takes three days to put files back; that' s what' s in those shopping carts up front. Sometimes the files can' t be found; they've been misplaced in some bureaucratic limbo. And even if the paper system were improved, its utility would still be dismally limited. It might work better for enforcement as applied to individuals, but would still be wholly useless for answering larger questions, answers that can and should be available to help manage the Service and, still more important, to create a basis for national immigration policies. Consider a famous example: How many Iranian students are there in the United States? President Carter asked that question of the Immigration and Naturalization Service (INS) in 1979 after the hostage crisis began. The hostages were released after 444 days and another 1,500 or so days have passed since then and still the INS cannot answer the question. Yet, at least in theory, the Service now has a computer entry in its Nonimmigrant Information System for every foreign student admitted to the United States. Consider an example more relevant to immigration policy: Do aliens pay more in taxes and unclaimable social security contributions than they use up in social services--and which levels of government come out ahead and behind? If the Service had been able even to offer an informed estimate in October 1984, the Simpson-Mazzoli immigration reform bill might not have died in the closing hours of the 98th Congress from concerns over a proposed ceiling of t1 billion on federal reimbursement to local governments. Consider a question of still wider public interest: Do immigrants, legal and illegal, take jobs away from unskilled natives, especially minorities and youth? We just don't know, and the underlying reason why we don't know is not that analysis has been inadequate, but that the data needed for a convincing analysis do not exist. It is not just a question of timeliness or of quality, issues that may be resolved by the INS's enthusiastic pursuit of electronic sophistication, but also a conceptual failure to understand what data are for, and what data can do for the Service and the nation. The INS and other government agencies produce masses of data, if not always timely and not always accurate, about immigrants, refugees, and the foreign-born, but the data are not what we need to answer the fundamental policy issues of the day. THE PROBLEM As a nation built by waves of immigrants from colonial times to the present, we know remarkably little about the composition and characteristics of the flow of new arrivals in any given year or about how they settle in to their new lives in the United States. This anomaly is particularly surprising in view of two further considerations. The first is that immigration and immigration policy affect the lives of enormous numbers of people, both in the United States and elsewhere. Residents of the United States may benefit or suffer from immigration: some may receive lower wages or even lose their jobs as a result of immigration; many more may benefit from lower overall prices

OCR for page 1
3 that may result from immigration. Would-be immigrants unable to migrate legally to the United States under present policies find their aspirations for a better life thwarted and may respond by attempted illegal entry. Past immigrants and their immediate relatives who may be unable to obtain early admission under existing policies may suffer substantial emotional costs, although remittances sent hone may improve the economic lot of those left behind at the cost of the United States' balance of payments and any lost multiplier effects. The problem is that we really do not know anything concrete about the costs and benefits involved because the basic statistics are so limited; policy has been made in a data vacuum that should naturally be abhorred by policy makers. This point leads to the second consideration, namely the substantial effort that goes into the collection of many of the other data series generated by government. Immigration for some reason is the Cinderella of the federal statistical system. In essence a history of neglect has afflicted record-keeping concerning one of the most fundamental processes underlying the development of American society: the arrival and integration of new populations into contemporary American social and economic structures. In recent years, the expressions of concern over inadequate, incomplete, and often unreliable information available for use in planning, implementing, or evaluating immigration policy have become both more numerous and more strident: for example, the Select Committee on Population of the House of Representatives (1978:1), in attempting to explore the role of immigration in future population growth, concluded that "immigration issues are clouded by faulty data and inflamed passions--not a good combination for rational policymaking." The Interagency Task Force on Immigration (1979:212) reported that "the level and accuracy of available immigration data have been repeatedly criticized as unresponsive to policy issues and inadequate for use in social science research." And even more recently, during the debate over aspects of the Simpson-Mazzoli legislation, a committee in the House (House Report 98-759, 1984) noted that, "The Committee is deeply concerned about the unavailability of accurate and current statistical information on immigration matters. . . . The Committee notes . . . that INS has not devoted sufficient resources and attention to this problem and, to a great extent, has ignored the statistical needs of Congress, as well as, the research needs of demographers and other outside users." These few examples amply demonstrate that the need for a high-quality, readily accessible data base of timely information on immigrant populations has been manifest for years. This need has been heightened recently by renewed controversy in the Congress and in public debate over both restrictions on the size and composition of new immigrant groups and the perceived threat posed by the supposed onslaught of illegal or undocumented migrants against our borders.

OCR for page 1
4 THE CHARGE TO THE PANEL Responding to the growing chorus of concerns in the late 1970s, the Immigration and Naturalization Service of the U.S. Department of Justice asked the Committee on National Statistics of the National Research Council to convene a conference to assess the feasibility of and need for a review of federal immigration statistics. Held in late 1980, the conference strongly supported the idea of a comprehensive review and, accordingly, in late 1982 the Panel on Immigration Statistics was established by the National Research Council, with the support of the Immigration and Naturalization Service. The panel's objectives were three: o To determine the data needs for immigration policy, for administration of immigration law, and for other purposes related to . . . Immigration; o To review existing data sources related to immigration, emigration, and the foreign stock and to assess their statistical adequacy; and o To identify major shortcomings and recommend appropriate remedies and actions. The panel's activities have included interviews with appropriate officials and experts, visits to regional, district, and Border Patrol offices of the INS, visits to ports of entry, and meetings at the U.N. and with officials from foreign governments. The panel also examined the "demand" side to determine the current and projected needs for data among the range of users, attempted to quantify the "supply" side of what exists both within and outside government, explored organizational responsibilities, and developed suggestions and recommendations for both change and improvement. The panel sponsored workshops on the measurement of illegal entries and the possible uses of administrative data collected for other purposes, such as social security or internal revenue, in immigration research. Finally, the panel supported the preparation of a number of methodological papers related to immigration data issues. This report is the culmination of a searching two-year study by the panel of the demand for information on immigrant populations and existing data resources to respond to that demand. The panel has taken a very broad view of the problems of immigrant populations and has included in its deliberations not only the statistical needs and products of the INS but also those of the Office of Refugee Resettlement, a wide variety of other federal offices, and a myriad of nongovernmental sources. The panel's recommendations are sweeping and comprehensive and will require major commitments at several levels in the government if real improvement in the immigration statistics area is to be realized. The purpose of this report is to specify the data requirements and to recommend the practical steps that must be taken to place future debate about immigrant populations on a solid base of statistical information.

OCR for page 1
5 PLAN OF THE REPORT The report is organized as follows. Chapter 1 provides an overview of the panel's work. In this chapter the reader will find a brief account of the background to the panel's work (discussion of the inadequacy of our knowledge of the processes underlying migration flows or of the adjustment of immigrants to life in the United States), the charge that the panel undertook, the way in which the panel carried out its work, and the major conclusions and recommendations that it arrived at. Thus Chapter 1 provides the essentials of the panel's study, though for more detail and justification the reader must consult chapters 2 through 9. The overall approach of the report is to examine data needs (Chapter 3), data availability (Chapters 4 through 7), and the balance, the areas in which data needs are not met by available data (Chapter 8~. Although detailed recommendations appear throughout the report, Chapter 9 presents the panel's major recommendations. First, however, Chapter 2 provides a historical overview of U.S. immigration policy in order to set the problems associated with immigration statistics in their proper perspective. This chapter contains definitions of the populations involved in congressional legislation--past and present--and includes a detailed description of the current visa allocation system. A presentation of immigration law and the role of the INS in implementing it clarifies the minimal information that must be collected to fulfill legislative requirements. The needs for statistics on immigration go well beyond what is necessary for the INS to fulfill its mandate as a law enforcement agency. Chapter 3 examines the diversity of data needs on immigration. The chapter starts with a description of the data collection required by law, that is, the irreducible minimum information on immigrant populations that should be available according to existing legislation. Given this basic, mandated data foundation, the chapter examines the wider data needs for those who make, execute, and evaluate public policy, including the data requirements, primarily within the INS, for program management purposes. The chapter concludes with a discussion of potential impediments to meeting the various data needs identified. Chapter 4 presents an in-depth analysis of the INS. Because of its central role in regulating the flow of international migration to the United States, the INS's statistical activities are examined in far greater detail than those of the other federal agencies involved in collecting immigration data. As the recommendations of the panel point out, it is from the core of information within the INS--for instance, in the File Room in New York--that we must build the statistical system that can respond to the diverse array of data needs documented in Chapter 3. Recommendations for improvement from the general (overall organization and attitude) to the specific (statistical procedures) are incorporated in the discussion. Other official government sources of information on immigration are described and evaluated, with recommendations for improvements, in Chapter 5. Complementary to these sources is a heterogeneous collection of data sources and analytical studies produced outside government. Such nongovernment work is discussed in general teems in Chapter 6 together

OCR for page 1
6 with recommendations for promoting more effective data collection and analysis on immigrant populations in the private sector in the future. Chapter 7 examines data production for a "best case," that of refugees, and highlights the deficiencies even of this relatively favored system. Chapter 8 weighs supply against demand, and costs against benefits, to identify the approaches to improvement that should be adopted to ensure an adequate statistical basis for the future. Chapter 9 presents the panel's major recommendations by the agency to which they are directed; the recommendations are keyed to the chapters in which they are discussed. The report is followed by a glossary of teems and abbreviations and a number of appendices concerned with particular aspects of data collection, data analysis, and the panel's work. Appendix A reproduces a selection of the forms that provide the basic raw information for the INS data system. Appendix B contains several methodological papers on immigration-related issues: (a) some new methods for assessing stocks and flows of migrants; (b) a review of estimates of the size of the illegal alien population; and (c) imputation and the treatment of missing data. Appendix C is a paper written for the panel by Douglas Massey on the settlement process among Mexican migrants. Its relevance to the report stems from the fact that it dramatically illustrates the economic and social dynamics of migrant populations that can be ascertained by a coordinated interplay of good survey data and qualitative ethnographic evidence. Appendix D presents the letter report submitted by the panel to the INS in early 1983 concerning selected statistical aspects of the Senate version of the Simpson-Mazzoli bill. Although the legislation died in the waning hours of the 98th Congress, the wording of the bill illustrates that important provisions regarding statistics on aliens have in fact made their way into contemporary pending legislation. Three final appendices complete the report. Appendix E outlines many of the diverse activities undertaken by panel and staff; Appendix F presents a description of key bibliographic sources and selected bibliographic examples; and Appendix G presents biographical sketches of panel members and staff. WHAT WE FOUND The one fact that struck the panel repeatedly was that a statistical system to produce immigration data does exist, but it does so in an atmosphere of almost total neglect. We found an extraordinary lack of concern with the situation on the part of many who are key to the operation of the statistical system, an almost total ignorance of its existence on the part of the top management that most needs its products, and, finally, that this neglect extends throughout almost all levels of responsibility and almost all the agencies most directly involved in the system. Further, the Congress, with its ultimate power of the purse, must share in the blame for having condoned the situation far too long. In its explorations into data needs and availability, the panel was surprised to find that the Immigration and Nationality Act, the legislative centerpiece in the immigration field, mandates very little

OCR for page 1
7 statistical compilation. Nonetheless, program needs (more than policy needs) have resulted in the establishment of administrative record-keeping systems that are the source of a variety of information on those entering or applying to enter the United States as immigrants, visitors, students, or in some other category. For the most part, requests for information for policy purposes are met from these record-keeping sources. The panel approached the issue of defining data needs by identifying layers of data users, starting with officials in the legislative and executive branches of the federal government who make and execute immigration policy. At this level, as the recent debates over immigration reform quickly reveal, the argument revolves largely around the issues of whether and why immigration is beneficial or, in teems of data, who benefits from immigration and who loses. Obviously, the data needs vary widely over time, often determined by legislative initiatives far removed from the immigration area, such as Medicaid use by illegal aliens or attempts to evaluate factors affecting the apportionment of funds. In the middle, the panel identified the data needs of program managers throughout agencies that deal directly with immigrants, either in a service or a law enforcement role. Their involvement is in attempting to achieve effective management and, for this purpose, information is a tool through which scarce resources are effectively allocated. Finally, there are those whose daily contact is with aliens, and who need access to files about individuals. The dilemma is that many of the statistics used by all three levels are generated initially by this last level, who have little if any stake in their use by the other levels. In fact, such program staff often have little or no knowledge about such use or the need for such use by the agency and thus have little commitment to quality, timeliness, accuracy, or adherence to statistical standards. Thus, irrespective of the ultimate data need, the systems must recognize and deal with the need to inform, educate, and motivate the data collector else, as the wisdom goes (and as experience shows), "garbage in, garbage out." Recognizing that it is not possible to foresee future needs in detail, nor to define every last piece of information that should be collected from or about a specific alien, the panel dealt with the issue by defining major areas in which data are needed and examining alternative strategies for obtaining them. Furthermore, given the awareness that this is a time of acute concern with government spending, the costs of the different strategies were carefully weighed against the value of the expected improvement in both data quality and quantity. Thus, the least expensive type of improvement is to be found in a more extensive presentation of data already collected and available, either in tabulated form or ready for tabulation. Next in cost is the processing of data that have been collected but not previously used, followed by the merging of existing data files to expand the core of knowledge. Modification of existing data collection mechanisms is more costly still, and superseded only by the institution of wholly new data collection systems. These, then, are the choices to be considered and weighed against the information gains and the importance of the needs.

OCR for page 1
8 The examination of data needs versus availability led directly to the agencies and offices that produce or use the data or, in many cases, do both. The INS is, of course, predominant in the collection, if not the dissemination, of data. Unfortunately, the panel also found it the most wanting and is highly critical of its performance in the statistical field. Even recognizing the legal and administrative missions of the agency and that the statistical activities of the Service are directly related to and controlled by the activities involved in carrying out its missions--monitoring entry into and exit from the United States and changes in legal status--the panel noted numerous examples of the agency's inability to meet its own needs, much less those from outside. Two examples illustrate this inability. The first concerns the G-23 report, which summarizes office workload activities on a monthly basis in a variety of areas and contains some 25,000 potential data entries. When "statistics" are discussed at INS, the point of reference is usually the G-23 report. Why then was the panel told repeatedly by staff at all levels of the agency that data reported on the G-23 are assumed to be inaccurate, invalid, and irrelevant to program evaluation and operational analysis? The second example is a quote from the latest INS Statistical Yearbook, that for 1981, which states, "Data processing problems have resulted in incomplete information on immigrants admitted in fiscal 1980 and 1981, the loss of all nonimmigrant information for fiscal year 1980 and incomplete nonimmigrant information for 1982." In fairness to the INS, it should be noted that a major effort is now under way in the agency to install automated systems that are intended to overcome many of the problems that have plagued it in the past. Planning for these systems was preceded by an extensive study of the information requirements of the agency; unfortunately, the needs for policy information of the executive branch or the Congress, while acknowledged, were considered to be outside the purview of the exercise and were omitted from examination. In sum, then, it would be naive to assume that automation will solve the problems that have been evident for too long a time in the statistical operations of the INS. Data vary in quality, refinement, consistency, accessibility, and timeliness. The agency's problems are fundamental and pervasive but center on the basic issue of quality. The Statistical Analysis Branch, the key organizational entity as far as statistics are concerned, does not even appear on an organizational chart of the agency and its influence is notable for its absence. Three factors can be said to characterize the low status of statistical programs in the INS: o A lack of understanding and commitment throughout the agency to the need for high-quality statistics; o The lack of emphasis on statistics in the current bureaucratic structure; and o The absence of standards for performance in the collection, processing, or publication of data. Taken together, these factors are significant warning signals and clearly demonstrate that immediate and direct action is necessary. What is

OCR for page 1
9 required is a fundamental change in the outlook of the institution toward statistics, a change that will recognize explicitly and unequivocally the role of statistics and statistical analysis in the mission of the INS and will ensure that this role is nurtured and supported. Not surprisingly, data on aliens are collected and produced by a variety of federal agencies or offices, either as a result of dealing directly with aliens, as in the case of the Bureau of Consular Affairs of the State Department, or as a by-product of their activities, as in the case of the Social Security Administration. The panel reviewed a total of 11 separate organizations within ~ different cabinet-level agencies; the number of organizations having some information on aliens, or the potential to obtain such information, is undoubtedly much greater. The overriding Impression that emerged from this review was of the need for coordination and direction. Each agency has followed its own institutional priorities in the area of what to collect, how to define it, whether to publish it, or how much to spend on it with little regard for the broader issues involved. Furthermore, there is obvious potential for the establishment of symbiotic relationships between agencies, whereby the use and interpretation of one data set could be greatly enhanced by ready access to others, but in the field of immigration statistics this potential has gone largely unrealized. Someone must bring together all the agencies concerned with immigration data, both those who produce and those who use, in order to ensure the best use of scarce resources, whether money, staff, time, or public tolerance. Progress must be monitored, as must adherence to standards, common definitions, timeliness in publication, and full disclosure of methods, methodology, procedures, and problems. Only through such coordination will significant improvement occur in the data base. Logic dictates that this coordinating role be played by the Office of Management and Budget (OMB), which, by statute and through its review of budget proposals, is ultimately responsible for establishing current statistical agenda and for monitoring progress. It is regrettable that OMB has not adequately exercised its coordinating authority in recent years, but action now is not too late to improve the data base for future policy deliberations. One segment of the data, emigration from the United States, is totally lacking and is generally ignored in discussions of immigration statistics. Yet estimates made during the past few decades indicate that more than 100,000 persons move out of the United States each year. A strategy should be devised for making accurate and timely estimates of emigration. Finally, as part of our task, the panel explored the possibilities of developing estimates of the illegal population in the United States. Following a review of existing methodologies and studies, as well as meeting with interested researchers to brainstorm new approaches, the panel concluded, albeit reluctantly, that it could not identify or contribute to any breakthroughs in methodology that would substantially narrow the current uncertainty in the estimates. Nonetheless, the brief review of the methods used to estimate the size of the illegal population, which appears in Appendix B. leads the panel to the view that, although all the studies suffer from uncertainties, the number of illegals currently in the United States is between 2 and 4 million and,

OCR for page 1
10 further, that the number has not been growing remarkably fast in recent years. SUMMARY OF MAJOR RECOMMENDATIONS The concluding objective of the panel's charge was to identify major shortcomings and recommend appropriate remedies and actions. The report documents the efforts of the panel in this regard and provides a wide range of suggestions and recommendations on both change and improvement. These are described in the individual chapters and are summarized in Chapter 9. Since many different organizations are involved in the area of immigration statistics, our comments are directed to many different places. The panel wishes to emphasize strongly, however, the need for these diverse groups to act in concert. The recommendations represent a range of different actions that will be fully successful only if implemented as a whole. Most of our recommendations are general in nature, concerned with process rather than the particular, and intentionally so. It is the panel's belief, after extensive study of the current situation and how it has arisen, that superficial local patching will not solve the problem. Without major changes in direction from the top policy-making levels and focused interest within the key agencies, the immigration statistics system will never produce reliable, accurate, and timely statistics that permit rational decision making concerning immigration policy. What follows is a brief listing of those recommendations that, in the panel's view, are of overriding importance, both because they require action and commitment at a high policy level and because each is fundamental to the accomplishment of the key goal--the ready availability of accurate, timely, and useful statistical information on international migration. Failure to implement any of these key recommendations will result in failure to improve the data system fully and cost-effectively. The panel recommends that Congress: o Strongly affirm the importance of reliable, accurate, and timely statistical information on immigration to the needs of the Congress and direct the Attorney General to reexamine the organizational structure of the Immigration and Naturalization Service as it relates to statistics, with a view to placing greater priority on this important task; o Require that the Attorney General prepare and submit by June 30 each year an annual report to the President and the Congress, presenting data on aliens admitted or excluded, naturalizations, asylees, and refugees, describing their characteristics, and containing an analysis of significant developments during the preceding fiscal year in the field of immigration and emigration; and 0 Mandate that a study be initiated and conducted among new immigrants over a 5-year period, in order to develop information for policy guidance on the adjustment experience of families and individuals to the labor market, use of educational and health facilities, reliance on social programs, mobility experience, and income history.

OCR for page 1
The panel recommends that the Attorney General: o Issue a strong policy directive asserting the importance of reliable, accurate, and timely statistical information on immigration to the.mission of the INS and unequivocally committing the INS to improving its existing capabilities. The panel recommends that the commissioner of the INS: o Issue an explicit statement clearly setting forth that the collection, cumulation, and tabulation of reliable, accurate, and timely statistical information on immigration is a basic responsibility and inherent in the mission of the INS; o Establish a Division of Immigration Statistics, reporting directly to an associate commissioner or an equivalent level, with overall responsibility: --for ensuring the use of appropriate statistical standards and procedures in the collection of data throughout the agency; --for ensuring the timely publication of a variety of statistical and analytic reports; --for providing statistical assistance and direction to all parts of the agency to help in carrying out their mission; --for directing statistical activities throughout the agency; o Direct and implement the recruitment of a full complement of competent, trained professionals with statistical capabilities and subject-area expertise; o Establish an advisory committee composed of experts in the use and production of immigration-related data, to advise the associate commissioner and the proposed Division of Immigration Statistics of needs for new or different types of data; to review existing data and data collection methodology; and to provide the Service with independent evaluation of its statistical products, plans, and performance; o Establish formal liaison with other federal and state agencies involved in the collection or analysis of immigration- and emigration-related data; and o Initiate a review of all data gathering activities to eliminate duplication, minimize burden and waste, review specific data needs and uses, improve question wording and format design, standardize definitions and concepts, document methodologies, introduce statistical standards and procedures, and promote efficiencies in the use of staff and resources. The panel recommends that the director, Office of Management and Budget (OMB): o Ensure that OMB exercise its responsibilities to monitor and review statistical activities and budgets concerning statistics on immigration and emigration, and particularly those of the INS, to minimize duplication and ensure that appropriate procedures are used, standards met, and priorities observed in the collection, production, and publication of such data;

OCR for page 1
12 o Require and establish an interagency review group responsible for direction and coordination in the field of immigration and emigration data; the group would examine consistency and comparability in concepts and definitions used by individual organizations in the collection of such information; and oversee the introduction and use of standardized approaches; and o Actively encourage and support the timely publication and dissemination of data on immigration, emigration, the ready availability of fully documented public-use data tapes, including samples of individual records without identifiers where feasible, and data summaries. In making its recommendations, the panel has been mindful of costs. Many of its recommendations fall within the scope and margin of administrative discretion, and, if they require additional funds, the amounts are relatively small. Two of the panel's major recommendations will require new funding, but in both cases implementation will be gradual, with expenditures spread over a number of years. The major recommendation for change in administrative structure concerns the establishment of a Division of Immigration Statistics within the INS, which will have increased authority, responsibility, and professional staff. We expect, however, that a period of 3-5 years will be required for the full development of such a division, in order to locate and integrate new staff and to acquire new responsibilities and demonstrate capability on a step-by-step basis. Thus, the initial cost implications are modest and the cost increments can be viewed in the light of some initial accomplishments. ~e , ~_ ~ The major recommendation for a new data collection 1nltlatlve, one longitudinal survey of immigrants, also requires new funding but, again, the estimated cost will be spread over a number of years and is amply justified in the view of the panel.